200-Bid_Prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "xmZKjC8xkH3X"
   },
   "source": [
    "## Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 16
    },
    "colab_type": "code",
    "id": "l6GH5t1wjrsB",
    "outputId": "8001469a-519a-4a95-aa8b-71b5861f4356"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "        <script type=\"text/javascript\">\n",
       "        window.PlotlyConfig = {MathJaxConfig: 'local'};\n",
       "        if (window.MathJax) {MathJax.Hub.Config({SVG: {font: \"STIX-Web\"}});}\n",
       "        if (typeof require !== 'undefined') {\n",
       "        require.undef(\"plotly\");\n",
       "        requirejs.config({\n",
       "            paths: {\n",
       "                'plotly': ['https://cdn.plot.ly/plotly-latest.min']\n",
       "            }\n",
       "        });\n",
       "        require(['plotly'], function(Plotly) {\n",
       "            window._Plotly = Plotly;\n",
       "        });\n",
       "        }\n",
       "        </script>\n",
       "        "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data manipulation\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "# Data visualization\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Plotly visualizations\n",
    "from plotly import tools\n",
    "import plotly.plotly as py\n",
    "import plotly.figure_factory as ff\n",
    "import plotly.graph_objs as go\n",
    "from plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot\n",
    "init_notebook_mode(connected=True)\n",
    "\n",
    "# Display properties\n",
    "pd.set_option('display.max_columns', 150)\n",
    "pd.set_option('display.max_rows', 150)\n",
    "from IPython.display import display, HTML\n",
    "\n",
    "# Date\n",
    "import datetime\n",
    "from datetime import date, timedelta\n",
    "\n",
    "\n",
    "# Maps\n",
    "from math import pi\n",
    "\n",
    "# Storing and Reading data (pickle)\n",
    "import pickle \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "0PPHBQrxjrsL"
   },
   "outputs": [],
   "source": [
    "#Sklearn packages\n",
    "\n",
    "# Import required packages for pre and post processing\n",
    "from sklearn.preprocessing import normalize\n",
    "\n",
    "# Perform the necessary imports\n",
    "from scipy.cluster.hierarchy import linkage, dendrogram\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "zfhBy7RNkPf0"
   },
   "source": [
    "## Call Pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('fifa_processed.pkl', 'rb') as f:\n",
    "    fifa = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('predictors_processed.pkl', 'rb') as f:\n",
    "    predictors_df = pickle.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Changing the categorical variables to dtype 'categorical'."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_df[['Preferred Foot','International Reputation','Weak Foot','Skill Moves','Real Face']]=predictors_df[['Preferred Foot',\n",
    "                                                                                                                  'International Reputation',\n",
    "                                                                                                                  'Weak Foot','Skill Moves',\n",
    "                                                                                                                  'Real Face']].astype('category')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_df[list(predictors_df.loc[:,'Body Type':'Position_Group_Midfielder'])]=predictors_df[list(predictors_df.loc[:,'Body Type':'Position_Group_Midfielder'])].astype('category')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.) Linear Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\statsmodels\\compat\\pandas.py:49: FutureWarning:\n",
      "\n",
      "The Panel class is removed from pandas. Accessing it from the top-level namespace will also be removed in the next version\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "import statsmodels.api as sm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we are seperating value of player out from the fifa data frame, as these are the values we would like to predict."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "target_val = fifa.Value"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we run a simple linear regression model to get a first impression of what is going on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Intercept: \n",
      " -14799392093692.605\n",
      "Coefficients: \n",
      " [-3.48640994e+05 -3.79861117e+16  3.79861117e+16 -4.18271163e+04\n",
      " -3.34419519e+04  3.42066805e+06  7.79750731e+04  1.29307541e+05\n",
      "  1.01594625e+06 -1.00820000e+04  1.72000000e+02  3.50699688e+04\n",
      "  2.64960438e+05 -6.15350000e+03  1.44800000e+03  5.82435658e+14\n",
      " -2.08458554e+14 -3.73977104e+14  1.91772874e+14 -5.51351973e+13\n",
      "  5.34971915e+13  1.63800592e+12 -1.91772874e+14  2.43524915e+15\n",
      " -1.15362983e+15 -1.28161931e+15 -2.10800106e+14 -1.05364314e+13\n",
      "  6.48714259e+12  4.04928875e+12  2.10800106e+14  4.28481925e+14\n",
      " -1.78365680e+13 -2.76531297e+11  1.81130992e+13 -4.28481925e+14\n",
      "  2.72935275e+12 -2.58876566e+14  1.39499087e+14  1.19377478e+14\n",
      " -2.72935278e+12  4.02817792e+13  4.02817792e+13  4.02817792e+13\n",
      "  4.02817792e+13  4.02817792e+13 -1.12403274e+14 -1.12403274e+14\n",
      " -1.12403274e+14 -1.12403274e+14 -1.12403274e+14  1.40108006e+14\n",
      "  1.40108006e+14  1.40108006e+14  1.40108006e+14  1.40108006e+14\n",
      "  1.41050468e+14  1.41050468e+14  1.41050468e+14  1.41050468e+14\n",
      "  1.41050468e+14 -1.16375105e+14 -1.16375105e+14 -1.16375105e+14\n",
      " -1.16375105e+14 -1.16375105e+14 -1.16375105e+14  1.00687888e+14\n",
      "  1.00687888e+14  1.00687888e+14  1.40338264e+14  1.40338264e+14\n",
      "  1.40338264e+14  1.40338264e+14  1.40338264e+14  4.94454795e+06\n",
      "  6.49866262e+06 -2.01408896e+14  5.62016372e+14 -7.00540028e+14\n",
      " -7.05252341e+14  6.98250632e+14 -3.02063665e+14 -7.01691318e+14\n",
      " -3.79861117e+16  1.02161250e+04  1.17748625e+05  2.23001711e+05\n",
      "  1.16898325e+12 -2.21836366e+12  1.16898345e+12  1.29538678e+12\n",
      "  1.16898372e+12  9.49045923e+12  1.16898427e+12 -2.21836395e+12\n",
      " -2.21836339e+12  1.16898408e+12  1.16898374e+12  1.29539528e+12\n",
      "  1.16898320e+12  1.29538771e+12  1.29538738e+12 -2.21836362e+12\n",
      "  1.16898340e+12 -2.21836399e+12 -2.21836363e+12  1.16898403e+12\n",
      "  1.16898370e+12  1.29539210e+12  1.16898325e+12  1.29538721e+12\n",
      "  1.29538718e+12 -2.21836333e+12  1.29538702e+12 -2.91354707e+12\n",
      " -2.91354684e+12 -2.91354688e+12 -2.91354719e+12 -2.91354667e+12\n",
      " -2.91354661e+12  1.99312712e+13  1.64175201e+13  8.22246063e+12\n",
      "  1.65439240e+13]\n"
     ]
    }
   ],
   "source": [
    "# with sklearn\n",
    "lre = LinearRegression()\n",
    "lre.fit(predictors_df, target_val)\n",
    "\n",
    "print('Intercept: \\n', lre.intercept_)\n",
    "print('Coefficients: \\n', lre.coef_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.) Baseline models and feature importance for Value prediction\n",
    "### 2.1) Splitting data (stratified)\n",
    "\n",
    "To make sure that our test set as well as all k-folds have a similar value distribution, we stratify our data when splitting it into test and training set. To identify reasonable bins for stratifying, we first identify outliers and how big the portion of outliers is in the dataset. We, therefore, only consider exceptionally high values to be outliers, as the value distribution does not follow a normal distribution (rather a gamma distribution)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "target_val_q25, target_val_q75 = np.percentile(target_val, 25), np.percentile(target_val, 75)\n",
    "target_val_iqr = target_val_q75 - target_val_q25\n",
    "\n",
    "cut_off = target_val_iqr * 1.5\n",
    "target_val_upper = target_val_q75 + cut_off\n",
    "\n",
    "target_val_outliers = [x for x in target_val if x > target_val_upper]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.1369568808855113"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(target_val_outliers)/len(target_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Over 13% of the data are considered to have exceptionally high values. This group is big enough to build its own bin. We are going to split all the rest into four bins, each containing 25% of the left portion of the data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "target_val_left=[x for x in target_val if x <= target_val_upper]\n",
    "q25, q50, q75=np.percentile(target_val_left,25), np.percentile(target_val_left,50), np.percentile(target_val_left,75)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_binned=[1 if x<=q25 else 2 if q25<x<=q50 else 3 if q50<x<=q75 else 4 if q75<x<=target_val_upper else 5 for x in target_val]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These bins can now be used for stratifying our training and test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "train_predictors_val, test_predictors_val, train_target_val, test_target_val = train_test_split(predictors_df,\n",
    "                                                                                                target_val,\n",
    "                                                                                                test_size=0.2,\n",
    "                                                                                                stratify=y_binned)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1.452700e+04\n",
       "mean     2.426096e+06\n",
       "std      5.692163e+06\n",
       "min      0.000000e+00\n",
       "25%      3.000000e+05\n",
       "50%      6.750000e+05\n",
       "75%      2.000000e+06\n",
       "max      1.185000e+08\n",
       "Name: Value, dtype: float64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_target_val.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    3.632000e+03\n",
       "mean     2.376276e+06\n",
       "std      5.222748e+06\n",
       "min      0.000000e+00\n",
       "25%      3.000000e+05\n",
       "50%      6.750000e+05\n",
       "75%      2.000000e+06\n",
       "max      7.200000e+07\n",
       "Name: Value, dtype: float64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_target_val.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([3.318e+03, 2.170e+02, 4.700e+01, 2.100e+01, 1.200e+01, 5.000e+00,\n",
       "        4.000e+00, 3.000e+00, 3.000e+00, 2.000e+00]),\n",
       " array([       0.,  7200000., 14400000., 21600000., 28800000., 36000000.,\n",
       "        43200000., 50400000., 57600000., 64800000., 72000000.]),\n",
       " <a list of 10 Patch objects>)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(test_target_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([1.396e+04, 3.840e+02, 1.050e+02, 3.600e+01, 1.400e+01, 1.400e+01,\n",
       "        8.000e+00, 3.000e+00, 1.000e+00, 2.000e+00]),\n",
       " array([0.0000e+00, 1.1850e+07, 2.3700e+07, 3.5550e+07, 4.7400e+07,\n",
       "        5.9250e+07, 7.1100e+07, 8.2950e+07, 9.4800e+07, 1.0665e+08,\n",
       "        1.1850e+08]),\n",
       " <a list of 10 Patch objects>)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(train_target_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The distribution of our test and training set look very similar. Just as we wanted them to be."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2) Baseline models\n",
    "First, we are applying some models to our data without doing any parameter tuning or dropping any features. The obtained cross-validation, test and training errors can then be used as a baseline as we progress.\n",
    "\n",
    "Before fitting the models, we scale all numeric features. Since the dataframe only has data of type categorical, float and int, we select the latter two for scaling."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 14527 entries, 11893 to 13818\n",
      "Columns: 125 entries, Age to Position_Group_Midfielder\n",
      "dtypes: category(45), float64(43), int64(37)\n",
      "memory usage: 9.6 MB\n"
     ]
    }
   ],
   "source": [
    "train_predictors_val.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "scaler = StandardScaler()\n",
    "train_predictors_val[train_predictors_val.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']).columns] = scaler.fit_transform(train_predictors_val.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']))\n",
    "\n",
    "test_predictors_val[test_predictors_val.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']).columns] = scaler.fit_transform(test_predictors_val.select_dtypes(include=['float64','float32','int64','int32'],exclude=['category']))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_predictors_val[train_predictors_val.select_dtypes(include='category').columns]=train_predictors_val[train_predictors_val.select_dtypes(include='category').columns].astype('int')\n",
    "test_predictors_val[test_predictors_val.select_dtypes(include='category').columns]=test_predictors_val[test_predictors_val.select_dtypes(include='category').columns].astype('int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Age</th>\n",
       "      <th>Overall</th>\n",
       "      <th>Potential</th>\n",
       "      <th>Special</th>\n",
       "      <th>Preferred Foot</th>\n",
       "      <th>International Reputation</th>\n",
       "      <th>Weak Foot</th>\n",
       "      <th>Skill Moves</th>\n",
       "      <th>Real Face</th>\n",
       "      <th>Jersey Number</th>\n",
       "      <th>Joined</th>\n",
       "      <th>Loaned From</th>\n",
       "      <th>Contract Valid Until</th>\n",
       "      <th>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>LS</th>\n",
       "      <th>ST</th>\n",
       "      <th>RS</th>\n",
       "      <th>LW</th>\n",
       "      <th>LF</th>\n",
       "      <th>CF</th>\n",
       "      <th>RF</th>\n",
       "      <th>RW</th>\n",
       "      <th>LAM</th>\n",
       "      <th>CAM</th>\n",
       "      <th>RAM</th>\n",
       "      <th>LM</th>\n",
       "      <th>LCM</th>\n",
       "      <th>CM</th>\n",
       "      <th>RCM</th>\n",
       "      <th>RM</th>\n",
       "      <th>LWB</th>\n",
       "      <th>LDM</th>\n",
       "      <th>CDM</th>\n",
       "      <th>RDM</th>\n",
       "      <th>RWB</th>\n",
       "      <th>LB</th>\n",
       "      <th>LCB</th>\n",
       "      <th>CB</th>\n",
       "      <th>RCB</th>\n",
       "      <th>RB</th>\n",
       "      <th>Crossing</th>\n",
       "      <th>Finishing</th>\n",
       "      <th>HeadingAccuracy</th>\n",
       "      <th>ShortPassing</th>\n",
       "      <th>Volleys</th>\n",
       "      <th>Dribbling</th>\n",
       "      <th>Curve</th>\n",
       "      <th>FKAccuracy</th>\n",
       "      <th>LongPassing</th>\n",
       "      <th>BallControl</th>\n",
       "      <th>Acceleration</th>\n",
       "      <th>SprintSpeed</th>\n",
       "      <th>Agility</th>\n",
       "      <th>Reactions</th>\n",
       "      <th>Balance</th>\n",
       "      <th>ShotPower</th>\n",
       "      <th>Jumping</th>\n",
       "      <th>Stamina</th>\n",
       "      <th>Strength</th>\n",
       "      <th>LongShots</th>\n",
       "      <th>Aggression</th>\n",
       "      <th>Interceptions</th>\n",
       "      <th>Positioning</th>\n",
       "      <th>Vision</th>\n",
       "      <th>Penalties</th>\n",
       "      <th>Composure</th>\n",
       "      <th>Marking</th>\n",
       "      <th>StandingTackle</th>\n",
       "      <th>SlidingTackle</th>\n",
       "      <th>GKDiving</th>\n",
       "      <th>GKHandling</th>\n",
       "      <th>GKKicking</th>\n",
       "      <th>GKPositioning</th>\n",
       "      <th>GKReflexes</th>\n",
       "      <th>Club_Reputation</th>\n",
       "      <th>Position_Potential</th>\n",
       "      <th>Attacking</th>\n",
       "      <th>Skill</th>\n",
       "      <th>Movement</th>\n",
       "      <th>Power</th>\n",
       "      <th>Mentality</th>\n",
       "      <th>Defending</th>\n",
       "      <th>GoalKeeping</th>\n",
       "      <th>Potential_Difference</th>\n",
       "      <th>Body Type</th>\n",
       "      <th>Attack_WR</th>\n",
       "      <th>Defense_WR</th>\n",
       "      <th>Position_CAM</th>\n",
       "      <th>Position_CB</th>\n",
       "      <th>Position_CDM</th>\n",
       "      <th>Position_CF</th>\n",
       "      <th>Position_CM</th>\n",
       "      <th>Position_GK</th>\n",
       "      <th>Position_LAM</th>\n",
       "      <th>Position_LB</th>\n",
       "      <th>Position_LCB</th>\n",
       "      <th>Position_LCM</th>\n",
       "      <th>Position_LDM</th>\n",
       "      <th>Position_LF</th>\n",
       "      <th>Position_LM</th>\n",
       "      <th>Position_LS</th>\n",
       "      <th>Position_LW</th>\n",
       "      <th>Position_LWB</th>\n",
       "      <th>Position_RAM</th>\n",
       "      <th>Position_RB</th>\n",
       "      <th>Position_RCB</th>\n",
       "      <th>Position_RCM</th>\n",
       "      <th>Position_RDM</th>\n",
       "      <th>Position_RF</th>\n",
       "      <th>Position_RM</th>\n",
       "      <th>Position_RS</th>\n",
       "      <th>Position_RW</th>\n",
       "      <th>Position_RWB</th>\n",
       "      <th>Position_ST</th>\n",
       "      <th>Continent_Africa</th>\n",
       "      <th>Continent_Asia</th>\n",
       "      <th>Continent_Europe</th>\n",
       "      <th>Continent_North America</th>\n",
       "      <th>Continent_Oceania</th>\n",
       "      <th>Continent_South America</th>\n",
       "      <th>Position_Group_Defender</th>\n",
       "      <th>Position_Group_Forward</th>\n",
       "      <th>Position_Group_GoalKeeper</th>\n",
       "      <th>Position_Group_Midfielder</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>11893</td>\n",
       "      <td>-0.026879</td>\n",
       "      <td>-0.326537</td>\n",
       "      <td>-0.377531</td>\n",
       "      <td>-1.957513</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.157975</td>\n",
       "      <td>0.724599</td>\n",
       "      <td>-0.274513</td>\n",
       "      <td>1.433110</td>\n",
       "      <td>1.012672</td>\n",
       "      <td>0.256345</td>\n",
       "      <td>-2.534018</td>\n",
       "      <td>-2.534018</td>\n",
       "      <td>-2.534018</td>\n",
       "      <td>-2.504813</td>\n",
       "      <td>-2.504943</td>\n",
       "      <td>-2.504943</td>\n",
       "      <td>-2.504943</td>\n",
       "      <td>-2.504813</td>\n",
       "      <td>-2.510670</td>\n",
       "      <td>-2.510670</td>\n",
       "      <td>-2.510670</td>\n",
       "      <td>-2.541207</td>\n",
       "      <td>-2.552899</td>\n",
       "      <td>-2.552899</td>\n",
       "      <td>-2.552899</td>\n",
       "      <td>-2.541207</td>\n",
       "      <td>-2.536604</td>\n",
       "      <td>-2.473143</td>\n",
       "      <td>-2.473143</td>\n",
       "      <td>-2.473143</td>\n",
       "      <td>-2.536604</td>\n",
       "      <td>-2.501955</td>\n",
       "      <td>-2.369425</td>\n",
       "      <td>-2.369425</td>\n",
       "      <td>-2.369425</td>\n",
       "      <td>-2.501955</td>\n",
       "      <td>-1.774264</td>\n",
       "      <td>-1.969235</td>\n",
       "      <td>-2.251654</td>\n",
       "      <td>-1.673917</td>\n",
       "      <td>-2.084089</td>\n",
       "      <td>-2.181190</td>\n",
       "      <td>-1.580779</td>\n",
       "      <td>-1.588286</td>\n",
       "      <td>-1.740636</td>\n",
       "      <td>-3.009445</td>\n",
       "      <td>-1.635844</td>\n",
       "      <td>-1.128942</td>\n",
       "      <td>-1.715828</td>\n",
       "      <td>0.016166</td>\n",
       "      <td>-0.908398</td>\n",
       "      <td>-1.532306</td>\n",
       "      <td>-0.933117</td>\n",
       "      <td>-1.141113</td>\n",
       "      <td>-0.347675</td>\n",
       "      <td>-1.767453</td>\n",
       "      <td>-2.168351</td>\n",
       "      <td>-1.140290</td>\n",
       "      <td>-2.089176</td>\n",
       "      <td>-2.710039</td>\n",
       "      <td>-1.304909</td>\n",
       "      <td>-2.761188</td>\n",
       "      <td>-1.316796</td>\n",
       "      <td>-1.408347</td>\n",
       "      <td>-1.387178</td>\n",
       "      <td>2.550967</td>\n",
       "      <td>2.685877</td>\n",
       "      <td>2.758203</td>\n",
       "      <td>2.665921</td>\n",
       "      <td>2.844348</td>\n",
       "      <td>-0.173253</td>\n",
       "      <td>-2.735504</td>\n",
       "      <td>-2.310606</td>\n",
       "      <td>-2.205011</td>\n",
       "      <td>-1.411056</td>\n",
       "      <td>-1.695130</td>\n",
       "      <td>-2.626452</td>\n",
       "      <td>-1.406341</td>\n",
       "      <td>2.735389</td>\n",
       "      <td>-0.012293</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2672</td>\n",
       "      <td>2.329249</td>\n",
       "      <td>0.977567</td>\n",
       "      <td>0.273397</td>\n",
       "      <td>0.670659</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.157975</td>\n",
       "      <td>-0.777071</td>\n",
       "      <td>-0.274513</td>\n",
       "      <td>-1.613802</td>\n",
       "      <td>-1.974715</td>\n",
       "      <td>-0.576948</td>\n",
       "      <td>0.630116</td>\n",
       "      <td>0.630116</td>\n",
       "      <td>0.630116</td>\n",
       "      <td>0.798782</td>\n",
       "      <td>0.816303</td>\n",
       "      <td>0.816303</td>\n",
       "      <td>0.816303</td>\n",
       "      <td>0.798782</td>\n",
       "      <td>0.947259</td>\n",
       "      <td>0.947259</td>\n",
       "      <td>0.947259</td>\n",
       "      <td>0.771154</td>\n",
       "      <td>0.860020</td>\n",
       "      <td>0.860020</td>\n",
       "      <td>0.860020</td>\n",
       "      <td>0.771154</td>\n",
       "      <td>0.249601</td>\n",
       "      <td>0.324490</td>\n",
       "      <td>0.324490</td>\n",
       "      <td>0.324490</td>\n",
       "      <td>0.249601</td>\n",
       "      <td>0.131050</td>\n",
       "      <td>-0.111701</td>\n",
       "      <td>-0.111701</td>\n",
       "      <td>-0.111701</td>\n",
       "      <td>0.131050</td>\n",
       "      <td>1.218376</td>\n",
       "      <td>0.895340</td>\n",
       "      <td>-0.533713</td>\n",
       "      <td>1.178301</td>\n",
       "      <td>1.083172</td>\n",
       "      <td>0.990126</td>\n",
       "      <td>1.408282</td>\n",
       "      <td>1.390602</td>\n",
       "      <td>1.260642</td>\n",
       "      <td>1.298534</td>\n",
       "      <td>-0.302430</td>\n",
       "      <td>-0.857631</td>\n",
       "      <td>0.441535</td>\n",
       "      <td>0.461159</td>\n",
       "      <td>1.695991</td>\n",
       "      <td>0.846964</td>\n",
       "      <td>-2.701805</td>\n",
       "      <td>-0.702148</td>\n",
       "      <td>-0.985416</td>\n",
       "      <td>1.038463</td>\n",
       "      <td>-0.964429</td>\n",
       "      <td>-0.562213</td>\n",
       "      <td>0.721929</td>\n",
       "      <td>1.881990</td>\n",
       "      <td>1.811280</td>\n",
       "      <td>1.257783</td>\n",
       "      <td>-0.264127</td>\n",
       "      <td>0.154115</td>\n",
       "      <td>-0.450233</td>\n",
       "      <td>-0.207226</td>\n",
       "      <td>-0.557767</td>\n",
       "      <td>-0.017901</td>\n",
       "      <td>-0.318153</td>\n",
       "      <td>-0.042934</td>\n",
       "      <td>0.692733</td>\n",
       "      <td>0.318732</td>\n",
       "      <td>0.901504</td>\n",
       "      <td>1.384870</td>\n",
       "      <td>0.312968</td>\n",
       "      <td>-0.379497</td>\n",
       "      <td>0.751068</td>\n",
       "      <td>-0.187339</td>\n",
       "      <td>-0.230296</td>\n",
       "      <td>-0.937608</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9565</td>\n",
       "      <td>0.187314</td>\n",
       "      <td>-0.036736</td>\n",
       "      <td>-0.214799</td>\n",
       "      <td>-1.990457</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.157338</td>\n",
       "      <td>-0.689495</td>\n",
       "      <td>-0.274513</td>\n",
       "      <td>0.671382</td>\n",
       "      <td>1.012672</td>\n",
       "      <td>1.217837</td>\n",
       "      <td>-2.534018</td>\n",
       "      <td>-2.534018</td>\n",
       "      <td>-2.534018</td>\n",
       "      <td>-2.504813</td>\n",
       "      <td>-2.504943</td>\n",
       "      <td>-2.504943</td>\n",
       "      <td>-2.504943</td>\n",
       "      <td>-2.504813</td>\n",
       "      <td>-2.510670</td>\n",
       "      <td>-2.510670</td>\n",
       "      <td>-2.510670</td>\n",
       "      <td>-2.541207</td>\n",
       "      <td>-2.552899</td>\n",
       "      <td>-2.552899</td>\n",
       "      <td>-2.552899</td>\n",
       "      <td>-2.541207</td>\n",
       "      <td>-2.536604</td>\n",
       "      <td>-2.473143</td>\n",
       "      <td>-2.473143</td>\n",
       "      <td>-2.473143</td>\n",
       "      <td>-2.536604</td>\n",
       "      <td>-2.501955</td>\n",
       "      <td>-2.369425</td>\n",
       "      <td>-2.369425</td>\n",
       "      <td>-2.369425</td>\n",
       "      <td>-2.501955</td>\n",
       "      <td>-1.991911</td>\n",
       "      <td>-1.662317</td>\n",
       "      <td>-1.965330</td>\n",
       "      <td>-2.420927</td>\n",
       "      <td>-1.914415</td>\n",
       "      <td>-2.181190</td>\n",
       "      <td>-2.015552</td>\n",
       "      <td>-1.588286</td>\n",
       "      <td>-2.393088</td>\n",
       "      <td>-2.171782</td>\n",
       "      <td>-2.102538</td>\n",
       "      <td>-2.146359</td>\n",
       "      <td>-1.041652</td>\n",
       "      <td>1.239896</td>\n",
       "      <td>-1.471509</td>\n",
       "      <td>-2.228678</td>\n",
       "      <td>-0.512001</td>\n",
       "      <td>-1.830915</td>\n",
       "      <td>0.210349</td>\n",
       "      <td>-1.715492</td>\n",
       "      <td>-1.996362</td>\n",
       "      <td>-1.573848</td>\n",
       "      <td>-1.833621</td>\n",
       "      <td>-0.378701</td>\n",
       "      <td>-1.750079</td>\n",
       "      <td>-0.052751</td>\n",
       "      <td>-1.366923</td>\n",
       "      <td>-1.638120</td>\n",
       "      <td>-1.527719</td>\n",
       "      <td>2.944994</td>\n",
       "      <td>2.567926</td>\n",
       "      <td>2.335752</td>\n",
       "      <td>2.724432</td>\n",
       "      <td>2.844348</td>\n",
       "      <td>-0.506325</td>\n",
       "      <td>-2.735504</td>\n",
       "      <td>-2.324046</td>\n",
       "      <td>-2.255395</td>\n",
       "      <td>-1.571017</td>\n",
       "      <td>-1.875354</td>\n",
       "      <td>-1.863786</td>\n",
       "      <td>-1.552621</td>\n",
       "      <td>2.723667</td>\n",
       "      <td>-0.197356</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17906</td>\n",
       "      <td>-1.312040</td>\n",
       "      <td>-2.210243</td>\n",
       "      <td>-0.540263</td>\n",
       "      <td>-0.555577</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0.909583</td>\n",
       "      <td>0.022703</td>\n",
       "      <td>-0.274513</td>\n",
       "      <td>0.671382</td>\n",
       "      <td>-2.422823</td>\n",
       "      <td>-2.756329</td>\n",
       "      <td>-0.210357</td>\n",
       "      <td>-0.210357</td>\n",
       "      <td>-0.210357</td>\n",
       "      <td>-0.110904</td>\n",
       "      <td>-0.146377</td>\n",
       "      <td>-0.146377</td>\n",
       "      <td>-0.146377</td>\n",
       "      <td>-0.110904</td>\n",
       "      <td>-0.061304</td>\n",
       "      <td>-0.061304</td>\n",
       "      <td>-0.061304</td>\n",
       "      <td>-0.092941</td>\n",
       "      <td>-0.129232</td>\n",
       "      <td>-0.129232</td>\n",
       "      <td>-0.129232</td>\n",
       "      <td>-0.092941</td>\n",
       "      <td>-0.148429</td>\n",
       "      <td>-0.215404</td>\n",
       "      <td>-0.215404</td>\n",
       "      <td>-0.215404</td>\n",
       "      <td>-0.148429</td>\n",
       "      <td>-0.167026</td>\n",
       "      <td>-0.303848</td>\n",
       "      <td>-0.303848</td>\n",
       "      <td>-0.303848</td>\n",
       "      <td>-0.167026</td>\n",
       "      <td>-0.413973</td>\n",
       "      <td>-0.434641</td>\n",
       "      <td>-0.590977</td>\n",
       "      <td>0.363382</td>\n",
       "      <td>-0.161109</td>\n",
       "      <td>-0.278401</td>\n",
       "      <td>-0.004729</td>\n",
       "      <td>-0.041556</td>\n",
       "      <td>0.151474</td>\n",
       "      <td>-0.795623</td>\n",
       "      <td>0.164265</td>\n",
       "      <td>0.431096</td>\n",
       "      <td>0.441535</td>\n",
       "      <td>-1.207563</td>\n",
       "      <td>1.907157</td>\n",
       "      <td>-0.313656</td>\n",
       "      <td>-0.343554</td>\n",
       "      <td>-0.764857</td>\n",
       "      <td>-2.500051</td>\n",
       "      <td>-0.572341</td>\n",
       "      <td>-0.735111</td>\n",
       "      <td>-0.995771</td>\n",
       "      <td>-0.095847</td>\n",
       "      <td>-0.378701</td>\n",
       "      <td>-0.032995</td>\n",
       "      <td>-0.576965</td>\n",
       "      <td>-0.113746</td>\n",
       "      <td>-0.213523</td>\n",
       "      <td>0.111934</td>\n",
       "      <td>-0.319805</td>\n",
       "      <td>-0.439816</td>\n",
       "      <td>-0.078251</td>\n",
       "      <td>-0.435175</td>\n",
       "      <td>-0.598181</td>\n",
       "      <td>-0.056678</td>\n",
       "      <td>0.318732</td>\n",
       "      <td>-0.321517</td>\n",
       "      <td>-0.214831</td>\n",
       "      <td>0.561796</td>\n",
       "      <td>-1.154459</td>\n",
       "      <td>-0.651692</td>\n",
       "      <td>-0.073566</td>\n",
       "      <td>-0.382683</td>\n",
       "      <td>2.208463</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15656</td>\n",
       "      <td>0.401508</td>\n",
       "      <td>-1.051039</td>\n",
       "      <td>-1.842120</td>\n",
       "      <td>0.084994</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.032071</td>\n",
       "      <td>1.431645</td>\n",
       "      <td>-0.274513</td>\n",
       "      <td>-0.852074</td>\n",
       "      <td>-0.481021</td>\n",
       "      <td>-0.448749</td>\n",
       "      <td>0.086280</td>\n",
       "      <td>0.086280</td>\n",
       "      <td>0.086280</td>\n",
       "      <td>0.080609</td>\n",
       "      <td>0.094293</td>\n",
       "      <td>0.094293</td>\n",
       "      <td>0.094293</td>\n",
       "      <td>0.080609</td>\n",
       "      <td>0.082777</td>\n",
       "      <td>0.082777</td>\n",
       "      <td>0.082777</td>\n",
       "      <td>0.099080</td>\n",
       "      <td>0.167544</td>\n",
       "      <td>0.167544</td>\n",
       "      <td>0.167544</td>\n",
       "      <td>0.099080</td>\n",
       "      <td>0.349108</td>\n",
       "      <td>0.275409</td>\n",
       "      <td>0.275409</td>\n",
       "      <td>0.275409</td>\n",
       "      <td>0.349108</td>\n",
       "      <td>0.379447</td>\n",
       "      <td>0.320629</td>\n",
       "      <td>0.320629</td>\n",
       "      <td>0.320629</td>\n",
       "      <td>0.379447</td>\n",
       "      <td>0.293378</td>\n",
       "      <td>-0.025416</td>\n",
       "      <td>-0.648242</td>\n",
       "      <td>-0.451538</td>\n",
       "      <td>-1.122599</td>\n",
       "      <td>-0.119835</td>\n",
       "      <td>-0.656888</td>\n",
       "      <td>0.588594</td>\n",
       "      <td>0.020983</td>\n",
       "      <td>-0.017793</td>\n",
       "      <td>-0.235759</td>\n",
       "      <td>0.295441</td>\n",
       "      <td>-0.232641</td>\n",
       "      <td>-1.430060</td>\n",
       "      <td>0.780935</td>\n",
       "      <td>0.730902</td>\n",
       "      <td>1.256687</td>\n",
       "      <td>0.802875</td>\n",
       "      <td>0.050913</td>\n",
       "      <td>0.570810</td>\n",
       "      <td>0.182163</td>\n",
       "      <td>0.401249</td>\n",
       "      <td>0.159708</td>\n",
       "      <td>0.045178</td>\n",
       "      <td>-0.414569</td>\n",
       "      <td>-0.926440</td>\n",
       "      <td>0.187016</td>\n",
       "      <td>0.659617</td>\n",
       "      <td>0.674101</td>\n",
       "      <td>-0.376095</td>\n",
       "      <td>-0.262890</td>\n",
       "      <td>-0.380002</td>\n",
       "      <td>-0.259641</td>\n",
       "      <td>-0.376082</td>\n",
       "      <td>-0.506325</td>\n",
       "      <td>0.318732</td>\n",
       "      <td>-0.442475</td>\n",
       "      <td>-0.051082</td>\n",
       "      <td>-0.078048</td>\n",
       "      <td>0.936136</td>\n",
       "      <td>-0.025217</td>\n",
       "      <td>0.527808</td>\n",
       "      <td>-0.335795</td>\n",
       "      <td>-0.752545</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Age   Overall  Potential   Special  Preferred Foot  \\\n",
       "11893 -0.026879 -0.326537  -0.377531 -1.957513               1   \n",
       "2672   2.329249  0.977567   0.273397  0.670659               0   \n",
       "9565   0.187314 -0.036736  -0.214799 -1.990457               1   \n",
       "17906 -1.312040 -2.210243  -0.540263 -0.555577               1   \n",
       "15656  0.401508 -1.051039  -1.842120  0.084994               0   \n",
       "\n",
       "       International Reputation  Weak Foot  Skill Moves  Real Face  \\\n",
       "11893                         1          2            1          0   \n",
       "2672                          2          2            3          1   \n",
       "9565                          1          2            1          0   \n",
       "17906                         1          3            2          0   \n",
       "15656                         1          3            2          0   \n",
       "\n",
       "       Jersey Number    Joined  Loaned From  Contract Valid Until    Height  \\\n",
       "11893       0.157975  0.724599    -0.274513              1.433110  1.012672   \n",
       "2672        0.157975 -0.777071    -0.274513             -1.613802 -1.974715   \n",
       "9565       -1.157338 -0.689495    -0.274513              0.671382  1.012672   \n",
       "17906       0.909583  0.022703    -0.274513              0.671382 -2.422823   \n",
       "15656      -1.032071  1.431645    -0.274513             -0.852074 -0.481021   \n",
       "\n",
       "         Weight        LS        ST        RS        LW        LF        CF  \\\n",
       "11893  0.256345 -2.534018 -2.534018 -2.534018 -2.504813 -2.504943 -2.504943   \n",
       "2672  -0.576948  0.630116  0.630116  0.630116  0.798782  0.816303  0.816303   \n",
       "9565   1.217837 -2.534018 -2.534018 -2.534018 -2.504813 -2.504943 -2.504943   \n",
       "17906 -2.756329 -0.210357 -0.210357 -0.210357 -0.110904 -0.146377 -0.146377   \n",
       "15656 -0.448749  0.086280  0.086280  0.086280  0.080609  0.094293  0.094293   \n",
       "\n",
       "             RF        RW       LAM       CAM       RAM        LM       LCM  \\\n",
       "11893 -2.504943 -2.504813 -2.510670 -2.510670 -2.510670 -2.541207 -2.552899   \n",
       "2672   0.816303  0.798782  0.947259  0.947259  0.947259  0.771154  0.860020   \n",
       "9565  -2.504943 -2.504813 -2.510670 -2.510670 -2.510670 -2.541207 -2.552899   \n",
       "17906 -0.146377 -0.110904 -0.061304 -0.061304 -0.061304 -0.092941 -0.129232   \n",
       "15656  0.094293  0.080609  0.082777  0.082777  0.082777  0.099080  0.167544   \n",
       "\n",
       "             CM       RCM        RM       LWB       LDM       CDM       RDM  \\\n",
       "11893 -2.552899 -2.552899 -2.541207 -2.536604 -2.473143 -2.473143 -2.473143   \n",
       "2672   0.860020  0.860020  0.771154  0.249601  0.324490  0.324490  0.324490   \n",
       "9565  -2.552899 -2.552899 -2.541207 -2.536604 -2.473143 -2.473143 -2.473143   \n",
       "17906 -0.129232 -0.129232 -0.092941 -0.148429 -0.215404 -0.215404 -0.215404   \n",
       "15656  0.167544  0.167544  0.099080  0.349108  0.275409  0.275409  0.275409   \n",
       "\n",
       "            RWB        LB       LCB        CB       RCB        RB  Crossing  \\\n",
       "11893 -2.536604 -2.501955 -2.369425 -2.369425 -2.369425 -2.501955 -1.774264   \n",
       "2672   0.249601  0.131050 -0.111701 -0.111701 -0.111701  0.131050  1.218376   \n",
       "9565  -2.536604 -2.501955 -2.369425 -2.369425 -2.369425 -2.501955 -1.991911   \n",
       "17906 -0.148429 -0.167026 -0.303848 -0.303848 -0.303848 -0.167026 -0.413973   \n",
       "15656  0.349108  0.379447  0.320629  0.320629  0.320629  0.379447  0.293378   \n",
       "\n",
       "       Finishing  HeadingAccuracy  ShortPassing   Volleys  Dribbling  \\\n",
       "11893  -1.969235        -2.251654     -1.673917 -2.084089  -2.181190   \n",
       "2672    0.895340        -0.533713      1.178301  1.083172   0.990126   \n",
       "9565   -1.662317        -1.965330     -2.420927 -1.914415  -2.181190   \n",
       "17906  -0.434641        -0.590977      0.363382 -0.161109  -0.278401   \n",
       "15656  -0.025416        -0.648242     -0.451538 -1.122599  -0.119835   \n",
       "\n",
       "          Curve  FKAccuracy  LongPassing  BallControl  Acceleration  \\\n",
       "11893 -1.580779   -1.588286    -1.740636    -3.009445     -1.635844   \n",
       "2672   1.408282    1.390602     1.260642     1.298534     -0.302430   \n",
       "9565  -2.015552   -1.588286    -2.393088    -2.171782     -2.102538   \n",
       "17906 -0.004729   -0.041556     0.151474    -0.795623      0.164265   \n",
       "15656 -0.656888    0.588594     0.020983    -0.017793     -0.235759   \n",
       "\n",
       "       SprintSpeed   Agility  Reactions   Balance  ShotPower   Jumping  \\\n",
       "11893    -1.128942 -1.715828   0.016166 -0.908398  -1.532306 -0.933117   \n",
       "2672     -0.857631  0.441535   0.461159  1.695991   0.846964 -2.701805   \n",
       "9565     -2.146359 -1.041652   1.239896 -1.471509  -2.228678 -0.512001   \n",
       "17906     0.431096  0.441535  -1.207563  1.907157  -0.313656 -0.343554   \n",
       "15656     0.295441 -0.232641  -1.430060  0.780935   0.730902  1.256687   \n",
       "\n",
       "        Stamina  Strength  LongShots  Aggression  Interceptions  Positioning  \\\n",
       "11893 -1.141113 -0.347675  -1.767453   -2.168351      -1.140290    -2.089176   \n",
       "2672  -0.702148 -0.985416   1.038463   -0.964429      -0.562213     0.721929   \n",
       "9565  -1.830915  0.210349  -1.715492   -1.996362      -1.573848    -1.833621   \n",
       "17906 -0.764857 -2.500051  -0.572341   -0.735111      -0.995771    -0.095847   \n",
       "15656  0.802875  0.050913   0.570810    0.182163       0.401249     0.159708   \n",
       "\n",
       "         Vision  Penalties  Composure   Marking  StandingTackle  \\\n",
       "11893 -2.710039  -1.304909  -2.761188 -1.316796       -1.408347   \n",
       "2672   1.881990   1.811280   1.257783 -0.264127        0.154115   \n",
       "9565  -0.378701  -1.750079  -0.052751 -1.366923       -1.638120   \n",
       "17906 -0.378701  -0.032995  -0.576965 -0.113746       -0.213523   \n",
       "15656  0.045178  -0.414569  -0.926440  0.187016        0.659617   \n",
       "\n",
       "       SlidingTackle  GKDiving  GKHandling  GKKicking  GKPositioning  \\\n",
       "11893      -1.387178  2.550967    2.685877   2.758203       2.665921   \n",
       "2672       -0.450233 -0.207226   -0.557767  -0.017901      -0.318153   \n",
       "9565       -1.527719  2.944994    2.567926   2.335752       2.724432   \n",
       "17906       0.111934 -0.319805   -0.439816  -0.078251      -0.435175   \n",
       "15656       0.674101 -0.376095   -0.262890  -0.380002      -0.259641   \n",
       "\n",
       "       GKReflexes  Club_Reputation  Position_Potential  Attacking     Skill  \\\n",
       "11893    2.844348        -0.173253           -2.735504  -2.310606 -2.205011   \n",
       "2672    -0.042934         0.692733            0.318732   0.901504  1.384870   \n",
       "9565     2.844348        -0.506325           -2.735504  -2.324046 -2.255395   \n",
       "17906   -0.598181        -0.056678            0.318732  -0.321517 -0.214831   \n",
       "15656   -0.376082        -0.506325            0.318732  -0.442475 -0.051082   \n",
       "\n",
       "       Movement     Power  Mentality  Defending  GoalKeeping  \\\n",
       "11893 -1.411056 -1.695130  -2.626452  -1.406341     2.735389   \n",
       "2672   0.312968 -0.379497   0.751068  -0.187339    -0.230296   \n",
       "9565  -1.571017 -1.875354  -1.863786  -1.552621     2.723667   \n",
       "17906  0.561796 -1.154459  -0.651692  -0.073566    -0.382683   \n",
       "15656 -0.078048  0.936136  -0.025217   0.527808    -0.335795   \n",
       "\n",
       "       Potential_Difference  Body Type  Attack_WR  Defense_WR  Position_CAM  \\\n",
       "11893             -0.012293          2          2           2             0   \n",
       "2672              -0.937608          1          2           2             1   \n",
       "9565              -0.197356          2          2           2             0   \n",
       "17906              2.208463          1          2           2             0   \n",
       "15656             -0.752545          2          3           2             0   \n",
       "\n",
       "       Position_CB  Position_CDM  Position_CF  Position_CM  Position_GK  \\\n",
       "11893            0             0            0            0            1   \n",
       "2672             0             0            0            0            0   \n",
       "9565             0             0            0            0            1   \n",
       "17906            0             0            0            1            0   \n",
       "15656            0             0            0            0            0   \n",
       "\n",
       "       Position_LAM  Position_LB  Position_LCB  Position_LCM  Position_LDM  \\\n",
       "11893             0            0             0             0             0   \n",
       "2672              0            0             0             0             0   \n",
       "9565              0            0             0             0             0   \n",
       "17906             0            0             0             0             0   \n",
       "15656             0            1             0             0             0   \n",
       "\n",
       "       Position_LF  Position_LM  Position_LS  Position_LW  Position_LWB  \\\n",
       "11893            0            0            0            0             0   \n",
       "2672             0            0            0            0             0   \n",
       "9565             0            0            0            0             0   \n",
       "17906            0            0            0            0             0   \n",
       "15656            0            0            0            0             0   \n",
       "\n",
       "       Position_RAM  Position_RB  Position_RCB  Position_RCM  Position_RDM  \\\n",
       "11893             0            0             0             0             0   \n",
       "2672              0            0             0             0             0   \n",
       "9565              0            0             0             0             0   \n",
       "17906             0            0             0             0             0   \n",
       "15656             0            0             0             0             0   \n",
       "\n",
       "       Position_RF  Position_RM  Position_RS  Position_RW  Position_RWB  \\\n",
       "11893            0            0            0            0             0   \n",
       "2672             0            0            0            0             0   \n",
       "9565             0            0            0            0             0   \n",
       "17906            0            0            0            0             0   \n",
       "15656            0            0            0            0             0   \n",
       "\n",
       "       Position_ST  Continent_Africa  Continent_Asia  Continent_Europe  \\\n",
       "11893            0                 0               0                 0   \n",
       "2672             0                 0               0                 1   \n",
       "9565             0                 0               0                 1   \n",
       "17906            0                 0               1                 0   \n",
       "15656            0                 0               0                 1   \n",
       "\n",
       "       Continent_North America  Continent_Oceania  Continent_South America  \\\n",
       "11893                        0                  0                        1   \n",
       "2672                         0                  0                        0   \n",
       "9565                         0                  0                        0   \n",
       "17906                        0                  0                        0   \n",
       "15656                        0                  0                        0   \n",
       "\n",
       "       Position_Group_Defender  Position_Group_Forward  \\\n",
       "11893                        0                       0   \n",
       "2672                         0                       0   \n",
       "9565                         0                       0   \n",
       "17906                        0                       0   \n",
       "15656                        1                       0   \n",
       "\n",
       "       Position_Group_GoalKeeper  Position_Group_Midfielder  \n",
       "11893                          1                          0  \n",
       "2672                           0                          1  \n",
       "9565                           1                          0  \n",
       "17906                          0                          1  \n",
       "15656                          0                          0  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_predictors_val.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that all numerical non-categorical data are scaled and categorical data are converted to integer (the models do not take dtype categorical as input) we proceed with fitting our baseline models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py:245: FutureWarning:\n",
      "\n",
      "The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "\n",
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:00:56] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:193: FutureWarning:\n",
      "\n",
      "The default value of gamma will change from 'auto' to 'scale' in version 0.22 to account better for unscaled features. Set gamma explicitly to 'auto' or 'scale' to avoid this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import sklearn.model_selection as cv\n",
    "from sklearn.metrics import mean_squared_error as MSE\n",
    "from sklearn.model_selection import cross_val_score, LeaveOneOut\n",
    "\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "import xgboost as xgb\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.svm import SVR\n",
    "\n",
    "dtr=DecisionTreeRegressor()\n",
    "rfr=RandomForestRegressor()\n",
    "xbr=xgb.XGBRegressor()\n",
    "lre=LinearRegression()\n",
    "svr=SVR()\n",
    "\n",
    "models = [dtr,rfr,xbr,lre,svr]\n",
    "\n",
    "for mod in models:\n",
    "    mod.fit(train_predictors_val, train_target_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_error = []\n",
    "\n",
    "# Sequentially fit all models and calculate test score\n",
    "for mod in models:\n",
    "    y_pred = mod.predict(test_predictors_val)\n",
    "    mse_mod = MSE(test_target_val, y_pred)\n",
    "    rmse_mod = round(mse_mod**(1/2))\n",
    "    test_error.append(rmse_mod)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "cv_error=[]\n",
    "\n",
    "for mod in models:\n",
    "    MSE_cv_error = -cross_val_score(mod, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "    RMSE_cv = round((MSE_cv_error.mean())**(1/2))\n",
    "    cv_error.append(RMSE_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_error=[]\n",
    "\n",
    "# Sequentially fit all models and calculate train score\n",
    "for mod in models:\n",
    "    y_pred = mod.predict(train_predictors_val)\n",
    "    mse_mod = MSE(train_target_val, y_pred)\n",
    "    rmse_mod = round(mse_mod**(1/2))\n",
    "    train_error.append(rmse_mod)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_names = ['dtr','rfr','xbr','lre','svc']\n",
    "\n",
    "error = pd.DataFrame(dict(graph=model_names,test_error=test_error, cv_error=cv_error, train_error=train_error)) \n",
    "\n",
    "ind = np.arange(len(error))\n",
    "width = 0.2\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12,6))\n",
    "ax.barh(ind, error.test_error, width, color='maroon', label='test')\n",
    "ax.barh(ind + width, error.cv_error, width, color='forestgreen', label='cv')\n",
    "ax.barh(ind + 2*width, error.train_error, width, color='slategrey', label='train')\n",
    "ax.set_xlim(0, max(error.cv_error)+min(error.cv_error))\n",
    "plt.xlabel('RMSE', fontsize=16)\n",
    "plt.ylabel('Algorithms', fontsize=16)\n",
    "plt.title('Test, Cross Validation and Train Error per Algorithm',fontsize=20)\n",
    "\n",
    "ax.set(yticks=ind + width, yticklabels=error.graph, ylim=[2*width - 1, len(error)])\n",
    "ax.legend()\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>graph</th>\n",
       "      <th>test_error</th>\n",
       "      <th>cv_error</th>\n",
       "      <th>train_error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>dtr</td>\n",
       "      <td>8.899300e+05</td>\n",
       "      <td>1582050.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>rfr</td>\n",
       "      <td>8.853950e+05</td>\n",
       "      <td>963003.0</td>\n",
       "      <td>412169.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>xbr</td>\n",
       "      <td>6.058860e+05</td>\n",
       "      <td>742355.0</td>\n",
       "      <td>423097.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>lre</td>\n",
       "      <td>6.421262e+16</td>\n",
       "      <td>3353246.0</td>\n",
       "      <td>3277148.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>svc</td>\n",
       "      <td>5.491672e+06</td>\n",
       "      <td>5952538.0</td>\n",
       "      <td>5954762.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  graph    test_error   cv_error  train_error\n",
       "0   dtr  8.899300e+05  1582050.0          0.0\n",
       "1   rfr  8.853950e+05   963003.0     412169.0\n",
       "2   xbr  6.058860e+05   742355.0     423097.0\n",
       "3   lre  6.421262e+16  3353246.0    3277148.0\n",
       "4   svc  5.491672e+06  5952538.0    5954762.0"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "While the Support Vector Regression is the most stable model, we can see that the Linear Regression has a very high test error (and, therefore, is overfit). The Decision Tree Regressor is also overfit, as its train error is zero. XGBoost and Random Forest also need some complexity reduction.<br>\n",
    "\n",
    "In general, the errors are fluctuating between 500K and 6M, which serves as a good first impression."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3) Feature selection based on feature importance\n",
    "Next, we are looking at the respective features to find out which ones have the most explanatory power and which ones should not be considered or could be transformed for further analysis. First we are looking at the features using SelectKBest from the sklearn package. As an evaluation criteria we are using the F-value for each respective feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Specs</th>\n",
       "      <th>Score</th>\n",
       "      <th>Expl_percent</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>International Reputation</td>\n",
       "      <td>10926.667128</td>\n",
       "      <td>1.020489e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>Overall</td>\n",
       "      <td>9247.932429</td>\n",
       "      <td>8.637050e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75</td>\n",
       "      <td>Club_Reputation</td>\n",
       "      <td>8635.202466</td>\n",
       "      <td>8.064794e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>Potential</td>\n",
       "      <td>7144.411856</td>\n",
       "      <td>6.672479e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>54</td>\n",
       "      <td>Reactions</td>\n",
       "      <td>5942.582653</td>\n",
       "      <td>5.550038e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>66</td>\n",
       "      <td>Composure</td>\n",
       "      <td>3632.566134</td>\n",
       "      <td>3.392613e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>Real Face</td>\n",
       "      <td>3565.014584</td>\n",
       "      <td>3.329523e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>Special</td>\n",
       "      <td>2416.453120</td>\n",
       "      <td>2.256832e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>64</td>\n",
       "      <td>Vision</td>\n",
       "      <td>2126.810904</td>\n",
       "      <td>1.986322e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>81</td>\n",
       "      <td>Mentality</td>\n",
       "      <td>1937.306594</td>\n",
       "      <td>1.809336e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>44</td>\n",
       "      <td>ShortPassing</td>\n",
       "      <td>1773.834349</td>\n",
       "      <td>1.656662e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>Skill Moves</td>\n",
       "      <td>1609.357013</td>\n",
       "      <td>1.503049e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>78</td>\n",
       "      <td>Skill</td>\n",
       "      <td>1594.888134</td>\n",
       "      <td>1.489536e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50</td>\n",
       "      <td>BallControl</td>\n",
       "      <td>1545.430054</td>\n",
       "      <td>1.443345e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>77</td>\n",
       "      <td>Attacking</td>\n",
       "      <td>1515.701463</td>\n",
       "      <td>1.415580e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>49</td>\n",
       "      <td>LongPassing</td>\n",
       "      <td>1483.227399</td>\n",
       "      <td>1.385251e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>80</td>\n",
       "      <td>Power</td>\n",
       "      <td>1445.259271</td>\n",
       "      <td>1.349791e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>47</td>\n",
       "      <td>Curve</td>\n",
       "      <td>1306.580608</td>\n",
       "      <td>1.220273e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>45</td>\n",
       "      <td>Volleys</td>\n",
       "      <td>1287.712881</td>\n",
       "      <td>1.202652e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>60</td>\n",
       "      <td>LongShots</td>\n",
       "      <td>1237.448985</td>\n",
       "      <td>1.155708e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>56</td>\n",
       "      <td>ShotPower</td>\n",
       "      <td>1231.491020</td>\n",
       "      <td>1.150143e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>46</td>\n",
       "      <td>Dribbling</td>\n",
       "      <td>1177.527261</td>\n",
       "      <td>1.099744e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>48</td>\n",
       "      <td>FKAccuracy</td>\n",
       "      <td>1112.727598</td>\n",
       "      <td>1.039225e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>63</td>\n",
       "      <td>Positioning</td>\n",
       "      <td>1053.322863</td>\n",
       "      <td>9.837444e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>79</td>\n",
       "      <td>Movement</td>\n",
       "      <td>1029.623299</td>\n",
       "      <td>9.616103e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>42</td>\n",
       "      <td>Finishing</td>\n",
       "      <td>1018.239114</td>\n",
       "      <td>9.509782e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>41</td>\n",
       "      <td>Crossing</td>\n",
       "      <td>996.793642</td>\n",
       "      <td>9.309493e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>27</td>\n",
       "      <td>LCM</td>\n",
       "      <td>943.087127</td>\n",
       "      <td>8.807904e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>28</td>\n",
       "      <td>CM</td>\n",
       "      <td>943.087127</td>\n",
       "      <td>8.807904e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>29</td>\n",
       "      <td>RCM</td>\n",
       "      <td>943.087127</td>\n",
       "      <td>8.807904e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25</td>\n",
       "      <td>RAM</td>\n",
       "      <td>939.067333</td>\n",
       "      <td>8.770362e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>24</td>\n",
       "      <td>CAM</td>\n",
       "      <td>939.067333</td>\n",
       "      <td>8.770362e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>23</td>\n",
       "      <td>LAM</td>\n",
       "      <td>939.067333</td>\n",
       "      <td>8.770362e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>CF</td>\n",
       "      <td>928.109609</td>\n",
       "      <td>8.668023e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>LF</td>\n",
       "      <td>928.109609</td>\n",
       "      <td>8.668023e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>21</td>\n",
       "      <td>RF</td>\n",
       "      <td>928.109609</td>\n",
       "      <td>8.668023e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>22</td>\n",
       "      <td>RW</td>\n",
       "      <td>869.721881</td>\n",
       "      <td>8.122714e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>LW</td>\n",
       "      <td>869.721881</td>\n",
       "      <td>8.122714e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>RS</td>\n",
       "      <td>865.959134</td>\n",
       "      <td>8.087572e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>ST</td>\n",
       "      <td>865.959134</td>\n",
       "      <td>8.087572e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>LS</td>\n",
       "      <td>865.959134</td>\n",
       "      <td>8.087572e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>65</td>\n",
       "      <td>Penalties</td>\n",
       "      <td>862.551767</td>\n",
       "      <td>8.055749e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>30</td>\n",
       "      <td>RM</td>\n",
       "      <td>859.391734</td>\n",
       "      <td>8.026236e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>26</td>\n",
       "      <td>LM</td>\n",
       "      <td>859.391734</td>\n",
       "      <td>8.026236e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>Contract Valid Until</td>\n",
       "      <td>830.171904</td>\n",
       "      <td>7.753339e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>58</td>\n",
       "      <td>Stamina</td>\n",
       "      <td>697.230413</td>\n",
       "      <td>6.511741e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>34</td>\n",
       "      <td>RDM</td>\n",
       "      <td>594.490257</td>\n",
       "      <td>5.552205e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>33</td>\n",
       "      <td>CDM</td>\n",
       "      <td>594.490257</td>\n",
       "      <td>5.552205e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32</td>\n",
       "      <td>LDM</td>\n",
       "      <td>594.490257</td>\n",
       "      <td>5.552205e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>35</td>\n",
       "      <td>RWB</td>\n",
       "      <td>579.239919</td>\n",
       "      <td>5.409776e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>31</td>\n",
       "      <td>LWB</td>\n",
       "      <td>579.239919</td>\n",
       "      <td>5.409776e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>53</td>\n",
       "      <td>Agility</td>\n",
       "      <td>563.470035</td>\n",
       "      <td>5.262494e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>43</td>\n",
       "      <td>HeadingAccuracy</td>\n",
       "      <td>540.462276</td>\n",
       "      <td>5.047614e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>61</td>\n",
       "      <td>Aggression</td>\n",
       "      <td>514.281218</td>\n",
       "      <td>4.803098e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>40</td>\n",
       "      <td>RB</td>\n",
       "      <td>496.009921</td>\n",
       "      <td>4.632454e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>36</td>\n",
       "      <td>LB</td>\n",
       "      <td>496.009921</td>\n",
       "      <td>4.632454e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>52</td>\n",
       "      <td>SprintSpeed</td>\n",
       "      <td>452.855522</td>\n",
       "      <td>4.229416e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>51</td>\n",
       "      <td>Acceleration</td>\n",
       "      <td>439.705422</td>\n",
       "      <td>4.106602e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>76</td>\n",
       "      <td>Position_Potential</td>\n",
       "      <td>437.654993</td>\n",
       "      <td>4.087452e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>Weak Foot</td>\n",
       "      <td>398.802887</td>\n",
       "      <td>3.724595e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>38</td>\n",
       "      <td>CB</td>\n",
       "      <td>362.545501</td>\n",
       "      <td>3.385971e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>39</td>\n",
       "      <td>RCB</td>\n",
       "      <td>362.545501</td>\n",
       "      <td>3.385971e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>37</td>\n",
       "      <td>LCB</td>\n",
       "      <td>362.545501</td>\n",
       "      <td>3.385971e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>86</td>\n",
       "      <td>Attack_WR</td>\n",
       "      <td>356.854766</td>\n",
       "      <td>3.332823e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>84</td>\n",
       "      <td>Potential_Difference</td>\n",
       "      <td>305.796991</td>\n",
       "      <td>2.855972e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>62</td>\n",
       "      <td>Interceptions</td>\n",
       "      <td>304.515571</td>\n",
       "      <td>2.844004e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>67</td>\n",
       "      <td>Marking</td>\n",
       "      <td>278.868261</td>\n",
       "      <td>2.604473e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>59</td>\n",
       "      <td>Strength</td>\n",
       "      <td>245.357389</td>\n",
       "      <td>2.291500e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>57</td>\n",
       "      <td>Jumping</td>\n",
       "      <td>220.658733</td>\n",
       "      <td>2.060829e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>55</td>\n",
       "      <td>Balance</td>\n",
       "      <td>194.010460</td>\n",
       "      <td>1.811949e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>82</td>\n",
       "      <td>Defending</td>\n",
       "      <td>193.679636</td>\n",
       "      <td>1.808859e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>Joined</td>\n",
       "      <td>189.819486</td>\n",
       "      <td>1.772807e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>68</td>\n",
       "      <td>StandingTackle</td>\n",
       "      <td>180.733899</td>\n",
       "      <td>1.687953e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>Jersey Number</td>\n",
       "      <td>135.724439</td>\n",
       "      <td>1.267590e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>116</td>\n",
       "      <td>Continent_Asia</td>\n",
       "      <td>134.369482</td>\n",
       "      <td>1.254936e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>69</td>\n",
       "      <td>SlidingTackle</td>\n",
       "      <td>116.897428</td>\n",
       "      <td>1.091756e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>99</td>\n",
       "      <td>Position_LF</td>\n",
       "      <td>101.207693</td>\n",
       "      <td>9.452230e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>Age</td>\n",
       "      <td>72.731146</td>\n",
       "      <td>6.792681e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>87</td>\n",
       "      <td>Defense_WR</td>\n",
       "      <td>52.734450</td>\n",
       "      <td>4.925101e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>123</td>\n",
       "      <td>Position_Group_GoalKeeper</td>\n",
       "      <td>47.677265</td>\n",
       "      <td>4.452789e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>93</td>\n",
       "      <td>Position_GK</td>\n",
       "      <td>47.677265</td>\n",
       "      <td>4.452789e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>109</td>\n",
       "      <td>Position_RF</td>\n",
       "      <td>39.097459</td>\n",
       "      <td>3.651483e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>121</td>\n",
       "      <td>Position_Group_Defender</td>\n",
       "      <td>38.073964</td>\n",
       "      <td>3.555894e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>107</td>\n",
       "      <td>Position_RCM</td>\n",
       "      <td>37.209815</td>\n",
       "      <td>3.475188e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>120</td>\n",
       "      <td>Continent_South America</td>\n",
       "      <td>37.076901</td>\n",
       "      <td>3.462774e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>89</td>\n",
       "      <td>Position_CB</td>\n",
       "      <td>34.999811</td>\n",
       "      <td>3.268786e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>124</td>\n",
       "      <td>Position_Group_Midfielder</td>\n",
       "      <td>34.718032</td>\n",
       "      <td>3.242469e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>101</td>\n",
       "      <td>Position_LS</td>\n",
       "      <td>32.375791</td>\n",
       "      <td>3.023717e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>122</td>\n",
       "      <td>Position_Group_Forward</td>\n",
       "      <td>31.844387</td>\n",
       "      <td>2.974087e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>97</td>\n",
       "      <td>Position_LCM</td>\n",
       "      <td>30.475299</td>\n",
       "      <td>2.846222e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>Weight</td>\n",
       "      <td>27.955588</td>\n",
       "      <td>2.610895e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>72</td>\n",
       "      <td>GKKicking</td>\n",
       "      <td>19.583735</td>\n",
       "      <td>1.829011e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>119</td>\n",
       "      <td>Continent_Oceania</td>\n",
       "      <td>19.137204</td>\n",
       "      <td>1.787307e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>117</td>\n",
       "      <td>Continent_Europe</td>\n",
       "      <td>17.971948</td>\n",
       "      <td>1.678479e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>118</td>\n",
       "      <td>Continent_North America</td>\n",
       "      <td>17.889371</td>\n",
       "      <td>1.670767e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>83</td>\n",
       "      <td>GoalKeeping</td>\n",
       "      <td>17.838895</td>\n",
       "      <td>1.666053e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>102</td>\n",
       "      <td>Position_LW</td>\n",
       "      <td>17.732047</td>\n",
       "      <td>1.656074e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>70</td>\n",
       "      <td>GKDiving</td>\n",
       "      <td>17.698213</td>\n",
       "      <td>1.652914e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>88</td>\n",
       "      <td>Position_CAM</td>\n",
       "      <td>17.513030</td>\n",
       "      <td>1.635619e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>71</td>\n",
       "      <td>GKHandling</td>\n",
       "      <td>17.426762</td>\n",
       "      <td>1.627562e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>73</td>\n",
       "      <td>GKPositioning</td>\n",
       "      <td>16.411120</td>\n",
       "      <td>1.532706e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>95</td>\n",
       "      <td>Position_LB</td>\n",
       "      <td>16.132401</td>\n",
       "      <td>1.506676e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>74</td>\n",
       "      <td>GKReflexes</td>\n",
       "      <td>16.081189</td>\n",
       "      <td>1.501893e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>105</td>\n",
       "      <td>Position_RB</td>\n",
       "      <td>15.477208</td>\n",
       "      <td>1.445484e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>111</td>\n",
       "      <td>Position_RS</td>\n",
       "      <td>11.702836</td>\n",
       "      <td>1.092979e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>92</td>\n",
       "      <td>Position_CM</td>\n",
       "      <td>10.519173</td>\n",
       "      <td>9.824317e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>112</td>\n",
       "      <td>Position_RW</td>\n",
       "      <td>8.628299</td>\n",
       "      <td>8.058347e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>106</td>\n",
       "      <td>Position_RCB</td>\n",
       "      <td>7.659450</td>\n",
       "      <td>7.153496e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>Preferred Foot</td>\n",
       "      <td>6.779774</td>\n",
       "      <td>6.331928e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>115</td>\n",
       "      <td>Continent_Africa</td>\n",
       "      <td>6.256469</td>\n",
       "      <td>5.843191e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>96</td>\n",
       "      <td>Position_LCB</td>\n",
       "      <td>6.101768</td>\n",
       "      <td>5.698709e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>98</td>\n",
       "      <td>Position_LDM</td>\n",
       "      <td>4.716522</td>\n",
       "      <td>4.404967e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>100</td>\n",
       "      <td>Position_LM</td>\n",
       "      <td>4.209737</td>\n",
       "      <td>3.931658e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>108</td>\n",
       "      <td>Position_RDM</td>\n",
       "      <td>4.072637</td>\n",
       "      <td>3.803614e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>110</td>\n",
       "      <td>Position_RM</td>\n",
       "      <td>3.523240</td>\n",
       "      <td>3.290508e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>90</td>\n",
       "      <td>Position_CDM</td>\n",
       "      <td>2.702496</td>\n",
       "      <td>2.523980e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>104</td>\n",
       "      <td>Position_RAM</td>\n",
       "      <td>1.612176</td>\n",
       "      <td>1.505682e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>91</td>\n",
       "      <td>Position_CF</td>\n",
       "      <td>1.286187</td>\n",
       "      <td>1.201226e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>114</td>\n",
       "      <td>Position_ST</td>\n",
       "      <td>1.056556</td>\n",
       "      <td>9.867639e-04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>113</td>\n",
       "      <td>Position_RWB</td>\n",
       "      <td>0.991997</td>\n",
       "      <td>9.264695e-04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>85</td>\n",
       "      <td>Body Type</td>\n",
       "      <td>0.847246</td>\n",
       "      <td>7.912803e-04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>Loaned From</td>\n",
       "      <td>0.130068</td>\n",
       "      <td>1.214757e-04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>94</td>\n",
       "      <td>Position_LAM</td>\n",
       "      <td>0.108806</td>\n",
       "      <td>1.016187e-04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>Height</td>\n",
       "      <td>0.075696</td>\n",
       "      <td>7.069613e-05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>103</td>\n",
       "      <td>Position_LWB</td>\n",
       "      <td>0.000003</td>\n",
       "      <td>2.870708e-09</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         Specs         Score  Expl_percent\n",
       "5     International Reputation  10926.667128  1.020489e+01\n",
       "1                      Overall   9247.932429  8.637050e+00\n",
       "75             Club_Reputation   8635.202466  8.064794e+00\n",
       "2                    Potential   7144.411856  6.672479e+00\n",
       "54                   Reactions   5942.582653  5.550038e+00\n",
       "66                   Composure   3632.566134  3.392613e+00\n",
       "8                    Real Face   3565.014584  3.329523e+00\n",
       "3                      Special   2416.453120  2.256832e+00\n",
       "64                      Vision   2126.810904  1.986322e+00\n",
       "81                   Mentality   1937.306594  1.809336e+00\n",
       "44                ShortPassing   1773.834349  1.656662e+00\n",
       "7                  Skill Moves   1609.357013  1.503049e+00\n",
       "78                       Skill   1594.888134  1.489536e+00\n",
       "50                 BallControl   1545.430054  1.443345e+00\n",
       "77                   Attacking   1515.701463  1.415580e+00\n",
       "49                 LongPassing   1483.227399  1.385251e+00\n",
       "80                       Power   1445.259271  1.349791e+00\n",
       "47                       Curve   1306.580608  1.220273e+00\n",
       "45                     Volleys   1287.712881  1.202652e+00\n",
       "60                   LongShots   1237.448985  1.155708e+00\n",
       "56                   ShotPower   1231.491020  1.150143e+00\n",
       "46                   Dribbling   1177.527261  1.099744e+00\n",
       "48                  FKAccuracy   1112.727598  1.039225e+00\n",
       "63                 Positioning   1053.322863  9.837444e-01\n",
       "79                    Movement   1029.623299  9.616103e-01\n",
       "42                   Finishing   1018.239114  9.509782e-01\n",
       "41                    Crossing    996.793642  9.309493e-01\n",
       "27                         LCM    943.087127  8.807904e-01\n",
       "28                          CM    943.087127  8.807904e-01\n",
       "29                         RCM    943.087127  8.807904e-01\n",
       "25                         RAM    939.067333  8.770362e-01\n",
       "24                         CAM    939.067333  8.770362e-01\n",
       "23                         LAM    939.067333  8.770362e-01\n",
       "20                          CF    928.109609  8.668023e-01\n",
       "19                          LF    928.109609  8.668023e-01\n",
       "21                          RF    928.109609  8.668023e-01\n",
       "22                          RW    869.721881  8.122714e-01\n",
       "18                          LW    869.721881  8.122714e-01\n",
       "17                          RS    865.959134  8.087572e-01\n",
       "16                          ST    865.959134  8.087572e-01\n",
       "15                          LS    865.959134  8.087572e-01\n",
       "65                   Penalties    862.551767  8.055749e-01\n",
       "30                          RM    859.391734  8.026236e-01\n",
       "26                          LM    859.391734  8.026236e-01\n",
       "12        Contract Valid Until    830.171904  7.753339e-01\n",
       "58                     Stamina    697.230413  6.511741e-01\n",
       "34                         RDM    594.490257  5.552205e-01\n",
       "33                         CDM    594.490257  5.552205e-01\n",
       "32                         LDM    594.490257  5.552205e-01\n",
       "35                         RWB    579.239919  5.409776e-01\n",
       "31                         LWB    579.239919  5.409776e-01\n",
       "53                     Agility    563.470035  5.262494e-01\n",
       "43             HeadingAccuracy    540.462276  5.047614e-01\n",
       "61                  Aggression    514.281218  4.803098e-01\n",
       "40                          RB    496.009921  4.632454e-01\n",
       "36                          LB    496.009921  4.632454e-01\n",
       "52                 SprintSpeed    452.855522  4.229416e-01\n",
       "51                Acceleration    439.705422  4.106602e-01\n",
       "76          Position_Potential    437.654993  4.087452e-01\n",
       "6                    Weak Foot    398.802887  3.724595e-01\n",
       "38                          CB    362.545501  3.385971e-01\n",
       "39                         RCB    362.545501  3.385971e-01\n",
       "37                         LCB    362.545501  3.385971e-01\n",
       "86                   Attack_WR    356.854766  3.332823e-01\n",
       "84        Potential_Difference    305.796991  2.855972e-01\n",
       "62               Interceptions    304.515571  2.844004e-01\n",
       "67                     Marking    278.868261  2.604473e-01\n",
       "59                    Strength    245.357389  2.291500e-01\n",
       "57                     Jumping    220.658733  2.060829e-01\n",
       "55                     Balance    194.010460  1.811949e-01\n",
       "82                   Defending    193.679636  1.808859e-01\n",
       "10                      Joined    189.819486  1.772807e-01\n",
       "68              StandingTackle    180.733899  1.687953e-01\n",
       "9                Jersey Number    135.724439  1.267590e-01\n",
       "116             Continent_Asia    134.369482  1.254936e-01\n",
       "69               SlidingTackle    116.897428  1.091756e-01\n",
       "99                 Position_LF    101.207693  9.452230e-02\n",
       "0                          Age     72.731146  6.792681e-02\n",
       "87                  Defense_WR     52.734450  4.925101e-02\n",
       "123  Position_Group_GoalKeeper     47.677265  4.452789e-02\n",
       "93                 Position_GK     47.677265  4.452789e-02\n",
       "109                Position_RF     39.097459  3.651483e-02\n",
       "121    Position_Group_Defender     38.073964  3.555894e-02\n",
       "107               Position_RCM     37.209815  3.475188e-02\n",
       "120    Continent_South America     37.076901  3.462774e-02\n",
       "89                 Position_CB     34.999811  3.268786e-02\n",
       "124  Position_Group_Midfielder     34.718032  3.242469e-02\n",
       "101                Position_LS     32.375791  3.023717e-02\n",
       "122     Position_Group_Forward     31.844387  2.974087e-02\n",
       "97                Position_LCM     30.475299  2.846222e-02\n",
       "14                      Weight     27.955588  2.610895e-02\n",
       "72                   GKKicking     19.583735  1.829011e-02\n",
       "119          Continent_Oceania     19.137204  1.787307e-02\n",
       "117           Continent_Europe     17.971948  1.678479e-02\n",
       "118    Continent_North America     17.889371  1.670767e-02\n",
       "83                 GoalKeeping     17.838895  1.666053e-02\n",
       "102                Position_LW     17.732047  1.656074e-02\n",
       "70                    GKDiving     17.698213  1.652914e-02\n",
       "88                Position_CAM     17.513030  1.635619e-02\n",
       "71                  GKHandling     17.426762  1.627562e-02\n",
       "73               GKPositioning     16.411120  1.532706e-02\n",
       "95                 Position_LB     16.132401  1.506676e-02\n",
       "74                  GKReflexes     16.081189  1.501893e-02\n",
       "105                Position_RB     15.477208  1.445484e-02\n",
       "111                Position_RS     11.702836  1.092979e-02\n",
       "92                 Position_CM     10.519173  9.824317e-03\n",
       "112                Position_RW      8.628299  8.058347e-03\n",
       "106               Position_RCB      7.659450  7.153496e-03\n",
       "4               Preferred Foot      6.779774  6.331928e-03\n",
       "115           Continent_Africa      6.256469  5.843191e-03\n",
       "96                Position_LCB      6.101768  5.698709e-03\n",
       "98                Position_LDM      4.716522  4.404967e-03\n",
       "100                Position_LM      4.209737  3.931658e-03\n",
       "108               Position_RDM      4.072637  3.803614e-03\n",
       "110                Position_RM      3.523240  3.290508e-03\n",
       "90                Position_CDM      2.702496  2.523980e-03\n",
       "104               Position_RAM      1.612176  1.505682e-03\n",
       "91                 Position_CF      1.286187  1.201226e-03\n",
       "114                Position_ST      1.056556  9.867639e-04\n",
       "113               Position_RWB      0.991997  9.264695e-04\n",
       "85                   Body Type      0.847246  7.912803e-04\n",
       "11                 Loaned From      0.130068  1.214757e-04\n",
       "94                Position_LAM      0.108806  1.016187e-04\n",
       "13                      Height      0.075696  7.069613e-05\n",
       "103               Position_LWB      0.000003  2.870708e-09"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.feature_selection import SelectKBest\n",
    "from sklearn.feature_selection import f_regression\n",
    "\n",
    "X = train_predictors_val  \n",
    "y = train_target_val\n",
    "bestfeatures = SelectKBest(score_func=f_regression, k='all')\n",
    "fit = bestfeatures.fit(X,y)\n",
    "dfscores = pd.DataFrame(fit.scores_)\n",
    "dfcolumns = pd.DataFrame(X.columns) \n",
    "featureScores = pd.concat([dfcolumns,dfscores],axis=1)\n",
    "featureScores.columns = ['Specs','Score']  #naming the dataframe columns\n",
    "featureScores['Expl_percent']=(featureScores['Score']/featureScores['Score'].sum())*100\n",
    "featureScores.nlargest(125,'Score')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the F-values, we can see that International Reputation, Overall score, Club Reputation, Potential, Reactions, Real Face, Composure, Special, Vision and Mentality are the ten most important features. Height, Loaned From, Body Type and a couple of position features are among the ten least important features. \n",
    "\n",
    "To confirm our findings, we are also running an ExtraTreeRegressor and extract the feature importance. Feature importance gives back a score. A more important feature is indicated by a higher score."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py:245: FutureWarning:\n",
      "\n",
      "The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAACVoAAAVxCAYAAAB8tqXVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3W203WV57/vvzwRrEt107E3aiq0NSC0CNguYFAgJhS2VJz0SSndK030Kp8jBsaWbMijQwPB4rLvACe5ubbU2xRZbIrZQYltRsLYGAkHCSlghkSLWahFTMJ5THwIRYnKdF/OOTiYrWQ+ssLLC9zPGHPP+34/Xf6718hrXnapCkiRJkiRJkiRJkiRJkrRrL5vsACRJkiRJkiRJkiRJkiRpb2eilSRJkiRJkiRJkiRJkiSNwEQrSZIkSZIkSZIkSZIkSRqBiVaSJEmSJEmSJEmSJEmSNAITrSRJkiRJkiRJkiRJkiRpBCZaSZIkSZIkSZIkSZIkSdIITLSSJEmSJEmSJEmSJEmSpBGYaCVJkiRJkiRJkiRJkiRJIzDRSpIkSZIkSZIkSZIkSZJGMH2yA9C+7YADDqg5c+ZMdhiSJEmSJEmSJEmSJEnSsNauXfvNqpo90jwTrbRHzZkzh8HBwckOQ5IkSZIkSZIkSZIkSRpWkn8dzTyvDpQkSZIkSZIkSZIkSZKkEZho9SJIsj3JUJKNSW5JMnMce9yQ5LDWXtI3tnqiYm37rUzyxSTrkzyQZKBn7KtJNrT3GUoybyLPliRJkiRJkiRJkiRJkvZGJlq9OLZW1UBVHQE8C1w01g2q6oKqerg9Lukb2xPJTourai7wIWBp39jJ7X0GqmpCk7wkSZIkSZIkSZIkSZKkvZGJVi++VcAhAEkubVWuNia5pPXNSnJ7qya1Mcmi1r8ySSfJtcCMVk1qeRvb0r6TZGlbt6Fn7Ult/a1JHkmyPElGGe99wGvG8oJJLkwymGRw8+bNY1kqSZIkSZIkSZIkSZIk7ZWmT3YALyVJpgOnA3ckORo4HzgWCHB/kruAg4FNVXVmW7N/7x5VdWWSd1bVAM93NjAAzAUOAB5IcncbOxI4HNgE3AucANwzirBPAz7R1/e5JNuBZ6rq2P4FVbUMWAbQ6XRqFGdIkiRJkiRJkiRJkiRJezUTrV4cM5IMtfYq4CPAO4AVVfUUQJLbgAXAHcD1Sa4DPllVq8Zwznzg5qraDjzZEreOAb4DrKmqx9tZQ8Acdp9otTzJLGAacFTf2MlV9c0xxCVJkiRJkiRJkiRJkiRNaV4d+OLYWlUD7XNxVT1Lt4rV81TVo8DRwAbgmiTvGsM5u7sO8Jme9nZGTrJbDBwEfAz44BhikCRJkiRJkiRJkiRJkvY5JlpNnruBs5LMbJWjFgKrkhwIPF1VNwHX8/xqUgDbkuy3iz0XJZmWZDZwIrBmvAFW1TbgauC4JG8Y7z6SJEmSJEmSJEmSJEnSVGei1SSpqnXAjXQToe4HbqiqB4E3Amva9X5XAe8dZvky4KEky/v6VwAPAeuBfwQur6onXmCcW4H3AZe9kH0kSZIkSZIkSZIkSZKkqSxVNdkxaB/W6XRqcHBwssOQJEmSJEmSJEmSJEmShpVkbVV1RppnRStJkiRJkiRJkiRJkiRJGsH0yQ5AkyfJCuCgvu4rqurOyYhHkiRJkiRJkiRJkiRJ2ltZ0eoFSLI9yVCSjUluSTJzHHvckOSw1l7SN7Z6omJt+61M8oMyZ1W1ECjgvKoaADrAbUl+rWfN2iRHJTkvyeb2vl9Icut43leSJEmSJEmSJEmSJEmaiky0emG2VtVAVR0BPAtcNNYNquqCqnq4PS7pG5s3ATGOZDWw85y5wBd3PieZBRwMrG/jf9ne93C677voRYhPkiRJkiRJkiRJkiRJmnQmWk2cVcAhAEkubVWuNia5pPXNSnJ7kvWtf1HrX5mkk+RaYEarGLW8jW1p30mytK3b0LP2pLb+1iSPJFmeJGOM+15+mGg1D/gwMNCefx5YV1XbexckmQ7MAv59uA2TXJhkMMng5s2bxxiOJEmSJEmSJEmSJEmStPcx0WoCtMSj04ENSY4GzgeOBY4D3p7kSOA0YFNVzW0VsO7o3aOqruSHFbIW9x1xNt3kp7nAKcDSJK9uY0cClwCH0a0+dcIYw++taDUPuBt4Jsmr2vO9PXMXJRkCvg78R+DvhtuwqpZVVaeqOrNnzx5jOJIkSZIkSZIkSZIkSdLex0SrF2ZGSzwaBB4DPgLMB1ZU1VNVtQW4DVgAbABOSXJdkgVV9e0xnDMfuLmqtlfVk8BdwDFtbE1VPV5VO4AhYM5YXqCqvgq8PMlPAIfSvTrwAbqJYvPoJmLt9JdVNQD8RHuf3x7LWZIkSZIkSZIkSZIkSdJUZaLVC7OzAtVAVV1cVc8Cw17dV1WPAkfTTVC6Jsm7xnDO7q4DfKanvR2YPoZ9d7oPOAf4t6oq4PN0K2P9fGs/R5vzd8CJ4zhLkiRJkiRJkiRJkiRJmnJMtJp4dwNnJZmZZBawEFiV5EDg6aq6CbgeOGqYtduS7LeLPRclmZZkNt0EpzUTGPO9wG/RTbiiff/vwBNV9a1drJkPfHkCY5AkSZIkSZIkSZIkSZL2WuOpfqTdqKp1SW7kh4lQN1TVg0lOBZYm2QFsA94xzPJlwENJ1lXV4p7+FcDxwHqggMur6okkh44jxNuTbGvt+6rql+kmWv0+LdGqqv4tyTSee20gdJO95tNN0HscOG8c50uSJEmSJEmSJEmSJElTTrq3wEl7RqfTqcHBwckOQ5IkSZIkSZIkSZIkSRpWkrVV1RlpnlcHSpIkSZIkSZIkSZIkSdIIvDpwH5RkBXBQX/cVVXXnZMQjSZIkSZIkSZIkSZIkTXVWtHoBkmxPMpRkY5Jbkswcxx43JDmstZf0ja0eT1xVtbCqBvo+dyZZmeSLSdYneSDJQN95RyapJKf29VeSv+h5np5kc5JPjic+SZIkSZIkSZIkSZIkaaox0eqF2dqSmI4AngUuGusGVXVBVT3cHpf0jc2bgBj7La6qucCHgKV9Y+cC97TvXk8BRySZ0Z5/Efj6HohNkiRJkiRJkiRJkiRJ2iuZaDVxVgGHACS5tFW52pjkktY3K8ntrZrUxiSLWv/KJJ0k1wIzWoWs5W1sS/tOkqVt3YaetSe19bcmeSTJ8iQZZbz3Aa/Z+dDWnQOcB7w5ySv65n8aOLO1zwVu3tXGSS5MMphkcPPmzaMMR5IkSZIkSZIkSZIkSdp7mWg1AZJMB04HNiQ5GjgfOBY4Dnh7kiOB04BNVTW3VcC6o3ePqrqSH1bIWtx3xNnAADAXOAVYmuTVbexI4BLgMOBg4IRRhn0a8Ime5xOAr1TVl4GVwBl98z8O/EpLwPo54P5dbVxVy6qqU1Wd2bNnjzIcSZIkSZIkSZIkSZIkae9lotULMyPJEDAIPAZ8BJgPrKiqp6pqC3AbsADYAJyS5LokC6rq22M4Zz5wc1Vtr6ongbuAY9rYmqp6vKp2AEPAnBH2Wp7kceAK4A96+s+lm0xF+37O9YFV9VDb+1zgU2OIXZIkSZIkSZIkSZIkSZrypk92AFPc1qoa6O3Y1dV9VfVoq3Z1BnBNks9U1XtGec7urgN8pqe9nZH/pouB9cC1wAeBs5NMA34J+N+SXNXO+09JXlVV3+1Z+7fA9cBJwH8aZeySJEmSJEmSJEmSJEnSlGdFq4l3N3BWkplJZgELgVVJDgSerqqb6CYrHTXM2m1J9tvFnouSTEsyGzgRWDPeAKtqG3A1cFySN9C9jnB9Vf1UVc2pqp8G/ho4q2/pnwLvqaoN4z1bkiRJkiRJkiRJkiRJmoqsaDXBqmpdkhv5YSLUDVX1YJJTgaVJdgDbgHcMs3wZ8FCSdVW1uKd/BXA83UpUBVxeVU8kOfQFxLk1yfuAy4Bp7Yxef91i/IueNY8D7x/vmZIkSZIkSZIkSZIkSdJUlaqa7Bi0D+t0OjU4ODjZYUiSJEmSJEmSJEmSJEnDSrK2qjojzfPqQEmSJEmSJEmSJEmSJEkagVcH7oOSrAAO6uu+oqrunIx4JEmSJEmSJEmSJEmSpKnOilYvgiTbkwwl2ZjkliQzx7HHDUkOa+0lfWOre5+ramFVDfR9Rp1klWS/JNcm+VKLeU2S09vYV5NsaO8zlGTeWN9FkiRJkiRJkiRJkiRJmmpMtHpxbG3JTkcAzwIXjXWDqrqgqh5uj0v6xiY62el3gVcDR7SY3wq8qmf85J4ErtXD7iBJkiRJkiRJkiRJkiTtQ0y0evGtAg4BSHJpqxi1McklrW9WktuTrG/9i1r/yiSdJNcCM1o1qeVtbEv7TpKlbd2GnrUntfW3JnkkyfIkGS64Vm3r7cDFVfUMQFU9WVV/NdoXTHJhksEkg5s3bx7v7yRJkiRJkiRJkiRJkiTtNaZPdgAvJUmmA6cDdyQ5GjgfOBYIcH+Su4CDgU1VdWZbs3/vHlV1ZZJ3VtXAMEecDQwAc4EDgAeS3N3GjgQOBzYB9wInAPcMs8chwGNV9Z3dvMrnkmwHnqmqY/sHq2oZsAyg0+nUbvaRJEmSJEmSJEmSJEmSpgQrWr04ZiQZAgaBx4CPAPOBFVX1VFVtAW4DFgAbgFOSXJdkQVV9ewznzAdurqrtVfUkcBdwTBtbU1WPV9UOYAiY8wLeZ+fVgc9LspIkSZIkSZIkSZIkSZL2RVa0enFs7a9Ataur+6rq0Vbt6gzgmiSfqar3jPKcYfdsnulpb2fXf/t/Bl6b5FVV9d1RnitJkiRJkiRJkiRJkiTt06xoNXnuBs5KMjPJLGAhsCrJgcDTVXUTcD1w1DBrtyXZbxd7LkoyLcls4ERgzViCqqqn6Vbc+kCSlwMkeXWSXxvLPpIkSZIkSZIkSZIkSdK+xESrSVJV64Ab6SZC3Q/cUFUPAm8E1rSrBq8C3jvM8mXAQ0mW9/WvAB4C1gP/CFxeVU+MI7yrgc3Aw0k2Ap9oz5IkSZIkSZIkSZIkSdJLUqpqsmPQPqzT6dTg4OBkhyFJkiRJkiRJkiRJkiQNK8naquqMNM+KVpIkSZIkSZIkSZIkSZI0gumTHYAmT5IVwEF93VdU1Z2TEY8kSZIkSZIkSZIkSZK0t7Ki1R6QZHuSoSQbk9ySZOY49rghyWGtvaRvbPVExFlVC6tqAPgWMAMI8N4kA+2c/57kf/Wc+8dJPtvzfHGSD0xELJIkSZIkSZIkSZIkSdLezESrPWNrVQ1U1RHAs8BFY92gqi6oqofb45K+sXkTEGO/xVU1F/gQsLT1rQZ6zxoA9k8yrT3PA+7dA7FIkiRJkiRJkiRJkiRJexUTrfa8VcAhAEkubVWuNia5pPXNSnJ7kvWtf1HrX5mkk+RaYEarkLW8jW1p30mytK3b0LP2pLb+1iSPJFmeJKOM9z7gNa39IPD6JDOS7A88DQwBb2zj8+gmYz1HkguTDCYZ3Lx585h/MEmSJEmSJEmSJEmSJGlvM32yA9iXJZkOnA7ckeRo4HzgWLpX9N2f5C7gYGBTVZ3Z1uzfu0dVXZnkne2Kv35n060yNRc4AHggyd1t7EjgcGAT3apTJwD3jCLs04BPtLO/n2QIOIbu1YL3A18C5iX5BpCq+lr/BlW1DFgG0Ol0ahRnSpIkSZIkSZIkSZIkSXs1E632jBktQQm6Fa0+ArwDWFFVTwEkuQ1YANwBXJ/kOuCTVbVqDOfMB26uqu3Aky1x6xjgO8Caqnq8nTUEzGH3iVbLk8wCpgFH9fTfS7dy1Qy61a6+RPcqw80MU81KkiRJkiRJkiRJkiRJ2hd5deCesbWqBtrn4qp6lm4Vq+epqkeBo4ENwDVJ3jWGc3Z3HeAzPe3tjJxUtxg4CPgY8MGe/tV0E62Op5to9U/AYa3v3jHEKkmSJEmSJEmSJEmSJE1ZJlq9eO4Gzkoys1WOWgisSnIg8HRV3QRcz3OrSe20Lcl+u9hzUZJpSWYDJwJrxhtgVW0DrgaOS/KG1r0aOA6YXVXfqKqiW83qbVjRSpIkSZIkSZIkSZIkSS8RJlq9SKpqHXAj3USo+4EbqupB4I3Amna931XAe4dZvgx4KMnyvv4VwEPAeuAfgcur6okXGOdW4H3AZe353+kmVn2hZ9p9wI+1cyVJkiRJkiRJkiRJkqR9XroFiqQ9o9Pp1ODg4GSHIUmSJEmSJEmSJEmSJA0rydqq6ow0z4pWkiRJkiRJkiRJkiRJkjQCE60mQJLtSYaSbExyS5KZ49jjhiSHtfaSvrHVExTnihbnliRfbO1T++a8P8nXk7ysp++8JJXkTT19C1vfORMRmyRJkiRJkiRJkiRJkrQ3M9FqYmytqoGqOgJ4FrhorBtU1QVV9XB7XNI3Nm8CYqSqFlbVADAILG4x37lzvCVXLQS+BpzYt3wDcG7P868A6yciLkmSJEmSJEmSJEmSJGlvZ6LVxFsFHAKQ5NJW5Wpjkkta36wktydZ3/oXtf6VSTpJrgVmtGpTy9vYlvadJEvbug09a09q629N8kiS5UkyjthPBjYCf8Rzk6p2vtfPJ9kvySvbOw6N4wxJkiRJkiRJkiRJkiRpypk+2QHsS5JMB04H7khyNHA+cCwQ4P4kdwEHA5uq6sy2Zv/eParqyiTvbJWn+p0NDABzgQOAB5Lc3caOBA4HNgH3AicA94zxFc4Fbgb+Bvi9JPtV1badoQGfBU4F9gf+FjhoF7/DhcCFAK997WvHGIIkSZIkSZIkSZIkSZK097Gi1cSYkWSI7pV8jwEfAeYDK6rqqaraAtwGLKB7Bd8pSa5LsqCqvj2Gc+YDN1fV9qp6ErgLOKaNramqx6tqB91KU3PG8gJJXg6cAXyiqr4D3A+8uW/ax+leGfgrdBOyhlVVy6qqU1Wd2bNnjyUMSZIkSZIkSZIkSZIkaa9kRauJsbW/AtWuru6rqkdbtaszgGuSfKaq3jPKc3Z3HeAzPe3tjP1vexrdSlUbWugzgaeB23dOqKo1SY6g+76Pju92QkmSJEmSJEmSJEmSJGnqsaLVnnM3cFaSmUlmAQuBVUkOBJ6uqpuA64Gjhlm7Lcl+u9hzUZJpSWYDJwJrJijec4ELqmpOVc2hey3gm5PM7Jv3O8CSCTpTkiRJkiRJkiRJkiRJmhKsaLWHVNW6JDfyw0SoG6rqwSSnAkuT7AC2Ae8YZvky4KEk66pqcU//CuB4YD1QwOVV9USSQ8cR4u1JtrX2fcCbgP+zJ/6nktwDvLXvvT49jrMkSZIkSZIkSZIkSZKkKS1VNdkxaB/W6XRqcHBwssOQJEmSJEmSJEmSJEmShpVkbVV1Rprn1YGSJEmSJEmSJEmSJEmSNAKvDtyHJVkBHNTXfUVV3TkZ8UiSJEmSJEmSJEmSJElTlRWt9qAk25MMJdmY5JYkM8exxw1JDmvtJX1jq3e3tqoWVtVA32eXSVZJpif5vSRfanEPJbmqZ3xLT/uMNu+1Y30nSZIkSZIkSZIkSZIkaaox0WrP2tqSm44AngUuGusGVXVBVT3cHpf0jc2bgBh7vRc4EHhjVQ0AC4D9+icleRPwB8BpVfXYBMcgSZIkSZIkSZIkSZIk7XVMtHrxrAIOAUhyaatytTHJJa1vVpLbk6xv/Yta/8oknSTXAjNalanlbWxL+06SpW3dhp61J7X1tyZ5JMnyJBkuuFZt6+3AxVX1PYCq+m5Vvbtv3gLgT4Azq+rLu9jrwiSDSQY3b978An82SZIkSZIkSZIkSZIkafJNn+wAXgqSTAdOB+5IcjRwPnAsEOD+JHcBBwObqurMtmb/3j2q6sok72yVpvqdDQwAc4EDgAeS3N3GjgQOBzYB9wInAPcMs8chwGNV9d3dvMqPAH8DnFRVj+xqUlUtA5YBdDqd2s1+kiRJkiRJkiRJkiRJ0pRgRas9a0aSIWAQeAz4CDAfWFFVT1XVFuA2ulf0bQBOSXJdkgVV9e0xnDMfuLmqtlfVk8BdwDFtbE1VPV5VO4AhYM5oNkxyfque9bUkP9W6twGrgd8YQ2ySJEmSJEmSJEmSJEnSlGei1Z61taoG2ufiqnqWbhWr56mqR4Gj6SZcXZPkXWM4Z9g9m2d62tvZdRWzfwZem+RVLZ4/a9Wzvg1Ma3N2AP8FOCbJkjHEJ0mSJEmSJEmSJEmSJE1pJlq9+O4GzkoyM8ksYCGwKsmBwNNVdRNwPXDUMGu3JdlvF3suSjItyWzgRGDNWIKqqqfpVtz6wySvAEgyDXj5MPPeAixOYmUrSZIkSZIkSZIkSZIkvSTsqrqR9pCqWpfkRn6YCHVDVT2Y5FRgaZIddK/oe8cwy5cBDyVZV1WLe/pXAMcD64ECLq+qJ5IcOsbwrgJ+F9iY5LvAVuCjwKa+d/j/kpwG3J3km1X1N2M8R5IkSZIkSZIkSZIkSZpSUlWTHYP2YZ1OpwYHByc7DEmSJEmSJEmSJEmSJGlYSdZWVWekeV4dKEmSJEmSJEmSJEmSJEkj8OrAl6AkK4CD+rqvqKo7JyMeSZIkSZIkSZIkSZIkaW9nRatxSrI9yVCSjUluSTJzHHvckOSw1l7SN7Z6omJt+61M8sUk64GfBM6rqgHgz4CVO5Oskvxxks/2rLs4yQdae+c7r0+yLsm8iYxRkiRJkiRJkiRJkiRJ2luZaDV+W6tqoKqOAJ4FLhrrBlV1QVU93B6X9I3tiSSmxVU1F/gQsLT1rQZ6zxoA9k8yrT3PA+5t7Z3vPBf4HeCaPRCjJEmSJEmSJEmSJEmStNcx0WpirAIOAUhyaatytTHJJa1vVpLbWyWojUkWtf6VSTpJrgVmtGpRy9vYlvadJEvbug09a09q629N8kiS5UkyynjvA17T2g8Cr08yI8n+wNPAEPDGNj6PbjJWv/8A/Ptwmye5MMlgksHNmzePMiRJkiRJkiRJkiRJkiRp7zV9sgOY6pJMB04H7khyNHA+cCwQ4P4kdwEHA5uq6sy2Zv/eParqyiTvbFf59TubbpWpucABwANJ7m5jRwKHA5voVp06AbhnFGGfBnyinf39JEPAMcAM4H7gS8C8JN8AUlVfa+tmtLmvAF4N/OfhNq+qZcAygE6nU6OIR5IkSZIkSZIkSZIkSdqrmWg1fjuTjqBb0eojwDuAFVX1FECS24AFwB3A9UmuAz5ZVavGcM584Oaq2g482RK3jgG+A6ypqsfbWUPAHHafaLU8ySxgGnBUT/+9dCtXzaBb7epLdK8y3Mxzq1lt3ZkMluR44M+THFFVJlNJkiRJkiRJkiRJkiRpn+bVgeO3taoG2ufiqnqWbhWr56mqR4GjgQ3ANUneNYZzdncd4DM97e2MnDi3GDgI+BjwwZ7+1XQTrY6nm2j1T8Bhre/e4TaqqvvoVtiaPcKZkiRJkiRJkiRJkiRJ0pRnotXEuhs4K8nMVjlqIbAqyYHA01V1E3A9z60mtdO2JPvtYs9FSaYlmQ2cCKwZb4BVtQ24GjguyRta92rgOGB2VX2jVajaDLyN51a0+oEkh9KtjPX/jjcWSZIkSZIkSZIkSZIkaarw6sAJVFXrktzIDxOhbqiqB5OcCixNsgPYRveKwX7LgIeSrKuqxT39K+hWmloPFHB5VT3REp3GG+fWJO8DLgN+o6r+Pclm4As90+4DTmjn7tR7XWKAX29XGkqSJEmSJEmSJEmSJEn7tHSLF0l7RqfTqcHBwckOQ5IkSZIkSZIkSZIkSRpWkrVV1RlpnlcHSpIkSZIkSZIkSZIkSdIIvDpwH5NkBXBQX/cVVXXnZMQjSZIkSZIkSZIkSZIk7QtGrGiVZHuSoSQbk9ySZOZYD0lyQ5LDWntJ39jqse43wlnTk/xeki+1uIeSXDWRZ4wxnp9PsrLFsy7J7UneOM69zkvyh6397iSXtfYrkvx9kv+rqhZW1UDfxyQrSZIkSZIkSZIkSZIk6QUYzdWBW1uyzhHAs8BFYz2kqi6oqofb45K+sXlj3W8E7wUOBN5YVQPAAmC//knp2qNXJyb5ceCvgCVV9TNVdRRwDfC6CTzj5cBfA2ur6v+eqH3HcL5V0SRJkiRJkiRJkiRJkrTPG2ui0SrgEIAkl7YqVxuTXNL6ZrWKTetb/6LWvzJJJ8m1wIxWZWp5G9vSvpNkaVu3oWftSW39rUkeSbI8SYYLrlXbejtwcVV9D6CqvltV727jc5L8U5IPAeuAn0pybjtvY5Lrevba0tM+J8mNrX1jkg8nWZXk0SRv2c3v9U7go1X1g6pdVXVPVX2i7fXTSf4hyUPt+7Wt/61J7k/yYJLPtoSt4UwHPg58qaqu7In315Ksab/zHyeZ1vrfnOS+VlnrliSvbP1fTXJdW7Mmyc6/8ewkf53kgfY5ofW/O8myJJ8B/nyYv8OFSQaTDG7evHk3P48kSZIkSZIkSZIkSZI0NYw60apVLjod2JDkaOB84FjgOODtSY4ETgM2VdXcVgHrjt49WjLQzgpZi/uOOBsYAOYCpwBLk7y6jR0JXAIcBhwMnLCLMA8BHquq7+7mVX4W+POqOhLYBlwH/Od29jFJzhrhpwCYA/wCcCbw4SSv2MW8w+kmdO3KH7ZYfg5YDnyg9d8DHNdi/Dhw+S7WXw58v6ou2dmR5A3AIuCEVtFrO7A4yQHA1cAprbLWIHBpz17fqaqfbzH9r9b3fuD3q+oY4JeAG3rmHw28rap+tT+oqlpWVZ2q6syePXs3ry9JkiRJkiRJkiRJkiRNDaNJtJqRZIhuYs5jwEeA+cCKqnqqqrYAt9G9om8DcEqrjrSgqr49hljmAzdX1faqehK4Czimja2pqseragcwRDfRaURJzm9Vnb6W5Kda979W1edb+xhgZVVtrqrv0012OnEUW/9VVe2oqi8B/wIcOsp47m8Vtd7fuo4HPtbaf0H3NwD4SeDOJBuA36absDWce4Djk7y+p+9NdJOgHmh/tzfRTU47jm6i2r2t/9eBn+5Zd3PP9/GtfQrwh23+3wL/Icmr2tjfVtXW0by3JEmSJEmSJEmSJEmSNNVNH8Wcra0y0g/s6uq+qnq0Vbs6A7gmyWeq6j2jjGXYPZtnetrb2XXc/wy8Nsmr2pWBfwb8WZKNwLQ256lRnlk97f6KVTXC805fAI4C/gagqo5Ncg6wq+sGd+7zB8D/rKqz8WPyAAAgAElEQVS/TXIS8O5dzL8b+Cjw6ZbYtonuO320qn6nd2KStwJ/X1XnjnB2b/tlwPH9CVXtz9/7O0qSJEmSJEmSJEmSJEn7tFFfHdjnbuCsJDOTzAIWAquSHAg8XVU3AdfTTTLqty3JfrvYc1GSaUlm060stWYsQVXV03Qrbv3hzuv8kkwDXr6LJfcDv5DkgDbvXLqVtACeTPKGJC9r79frl5O8LMnr6FaL+uIu9v8gcF6SeT19M3vaq4Ffae3FdCtUAewPfL21f30XewNQVX8NLAXuSPKjwD8A5yT5MYAk/zHJTwOfB05Ickjrn9lXCWtRz/d9rf0Z4J07JyR5TsKdJEmSJEmSJEmSJEmS9FIxmopWz1NV65LcyA8ToW6oqgeTnAosTbID2Aa8Y5jly4CHkqyrqsU9/SvoXlm3nm5Fpcur6okko7qWr8dVwO8CG5N8F9hKt+rTJuDAvvf4tyS/A3yObiWoT1XV37ThK4FPAl8DNgKv7Fn6RboJWT8OXFRV3xsukBb/IuC6JK8BvgF8E9hZ5es3gT9N8tvAZuD81v9u4JYkX6ebIHXQ7l64qj6c5CfoXu/3ZuBq4DMtSWwb8N+q6vNJzgNuTvIjbenVwKOt/SNJ7qebfLez6tVvAh9M8hDd/5W7gYt2F4skSZIkSZIkSZIkSZK0L0rVrm6903Bagtknq+rWyY5loiT5KtCpqm9O9N6dTqcGBwcneltJkiRJkiRJkiRJkiRpQiRZW1WdkeaN9+pASZIkSZIkSZIkSZIkSXrJGNfVgXuDJCt4/pV6V1TVnXvy3Ko6b5hYTgWu6+v+SlUt3JOxTJSqmjPZMUiSJEmSJEmSJEmSJEl7sylb0aqqFlbVQN9nwpOskvxEko8n+XKSh5N8Ksnr+2K5c5hYFvbtc0mSmT3Pn0ryoxMc65wkvzrKue9P8vUku/0fSHJgkn3mmkRJkiRJkiRJkiRJkiRpPKZsotWLIUmAFcDKqnpdVR0GLAF+fBzbXQL8INGqqs6oqm9NTKQ/MAcYMdGqJVctBL4GnLi7uVW1qarOmZDoJEmSJEmSJEmSJEmSpCnKRKvdOxnYVlUf3tlRVUPAPUmWJtmYZEOSRQBJTkqyMsmtSR5JsjxdvwkcCHwuyefa3K8mOaBVofqnJH+S5AtJPpNkRpvzuiR3JFmbZFWSQ1v/jUk+kGR1kn9JsjMR6lpgQZKhJL81wnttBP4IOHdnZ5JfaGuHkjyY5FUtvo1tfE6LY137zBtu8yQXJhlMMrh58+bx/O6SJEmSJEmSJEmSJEnSXsVEq907Alg7TP/ZwAAwFzgFWJrk1W3sSLrVqw4DDgZOqKoPAJuAk6vq5GH2+xngg1V1OPAt4Jda/zLg4qo6GrgM+FDPmlcD84G30E2wArgSWNWuLvz93bzXucDNdKt1vSXJfq3/MuC/VdUAsADY2rfuG8AvVtVRwCLgA8NtXlXLqqpTVZ3Zs2fvJgxJkiRJkiRJkiRJkiRpajDRanzmAzdX1faqehK4Czimja2pqseragcwRPc6v5F8pVXKgm5i15wkrwTmAbckGQL+mG5y1U6fqKodVfUwY7jKMMnLgTPa+u8A9wNvbsP3Av+zVeD60ar6ft/y/YA/SbIBuIVuMpkkSZIkSZIkSZIkSZK0z5s+2QHs5b4AnDNMf3az5pme9nZG9xv3r5lBNwnuW6261EhrdhdPv9OA/YENSQBmAk8Dt1fVtUlup5uI9fkkpwDf61n7W8CTdCt5vaxvTJIkSZIkSZIkSZIkSdpnWdFq9/4R+JEkb9/ZkeQY4N+BRUmmJZkNnAisGWGv7wKvGu3BrdrUV5L8cjs3SeZOwBnnAhdU1ZyqmgMcBLw5ycwkr6uqDVV1HTAIHNq3dn/g31q1rv8KTBvt+0iSJEmSJEmSJEmSJElTmYlWu1FVBSwEfjHJl5N8AXg38DHgIWA93WSsy6vqiRG2WwZ8OsnnxhDCYuA3kqynW13rbSPMfwj4fpL1SX6rfzDJTOBU4PadfVX1FHAP8FbgkiQb23lbgU/3bfEh4NeTfB54PfDUGN5FkiRJkiRJkiRJkiRJmrLSzSWS9oxOp1ODg4OTHYYkSZIkSZIkSZIkSZI0rCRrq6oz0jwrWkmSJEmSJEmSJEmSJEnSCKZPdgDaM5KcClzX1/2Vqlo4GfFIkiRJkiRJkiRJkiRJU5kVrV4ESbYnGUqyMcktSWaOY48bkhzW2kv6xlb3z6+qO6tqoO8zqiSrJCuTfDHJ+iQPJBnoGfs/kmxI8lB7n7eN9V0kSZIkSZIkSZIkSZKkqcZEqxfH1pbodATwLHDRWDeoqguq6uH2uKRvbN4ExNhvcVXNBT4ELAVI8pPAVcD8qvo54DjgoT1wtiRJkiRJkiRJkiRJkrRXMdHqxbcKOAQgyaWtKtTGJJe0vllJbm/VpDYmWdT6VybpJLkWmNEqZC1vY1vad5Isbes29Kw9qa2/NckjSZYnySjjvQ94TWv/GPBdYAtAVW2pqq/0L0hyYZLBJIObN28e588kSZIkSZIkSZIkSZIk7T2mT3YALyVJpgOnA3ckORo4HzgWCHB/kruAg4FNVXVmW7N/7x5VdWWSd1bVAM93NjAAzAUOAB5IcncbOxI4HNgE3AucANwzirBPAz7R2uuBJ4GvJPkH4Laq+rv+BVW1DFgG0Ol0ahRnSJIkSZIkSZIkSZIkSXs1K1q9OGYkGQIGgceAjwDzgRVV9VRVbQFuAxYAG4BTklyXZEFVfXsM58wHbq6q7VX1JHAXcEwbW1NVj1fVDmAImDPCXsuTPA5cAfwBQFVtp5t4dQ7wKPD7Sd49hvgkSZIkSZIkSZIkSZKkKclEqxfH1qoaaJ+Lq+pZulWsnqeqHgWOpptwdU2Sd43hnN1dB/hMT3s7I1czWwwcBHwM+GBPfFVVa6rqGuBXgF8aQ3ySJEmSJEmSJEmSJEnSlGSi1eS5Gzgrycwks4CFwKokBwJPV9VNwPXAUcOs3ZZkv13suSjJtCSzgROBNeMNsKq2AVcDxyV5Q5IDk/TGMwD863j3lyRJkiRJkiRJkiRJkqaKkaoaaQ+pqnVJbuSHiVA3VNWDSU4FlibZAWwD3jHM8mXAQ0nWVdXinv4VwPHAeqCAy6vqiSSHvoA4tyZ5H3AZ8B7g+pYM9j1gM3DRePeWJEmSJEmSJEmSJEmSpopU1WTHoH1Yp9OpwcHByQ5DkiRJkiRJkiRJkiRJGlaStVXVGWmeVwdKkiRJkiRJkiRJkiRJ0gi8OvAlLMkK4KC+7iuq6s7JiEeSJEmSJEmSJEmSJEnaW014RaskP5Hk40m+nOThJJ9K8vpx7nVJkpk9z59K8qMTFy0kmZPkV0eYc1KSSvLWnr5PJjlpjGedl+TAnuevJjlglGvfn+TrSSbsb1ZVC6tqYOcH+DDw4xO1vyRJkiRJkiRJkiRJkrSvmNBEqyQBVgArq+p1VXUYsITxJ+9cAvwg0aqqzqiqb73wSJ9jDrDbRKvmceCq8R6SZBpwHnDgCFOHW/syYCHwNeDE8cYwwhnTq+rDVfXne2J/SZIkSZIkSZIkSZIkaSqb6IpWJwPbqurDOzuqaqiqVqVraZKNSTYkWQQ/qBa1MsmtSR5JsrzN/U26SUmfS/K5NverSQ5oVaj+KcmfJPlCks8kmdHmvC7JHUnWJlmV5NDWf2OSDyRZneRfkpzTQrwWWJBkKMlv7ebd1gPfTvKL/QNJ3pTkwfZef5rkR3rifVeSe4BzgQ6wvJ01oy2/OMm6tvbQ3fyuG4E/avvsPPfdST7a3v+rSc5O8v+0ve5Isl+bd3SSu9pvcmeSV7f+lUl+L8ldwH9v+13Wxg5J8tkk61t8r0vyyiT/0BPv24YLNsmFSQaTDG7evHk3P6kkSZIkSZIkSZIkSZI0NUx0otURwNpdjJ0NDABzgVOApTsTfoAj6VavOgw4GDihqj4AbAJOrqqTh9nvZ4APVtXhwLeAX2r9y4CLq+po4DLgQz1rXg3MB95CN8EK4EpgVbs+7/dHeL/3Alf3diR5BXAjsKiq3ghMB97RM+V7VTW/qm4CBoHF7aytbfybVXUU3SSqy3Zx7rnAzXSrhb1lZwJV8zrgTOBtwE3A51ocW4Ez29w/AM5pv8mfAv+jZ/2PVtUvVNX7+s5cTvf3nQvMA/4N+B6wsMV7MvC+VsXsOapqWVV1qqoze/bsXbySJEmSJEmSJEmSJEmSNHVMdKLV7swHbq6q7VX1JHAXcEwbW1NVj1fVDmCI7nV+I/lKVQ219lpgTpJX0k0KuiXJEPDHdJOrdvpEVe2oqocZx3WGVbUKIMmCnu6fbbE82p4/ynOv9/vLEba9rfcd+geTvBw4o8X+HeB+4M09Uz5dVduADcA04I7Wv6Ht97N0E+D+vv0mVwM/ubv4krwKeE1VrQCoqu9V1dNAgN9L8hDwWeA1jP9aSEmSJEmSJEmSJEmSJGnKmD7B+30BOGcXY8+rfNTjmZ72dkYXV/+aGXQTx75VVQOjWLO7eHbnfwBXAd8f5T5PjTC+M6ZdvfdpwP7AhlY8aibwNHB77/qq2pFkW1VV69/R9gvwhao6fgzx7eqdFgOzgaOraluSrwKv2MVcSZIkSZIkSZIkSZIk6f9n7/6j9C7Le9+/PyVok0DprkSFc2oD0kqBygQeRCChcZdK0bolFBspdld2kY2rapHTAo0el8V2Azuc9tStrU2DYks2WpB022KB2koSfkgYwoQERKxVKUVouo+iCRHS5Dp/PHfK08cZJjOZZGbC+7XWs+b7vX9e9zf581rXvc+Y6IpWfwe8OMk7djYkOSHJTwOrgcVJ9ksyh27Vp7WjrPdd4MBd3bxVfPpakre0vZPk2Ane4zbgP9C9AhHgYbrVtI5o779Mt1rXbu/VnAOcX1Vzq2oucBjw+iSzdnH+l4E5SU4CSLJ/kqOfb0L7jo8lObPNeXHb7yDgn1uS1euAHxvjWSRJkiRJkiRJkiRJkqRpaUITrVo1pUXAzyb5apIHgQ8CjwMrgQeA9XQTsi6pqidGWXIZ8NdJvjCGMM4FfjXJeroVtt48yvgHgH9Nsj7Je3dxj9+lXb9XVd8DzqN7XeEGupWkPjbCvGuBjyUZSjJztE1actPpPFe9iqraAtwBvGlXAq2qZ+lWGbuqfZMhutcrjuaXgfe0awLvAl4OrAA6SQbpfueHdyUGSZIkSZIkSZIkSZIkabrLczfNSROv0+nU4ODgZIchSZIkSZIkSZIkSZIkDSvJfVXVGW3cRF8dKEmSJEmSJEmSJEmSJEn7nBmTHcBUkuR04Kq+5q9V1aLJiEeSJEmSJEmSJEmSJEnS1GBFqx5VdWtVDfT9Rk2ySrI9yVCSjUluSDJrrHsnWZ7kqPa8pK/vrrGuN8petyf5vnJnSV6TZHWSLyd5uMU0K8nbk1SSn+kZu6i1nT2RsUmSJEmSJEmSJEmSJElTkYlWE2NrS8o6BngWuHCsC1TV+VX1UHtd0td38gTE+LySvAy4Abi0ql4F/CRwC3BgG7IBOKdnyluB9Xs6LkmSJEmSJEmSJEmSJGkqMNFq4q0BjgBIcnGrcrUxyUWtbXaSm5Osb+2LW/vtSTpJrgRmtgpZK1rf5vY3SZa2eRt65i5s829slahWJMkY4/414JNVdTdAdd1YVU/2nOs1SfZPckA749BwCyW5IMlgksFNmzaNMQxJkiRJkiRJkiRJkiRp6pkx2QHsS5LMAM4AbklyPHAecCIQ4J4kq4DDgcer6o1tzkG9a1TVZUneVVUDw2xxFjAAHAscDNybZHXrmwccDTwO3AmcAtwxhvCPAT75PP0FfB44HTgI+Cxw2LADq5YBywA6nU6NIQZJkiRJkiRJkiRJkiRpSrKi1cSYmWQIGAQeBa4B5gMrq2pLVW0GbgIW0L2C77QkVyVZUFVPjWGf+cD1VbW9VZpaBZzQ+tZW1WNVtYNupam5E3Kyf+9TdK8MfCtw/R5YX5IkSZIkSZIkSZIkSZqSTLSaGFuraqD93l1Vz9KtYvV9quoR4Hi6CVdXJPnAGPZ5vusAn+l53s7Yq5U92OIaUVWtpVv56uB2DkmSJEmSJEmSJEmSJOkFwUSrPWc1cGaSWUlmA4uANUkOBZ6uquuAq4Hjhpm7Lcn+I6y5OMl+SeYApwJrJyjejwC/kuTEnQ1J3pbk5X3jfgtYMkF7SpIkSZIkSZIkSZIkSdPCWKseaRdV1bok1/JcItTyqro/yenA0iQ7gG3AO4eZvgx4IMm6qjq3p30lcBKwHijgkqp6IsmR4wjx5iTb2vPdVfWWJG8Frk7yUmAH3cSum/rO9dfj2EuSJEmSJEmSJEmSJEma1lJVkx2D9mGdTqcGBwcnOwxJkiRJkiRJkiRJkiRpWEnuq6rOaOO8OlCSJEmSJEmSJEmSJEmSRmGi1W5Ksj3JUJKNSW5IMmscayxPclR7XtLXd9duxLayxdb7W5+k0zduYZKnktyf5MtJVif5+Z7+DyapJEf0tL23tY2azSdJkiRJkiRJkiRJkiRNdyZa7b6tVTVQVccAzwIXjnWBqjq/qh5qr0v6+k4eb2BVtajF9m8/4FsjDF9TVfOq6lXAe4CPJPmZnv4NwFt73s8GHkKSJEmSJEmSJEmSJEl6ATDRamKtAY4ASHJxq3K1MclFrW12kptbVamNSRa39tuTdJJcCcxsladWtL7N7W+SLG3zNvTMXdjm35jk4SQrkmR3DlFVQ8DlwLt6mv8CeHPb83DgKWDT7uwjSZIkSZIkSZIkSZIkTRczJjuAfUWSGcAZwC1JjgfOA04EAtyTZBVwOPB4Vb2xzTmod42quizJu1rlqX5nAQPAscDBwL1JVre+ecDRwOPAncApwB27eaR1wG/2vH8H+Mckx9BNuPp0O+P3SXIBcAHAK17xit0MQ5IkSZIkSZIkSZIkSZp8VrTafTOTDAGDwKPANcB8YGVVbamqzcBNwAK61++dluSqJAuq6qkx7DMfuL6qtlfVk8Aq4ITWt7aqHquqHcAQMHcCzjVcVaxP0b0+8Exg5UgTq2pZVXWqqjNnzpwJCEWSJEmSJEmSJEmSJEmaXFa02n1b+ytQjXR1X1U90qpdvQG4IsltVXX5Lu7zfNcBPtPzvJ2J+XedB3ypr+0vgaXAYFV9ZzdvKJQkSZIkSZIkSZIkSZKmDSta7RmrgTOTzEoyG1gErElyKPB0VV0HXA0cN8zcbUn2H2HNxUn2SzIHOBVYuyeCT/Jq4P8GPtrbXlVbgUuB390T+0qSJEmSJEmSJEmSJElTlRWt9oCqWpfkWp5LhFpeVfcnOR1YmmQHsA145zDTlwEPJFlXVef2tK8ETgLWAwVcUlVPJDlyHCHenGRbe76bbkLVgiT3A7OAfwbeU1V/O8zZPjWO/SRJkiRJkiRJkiRJkqRpLVU12TFoH9bpdGpwcHCyw5AkSZIkSZIkSZIkSZKGleS+quqMNs6rAyVJkiRJkiRJkiRJkiRpFF4duI9KshI4rK/50qq6dTLikSRJkiRJkiRJkiRJkqYzK1qNU5LtSYaSbExyQ5JZ41hjeZKj2vOSvr67die+qlpUVQM7f8AJwOuSfKXFvDbJGW2vryfZ0H4PJfmdJC9ufXOTVJIP9cR2cJJtST6yOzFKkiRJkiRJkiRJkiRJ04WJVuO3tSUxHQM8C1w41gWq6vyqeqi9LunrO3kCYuz1IeAQ4JgW85uAA3v6X1dVPwW8BjgcWNbT9w/Az/e8vwV4cILjkyRJkiRJkiRJkiRJkqYsE60mxhrgCIAkF7eKURuTXNTaZie5Ocn61r64td+epJPkSmBmq5C1ovVtbn+TZGmbt6Fn7sI2/8YkDydZkSTDBdeqbb0DeHdVPQNQVU9W1Z/3j62qzXSTxs5M8iOteSvwpSSd9r4Y+L65PftdkGQwyeCmTZvG8h0lSZIkSZIkSZIkSZKkKWnGZAcw3SWZAZwB3JLkeOA84EQgwD1JVtGtEPV4Vb2xzTmod42quizJu9oVf/3OAgaAY4GDgXuTrG5984CjgceBO4FTgDuGWeMI4NGq+s6unKmqvpPka8CPA0+25k8Bb03yBLC97XnoCPOX0SpidTqd2pU9JUmSJEmSJEmSJEmSpKnMilbjNzPJEDAIPApcA8wHVlbVllYZ6iZgAbABOC3JVUkWVNVTY9hnPnB9VW2vqieBVcAJrW9tVT1WVTuAIWDuhJysq7861i3AzwLnAJ+ewH0kSZIkSZIkSZIkSZKkKc9Eq/HbWlUD7ffuqnqW709OAqCqHgGOp5twdUWSD4xhn2HXbJ7ped7OyBXK/h54RZIDd2nD7ri5wCM729r57gP+L+Azu7KOJEmSJEmSJEmSJEmStK8w0WpirQbOTDIryWxgEbAmyaHA01V1HXA1cNwwc7cl2X+ENRcn2S/JHOBUYO1Ygqqqp+lW3PpwkhcBJDkkydv6xyY5APhD4C+q6lt93f8PcGlV/e+x7C9JkiRJkiRJkiRJkiRNdyNVQNI4VNW6JNfyXCLU8qq6P8npwNIkO4BtwDuHmb4MeCDJuqo6t6d9JXASsB4o4JKqeiLJkWMM7/3A7wAPJfkesAXoraz1hSShm3y3EvjQMOd7EHhwjPtKkiRJkiRJkiRJkiRJ016qarJj0D6s0+nU4ODgZIchSZIkSZIkSZIkSZIkDSvJfVXVGW2cVwdKkiRJkiRJkiRJkiRJ0ii8OnAfk2QlcFhf86VVdetkxCNJkiRJkiRJkiRJkiTtC/a5ilZJXp7kU0m+muShJJ9L8hPjWOeiJLN63j+X5IcnONa5SX5pF8bNT7I2ycPtd8FIY6tqUVUN9P3GnWSV5D8luWy88yVJkiRJkiRJkiRJkqR9wT6VaJUkwErg9qp6ZVUdBSwBXjaO5S4C/i3RqqreUFXfnphI/81c4HkTrZK8HPifwIVVdSQwH/ivSd44wbEMq6o+W1VX7o29JEmSJEmSJEmSJEmSpKlqn0q0Al4HbKuqj+1sqKoh4I4kS5NsTLIhyWKAJAuT3J7kxlYpakW63gMcCnwhyRfa2K8nObhVofpSkj9J8mCS25LMbGNemeSWJPclWZPkyNZ+bZIPJ7kryT8kObuFdyWwIMlQkveOcKZfA66tqnXtPP8CXAJc1tZ+WZKVSda338mt/W2tCtZQkj9Osl9r/6Mkgy323965STvfbydZ177RztjfnuQj7flNSe5Jcn+SzycZNoEtyQVtj8FNmzaN7V9QkiRJkiRJkiRJkiRJmoL2tUSrY4D7hmk/CxgAjgVOA5YmOaT1zaNbveoo4HDglKr6MPA48Lqqet0w6/048NGqOhr4NvALrX0Z8O6qOh74DeAPe+YcQrca1c/TTbCCbrLUmna93++PcKajhznTYGsH+DCwqqqOBY4DHkzyk8DidpYBYDtwbhv/vqrqAK8GfjrJq3vW/ZeqOg74oxZ/vzuA11bVPOBTdBO+vk9VLauqTlV15syZM8KxJEmSJEmSJEmSJEmSpOljxmQHsJfMB66vqu3Ak0lWAScA3wHWVtVjAEmG6F7nd8co632tVcqCbhLU3CQHACcDN3RvMATgxT1z/qKqdgAPjVQJagQBapj2nW3/EfjPAO18TyX5ZeB44N4Wy0zgn9v4X0xyAd1/+0PoJpg90Ppu6jnTWcPs+X8Cn25Jai8CvjaGc0iSJEmSJEmSJEmSJEnT1r6WaPUgcPYw7Rmmbadnep63s2vfpH/OTLrVwb7dKkiNNuf54un3INABPtvTdjzw0PPMCfDJqvqtf9eYHEa3UtUJVfWtJNcCPzhMjCN9h/8B/F5VfTbJQuCDu34MSZIkSZIkSZIkSZIkafra164O/DvgxUnesbMhyQnAt4DFSfZLMgc4FVg7ylrfBQ7c1Y2r6jvA15K8pe2bJMdOwB4fBd6eZKCt+xLgKuC/t/6/Bd7Z+vZL8kOt7ewkL23tP5Lkx4AfArbQrXr1MuCMXT1fcxDwT+35V8Y4V5IkSZIkSZIkSZIkSZq29qlEq6oqYBHws0m+muRBulWX/ifd6/HW003GuqSqnhhluWXAXyf5whhCOBf41STr6VaievMo4x8A/jXJ+iTvHW5AVX0TeBvwJ0keBu4CPl5Vf9mG/DrwuiQb6F75d3RVPQS8H7gtyQPA3wCHVNV64P4W28eBO8dwNuh+yxuSrAH+ZYxzJUmSJEmSJEmSJEmSpGkr3dwkac/odDo1ODg42WFIkiRJkiRJkiRJkiRJw0pyX1V1Rhu3T1W0kiRJkiRJkiRJkiRJkqQ9YcZkB6CuJKcDV/U1f62qFk1GPJIkSZIkSZIkSZIkSZKeY0WrvSDJ9iRDSTYmuSHJrP4xVXVrVQ30/Rb1rLE8yVHteUnf+ndNcLw/n+T+JOuTPJTkvyZ5XzvDUM95hpK8ZyL3liRJkiRJkiRJkiRJkqaiVNVkx7DPS7K5qg5ozyuA+6rq9yZivYmWZH/gG8BrquqxJC8G5lbVl8ezf6fTqcHBwT0RqiRJkiRJkiRJkiRJkrTbktxXVZ3RxlnRau9bAxwBkOTiVuVqY5KLWtvsJDe3alIbkyxu7bcn6SS5EpjZqkmtaH2b298kWdrmbeiZu7DNvzHJw0lWJMkI8R1I90rJ/w1QVc/0JlntiiQXJBlMMrhp06YxfyBJkiRJkiRJkiRJkiRpqpkx2QG8kCSZAZwB3JLkeOA84EQgwD1JVgGHA49X1RvbnIN616iqy5K8q6oGhtniLGAAOBY4GLg3yerWNw84GngcuBM4Bbijf4Gq+v+SfBb4RpK/Bf4KuL6qduzqOatqGbAMuhWtdnWeJEmSJEmSJEmSJEmSNFVZ0WrvmJlkCBgEHgWuAeYDK6tqS1VtBm4CFgAbgNOSXJVkQVU9NYZ95tNNitpeVU8Cq4ATWt/aqnqsJUwNAXNHWqSqzgd+BlgL/Abw8THEIEmSJEmSJEmSJEmSJO1zrGi1d2ztr0A10tV9VfVIq3b1BlUdHOEAACAASURBVOCKJLdV1eW7uM9I1wECPNPzvJ1R/u2ragOwIcmfAV8D3r6LMUiSJEmSJEmSJEmSJEn7HCtaTZ7VwJlJZiWZDSwC1iQ5FHi6qq4DrgaOG2butiT7j7Dm4iT7JZkDnEq3KtUuS3JAkoU9TQPAN8ayhiRJkiRJkiRJkiRJkrSvsaLVJKmqdUmu5blEqOVVdX+S04GlSXYA24B3DjN9GfBAknVVdW5P+0rgJGA9UMAlVfVEkiPHEFqAS5L8MbAV2ILVrCRJkiRJkiRJkiRJkvQCl6qa7Bi0D+t0OjU4ODjZYUiSJEmSJEmSJEmSJEnDSnJfVXVGG+fVgZIkSZIkSZIkSZIkSZI0Cq8OfAFLshI4rK/50qq6dTLikSRJkiRJkiRJkiRJkqYqK1rtIUm2JxlKsjHJDUlmjWON5UmOas9L+vru2t0Yq2pRVQ1U1QDwbeD83iSrJPcnGWjPM5JsSfK2nv77khy3u3FIkiRJkiRJkiRJkiRJU52JVnvO1pbEdAzwLHDhWBeoqvOr6qH2uqSv7+QJiHE0dwE79zkW+PLO9ySzgcOB9XshDkmSJEmSJEmSJEmSJGlSmWi1d6wBjgBIcnGrcrUxyUWtbXaSm5Osb+2LW/vtSTpJrgRmtgpZK1rf5vY3SZa2eRt65i5s829M8nCSFUkyxrjv5LlEq5OBjwED7f01wLqq2t4/KckFSQaTDG7atGmMW0qSJEmSJEmSJEmSJElTj4lWe1iSGcAZwIYkxwPnAScCrwXekWQe8HPA41V1bKuAdUvvGlV1Gc9VyDq3b4uz6CY/HQucBixNckjrmwdcBBxFt/rUKWMMv7ei1cnAauCZJAe29zuHm1RVy6qqU1WdOXPmjHFLSZIkSZIkSZIkSZIkaeox0WrPmZlkCBgEHgWuAeYDK6tqS1VtBm4CFgAbgNOSXJVkQVU9NYZ95gPXV9X2qnoSWAWc0PrWVtVjVbUDGALmjuUAVfV14EVJXg4cSffqwHvpJoqdTDcRS5IkSZIkSZIkSZIkSdrnzZjsAPZhW6tqoLdhpKv7quqRVu3qDcAVSW6rqst3cZ/nuw7wmZ7n7Yzv3/tu4Gzgm1VVSb5ItzLWa4AvjmM9SZIkSZIkSZIkSZIkadqxotXetRo4M8msJLOBRcCaJIcCT1fVdcDVwHHDzN2WZP8R1lycZL8kc4BTgbUTGPOdwHvpJlzR/v5n4Imq+vYE7iNJkiRJkiRJkiRJkiRNWVa02ouqal2Sa3kuEWp5Vd2f5HRgaZIdwDbgncNMXwY8kGRdVZ3b074SOAlYDxRwSVU9keTIcYR4c5Jt7fnuqnoL3USr36clWlXVN5Psh9cGSpIkSZIkSZIkSZIk6QUkVTXZMWgf1ul0anBwcLLDkCRJkiRJkiRJkiRJkoaV5L6q6ow2zqsDJUmSJEmSJEmSJEmSJGkUXh34ApNkJXBYX/OlVXXrZMQjSZIkSZIkSZIkSZIkTQdWtNoLkmxPMpRkY5IbkswaxxrLkxzVnpf09d21q+tU1aKqGuj7/bskqyS3J+n0tS1M8lQ7xwNJPp/kpWM9hyRJkiRJkiRJkiRJkjQdmWi1d2xtCU3HAM8CF451gao6v6oeaq9L+vpOnoAYd8Wado5XA/cCv7aX9pUkSZIkSZIkSZIkSZImlYlWe98a4AiAJBe3Klcbk1zU2mYnuTnJ+ta+uLXfnqST5EpgZqsstaL1bW5/k2Rpm7ehZ+7CNv/GJA8nWZEk4z1Am3sg8K0R+i9IMphkcNOmTePdRpIkSZIkSZIkSZIkSZoyZkx2AC8kSWYAZwC3JDkeOA84EQhwT5JVwOHA41X1xjbnoN41quqyJO+qqoFhtjgLGACOBQ4G7k2yuvXNA44GHgfuBE4B7hjjERYkGQJeAmyhr7JWT4zLgGUAnU6nxriHJEmSJEmSJEmSJEmSNOVY0WrvmNkSlAaBR4FrgPnAyqraUlWbgZuABcAG4LQkVyVZUFVPjWGf+cD1VbW9qp4EVgEntL61VfVYVe0AhoC54zjHzqsDfxT4BPDfx7GGJEmSJEmSJEmSJEmSNO1Y0Wrv2NpfgWqkq/uq6pFW7eoNwBVJbquqy3dxn+e7DvCZnuft7P6//WeBz+zmGpIkSZIkSZIkSZIkSdK0YEWrybMaODPJrCSzgUXAmiSHAk9X1XXA1cBxw8zdlmT/EdZcnGS/JHOAU4G1eyj++cBX99DakiRJkiRJkiRJkiRJ0pRiRatJUlXrklzLc4lQy6vq/iSnA0uT7AC2Ae8cZvoy4IEk66rq3J72lcBJwHqggEuq6okkR44jxJuTbGvPdwMfBRa0KxADPAWcP451JUmSJEmSJEmSJEmSpGknVTXZMWgf1ul0anBwcLLDkCRJkiRJkiRJkiRJkoaV5L6q6ow2zqsDJUmSJEmSJEmSJEmSJGkUXh34ApZkJXBYX/OlVXXrZMQjSZIkSZIkSZIkSZIkTVW7XdEqyfYkQ0k2JrkhyaxxrLE8yVHteUlf3127G2PfejOS/LckX2lxDyV530TuMYZYFiapJL/a0zavtf1Ge788yWkjzP2r9vziJJ9vZ1nc+z2fZ+/bgd+tqoHeH3BIko9M6EElSZIkSZIkSZIkSZKkaW4irg7c2pJ0jgGeBS4c6wJVdX5VPdRel/T1nTwBMfb6HeBQ4KdaYtECYP/+QenaG1crbgAW97y/FVi/86WqPlBVnx9ljXnA/u3f4dN933OPSmJVNEmSJEmSJEmSJEmSJO3zJjqRaA1wBECSi1uVq41JLmpts5PcnGR9a1/c2m9P0klyJTCzVWZa0fo2t79JsrTN29Azd2Gbf2OSh5OsSJLhgmvVtt4BvLuqvgdQVd+tqg+2/rlJvpTkD4F1wI8mOafttzHJVT1rbe55PjvJte352iQfS7ImySNJfn6Ub/Yo8INJXtbi/jngr3vWvjbJ2e3559oZ7wDOam0vBa4DBtp3e+XO79n6X5/k7iTrWsWxA4b5Lue1WFcBp/S0z0nymST3tt8prf2DSZYluQ3402HWuyDJYJLBTZs2jXJ8SZIkSZIkSZIkSZIkaeqbsESrVtnoDGBDkuOB84ATgdcC70gyj24S0eNVdWyrgHVL7xpVdRnPVcg6t2+Ls4AB4FjgNGBpkkNa3zzgIuAo4HB6koX6HAE8WlXffZ6jvAr406qaB2wDrgL+Y9v7hCRnjvIpAOYCPw28EfhYkh8cZfyNwFuAk+kmeD3TP6Ct8SfAm+hW4Xo5QFX9M3A+sKZ9t6/2zDkYeD9wWlUdBwwCF/etewjw23S/2c/S/YY7/QHw+1V1AvALwPKevuOBN1fVL/XHWlXLqqpTVZ05c+aMcnRJkiRJkiRJkiRJkiRp6puIRKuZSYboJvE8ClwDzAdWVtWWqtoM3EQ3OWgDcFqSq5IsqKqnxrDPfOD6qtpeVU8Cq4ATWt/aqnqsqnYAQ3QTnUbVKjkNJfnHJD/amr9RVV9szycAt1fVpqr6V2AFcOouLP3nVbWjqr4C/ANw5Gjj6SZanQNcP8KYI4GvVdVXqqroVrEazWvpJk7d2f6NfgX4sb4xJ/LcGZ8FPt3TdxrwkTb3s8APJTmw9X22qrbuQgySJEmSJEmSJEmSJEnStDdjAtbYWlUDvQ0jXd1XVY+0aldvAK5IcltVXb6L+wy7ZtNbAWo7I5/r74FXJDmwXRn4CeATSTYC+7UxW3Zxz+p57q9YVaO8//vOqieSbKNbUerX6Va2Gm3PXRHgb6rqnFHGjbTuDwAn9SdUtX/eLcPOkCRJkiRJkiRJkiRJkvZBE3Z1YJ/VwJlJZiWZDSwC1iQ5FHi6qq4DrgaOG2butiT7j7Dm4iT7JZlDt7LU2rEEVVVP06249ZGd1/kl2Q940QhT7gF+OsnBbdw5dCtpATyZ5CeT/EA7X6+3JPmBJK+ke5Xhl3chvA8Al1bV9hH6HwYOa2vSYhnNF4FTkhwB0P49fqJvzD3AwiQvad/9LT19twHv2vmSZABJkiRJkiRJkiRJkiTpBWgiKlp9n6pal+RankuEWl5V9yc5HViaZAewDXjnMNOXAQ8kWVdV5/a0rwROAtbTrcB0SasENdq1fP3eB3wI2Jjku8BW4JPA48Chfef4ZpLfAr5AtzrU56rqf7Xuy4C/Av4R2Agc0DP1y3QTsl4GXFhV3xstqKq6a5T+7yW5ALg5yb8AdwDHjDJnU5K3A9cneXFrfj/wSN8ZPwjcDXwTWMdz1b3eA3w0yQN0/6+sBi4c7SySJEmSJEmSJEmSJEnSviZVY72NTs+nJZj9VVXdONmxTAWdTqcGBwcnOwxJkiRJkiRJkiRJkiRpWEnuq6rOaOP21NWBkiRJkiRJkiRJkiRJkrTP2GcTrZKsTDLU9zt9N9bb3tbYmOSGJLOGG1dVb++vZpXk9CRfSbK1/dYnWTneWHYh1rlJNo7QvrXvm7xoHOsvTHLyxEQrSZIkSZIkSZIkSZIkTX0zJjuAPaWqFk3wkluragAgyQrgQuD3dnYmCd2rGHcME8utSf4WuKeqPrErmw23XpL9qmr7bp7jqzvPsRsWApuBu3ZzHUmSJEmSJEmSJEmSJGla2GcrWu1ha4AjWoWoLyX5Q2Ad8KNJXp/k7iTrWuWrA5KcD/wi8IGWpEWS30xyb5IHkvx2axtuvc1JLk9yD3BSkuOTrEpyX5JbkxzS5h7fKmXdDfzaWA6T5EeS/EWL5YtJXj1Se5K5dJPM3tsqYi2YgO8pSZIkSZIkSZIkSZIkTWkmWo1RkhnAGcCG1vQq4E+rah6wBXg/cFpVHQcMAhdX1XLgs8BvVtW5SV4P/DjwGmAAOD7Jqf3rVdU3gNnAxqo6EbgH+B/A2VV1PPBx4HfbvE8A76mqk0Y5wit7rg38aGv7beD+qno1sAT405Haq+rrwMeA36+qgapaM8w3uiDJYJLBTZs2jRKOJEmSJEmSJEmSJEmSNPXts1cH7gEzkwy15zXANcChwDeq6out/bXAUcCd3Zv/eBFw9zBrvb797m/vB9BNvHq0bz2A7cBn2vOrgGOAv2nr7wd8M8lBwA9X1ao27s/oJoMNZ7irA+cDvwBQVX+X5CVtzZHan1dVLQOWAXQ6nRptvCRJkiRJkiRJkiRJkjTVmWi167b2Jyi1ZKctvU3A31TVOaOsFeCKqvrjvvXm9q0H8L2q2t4z78H+qlVJfhjYnYSmDNNWz9MuSZIkSZIkSZIkSZIkvaB4deDE+iJwSpIjAJLMSvITw4y7FfgvSQ5o4/6PJC/dhfW/DMxJclKbt3+So6vq28BTSea3ceeOMe7VO+ckWQj8S1V953navwscOMY9JEmSJEmSJEmSJEmSpGnLilYTqKo2JXk7cH2SF7fm9wOP9I27LclPAne3qlibgbfRvSbw+dZ/NsnZwIfbFX4zgP8XeBA4D/h4kqfpJnKNxQeBTyR5AHga+JVR2v8SuDHJm4F3V9WaMe4nSZIkSZIkSZIkSZIkTSup8iY47TmdTqcGBwcnOwxJkiRJkiRJkiRJkiRpWEnuq6rOaOO8OlCSJEmSJEmSJEmSJEmSRmGilSRJkiRJkiRJkiRJkiSNwkSrvSTJ9iRDSTYmuSHJrHGssTzJUe15SV/fXRMVa1vv9iSdvrZZSVYk2dDOcUeSAyZyX0mSJEmSJEmSJEmSJGkqMtFq79laVQNVdQzwLHDhWBeoqvOr6qH2uqSv7+QJiHE0vw48WVU/1c7xq8C2vbCvJEmSJEmSJEmSJEmSNKlMtJoca4AjAJJc3KpDbUxyUWubneTmJOtb++LWfnuSTpIrgZmtQtaK1re5/U2SpW3ehp65C9v8G5M83CpTZYxxHwL8086XqvpyVT2zux9DkiRJkiRJkiRJkiRJmupmTHYALzRJZgBnALckOR44DzgRCHBPklXA4cDjVfXGNueg3jWq6rIk76qqgWG2OAsYAI4FDgbuTbK69c0DjgYeB+4ETgHuGEP4HwduS3I28LfAJ6vqK8Oc8QLgAoBXvOIVY1hekiRJkiRJkiRJkiRJmpqsaLX3zEwyBAwCjwLXAPOBlVW1pao2AzcBC4ANwGlJrkqyoKqeGsM+84Hrq2p7VT0JrAJOaH1rq+qxqtoBDAFzx3KAqhqimwS2FPgRuklcPznMuGVV1amqzpw5c8ayhSRJkiRJkiRJkiRJkjQlWdFq79naX4FqpKv7quqRVu3qDcAVSW6rqst3cZ/nuw6w95q/7Yzj378nIeymJDtajF8a6zqSJEmSJEmSJEmSJEnSdGJFq8m1Gjgzyawks4FFwJokhwJPV9V1wNXAccPM3ZZk/xHWXJxkvyRzgFOBtRMRbJJTkvyH9vwi4CjgGxOxtiRJkiRJkiRJkiRJkjSVWdFqElXVuiTX8lwi1PKquj/J6cDSVjFqG/DOYaYvAx5Isq6qzu1pXwmcBKwHCrikqp5IcuQ4Qrw5ybb2fDfwl8AftUpcPwDcDHxmHOtKkiRJkiRJkiRJkiRJ00qqarJj0D6s0+nU4ODgZIchSZIkSZIkSZIkSZIkDSvJfVXVGW2cVwdKkiRJkiRJkiRJkiRJ0ii8OvAFLslK4LC+5kur6tbJiEeSJEmSJEmSJEmSJEmaiqxotRuSbE8ylGRjkhuSzBrHGsuTHNWel/T13TVRsbb19k9yZZKvtJjXAsuqaqCqBoDzgGOHmVdJ/qznfUaSTUn+aiLjkyRJkiRJkiRJkiRJkqYqE612z9aWpHQM8Cxw4VgXqKrzq+qh9rqkr+/kCYix14eAQ4BjWsxvAg7s6T8HuKP97bUFOCbJzPb+s8A/TXBskiRJkiRJkiRJkiRJ0pRlotXEWQMcAZDk4lYxamOSi1rb7CQ3J1nf2he39tuTdJJcCcxsFbJWtL7N7W+SLG3zNvTMXdjm35jk4SQrkmS44Fq1rXcA766qZwCq6smq+vOdewBnA28HXp/kB/uW+Gvgje35HOD6kT5EkguSDCYZ3LRp09i+oiRJkiRJkiRJkiRJkjQFmWg1AZLMAM4ANiQ5nu4VfCcCrwXekWQe8HPA41V1bKsmdUvvGlV1Gc9VyDq3b4uzgAG61/qdBixNckjrmwdcBBwFHA6cMkKYRwCPVtV3Rug/BfhaVX0VuB14Q1//p4C3tgSsVwP3jLAOVbWsqjpV1ZkzZ85IwyRJkiRJkiRJkiRJkqRpw0Sr3TMzyRAwCDwKXAPMB1ZW1Zaq2gzcBCwANgCnJbkqyYKqemoM+8wHrq+q7VX1JLAKOKH1ra2qx6pqBzAEzB3nWc6hm0xF+/vvrg+sqgfa2ucAnxvnHpIkSZIkSZIkSZIkSdK0NGOyA5jmtlbVQG/DSFf3VdUjrdrVG4ArktxWVZfv4j7Drtk80/O8nZH/Tf8eeEWSA6vqu30x7wf8AvCfkryv7feSYcZ+FrgaWAi8ZBdjlyRJkiRJkiRJkiRJkqY9K1pNvNXAmUlmJZkNLALWJDkUeLqqrqObrHTcMHO3Jdl/hDUXJ9kvyRzgVGDtWIKqqqfpVtz6cJIXASQ5JMnb6F5HuL6qfrSq5lbVjwGfAc7sW+bjwOVVtWEse0uSJEmSJEmSJEmSJEnTnYlWE6yq1gHX0k2EugdYXlX3Az8FrG1XDb4P+J1hpi8DHkiyoq99JfAAsB74O+CSqnpiHOG9H9gEPJRkI/AX7f2ctkevzwC/1He2x6rqD8axryRJkiRJkiRJkiRJkjStpaomOwbtwzqdTg0ODk52GJIkSZIkSZIkSZIkSdKwktxXVZ3RxlnRSpIkSZIkSZIkSZIkSZJGMWOyA9DES7ISOKyv+dKqunUy4pEkSZIkSZIkSZIkSZKmOyta7WFJticZSrIxyQ1JZo1jjeVJjmrPS/r67uofX1WLqmqg77dLSVZJ9k9yZZKvtJjXJjmj9X09yYZ2ng1J3jzWs0iSJEmSJEmSJEmSJEnTkYlWe97Wluh0DPAscOFYF6iq86vqofa6pK/v5AmIsdeHgEOAY1rMbwIO7Ol/XVUNAGcDH57gvSVJkiRJkiRJkiRJkqQpyUSrvWsNcARAkotbxaiNSS5qbbOT3JxkfWtf3NpvT9JJciUws1WUWtH6Nre/SbK0zdvQM3dhm39jkoeTrEiS4YJr1bbeAby7qp4BqKonq+rPhxn+Q8C3RljngiSDSQY3bdo0/q8lSZIkSZIkSZIkSZIkTREzJjuAF4okM4AzgFuSHA+cB5wIBLgnySrgcODxqnpjm3NQ7xpVdVmSd7WKUv3OAgaAY4GDgXuTrG5984CjgceBO4FTgDuGWeMI4NGq+s7zHOULLVHrcOAXhxtQVcuAZQCdTqeeZy1JkiRJkiRJkiRJkiRpWrCi1Z43M8kQMAg8ClwDzAdWVtWWqtoM3AQsADYApyW5KsmCqnpqDPvMB66vqu1V9SSwCjih9a2tqseqagcwBMzdjfO8rl0p+FPAR5IcsBtrSZIkSZIkSZIkSZIkSdOCiVZ73taqGmi/d1fVs3SrWH2fqnoEOJ5uwtUVST4whn2GXbN5pud5OyNXMvt74BVJDhxts6r6KvAkcNQuRyhJkiRJkiRJkiRJkiRNUyZaTY7VwJlJZiWZDSwC1iQ5FHi6qq4DrgaOG2butiT7j7Dm4iT7JZkDnAqsHUtQVfU03YpbH07yIoAkhyR5W//YJC8FDgO+MZY9JEmSJEmSJEmSJEmSpOlopMpG2oOqal2Sa3kuEWp5Vd2f5HRgaZIdwDbgncNMXwY8kGRdVZ3b074SOAlYDxRwSVU9keTIMYb3fuB3gIeSfA/YAvRW1vpCku3A/sBl7ZpCSZIkSZIkSZIkSZIkaZ+WqprsGLQP63Q6NTg4ONlhSJIkSZIkSZIkSZIkScNKcl9VdUYb59WBkiRJkiRJkiRJkiRJkjQKrw58gUqyEjisr/nSqrp1MuKRJEmSJEmSJEmSJEmSprIXfEWrJC9P8qkkX03yUJLPJfmJcaxzUZJZPe+fS/LDExzr3CS/tItj/yDJPyX5gZ62Fyf5fJIh4FNVNdD7A05JctpExixJkiRJkiRJkiRJkiTtC17QiVZJAqwEbq+qV1bVUcAS4GXjWO4i4N8SrarqDVX17YmJ9N/MBUZNtGrJVYuAfwRO7emaB+zfEqs+3Tdnv6r6QFV9fgLjlSRJkiRJkiRJkiRJkvYJL+hEK+B1wLaq+tjOhqoaAu5IsjTJxiQbkiwGSLIwye1JbkzycJIV6XoPcCjwhSRfaGO/nuTgVoXqS0n+JMmDSW5LMrONeWWSW5Lcl2RNkiNb+7VJPpzkriT/kOTsFt6VwIIkQ0neO8q5NgJ/BJzT1nwpcB0w0Oa/ssX4gSR3AG9p+57dxp/Q9v//2bvzML/r8t7/zxcJQsISF6ICtUaQymoG8kV2BEVRq1XENiC/HjfMwaMiWqoUraXqcUMPP1E5NsW6AlrQ8EOtgFX2xTCBSSZwkBah/oCKIBhZIstwnz++n5Gv09kzk5lJno/rmms+n/dyv+/PJH/e1/1emWR5kq2ab7kiyfXNz/6DHZ5kSZLuJN333HPPuP5hJEmSJEmSJEmSJEmSpOlkYy+02h1YMcj464EuYCFwGHBqkm2buT1pd6/aFdgBOKCqTgfuAg6tqkMHibcT8MWq2g34DXBkM74UeHdVLQJOBM7o2LMtcCDwatoFVgAnAVc0HalOG+a7jgbOod2t69VJNq2qXwHHduy/tVn7u6o6sKq+1b85yVOAbwPvqar+v8Fa4FfAy6pqL2AxcPpgh1fV0qpqVVVr/vz5w6QpSZIkSZIkSZIkSZIkzQyzpzqBaepA4Jyq6gPuTnIZsDfwW2B5Vd0BkKSH9nV+V44Q77amUxa0C7sWJNkS2B84t32DIQCbdew5v6qeAG5KMuqrDJsiqVcB762qB5L8FHg58IMhtnx7kLEXAP9ZVdcBVNVvm9hbAF9I0gX0AX8y2rwkSZIkSZIkSZIkSZKkmWxjL7S6EXjDIOMZZKzfIx3PfYzubzhwzxza3cR+U1Vdo9gzXD4DvQKYB/Q2BVxzgYcZutDqoUHGAtQg4+8F7qbd6WsT4HdjyEuSJEmSJEmSJEmSJEmasTb2qwN/AmyW5O39A0n2Bu4HFieZlWQ+cDCwfIRYDwBbjfbgpkvUbUn+vDk3SRZOwBlHA8dW1YKqWgA8D3h5krmjzQ24Gdiu+VuQZKsks2kXcP1n02nrL4FZY4gpSZIkSZIkSZIkSZIkzVgbdaFVVRVwBPCyJLcmuRE4BTgbWAWspF2M9f6q+uUI4ZYCP0xyyRhSOAZ4W5KVtLtrvXaE9auAx5OsTPLegZNNMdXhdHSvqqqHaF9t+JrRJlVVjwKLgc83uf0I2Bw4A3hTkmtpXxs4WDcsSZIkSZIkSZIkSZIkaYOTdq2RNDlarVZ1d3dPdRqSJEmSJEmSJEmSJEnSoJKsqKrWSOs26o5WkiRJkiRJkiRJkiRJkjQas6c6AY1PksOBTw0Yvq2qjpiKfCRJkiRJkiRJkiRJkqQNmR2tJkCSviQ9SVYnOTfJ3HHEODPJrs3zyQPmrh64vqouqqquAT+jKrJKcmmSnyVZmeS6JF0dc1sm+Ycktya5McnlSfYZ8J0rk1yfZP+xfqckSZIkSZIkSZIkSZI0E1loNTHWNoVOuwOPAseNNUBVHVtVNzWvJw+Ym4yCpmOqaiFwBnBqx/iZwH3ATlW1G/BmYJtmrv87FwJ/A3xiEvKSJEmSJEmSJEmSJEmSph0LrSbeFcDzAZK8r+lytTrJCc3YFkl+0HSFWp1kcTN+aZJWkk8Cc5rOUWc1cw82v5Pk1GZfb8feQ5r95yW5OclZSTLKfK8Btm/i7AjsA3yoqp4AqKqfV9UPBtm3NXD/YAGTLEnSnaT7nnvuGWUakiRJkiRJkiRJkiRJ0vQ1e6oT2JAkmQ28ErgwySLgLbQLlwL8NMllwA7AXVX1p82eeZ0xquqkJO+qqi7+q9cDXcBC2l2mrktyeTO3J7AbcBdwecnbnQAAIABJREFUFXAAcOUo0n4FcH7zvBvQU1V9Q6ydk6QH2BzYFnjJYIuqaimwFKDVatUocpAkSZIkSZIkSZIkSZKmNTtaTYz+AqRu4BfAl4EDgWVV9VBVPQh8FzgI6AUOS/KpJAdV1ZoxnHMgcE5V9VXV3cBlwN7N3PKquqPpRNUDLBgh1llJ7gA+AHx+lOf3Xx24M+0Cra+PoXOWJEmSJEmSJEmSJEmSNGNZaDUx+guQuqrq3VX1KO0uVv9FVd0CLKJdcPWJJB8ewznDFTU90vHcx8jdyo4BngecDXyxGbsRWJhkxP8XVXUN7a5a80daK0mSJEmSJEmSJEmSJM10FlpNnsuB1yWZm2QL4AjgiiTbAQ9X1TeBzwB7DbL3sSSbDhFzcZJZSeYDBwPLx5tgVT0GfAjYN8kuVXUr7a5cf9/fqSrJTkleO3Bvkp2BWcCvx3u+JEmSJEmSJEmSJEmSNFOM1PVI41RV1yf5Kk8WQp1ZVTckORw4NckTwGPAOwbZvhRYleT6qjqmY3wZsB+wEijg/VX1y6boabx5rk3yWeBE4G3AscBngX9P8jDtQqq/bpb3X5EI7e5ab6qqvvGeLUmSJEmSJEmSJEmSJM0UqaqpzkEbsFarVd3d3VOdhiRJkiRJkiRJkiRJkjSoJCuqqjXSOq8OlCRJkiRJkiRJkiRJkqQRWGg1gyX5YJIbk6xK0pNknwHzy5rxzp/DRxn7X5I8dYQ1tyfZZl2+QZIkSZIkSZIkSZIkSZoJZk91AhqfJPsBrwb2qqpHmoKnp3Suqaojxhu/ql61jilKkiRJkiRJkiRJkiRJGww7Ws1c2wL3VtUjAFV1b1Xd1XSZ+lSS5c3P8wGSzE/ynSTXNT8HNONbJvlKkt6mM9aRzfjvu1UlOT/JiqZ71pIp+l5JkiRJkiRJkiRJkiRpylhoNXNdDDwnyS1Jzkjy4o6531bVi4AvAP9vM/Y54LSq2hs4EjizGf9bYE1V7VFVLwR+MshZb62qRUALOD7JM4ZLLMmSJN1Juu+5557xf6EkSZIkSZIkSZIkSZI0TXh14AxVVQ8mWQQcBBwKfDvJSc30OR2/T2ueDwN2TdIfYuskWzXjR3XEvX+Q445P0n8N4XOAnYBfD5PbUmApQKvVqjF+miRJkiRJkiRJkiRJkjTtWGg1g1VVH3ApcGmSXuBN/VOdy5rfmwD7VdXazhhpV14NWQyV5BDaxVj7VdXDSS4FNp+I/CVJkiRJkiRJkiRJkqSZwqsDZ6gkL0iyU8dQF/AfzfPijt/XNM8XA+/q2N81xPjTBhw1D7i/KbLaGdh3Yr5AkiRJkiRJkiRJkiRJmjkstJq5tgS+luSmJKuAXYFTmrnNkvwUeA/w3mbseKCVZFWSm4DjmvGPAU9LsjrJStrXEHa6EJjdnPFR4NpJ+yJJkiRJkiRJkiRJkiRpmkrVkLfGaQZKcjvQqqp7pzoXgFarVd3d3VOdhiRJkiRJkiRJkiRJkjSoJCuqqjXSOjtaSZIkSZIkSZIkSZIkSdIIZk91AppYVbVgqnOQJEmSJEmSJEmSJEmSNjR2tJIkSZIkSZIkSZIkSZKkEVhoNcmS9CXpSbI6yblJ5o4jxplJdm2eTx4wd/VE5drEuzTJz5KsTHJdkq6OuduT9Dbf05vktRN5tiRJkiRJkiRJkiRJkjRdWWg1+dZWVVdV7Q48Chw31gBVdWxV3dS8njxgbv8JyHGgY6pqIXAGcOqAuUOrqgt4A3D6JJwtSZIkSZIkSZIkSZIkTTsWWq1fVwDPB0jyvqbL1eokJzRjWyT5QdNNanWSxc34pUlaST4JzGk6Sp3VzD3Y/E6SU5t9vR17D2n2n5fk5iRnJcko870G2H6Iua2B+webSLIkSXeS7nvuuWeUR0mSJEmSJEmSJEmSJEnT1+ypTmBjkWQ28ErgwiSLgLcA+wABfprkMmAH4K6q+tNmz7zOGFV1UpJ3NR2lBno90AUsBLYBrktyeTO3J7AbcBdwFXAAcOUo0n4FcP6AsUuaQq0dgL8YbFNVLQWWArRarRrFOZIkSZIkSZIkSZIkSdK0ZkeryTcnSQ/QDfwC+DJwILCsqh6qqgeB7wIHAb3AYUk+leSgqlozhnMOBM6pqr6quhu4DNi7mVteVXdU1RNAD7BghFhnJbkD+ADw+QFzhzbXIO4BfCHJlmPIUZIkSZIkSZIkSZIkSZqRLLSafGurqqv5eXdVPUq7i9V/UVW3AItoF1x9IsmHx3DOcNcBPtLx3MfIncyOAZ4HnA18cbAFVXUrcDew6xhylCRJkiRJkiRJkiRJkmYkC62mxuXA65LMTbIFcARwRZLtgIer6pvAZ4C9Btn7WJJNh4i5OMmsJPOBg4Hl402wqh4DPgTsm2SXgfNJnkm7GOs/xnuGJEmSJEmSJEmSJEmSNFOM1NlIk6Cqrk/yVZ4shDqzqm5IcjhwapIngMeAdwyyfSmwKsn1VXVMx/gyYD9gJVDA+6vql0l2Xoc81yb5LHAi8LZm+JIkfcCmwEnNNYWSJEmSJEmSJEmSJEnSBi1VNdU5aAPWarWqu7t7qtOQJEmSJEmSJEmSJEmSBpVkRVW1Rlrn1YGSJEmSJEmSJEmSJEmSNAILraZAkiOSVP+1fkkWJHljx3xXkletQ/zbk2wzwpplSX6V5M4kPc1Pd5IzO9Z8Nsn7mvzWNmtuSvL1JJuOJpfeO9eM9zMkSZIkSZIkSZIkSZKkacNCq6lxNHAlcFTzvgB4Y8d8FzDuQqvRqKojgHcCV1VVF7AXEGC3jmX7A1c1z7c26/YA/gj4i8nMT5IkSZIkSZIkSZIkSZpOLLRaz5JsCRwAvI0nC60+CRzUdIz6APARYHHzvjjJi5JcneSG5vcLmlizknwmSW+SVUnePeCsOUkuTPL2IdK5inYxFbQLrFYDDyR5WpLNgF2AGzo3VFUfsBzYfl3/FpIkSZIkSZIkSZIkSdJMMXuqE9gIvQ64sKpuSXJfkr2Ak4ATq+rVAEnuBlpV9a7mfWvg4Kp6PMlhwMeBI4ElwPOAPZu5p3ecsyXwLeDrVfX1wRKpqruSPJ7kj2kXXF1Du4BqP2ANsKqqHk3y+z1JNgf2Ad4z1AcmWdLkxqyt54/xzyNJkiRJkiRJkiRJkiRNP3a0Wv+Opl0ARfP76FHsmQecm2Q1cBpPXu93GPClqnocoKru69jz/wFfGarIqkN/V6v+QqtrOt6v7li3Y5Ie4NfAL6pq1VABq2ppVbWqqjVr7rxRfJ4kSZIkSZIkSZIkSZI0vVlotR4leQbwEuDMJLcDfw0sBjLcPuCjwCVVtTvwGmDz/pBADbHnKuCV6WxHNbiraRdV7UH76sBraXe02r+J0e/WquoCng/sm+TPRogLwB7bW2glSZIkSZIkSZIkSZKkmc9Cq/XrDbSv8ntuVS2oqucAtwFPAFt1rHtgwPs84M7m+c0d4xcDxyWZDTDg6sAP0+4+dcYIOV0FvBq4r6r6mq5YT6VdbHXNwMVV9Z+0rzr8mxHiSpIkSZIkSZIkSZIkSRsMC63Wr6OBZQPGvgMcBTyeZGWS9wKXALsm6UmyGPg08IkkVwGzOvaeCfwCWJVkJfDGAbFPADZP8ulhcuoFtqHdyapzbE1V3TvEnvOBuUkOGiauJEmSJEmSJEmSJEmStMFI1VA3z0nrrtVqVXd391SnIUmSJEmSJEmSJEmSJA0qyYqqao20zo5WkiRJkiRJkiRJkiRJkjSC2VOdgCZfkmcAPx5k6qVV9ev1nY8kSZIkSZIkSZIkSZI009jRapIk6UvSk2R1knOTzB1HjDOT7No8nzxg7urRxqmqX1dV1yA/vy+ySnJpkp8lWZnkuiRdHXO3J7liwPk9SVaPdHbvnWtGm6YkSZIkSZIkSZIkSZI0bVloNXnWNsVMuwOPAseNNUBVHVtVNzWvJw+Y238CchzomKpaCJwBnDpgbqskzwFIsssknC1JkiRJkiRJkiRJkiRNWxZarR9XAM8HSPK+psvV6iQnNGNbJPlB001qdZLFzfilSVpJPgnMabpIndXMPdj8TpJTm329HXsPafafl+TmJGclySjzvQbYfsDYPwOLm+ejgXPG/+eQJEmSJEmSJEmSJEmSZhYLrSZZktnAK4HeJIuAtwD7APsCb0+yJ/AK4K6qWth0wLqwM0ZVncSTHbKOGXDE64EuYCFwGHBqkm2buT2BE4BdgR2AA0aZ9iuA8weMndecBfAa4HvDfPOSJN1Juvse9upASZIkSZIkSZIkSZIkzXwWWk2eOUl6gG7gF8CXgQOBZVX1UFU9CHwXOAjoBQ5L8qkkB1XVWKqTDgTOqaq+qrobuAzYu5lbXlV3VNUTQA+wYIRYZyW5A/gA8PkBc/cB9yc5Cvg/wMNDBamqpVXVqqrWrLnzxvApkiRJkiRJkiRJkiRJ0vRkodXk6e9A1VVV766qR4FBr+6rqluARbQLrj6R5MNjOGe46wAf6XjuA2aPEOsY4HnA2cAXB5n/djM+6msD99jeQitJkiRJkiRJkiRJkiTNfBZarV+XA69LMjfJFsARwBVJtgMerqpvAp8B9hpk72NJNh0i5uIks5LMBw4Glo83wap6DPgQsG+SXQZMLwM+DVw03viSJEmSJEmSJEmSJEnSTDRShyNNoKq6PslXebIQ6syquiHJ4cCpSZ4AHgPeMcj2pcCqJNdX1TEd48uA/YCVQAHvr6pfJtl5HfJcm+SzwInA2zrGHwA+BZAM10hLkiRJkiRJkiRJkiRJ2rCkqqY6B23AWq1WdXd3T3UakiRJkiRJkiRJkiRJ0qCSrKiq1kjrvDpQkiRJkiRJkiRJkiRJkkbg1YEbmSTLgOcNGP5AVV00FflIkiRJkiRJkiRJkiRJM8FG09EqybOTfCvJrUluSvIvSf5kHHFOSDK34/1fkjx1gnNdkOSNI6w5JMmaJD0dP4eNFLuqjqiqrgE/k1Zk1XvnmskKLUmSJEmSJEmSJEmSJK03G0WhVZIAy4BLq2rHqtoVOBl41jjCnQD8vtCqql5VVb+ZmEx/bwEwbKFV44oBBVP/OtoDktjNTJIkSZIkSZIkSZIkSRqljaLQCjgUeKyqvtQ/UFU9wJVJTk2yOklvksXw+25RlyY5L8nNSc5K2/HAdsAlSS5p1t6eZJumC9X/SfKPSW5McnGSOc2aHZNcmGRFkiuS7NyMfzXJ6UmuTvLzJG9o0vskcFDTpeq9Y/nQJo/VHe8nJjmleb40yceTXAa8J8lzk/w4yarm9x935PWlJtdbkry6GZ/V/L2ua/b89yFyWJKkO0l338N2tJIkSZIkSZIkSZIkSdLMt7F0NdodWDHI+OuBLmAhsA1wXZLLm7k9gd2Au4CrgAOq6vQk7wMOrap7B4m3E3B0Vb09yT8DRwLfBJYCx1XVvyXZBzgDeEmzZ1vgQGBn4ALgPOAk4MSqevUI33VQkp6O9yOBvhH2PLWqXgyQ5HvA16vqa0neCpwOvK5ZtwB4MbAj7cKy5wP/DVhTVXsn2Qy4KsnFVXVb5wFVtbT5ZjbbdqcaIR9JkiRJkiRJkiRJkiRp2ttYCq2GciBwTlX1AXc3nZ72Bn4LLK+qOwCaYqYFwJUjxLut6ZQF7cKuBUm2BPYHzm3fYAjAZh17zq+qJ4Cbkoz1KsMrBhZjJVkwwp5vdzzvR7vYDOAbwKc75v65yevfkvycdiHYy4EXdnTemke7uOwPCq0kSZIkSZIkSZIkSZKkDc3GUmh1I/CGQcYzyFi/Rzqe+xjd32rgnjm0r2f8TVV1jWLPcPmM1uP84ZWQmw+Yf2iYvTXEc/97gHdX1UWjTWaP7eeNdqkkSZIkSZIkSZIkSZI0bW0y8pINwk+AzZK8vX8gyd7A/cDiJLOSzAcOBpaPEOsBYKvRHlxVvwVuS/LnzblJsnAizxjgbuCZSZ7RXO833PWDVwNHNc/H8Icdu/48ySZJdgR2AH4GXAS8I8mmAEn+JMkW48xTkiRJkiRJkiRJkiRJmjE2ikKrqirgCOBlSW5NciNwCnA2sApYSbsY6/1V9csRwi0FfpjkkjGkcAzwtiQraXfXeu0I61cBjydZmeS9w6w7KElPx88bquox4CPAT4HvAzcPs/944C1JVgF/CbynY+5nwGXAD4Hjqup3wJnATcD1SVYD/8DG0xVNkiRJkiRJkiRJkiRJG7G0a5CkJyX5KvD9qjpvXWO1Wq3q7u5e96QkSZIkSZIkSZIkSZKkSZBkRVW1Rlq3UXS0kiRJkiRJkiRJkiRJkqR1YaHVBEnywSQ3JlnVXOO3T5Lbk2yzDjG7knyo42rAXyR5PMlvktyU5O0TmP+ZSXYFqKo3T0Q3K0mSJEmSJEmSJEmSJGlDMXuqE9gQJNkPeDWwV1U90hRXPWUdY84GuoBnV1VXM/ZmoFVV70ryTODGJBdU1d3r9gVQVceua4zB9N65ZjLCSpIkSZIkSZIkSZIkSeuVHa0mxrbAvVX1CEBV3VtVdzVz705yfZLeJDsDJHl6kvOb7lfXJnlhM35KkqVJLga+DnwEWNx0s1rceWBV/Qq4FXhukhcluTrJDc3vFzTxdkuyvNm/KslOSbZI8oMkK5Os7o+b5NIkreb5wST/s1lzbZJnNeM7Nu/XJflIkgcn+e8qSZIkSZIkSZIkSZIkTQsWWk2Mi4HnJLklyRlJXtwxd29V7QX8b+DEZuzvgRuq6oXAybSLqvotAl5bVW8EPgx8u6q6qurbnQcm2QHYAfh34Gbg4Kras9nz8WbZccDnmo5YLeAO4BXAXVW1sKp2By4c5Hu2AK6tqoXA5UD/FYWfa+LtDdw1yL7+3JYk6U7S3fewHa0kSZIkSZIkSZIkSZI081loNQGq6kHaBVJLgHuAbzfX/AF8t/m9AljQPB8IfKPZ+xPgGUnmNXMXVNXaYY5bnKQHOAf471V1HzAPODfJauA0YLdm7TXAyUk+ADy3idsLHJbkU0kOqqrBKqEeBb4/SN77Aec2z2cPlWBVLa2qVlW1Zs2dN9QySZIkSZIkSZIkSZIkacaw0GqCVFVfVV1aVX8HvAs4spl6pPndB8xunjNYiOb3QyMc1d/hap+qWtaMfRS4pOlQ9Rpg8yans4E/A9YCFyV5SVXdQrsorBf4RJIPD3LGY1XVn09n3pIkSZIkSZIkSZIkSdJGyUKrCZDkBUl26hjqAv5jmC2XA8c0ew+hfb3gbwdZ9wCw1ShSmAfc2Ty/uSOvHYCfV9XpwAXAC5NsBzxcVd8EPgPsNYr4/a7lyQKyo0azYY/t7WglSZIkSZIkSZIkSZKkmc9Cq4mxJfC1JDclWQXsCpwyzPpTgFaz9pPAm4ZYdwmwa5KeJIuHifdp2t2prgJmdYwvBlY3Vw3uDHwd2ANY3ox9EPjYSB/X4QTgfUmWA9sCg107KEmSJEmSJEmSJEmSJG1w8uQNcdLwkswF1lZVJTkKOLqqXjvcnlarVd3d3esnQUmSJEmSJEmSJEmSJGmMkqyoqtZI62avj2S0wVgEfCFJgN8Ab53ifCRJkiRJkiRJkiRJkqT1wkIrjVpVXQEsnOo8JEmSJEmSJEmSJEmSpPVtk6lOQNNLkr4kPUlWJ/lekqc24wuSrG3mVia5OskLRorXe+eayU9akiRJkiRJkiRJkiRJmmQWWmmgtVXVVVW7A/cB7+yYu7WZWwh8DTh5SjKUJEmSJEmSJEmSJEmS1jMLrTSca4Dth5jbGrh/PeYiSZIkSZIkSZIkSZIkTZnZU52Apqcks4CXAl/uGN4xSQ+wFTAX2GeIvUuAJQCztp4/yZlKkiRJkiRJkiRJkiRJk8+OVhpoTlNM9Wvg6cCPOub6rw7cETgBWDpYgKpaWlWtqmrNmjtv8jOWJEmSJEmSJEmSJEmSJpmFVhpobVV1Ac8FngK8c4h1FwAHr7esJEmSJEmSJEmSJEmSpClkoZUGVVVrgOOBE5NsOsiSA4FbR4qzx/Z2tJIkSZIkSZIkSZIkSdLMN3uqE9D0VVU3JFkJHAVcAezYXCsY4FHg2KnMT5IkSZIkSZIkSZIkSVpfLLTSH6iqLQe8v6bjdc56TkeSJEmSJEmSJEmSJEmaFrw6UJIkSZIkSZIkSZIkSZJGYKGVJEmSJEmSJEmSJEmSJI3AQqtpIElfkp4kq5N8L8lT1yHW7Um2GWK8tzmnJ8n+65b16PTeuWZ9HCNJkiRJkiRJkiRJkiRNKgutpoe1VdVVVbsD9wHvnKRzDm3O6aqqqyfpDEmSJEmSJEmSJEmSJGmDY6HV9HMNsH3/S5K/TnJdklVJ/r5j/PwkK5LcmGTJeA5KsmWSHye5vul29dqOuf/WnLkyyTeasflJvtPkc12SA9bhOyVJkiRJkiRJkiRJkqQZY/ZUJ6AnJZkFvBT4cvP+cmAn4EVAgAuSHFxVlwNvrar7kswBrkvynar69QhHXJKkD3ikqvYBfgccUVW/ba4bvDbJBcCuwAeBA6rq3iRPb/Z/Djitqq5M8sfARcAug3zHEmAJwKyt56/DX0SSJEmSJEmSJEmSJEmaHiy0mh7mJOkBFgArgB814y9vfm5o3rekXXh1OXB8kiOa8ec04yMVWh1aVfd2vAf4eJKDgSdod9J6FvAS4Lz+tVV1X7P+MGDXJP37t06yVVU90HlIVS0FlgJstu1ONeLXS5IkSZIkSZIkSZIkSdOchVbTw9qq6koyD/g+8E7gdNqFUJ+oqn/oXJzkENpFT/tV1cNJLgU2H8e5xwDzgUVV9ViS25s4AQYrkNqkOXPtOM6SJEmSJEmSJEmSJEmSZqxNpjoBPamq1gDHAycm2ZT21XxvTbIlQJLtkzwTmAfc3xRZ7QzsO84j5wG/aoqsDgWe24z/GPiLJM9ozu2/OvBi4F39m5N0jXTAHtvPG2dqkiRJkiRJkiRJkiRJ0vRhodU0U1U3ACuBo6rqYuBs4JokvcB5wFbAhcDsJKuAjwLXjvO4s4BWkm7a3a1ubnK4EfifwGVJVgL/q1l/fLN+VZKbgOPGea4kSZIkSZIkSZIkSZI0o6RqsBvipInRarWqu7t7qtOQJEmSJEmSJEmSJEmSBpVkRVW1RlpnRytJkiRJkiRJkiRJkiRJGoGFVpIkSZIkSZIkSZIkSZI0AgutpliSZyU5O8nPk6xIck2SI5IckuT7Hes+luSiJJsluTRJqxlfkOTfkhyepJXk9BHOe3CQse2SnDfxXwe9d66ZjLCSJEmSJEmSJEmSJEnSejV7qhPYmCUJcD7wtap6YzP2XODPgPs71n0QOAB4VVU90t4GSf4IuAj4q6q6qFnePdY8quou4A3r8CmSJEmSJEmSJEmSJEnSBs2OVlPrJcCjVfWl/oGq+o+q+nz/e5K/Al4FvKaq1nbsfTZwMfChqrqgWfv7LlhJtkzylSS9SVYlObLz4CTbNN2z/rTpirW6GX9zku8mubDplPXpjj1vS3JL01HrH5N8YeL/JJIkSZIkSZIkSZIkSdL0Y0erqbUbcP0w8wcALwAWVdXAK/++TrvI6twh9v4tsKaq9gBI8rT+iSTPAi5o9v8oyYIBe7uAPYFHgJ8l+TzQ18TcC3gA+AmwcrCDkywBlgDM2nr+MJ8nSZIkSZIkSZIkSZIkzQx2tJpGknwxycok1zVD/w4EePkgy/8V+Mskc4cIdxjwxf6Xquq/inBT4MfA+6vqR0Ps/XFVramq3wE3Ac8FXgRcVlX3VdVjwFAFXlTV0qpqVVVr1tx5Qy2TJEmSJEmSJEmSJEmSZgwLrabWjbQ7RAFQVe8EXgr0t4G6m/a1gaclOXTA3k8DPwXOTTJYZ7IANcj448AK4PBh8nqk47mPduezDLNekiRJkiRJkiRJkiRJ2qBZaDW1fgJsnuQdHWN/0KGqqm4BXg98M0nXgP3vBX4LfDnJwEKoi4F39b90XB1YwFuBnZOcNIZclwMvTvK0prDryNFs2mN7O1pJkiRJkiRJkiRJkiRp5rPQagpVVQGvo13AdFuS5cDXgA8MWHcd8BbggiQ7Dtj/JmBb2h2uOn0MeFqS1UlWAod27OsDjgIOTfI/RpnrncDHaXfR+lfaVwquGcPnSpIkSZIkSZIkSZIkSTNW2rU60siSbFlVDzYdrZYB/1RVy4bb02q1qru7e/0kKEmSJEmSJEmSJEmSJI1RkhVV1RppnR2tNBanJOkBVgO3AedPcT6SJEmSJEmSJEmSJEnSejF7qhPQzFFVJ051DpIkSZIkSZIkSZIkSdJUsKOVJlXvnWumOgVJkiRJkiRJkiRJkiRpnVlotQ6S9CXpSbI6yblJ5o4jxplJdm2eTx4wd/VE5drEuzTJz5KsTHJdkq6OuS2T/EOSW5PcmOTyJPs0c5XkGx1rZye5J8n3JzI/SZIkSZIkSZIkSZIkabqy0GrdrK2qrqraHXgUOG6sAarq2Kq6qXk9ecDc/hOQ40DHVNVC4Azg1I7xM4H7gJ2qajfgzcA2zdxDwO5J5jTvLwPunITcJEmSJEmSJEmSJEmSpGnJQquJcwXwfIAk72u6XK1OckIztkWSHzTdpFYnWdyMX5qkleSTwJymQ9ZZzdyDze8kObXZ19ux95Bm/3lJbk5yVpKMMt9rgO2bODsC+wAfqqonAKrq51X1g471PwT+tHk+GjhnqMBJliTpTtLd97BXB0qSJEmSJEmSJEmSJGnms9BqAiSZDbwS6E2yCHgL7cKlfYG3J9kTeAVwV1UtbDpgXdgZo6pO4skOWccMOOL1QBewEDgMODXJts3cnsAJwK7ADsABo0z7FcD5zfNuQE9V9Q2z/lvAUUk2B14I/HSohVW1tKpaVdWaNXfeKNORJEmSJEmSJEmSJEmSpi8LrdbNnCQ9QDfwC+DLwIHAsqp6qKoeBL4LHAT0Aocl+VSSg6pqLK2eDgTOqaq+qrobuAzYu5lbXlV3NJ2oeoAFI8Q6K8kdwAeAz482gapa1cQ+GviX0e7bY3sLrSRJkiRJkiRJkiRJkjTzWWi1bvoPx3BsAAAgAElEQVQ7UHVV1bur6lFg0Kv7quoWYBHtgqtPJPnwGM4Z7jrARzqe+4DZI8Q6BngecDbwxWbsRmBhkpH+P1wAfIZhrg2UJEmSJEmSJEmSJEmSNkQWWk28y4HXJZmbZAvgCOCKJNsBD1fVN2kXK+01yN7Hkmw6RMzFSWYlmQ8cDCwfb4JV9RjwIWDfJLtU1a20u3L9fZIAJNkpyWsHbP0n4CNV1TvesyVJkiRJkiRJkiRJkqSZyEKrCVZV1wNfpV0I9VPgzKq6AdgDWN5cNfhB4GODbF8KrEpy1oDxZcAqYCXwE+D9VfXLdcxzLfBZ4MRm6Fjg2cC/J+kF/hG4a8CeO6rqc+tyriRJkiRJkiRJkiRJkjQTpaqmOgdtwFqtVnV3d091GpIkSZIkSZIkSZIkSdKgkqyoqtZI6+xoJUmSJEmSJEmSJEmSJEkjsNBqCiTpS9KTZGWS65PsP8b9X03yhmHmlzXxe5L8JsnaJHcmWdMxPqYzx6v3zjXr4xhJkiRJkiRJkiRJkiRpUs2e6gQ2UmurqgsgyeHAJ4AXT1Twqjpi4FiSQ4ATq+rVE3WOJEmSJEmSJEmSJEmStLGwo9XU2xq4HyBtpyZZnaQ3yeKO8S8kuSnJD4BnNuMvTbKsP1CSlyX57mgOTXJ4knM73l+Z5J+TzG66YJ3WdNv6UZJnNGt2SnJRkhVJLk/yJxP3Z5AkSZIkSZIkSZIkSZKmLwutpsac5vq+m4EzgY82468HuoCFwGHAqUm2BY4AXgDsAbwd6L/27yfALknmN+9vAb4yyhx+BLywv4hqwN55wLVVtRdwDfC3zfhS4H9U1SLgb4AvDBY4yZIk3Um6+x726kBJkiRJkiRJkiRJkiTNfBZaTY21VdVVVTsDrwC+niTAgcA5VdVXVXcDlwF7Awd3jN9Fu8CKqirgG8D/k+SpwH7AD0eTQFU9AZwNvDHJ04FFwMXN9ONAf7erbwIHNvH3Bb6TpAf4IrDdELGXVlWrqlqz5s4bw59FkiRJkiRJkiRJkiRJmp5mT3UCG7uquibJNsB8IMMtHWL8K8D3gN8B51bV42M4/p+A7zTP366qviSzBzmrmtzuraquMcRnj+0ttJIkSZIkSZIkSZIkSdLMZ0erKZZkZ2AW8GvgcmBxklnNdYAHA8ub8aOa8W2BQ/v3Nx2u7gI+BHx1LGdX1f8P3AucNGDvprSvMQR4I3BlVd0P/GeSI5q8N0mycGxfK0mSJEmSJEmSJEmSJM1MdrSaGnOa6/eg3SnqTU03qWW0r/9bSbuL1Pur6pfN+EuAXuAW2lcKdjoLmF9VN40jl7OBravqlo6xNcBeSU4G7gMWN+NHAf87ySnAU2hfK7hyHGdKkiRJkiRJkiRJkiRJM0qqhrqRTjNFki8AN1TVl8ex90vANVX1teZ9Nu0rAp86Ebm1Wq3q7u6eiFCSJEmSJEmSJEmSJEnShEuyoqpaI62zo9UMl2QF8BDwV+PY2wPcDxw/0XlJkiRJkiRJkiRJkiRJG5JNpjoBrZuqWgTsVVWP9I8leXPT5WpISf4M+FZVHVpVj3bEe7y/m1WSQ5J8f4j9JySZOzFfIUmSJEmSJEmSJEmSJE1vFlptpKrqgqr65DqEOAEYsdCq984163CEJEmSJEmSJEmSJEmSND1YaLWBSzI/yXeSXNf8HNCM/77rVZIdk1zbzH8kyYMdIbZMcl6Sm5Oclbbjge2AS5JcMgWfJUmSJEmSJEmSJEmSJK1Xs6c6AU2IOUl6Ot6fDlzQPH8OOK2qrkzyx8BFwC4D9n8O+FxVnZPkuAFzewK7AXcBVwEHVNXpSd4HHFpV9070x0iSJEmSJEmSJEmSJEnTjYVWG4a1VdXV/5LkzUCreT0M2DVJ//TWSbYasH8/4HXN89nAZzrmllfVHU3cHmABcOVwySRZAiwBmLX1/DF+iiRJkiRJkiRJkiRJkjT9WGi14dsE2K+q1nYOdhRejeSRjuc+RvF/pqqWAksBNtt2pxrtQZIkSZIkSZIkSZIkSdJ0tclUJ6BJdzHwrv6XJF2DrLkWOLJ5PmqUcR8ABnbG+i/22H7eKMNJkiRJkiRJkiRJkiRJ05eFVhu+44FWklVJbgKOG2TNCcD7kiwHtgXWjCLuUuCHSS6ZuFQlSZIkSZIkSZIkSZKk6SlV3uy2sUsyF1hbVZXkKODoqnrtRMRutVrV3d09EaEkSZIkSZIkSZIkSZKkCZdkRVW1Rlo3e30ko2lvEfCFJAF+A7x1ivORJEmSJEmSJEmSJEmSphULrURVXQEsnOo8JEmSJEmSJEmSJEmSpOlqk6lOQOtXkgcHvD81ya+bblYk2S9JJfmj5n1ekvuSbJLkq0luS9KT5OYkfzcV3yBJkiRJkiRJkiRJkiStbxZabeSq6jfAL4FdmqH9gRua3wD7Aj+tqiea97+uqi6gC3hTkucNF7/3zjUTn7QkSZIkSZIkSZIkSZK0nlloJYCreLKwan/gtAHvVw+yZ/Pm90OTm5okSZIkSZIkSZIkSZI09Sy0ErQLqfoLq3YAzgVazfv+tAux+p2apAe4A/hWVf1qYLAkS5J0J+nue9iOVpIkSZIkSZIkSZIkSZr5LLQSNB2tmmsAb6+q3wFJsiWwCFjesbb/6sBnAy9Nsv/AYFW1tKpaVdWaNXfe+shfkiRJkiRJkiRJkiRJmlQWWomq+jfgacBrgGua4RXAW4Db/i97dx6td1Xe/f/9IUEggMGBUqBqVFAEggEOAhIQKg44PIJT4Ie/n9BqHqyW8lhEl1Ox+iiK1VWcakQFC0IEwSoqYpXIPJyEJCcM0qpUCRVxiobEgOH6/XHvlNvDOTkn4YzJ+7XWve7vvfe1976+J/nzWteuqpUDrFkJLABmj1GakiRJkiRJkiRJkiRJ0rix0ErrXA/8HQ8XWl0PnELnWsFHSDIVOBD40fo2nbmrHa0kSZIkSZIkSZIkSZI0+VlotfmZluTurs9b2/i1wJOA3vb7euBpPLLQ6swki4GlQB9wyVgkLUmSJEmSJEmSJEmSJI2nqeOdgMZWVQ1YXFdVZwJndv2+C0i/mBNGMzdJkiRJkiRJkiRJkiRporKjlSRJkiRJkiRJkiRJkiQNwUIrSZIkSZIkSZIkSZIkSRqChVYjIMnKCZDDCUk+Ocj4fUkWt8+XxiM/SZIkSZIkSZIkSZIkaTKbOt4JaEzMr6q3DDaZZGpV/XE0Du5bvmI0tpUkSZIkSZIkSZIkSZLGlB2tRkmSpyT5XpKl7fvJbfzlSW5MckuSf0+yUxs/PckXkixI8uMkJ3ft9bokN7WOVJ9NMqWNn5jkziQ/AA7ZwPwWJPlgW/t368n3nCSfSXJly+t5Lc/bk5wzQn8uSZIkSZIkSZIkSZIkaUKz0Gr0fBL4UlXtA5wPnNXGrwEOqqp9gQuB07rW7AG8CHgO8A9JtkzyLGAOcEhVzQLWAscn2Rl4H50CqxcAe64nlzldVwee2DW+Q1U9r6r+aT35AjwO+Evg/wDfAD4O7AXMTDKr/2FJ5ibpTdK7dpUdrSRJkiRJkiRJkiRJkjT5eXXg6DkYeGV7/lfgI+35L4D5rVDqMcBPutZ8s6rWAGuS/ALYCXg+sD9wcxKAbYBfAAcCC6rqPoAk84FnDJLLYFcHzh9GvgDfqKpK0gfcW1V97cxbgRnA4u5Nq2oeMA9gq513r0FykiRJkiRJkiRJkiRJkiYNO1qNnXUFR58APllVM4H/DWzdFbOm63ktnUK4AOdW1az2eWZVnd5vz411/zDy7c7roX45PoTFepIkSZIkSZIkSZIkSdoMWGg1eq4Djm3Px9O5MhBgOrC8Pb9+GPt8D3h1kj8DSPL4JE8BbgQOT/KEJFsCrxmlfB+VmbtOH4ltJEmSJEmSJEmSJEmSpHFlN6KRMS3J3V2/PwacDHwhyduA+4AT29zpwEVJlgM3AE9d38ZVdVuSdwNXJNkCeBB4c1XdkOR04Hrgv4FFwJRH8Q6D5StJkiRJkiRJkiRJkiRt9lL1aG+fkwbX09NTvb29452GJEmSJEmSJEmSJEmSNKAkC6uqZ6g4rw6UJEmSJEmSJEmSJEmSpCFYaCVJkiRJkiRJkiRJkiRJQ7DQSqOqb/mK8U5BkiRJkiRJkiRJkiRJetQstNpISdYmWZxkWZKLkkzbiD3OTrJne35nv7nrRirXtt+CJD9MsiTJzUlmdc3dlaSvfW5L8oEkW7W5GUkqyfu74p+Y5MEknxzJHCVJkiRJkiRJkiRJkqSJykKrjbe6qmZV1d7AA8BJG7pBVb2hqm5rP9/Zb+65I5Bjf8dX1bOBTwNn9ps7oqpmAs8BngbM65r7MfCyrt+vAW4dhfwkSZIkSZIkSZIkSZKkCclCq5FxNbAbQJK3ti5Xy5Kc0sa2TfLN1k1qWZI5bXxBkp4kZwDbtA5Z57e5le07Sc5s6/q61h7e1l+c5I4k5yfJMPO9Hth1oImqWkmnaOzoJI9vw6uB25P0tN9zgK8MtnmSuUl6k/SuXeXVgZIkSZIkSZIkSZIkSZr8po53ApNdkqnAUcDlSfYHTgQOBALcmOQHdDpE3VNVL21rpnfvUVXvSPKWqprFI70SmAU8G3gicHOSq9rcvsBewD3AtcAhwDXDSPvFwNcGm6yq3yX5CbA7cG8bvhA4NsnPgbXtzF0GWT+P1hFrq513r2HkI0mSJEmSJEmSJEmSJE1odrTaeNskWQz0Aj8FPg/MBi6tqvtbZ6hLgEOBPuDIJB9OcmhVbUibp9nABVW1tqruBX4AHNDmbqqqu6vqIWAxMGOIvc5PcjfwduATQ8T27451OfAC4Dhg/gbkL0mSJEmSJEmSJEmSJE16FlptvNVVNat9/raqHuCRxUkAVNWdwP50Cq4+lOS9G3DO+q4DXNP1vJahO5QdDzwV+DLwqUEPTLanU7R157qx9n4Lgb8HvjrEOf9j5q7Thw6SJEmSJEmSJEmSJEmSJjgLrUbWVcDRSaYl2RY4Brg6yS7Aqqo6D/gosN8Aax9MsuUge85JMiXJjsBhwE0bm2BVPQi8GzgoybP6zyfZDvg08LWq+k2/6X8C3l5Vv9rY8yVJkiRJkiRJkiRJkqTJaKgOSNoAVbUoyTk8XAh1dlXdkuRFwJlJHgIeBN40wPJ5wNIki6rq+K7xS4GDgSVAAadV1c+T7PEo8lyd5J+AU4G/bsNXJgmd4rtLgfcPsO5W4NaNPVeSJEmSJEmSJEmSJEmarFJV452DNmE9PT3V29s73mlIkiRJkiRJkiRJkiRJA0qysKp6horz6kBJkiRJkiRJkiRJkiRJGoKFVuMgSSX5167fU5Pcl+SyEdj70iSL+31e9Gj37XfGrCQvGU5s3/IVI3m0JEmSJEmSJEmSJEmSNC6mjncCm6n7gb2TbFNVq4EXAMtHYuOqOmYk9hnCLKAH+NYYnCVJkiRJkiRJkiRJkiSNOztajZ9vAy9tz8cBF6ybSPL4JF9LsjTJDUn2SbJFkruS7NAV959JdkqyY5KvJrm5fQ5p86cnOTfJFW3tK5N8JElfksuTbNni9k/ygyQLk3wnyc5tfEGSDye5KcmdSQ5N8hjgH4E5rVvWnDH6e0mSJEmSJEmSJEmSJEnjxkKr8XMhcGySrYF9gBu75t4H3FJV+wDvBL5UVQ8B/wYcA5DkQOCuqroX+Gfg41V1APAq4OyuvZ5Op6DrFcB5wJVVNRNYDby0FVt9Anh1Ve0PfAH4v13rp1bVc4BTgH+oqgeA9wLzq2pWVc3v/2JJ5ibpTdK7dpVXB0qSJEmSJEmSJEmSJGny8+rAcVJVS5PMoNPNqv8VfLPpFExRVd9P8oQk04H5dIqcvggc234DHAnsmWTd+scm2b49f7uqHkzSB0wBLm/jfcAM4JnA3sB32/opwH935XJJ+17Y4ofzbvOAeQBb7bx7DWeNJEmSJEmSJEmSJEmSNJFZaDW+vg58FDgceELXeAaILeB6YLckOwJHAx9oc1sAB1fV6u4FrXBqDUBVPZTkwapaV/j0EJ1//wC3VtXBg+S4pn2vxf8vkiRJkiRJkiRJkiRJ2kx5deD4+gLwj1XV12/8KuB4gCSHA7+sqt+1IqlLgY8Bt1fVr1r8FcBb1i1OMmsDcvghsGOSg9vaLZPsNcSa3wPbDxEDwMxdp29AKpIkSZIkSZIkSZIkSdLEZKHVOKqqu6vqnweYOh3oSbIUOAN4fdfcfOB1PHxtIMDJ6+KT3AactAE5PAC8GvhwkiXAYuC5Qyy7ks5VhYuTzBnuWZIkSZIkSZIkSZIkSdJklYdvkpNGXk9PT/X29o53GpIkSZIkSZIkSZIkSdKAkiysqp6h4uxoJUmSJEmSJEmSJEmSJElDsNBKkiRJkiRJkiRJkiRJkoZgoZVGVd/yFeOdgiRJkiRJkiRJkiRJkvSoWWg1QpKsTbI4ybIkFyWZthF7nJ1kz/b8zn5z141Urm2/BUkecbdkkuckuSrJD5Pc0XKaluSEJPe1d7w1ycUb846SJEmSJEmSJEmSJEnSZGSh1chZXVWzqmpv4AHgpA3doKreUFW3tZ/v7Df33BHIcb2S7ARcBLy9qp4JPAu4HNi+hcxv77gXnXecM9o5SZIkSZIkSZIkSZIkSROBhVaj42pgN4Akb21drpYlOaWNbZvkm0mWtPE5bXxBkp4kZwDbtO5R57e5le07Sc5s6/q61h7e1l/cOlGdnyQbmPebgXOr6nqA6ri4qu7tDkoyFdgW+M1AmySZm6Q3Se/aVV4dKEmSJEmSJEmSJEmSpMlv6ngnsKlpRUhHAZcn2R84ETgQCHBjkh8ATwPuqaqXtjXTu/eoqnckeUtVzRrgiFcCs4BnA08Ebk5yVZvbF9gLuAe4FjgEuGYD0t8bOHc983OSzAZ2Bu4EvjFQUFXNA+YBbLXz7rUB50uSJEmSJEmSJEmSJEkTkh2tRs42SRYDvcBPgc8Ds4FLq+r+qloJXAIcCvQBRyb5cJJDq2pD2j7NBi6oqrWt09QPgAPa3E1VdXdVPQQsBmaMyJs9bH4r/vpzOu/wthHeX5IkSZIkSZIkSZIkSZqQLLQaOauralb7/G1VPUCni9UjVNWdwP50ipU+lOS9G3DO+q4DXNP1vJYN71h2a8trvaqq6HSzOmyo2Jm7Th8qRJIkSZIkSZIkSZIkSZrwLLQaXVcBRyeZlmRb4Bjg6iS7AKuq6jzgo8B+A6x9MMmWg+w5J8mUJDvSKXa6aYTy/STw+iQHrhtI8rokfz5A7GzgRyN0riRJkiRJkiRJkiRJkjShbWjHI22AqlqU5BweLoQ6u6puSfIi4MwkDwEPAm8aYPk8YGmSRVV1fNf4pcDBwBKggNOq6udJ9tiIFL+Z5MH2fH1VvSbJscBHk/wZ8BCdwq5LWsycJLPpFOjdDZywEWdKkiRJkiRJkiRJkiRJk046t8BJo6Onp6d6e3vHOw1JkiRJkiRJkiRJkiRpQEkWVlXPUHFeHShJkiRJkiRJkiRJkiRJQ7DQagQlWTnK+y9I8sMkS5Jcm+SZw1hzaZLF/T4v6hezS5KLRyPnvuUrRmNbSZIkSZIkSZIkSZIkaUxNHe8EtMGOr6reJHOBM4H/tb7gqjpmqA2r6h7g1SOUnyRJkiRJkiRJkiRJkrTJsaPVKEvylCTfS7K0fT+5jZ+T5Kwk1yX5cZJXt/Etknw6ya1JLkvyrXVz/VwF7NbWvDfJzUmWJZmXJG385CS3tbMvbGPP6+psdUuS7ZPMSLKszZ+Q5JIklyf5jyQf6XqXv05yZ+us9bkknxzdv54kSZIkSZIkSZIkSZI0MVhoNfo+CXypqvYBzgfO6prbGZgNvAw4o429EpgBzATeABw8yL4vB/rWnVFVB1TV3sA2bT+AdwD7trNPamOnAm+uqlnAocDqAfaeBcxpOcxJ8qQkuwDvAQ4CXgDsMdgLJ5mbpDdJ79pVXh0oSZIkSZIkSZIkSZKkyc9Cq9F3MPDl9vyvdAqr1vlaVT1UVbcBO7Wx2cBFbfznwJX99js/yWLgEDpFUwBHJLkxSR/wl8BebXxpi38d8Mc2di3wsSQnAztU1R95pO9V1Yqq+gNwG/AU4DnAD6rq11X1IHDRYC9cVfOqqqeqeqZMmz74X0aSJEmSJEmSJEmSJEmaJCy0GnvV9bym6zn9vgdzfFXNqqqjq+pnSbYGPg28uqpmAp8Dtm6xLwU+BewPLEwytarOoNMpaxvghiQDdabqzmstMHUYeQ1o5q4WWkmSJEmSJEmSJEmSJGnys9Bq9F0HHNuejweuGSL+GuBVSbZIshNw+BDx64qqfplkO+DVAEm2AJ5UVVcCpwE7ANsleXpV9VXVh4Fe1nMFYD83Ac9L8rgkU4FXDXOdJEmSJEmSJEmSJEmSNOlNHe8ENjHTktzd9ftjwMnAF5K8DbgPOHGIPb4KPB9YBtwJ3AisGCy4qn6b5HNAH3AXcHObmgKcl2Q6nW5UH2+x709yBJ1OVbcB3wZ2HurFqmp5kg+2fO5pawfNS5IkSZIkSZIkSZIkSdqUpKqGjtKYSrJdVa1M8gQ6naQOqaqfT6C8pgKXAl+oqkvXt6anp6d6e3vHJkFJkiRJkiRJkiRJkiRpAyVZWFU9Q8XZ0WpiuizJDsBjgPdPhCKr5vQkR9K5rvAK4GvjnI8kSZIkSZIkSZIkSZI0Jiy0moCq6vDxzmEgVXXqeOcgSZIkSZIkSZIkSZIkjYctxjuBsZJkZb/fJyT55AjtvSBJT3v+VutG9Wj2++cky5NM+n+fvuUrxjsFSZIkSZIkSZIkSZIk6VGb9IU8E01VvaSqfrux61tx1THAz4DDRiyxR56TTaGQS5IkSZIkSZIkSZIkSRoLFtoASXZM8tUkN7fPIW38OUmuS3JL+35mG98myYVJliaZD2zTtdddSZ6YZEaS25N8LsmtSa5Isk2LOaCtvT7JmUmWdaVzBLAM+AxwXNe+2yX5YpK+tvZVbfzFSRYlWZLke23s9CSndq1d1vJZl9OngUXAk5J8Jklvy/F9XWsOaO+8JMlNSbZPcnWSWV0x1ybZZwT/KSRJkiRJkiRJkiRJkqQJaep4JzCGtkmyuOv344Gvt+d/Bj5eVdckeTLwHeBZwB3AYVX1xyRHAh8EXgW8CVhVVfu0QqNFg5y5O3BcVb0xyVfa2vOALwJzq+q6JGf0W3MccAHwb8AHk2xZVQ8C7wFWVNVMgCSPS7Ij8LmW40+SPH4Yf4dnAidW1d+0fd5VVb9OMgX4XnufO4D5wJyqujnJY4HVwNnACcApSZ4BbFVVS/sfkGQuMBdgymN3HEZKkiRJkiRJkiRJkiRJ0sS2ORVara6q7m5MJwA97eeRwJ5J1k0/Nsn2wHTg3CS7AwVs2eYPA84CqKqlSR5RbNT8pKrWFXctBGYk2QHYvqqua+NfBl7WcnoM8BLg/1TV75PcCLwQ+GbL8dh1G1fVb5K8HLiqqn7Sxn49jL/Df1XVDV2/X9sKo6YCOwN7tnf976q6ue37u5bfRcB7krwN+CvgnIEOqKp5wDyArXbevYaRkyRJkiRJkiRJkiRJkjShbU6FVuuzBXBwVa3uHkzyCeDKqjomyQxgQdf0cAqI1nQ9r6VzxWAGiQV4MZ3irr5W9DUNWEWn0CoDnDnQGMAf+dNrIbfuer7/fxYnTwVOBQ5ohVvntNgB962qVUm+C7wCeC0PF6oNauau04cKkSRJkiRJkiRJkiRJkia8LYYO2SxcAbxl3Y8k6zpfTQeWt+cTuuKvAo5vsXsD+wz3oKr6DfD7JAe1oWO7po8D3lBVM6pqBvBU4IVJpg2Q4+OA64HntYIpuq4OvAvYr43t1/YZyGPpFF6tSLITcFQbvwPYJckBbY/tk6wryjubTjevm4fZQUuSJEmSJEmSJEmSJEma9Cy06jgZ6EmyNMltwElt/CPAh5JcC0zpiv8MsF27MvA04KYNPO+vgXlJrqfTPWpFK6Z6EZ3uVQBU1f3ANcDLgQ8Aj0uyLMkS4Iiqug+YC1zSxua3pV8FHp9kMfAm4M6BkqiqJcAtwK3AF4Br2/gDwBzgE23f79K6YlXVQuB3wBc38J0lSZIkSZIkSZIkSZKkSStVw7kBTyMpyXZVtbI9vwPYuar+bpzTGpYku9C5QnGPqnpoqPienp7q7e0d9bwkSZIkSZIkSZIkSZKkjZFkYVX1DBVnR6vx8dIki5MsAw6l061qwkvy/wE3Au8aTpGVJEmSJEmSJEmSJEmStKmw0OpRSrJyGDGntKsBAaiq+VU1q6r2rqqXtisARzPHPzk/ybeS7LCh+1TVl6rqSVV10chmKEmSJEmSJEmSJEmSJE1sFlqNjVOAaUNGdUkyZbTOr6qXVNVvR3D/QfUtXzEWx0iSJEmSJEmSJEmSJEmjykKrEZLk8CQLklyc5I4k56fjZGAX4MokV7bYFya5PsmiJBcl2a6N35XkvUmuAV6TZLck/55kSYt9eot7W5KbkyxN8r42NqOde24bvzjJtEHOvyvJE9vzW5Msa59Tuva6Pcnnktya5Iok27S5k5Pc1s64cEz/yJIkSZIkSZIkSZIkSdI4sdBqZO1Lp3vUnsDTgEOq6izgHuCIqjqiFTi9GziyqvYDeoG3du3xh6qaXVUXAucDn6qqZwPPBf47yQuB3YHnALOA/ZMc1tY+E5hXVfsAvwP+pv/53ckm2R84ETgQOAh4Y5J92/Tu7ey9gN8Cr2rj7wD2bWecNNAfIcncJL1JeteusqOVJEmSJEmSJEmSJEmSJj8LrUbWTVV1d1U9BCwGZgwQcxCdQqxrkywGXg88pWt+PkCS7YFdq+pSgKr6Q1WtAl7YPrcAi4A96BRFAfysqq5tz+cBs4fIdzZwaVXdX1UrgUuAQ9vcT6pqcXte2PUuS4Hzk7wO+ONAm1bVvKrqqaqeKdOmD5GCJEmSJEmSJEmSJEmSNPFNHe8ENjFrup7XMvDfN8B3q+q4Qfa4vytuIAE+VFWf/ZPBZLdAySkAACAASURBVAZQ/WL7/x5or8H0f5dt2vNLgcOA/wW8J8leVTVgwZUkSZIkSZIkSZIkSZK0qbCj1dj4PbB9e74BOCTJbgBJpiV5Rv8FVfU74O4kR7e4rZJMA74D/FWS7dr4rkn+rC17cpKD2/NxwDUDnN/tKuDolsO2wDHA1YO9RJItgCdV1ZXAacAOwHbre/GZu9rRSpIkSZIkSZIkSZIkSZOfhVZjYx7w7SRXVtV9wAnABUmW0im82mOQdf8vcHKLuw7486q6AvgycH2SPuBiHi6iuh14fYt/PPCZ/ud3b15Vi4BzgJuAG4Gzq+qW9bzHFOC8du4twMer6rfD/BtIkiRJkiRJkiRJkiRJk1aqhrpdTpNBuzrwsqrae5xT+RM9PT3V29s73mlIkiRJkiRJkiRJkiRJA0qysKp6hoqzo5UkSZIkSZIkSZIkSZIkDWHqeCegkVFVdwETqpuVJEmSJEmSJEmSJEmStKmw0GoTk2Qt0Efn3/Z24PVVtWp8s5IkSZIkSZIkSZIkSZImN68O3PSsrqpZVbU38ABw0mgfmGTKYHN9y1eM9vGSJEmSJEmSJEmSJEnSqLPQatN2NbAbQJK3JlnWPqe0sdOSnNyeP57k++35+UnOa88vTHJ9kkVJLkqyXRu/K8l7k1wDvGY8Xk6SJEmSJEmSJEmSJEkaKxZabaKSTAWOAvqS7A+cCBwIHAS8Mcm+wFXAoW1JD7Bdki2B2cDVSZ4IvBs4sqr2A3qBt3Yd84eqml1VF/Y7e26S3iS9a1fZ0UqSJEmSJEmSJEmSJEmTn4VWm55tkiymUxT1U+DzdAqnLq2q+6tqJXAJnQKrhcD+SbYH1gDX0ym4OpRON6yDgD2Ba9uerwee0nXW/IESqKp5VdVTVT1Tpk0fjXeUJEmSJEmSJEmSJEmSxtTU8U5AI251Vc3qHkiSgQKr6sEkd9HpdnUdsBQ4Ang6cHv7/m5VHTfIWfePVNKSJEmSJEmSJEmSJEnSRGZHq83DVcDRSaYl2RY4hk7HqnVzp7bvq4GTgMVVVcANwCFJdgNo65+xIQfP3NWOVpIkSZIkSZIkSZIkSZr8LLTaDFTVIuAc4CbgRuDsqrqlTV8N7AxcX1X3An9oY1TVfcAJwAVJltIpvNpjTJOXJEmSJEmSJEmSJEmSJoB0GhdJo6Onp6d6e3vHOw1JkiRJkiRJkiRJkiRpQEkWVlXPUHF2tJIkSZIkSZIkSZIkSZKkIVhoJUmSJEmSJEmSJEmSJElDsNBK/yPJnye5MMmPktyW5FtJnpGkkry/K+6JSR5M8snxzFeSJEmSJEmSJEmSJEkaKxZaCYAkAS4FFlTV06tqT+CdwE7Aj4GXdYW/Brh1OPv2LV8x0qlKkiRJkiRJkiRJkiRJY85CK61zBPBgVf3LuoGqWgz8DFgN3J6kp03NAb4y9ilKkiRJkiRJkiRJkiRJ48NCK62zN7BwPfMXAscm+QtgLXDPYIFJ5ibpTdK7dpUdrSRJkiRJkiRJkiRJkjT5WWil4boceAFwHDB/fYFVNa+qeqqqZ8q06WOSnCRJkiRJkiRJkiRJkjSaLLTSOrcC+w82WVUP0Ol49ffAV8cqKUmSJEmSJEmSJEmSJGkisNBK63wf2CrJG9cNJDkAeEpXzD8Bb6+qXw1305m72tFKkiRJkiRJkiRJkiRJk5+FVgKgqgo4BnhBkh8luRU4HbinK+bWqjp3nFKUJEmSJEmSJEmSJEmSxs3U8U5AE0dV3QO8doCpvQeIPQc4Z5RTkiRJkiRJkiRJkiRJkiYEO1pJkiRJkiRJkiRJkiRJ0hAstJIkSZIkSZIkSZIkSZKkIVhoNcEk2SnJl5P8OMnCJNcnOSbJ4Uku64r7QJLvJNkqyYIkP0yyNMkdST6ZZIeu2OuGOLMnyVmj8T59y1eMxraSJEmSJEmSJEmSJEnSmLLQagJJEuBrwFVV9bSq2h84FviLfnHvAg4Bjq6qNW34+KraB9gHWAP827r4qnru+s6tqt6qOnnk3kSSJEmSJEmSJEmSJEnatFhoNbH8JfBAVf3LuoGq+q+q+sS630n+HngJ8PKqWt1/g6p6ADgNeHKSZ7c1K9v3/CQv6drrnCSv6u6WleT0JF9oXbJ+nOTkrvj3tI5Z301yQZJTR/wvIEmSJEmSJEmSJEmSJE1AFlpNLHsBi9YzfwhwEnBUVa0cLKiq1gJLgD36TV0IzAFI8hjg+cC3BthiD+BFwHOAf0iyZZIe4FXAvsArgZ7Bzk8yN0lvkt61q7w6UJIkSZIkSZIkSZIkSZOfhVYTWJJPJVmS5OY29J9AgBcOZ/kAY98G/jLJVsBRdK4ofERXLOCbVbWmqn4J/ALYCZgN/FtVra6q3wPfGOzgqppXVT1V1TNl2vRhpCpJkiRJkiRJkiRJkiRNbBZaTSy3Avut+1FVb6bTdWrHNnQvnWsDP57kiME2STIFmAnc3j1eVX8AFtDpVjWHToergazpel4LTGXgwi1JkiRJkiRJkiRJkiRps2Ch1cTyfWDrJG/qGpvWHVBVd9K5uu+8JLP6b5BkS+BDwM+qaukAZ1wInAgcCnxnA3K7Bnh5kq2TbAe8dDiLZu5qRytJkiRJkiRJkiRJkiRNfhZaTSBVVcDRwPOS/CTJTcC5wNv7xd1Mp1jq60me3obPT7IUWAZsC7xikGOuAA4D/r2qHtiA3G4Gvg4sAS4BeoEVw10vSZIkSZIkSZIkSZIkTWbp1PZIQ0uyXVWtTDINuAqYW1WL1remp6enent7xyZBSZIkSZIkSZIkSZIkaQMlWVhVPUPFTR2LZLTJmJdkT2Br4NyhiqwkSZIkSZIkSZIkSZKkTYWFVhq2qvp/xjsHSZIkSZIkSZIkSZIkaTxsMd4JTFRJ3pXk1iRLkyxOcmCSU9q1eSN1xl1Jntier9vIPU5s+S1O8kCSvvZ8xkbsdXeSHfqNTU3y243JDaBv+YqNXSpJkiRJkiRJkiRJkiRNGHa0GkCSg4GXAftV1ZpWDPUYYD5wHrBqpM+squdu5LovAl+ETuEWcERV/XIEU5MkSZIkSZIkSZIkSZI2e3a0GtjOwC+rag1AK1x6NbALcGWSKwGSfCZJb+t89b51i1unqvclWdQ6TO3Rxp+Q5IoktyT5LJCuNSvb9+FJFiS5OMkdSc5Pkjb3kjZ2TZKzkly2vpdIclCS69t51ybZvY1PTfLxJMtax66/6bduWsvzxAH2fEeSm9q6927E31aSJEmSJEmSJEmSJEmadCy0GtgVwJOS3Jnk00meV1VnAffQ6Rh1RIt7V1X1APsAz0uyT9cev6yq/YDPAKe2sX8ArqmqfYGvA08e5Px9gVOAPYGnAYck2Rr4LHBUVc0GdhzGe9wOzG7nvR/4QBt/E52isWdX1T7AhV1rtgcuA85t3bL+R5KXtJwPBGYBz03yiE5cSea2ArTetau8OlCSJEmSJEmSJEmSJEmTn4VWA6iqlcD+wFzgPmB+khMGCH1tkkXALcBedAqj1rmkfS8EZrTnw+hcPUhVfRP4zSAp3FRVd1fVQ8Ditn4P4MdV9ZMWc8EwXmUH4JIky4CPthwBjgT+parWtlx+3bXmMuCzVXX+APu9EDiKzvsuAnYDntE/qKrmVVVPVfVMmTZ9GGlKkiRJkiRJkiRJkiRJE9vU8U5gompFSAuABUn6gNd3zyd5Kp1OVQdU1W+SnANs3RWypn2v5U//zjWM49d0Pa9bn0Fi1+f/At+pqk8n2Q24fF3668njWuCoJF+pqv4xAT5QVZ/fiFwkSZIkSZIkSZIkSZKkScuOVgNI8swku3cNzQL+C/g9nav1AB4L3A+sSLITnU5PQ7kKOL6dcRTwuA1I6w7gaUlmtN9zhrFmOrC8PZ/QNX4F8KYkU1ouj++aeyed9zprgP2+A/x1km3bur9I8sT1JTBzVztaSZIkSZIkSZIkSZIkafKz0Gpg2wHnJrktyVI6VwKeDswDvp3kyqpaQucKvVuBL9DpBDWU9wGHtesGXwj8dLgJVdVq4G+Ay5NcA9wLrBhi2YeBM5P0z+2zwM+BpUmWAK/tN/8WYHqSD/bL4VvAxcANrcvXV+j8rSRJkiRJkiRJkiRJkqRNWh55O5wmqiTbVdXKJAE+BfxHVX18vPNan56enurt7R3vNCRJkiRJkiRJkiRJkqQBJVlYVT1DxdnRanJ5Y5LFdLpoTafTmUqSJEmSJEmSJEmSJEnSKJs63glo+Fr3qgndwUqSJEmSJEmSJEmSJEnaFNnRajORZG2SxUmWJflGkh3a+KVJju6K+2GSd3f9/mqSVyY5PMmKtsfSJP+e5M+GOrdv+YrReSFJkiRJkiRJkiRJkiRpDFlotflYXVWzqmpv4NfAm9v4dcBzAZI8AVgJHNy17uAWA3B122Mf4OauPSRJkiRJkiRJkiRJkqRNmoVWm6frgV3b87W0Qqv2fRmwYzqeSqdA6+fdi5ME2B74zRjlK0mSJEmSJEmSJEmSJI2rqeOdgMZWkinA84HPt6GFwN5JHkOn0OoHwNOAZwH70inEWufQJIuBJwD3A+8c5Iy5wFyAKY/dcRTeQpIkSZIkSZIkSZIkSRpbdrTafGzTiqR+BTwe+C5AVa0BbgX2Aw4CbqTT8eq57XNd1x7rrg58EvBF4CMDHVRV86qqp6p6pkybPlrvI0mSJEmSJEmSJEmSJI0ZC602H6urahbwFOAxwJu75q4DDgO2r6rfADfwcKHVtf03ar7e1kiSJEmSJEmSJEmSJEmbPAutNjNVtQI4GTg1yZZt+FrgfwNL2u+ldLpbPZlOt6uBzAZ+NNR5M3e1o5UkSZIkSZIkSZIkSZImv6njnYDGXlXdkmQJcCzwr3Q6Wj0N+FCb/2OSXwA/q6qHupYe2q4fDLACeMPYZi5JkiRJkiRJkiRJkiSNDwutNhNVtV2/3y/vev4FneKp7vnD+/1eANieSpIkSZIkSZIkSZIkSZslrw6UJEmSJEmSJEmSJEmSpCFYaCVJkiRJkiRJkiRJkiRJQ7DQSqOqb/mK8U5BkiRJkiRJkiRJkiRJetQstBolSdYmWZxkWZKLkkzbiD3OTrJne35nv7nrRirXtt+WSc5I8h8t55uSHNXm7kpydb/4xUmWjWQOkiRJkiRJkiRJkiRJ0kRlodXoWV1Vs6pqb+AB4KQN3aCq3lBVt7Wf7+w399wRyLHb+4Gdgb1bzi8Htu+a3z7JkwCSPGuEz5YkSZIkSZIkSZIkSZImNAutxsbVwG4ASd7aOkYtS3JKG9s2yTeTLGnjc9r4giQ9Sc4AtmldpM5vcyvbd5Kc2db1da09vK2/OMkdSc5PkoGSa9223gj8bVWtAaiqe6vqK11hXwHmtOfjgAsGe9kkc5P0Juldu8qrAyVJkiRJkiRJkiRJkjT5WWg1ypJMBY4C+pLsD5wIHAgcBLwxyb7Ai4F7qurZrZvU5d17VNU7eLhD1vH9jnglMAt4NnAkcGaSndvcvsApwJ7A04BDBklzN+CnVfW79bzKxe0s6HS7+sZggVU1r6p6qqpnyrTp69lSkiRJkiRJkiRJkiRJmhwstBo92yRZDPQCPwU+D8wGLq2q+6tqJXAJcCjQBxyZ5MNJDq2qDWkDNRu4oKrWVtW9wA+AA9rcTVV1d1U9BCwGZjyK9/k18JskxwK3A6uGs2jmrhZaSZIkSZIkSZIkSZIkafKbOt4JbMJWV9Ws7oHBru6rqjtbt6uXAB9KckVV/eMwzxlwz2ZN1/NaBv/3/k/gyUm2r6rfr2e/+cCngBOGmZskSZIkSZIkSZIkSZK0SbCj1di6Cjg6ybQk2wLHAFcn2QVYVVXnAR8F9htg7YNJthxkzzlJpiTZETgMuGlDkqqqVXQ6bp2V5DEASXZO8rp+oZcCHwG+syH7S5IkSZIkSZIkSZIkSZOdHa3GUFUtSnIODxdCnV1VtyR5EXBmkoeAB4E3DbB8HrA0yaKqOr5r/FLgYGAJUMBpVfXzJHtsYHrvBj4A3JbkD8D9wHv75f974MMAgzTnkiRJkiRJkiRJkiRJkjZJqarxzkGbsJ6enurt7R3vNCRJkiRJkiRJkiRJkqQBJVlYVT1DxXl14CSUZEHrgtU9dkqSTw8SPyPJsvZ8eJLLxiJPSZIkSZIkSZIkSZIkaVNhodXkdAFwbL+xY9v4UN4PHJpkcdfnRUOu2kh9y1eM1taSJEmSJEmSJEmSJEnSmLHQanK6GHhZkq2g07EK2AW4JsmZSZYl6UsyZ4C17wGurqpZwCHAIuADSW5J8oq239VJZq1bkOTaJPskeV5XcdYtSbYf3deUJEmSJEmSJEmSJEmSJgYLrSahqvoVcBPw4jZ0LDAfeCUwC3g2cCRwZpKd17PVu4DvV9UBwBEtflvgbOAEgCTPALaqqqXAqcCbW5HWocDqEX41SZIkSZIkSZIkSZIkaUKy0Gry6r4+cN21gbOBC6pqbVXdC/wAOGA9e7wQeEeSxcACYGvgycBFdDpmbQn8FXBOi78W+FiSk4EdquqPA22aZG6S3iS9a1d5daAkSZIkSZIkSZIkSZImPwutJq+vAc9Psh+wTVUtArKBewR4VVXNap8nV9XtVbUK+C7wCuC1wJcBquoM4A3ANsANSfYYaNOqmldVPVXVM2Xa9I17O0mSJEmSJEmSJEmSJGkCsdBqkqqqlXS6UH2BTjcrgKuAOUmmJNkROIzOFYOD+Q7wt0kCkGTfrrmzgbOAm6vq123+6VXVV1UfBnqBAQutus3c1UIrSZIkSZIkSZIkSZIkTX4WWk1uFwDPBi5svy8FlgJLgO8Dp1XVz9ez/v3AlsDSJMvabwCqaiHwO+CLXfGnJFmWZAmwGvj2SL2IJEmSJEmSJEmSJEmSNJGlqsY7B01ASXah0zFrj6p6aGP36enpqd7e3hHLS5L0/7N37+F6luWd978/EgQCGqeCvphXzSC7ApEID1q2guIOqAWlDchbC25ScDdoncrAjKVuKhRnHLUymFJEKqCCogxYwCphL8kKhCQ4gC2ICorbiURCwHC+fzxX6tPlWlmbrKxN8v0cxzqe+742533ed/LneZyXJEmSJEmSJEmSJGksJVlSVZ2h1tnRSr8jyZuA24EzNqTISpIkSZIkSZIkSZIkSdpUTJ/oBDT5VNVFwEUTnYckSZIkSZIkSZIkSZI0WdjRapwl+XiSU3vur01yfs/9f0/y3lHGXjWMNWuTLO35mz2K58xNcsRocpQkSZIkSZIkSZIkSZKmIgutxt+twAEASbYAtgf27Jk/ALhlIz5/dVXN7fn73ihizAWGVWi1/KGVowgvSZIkSZIkSZIkSZIkTS4WWo2/W2iFVnQLrFYAjyb5D0m2An4fuBMgyX9OsjjJsiR/vS5Akq8mWZLk7iTz+z8gyfZJbkty5HASSrJ1ks8mWZ7kziSHDTae5GnAB4F5rSPWvA35GJIkSZIkSZIkSZIkSdJUMH2iE9jcVNXDSX6T5Pl0C65uA2YB+wMrgWVV9USSVwG7AC8BAlyZ5JCquhF4c1X9Isk2wOIkX66qnwMkeQ5wJfBfq+obA6SwTZKl7fqBqjoGeEfLbU6S3YHrkuw60DiwK/ABoFNV7xzoHVvx13yAac/YYUM+lyRJkiRJkiRJkiRJkjQpWGg1MdZ1tToA+B90C60OoFtodWtb86r2d2e7345u4dWNwLuTHNPGn9fGfw5sCXwTeEdV3TDIs1dX1dx+YwcBnwKoqnuSPEi3oGqw8fWqqgXAAoCtdtylhlovSZIkSZIkSZIkSZIkTXYWWk2MW+kWVs2he3TgD4C/AH4FXNDWBPhoVX2md2OSQ4HDgf2r6rEkC4Gt2/RvgCXAq4HBCq0GkhGOS5IkSZIkSZIkSZIkSZuVLSY6gc3ULcBRwC+qam1V/QJ4Jt3jA29ra64F3pxkO4Aks5I8G5gJ/LIVWe0O/EFP3ALeDOye5LQR5HMjcEJ7zq7A84F71zP+KPD04QSeM2vmCNKQJEmSJEmSJEmSJEmSJicLrSbGcmB74Nv9xlZW1c8Aquo64BLgtiTLgcvpFjddA0xPsgz4UL8YVNVa4DjgsCRvH2Y+5wLT2nO+CJxYVWvWM349sEeSpUnmjfz1JUmSJEmSJEmSJEmSpKklVTXROWgT1ul0qq+vb6LTkCRJkiRJkiRJkiRJkgaUZElVdYZaZ0crSZIkSZIkSZIkSZIkSRqChVaSJEmSJEmSJEmSJEmSNAQLrSaRJMckqSS7t/vZSd7YMz83yREbEP97SbYfYPzW0caUJEmSJEmSJEmSJEmSNgcWWk0uxwM3A8e1+9nAG3vm5wKjLrQaTFUdMNYx11n+0MqNFVqSJEmSJEmSJEmSJEkaNxZaTRJJtgMOBN7CbwutzgIOTrI0yfuBDwLz2v28JC9JcmuSO9vvbi3WtCQfS7I8ybIk7+r3rG2SXJPkbe1+Vfs9NMnCJJcnuSfJxUnS5o5oYzcn+WSSq8blw0iSJEmSJEmSJEmSJEmTwPSJTkD/5mjgmqq6L8kvkuwDnAa8r6qOAkjyCNCpqne2+2cAh1TVb5IcDvwN8AZgPvAfgRe3ud/rec52wBeAi6rqogHyeDGwJ/AwcAtwYJI+4DPtWQ8kuXR9L5JkfsuBac/YYVQfQ5IkSZIkSZIkSZIkSZpM7Gg1eRxPtwCK9nv8MPbMBC5LsgL4ON0CKYDDgfOq6jcAVfWLnj1fAz47SJEVwKKq+mFVPQUspXt84e7A/VX1QFuz3kKrqlpQVZ2q6kybMXMYryFJkiRJkiRJkiRJkiRNbna0mgSSPAt4ObBXkgKmAQV8fYitHwKur6pjkswGFq4L2fYP5BbgtUkuqaqB1qzpuV5L9/9IhvEakiRJkiRJkiRJkiRJ0ibLjlaTw7F0j/J7QVXNrqrnAQ8ATwFP71n3aL/7mcBD7frEnvHrgJOTTAfod3TgB4CfA+eOIL97gJ1aMRfAvOFunDPLjlaSJEmSJEmSJEmSJEma+iy0mhyOB67oN/Zl4DjgN0nuSvIe4HpgjyRLk8wD/hb4aJJb6HbBWud84PvAsiR3AW/sF/tUYOskfzuc5KpqNfB24JokNwOPACtH9IaSJEmSJEmSJEmSJEnSFJaBT4+T/r0k21XVqiQBPg18t6o+PtS+TqdTfX19Gz9BSZIkSZIkSZIkSZIkaRSSLKmqzlDr7Gil4XpbkqXA3XSPLPzMBOcjSZIkSZIkSZIkSZIkjZvpE52ApobWvWrIDlaSJEmSJEmSJEmSJEnSpsiOVtqolj+0cqJTkCRJkiRJkiRJkiRJkjaYhVYbWZK1SZYmWZHksiQzRhHj/CR7tOvT+83dOla5tngLk3T6jR2apJK8pWfsxW3sfWP5fEmSJEmSJEmSJEmSJGkystBq41tdVXOrai/gCeDkkQaoqrdW1Xfa7en95g4YgxyHYzkwr+f+OOCucXq2JEmSJEmSJEmSJEmSNKEstBpfNwE7AyR5b+tytSLJqW1s2yRXJ7mrjc9r4wuTdJKcBWzTOmRd3OZWtd8kOaftW96z99C2//Ik9yS5OElGkfv3ga2TPKftfw3wTwMtTDI/SV+SvrWPeXSgJEmSJEmSJEmSJEmSpr7pE53A5iLJdOC1wDVJ9gVOAl4KBLg9yQ3ATsDDVXVk2zOzN0ZVnZbknVU1d4BHvB6YC+wNbA8sTnJjm3sxsCfwMHALcCBw8yhe43Lgj4E7gTuANQMtqqoFwAKArXbcpUbxHEmSJEmSJEmSJEmSJGlSsaPVxrdNkqVAH92uUP8AHARcUVW/rqpVwFeAg+kez3d4krOTHFxVI2kHdRBwaVWtrapHgBuA/drcoqr6YVU9BSwFZo/yXb5Et9DqeODSUcaQJEmSJEmSJEmSJEmSphwLrTa+1VU1t/29q6qeoNvF6ndU1X3AvnQLrj6a5AMjeM76jgPs7Ty1llF2MquqHwNPAq8EvjmcPXNmzRx6kSRJkiRJkiRJkiRJkjTJWWg1MW4Ejk4yI8m2wDHATUmeCzxWVZ8HPgbsM8DeJ5NsOUjMeUmmJdkBOARYtBFy/wDw/qpauxFiS5IkSZIkSZIkSZIkSZPSqDobacNU1R1JLuS3hVDnV9WdSV4NnJPkKbqdo04ZYPsCYFmSO6rqhJ7xK4D9gbuAAv6yqn6cZPdRpHh1kifb9W3Ap3tyv3UU8SRJkiRJkiRJkiRJkqQpLVU10TloE9bpdKqvr2+i05AkSZIkSZIkSZIkSZIGlGRJVXWGWufRgZIkSZIkSZIkSZIkSZI0BAutRiHJc5JckuT+JEuS3JbkmCSHJrmqZ92Hk1ybZKskC5Pcm+SuJLck2W0Uzz05yZva9YlJntszd36SPUYQ64okS9vfD5I82I4uHFPLH1o51iElSZIkSZIkSZIkSZKkcTd9ohOYapIE+Crwuap6Yxt7AfA64Jc9684ADgSOqKo13W2cUFV9SeYD57Q9w1ZV5/XcngisAB5uc28dYaxjRrJekiRJkiRJkiRJkiRJ2pzZ0WrkXg480Vv0VFUPVtWn1t0n+QvgCOAPq2r1ADFuBHZua1+R5M4ky5NckGSrNn5Wku8kWZbkY23szCTvS3Is0AEubh2ptmkdszpt3fEt3ookZ/fktSrJR1pXrW8neU5v3Ha9MMnZSRYluS/JwW18RpIvtXy+mOT2dc+TJEmSJEmSJEmSJEmSNnUWWo3cnsAd65k/EDgZeG1VrRpkzR8Cy5NsDVwIzKuqOXQ7jJ2S5PeAY4A9q+pFwId7N1fV5UAf3Q5Zc3uLudpxgmfTLQibC+yX5Og2vS3w7aram26x19sGyW96Vb0EOBX4qzb2duCXLZ8PAfsO9gGSzE/Sl6Rv7WMeHShJkiRJkiRJkiRJkqSpz0KrDZTk061D1OI29C9AgFcNsPziJEvpFmO9D9gNeKCq7mvznwMOAX4FPA6cn+T1wGMjSGk/YGFV/bSqfgNc3GICPAFc1a6XALMHifGVAdYcBHwBoKpWiUOIGgAAIABJREFUAMsGS6CqFlRVp6o602bMHEHqkiRJkiRJkiRJkiRJ0uRkodXI3Q3ss+6mqt4BvALYoQ09QvfYwI8nOazf3nUdqI6uqh/QLcj6Ha1A6iXAl4GjgWtGkN+AMZsnq6ra9Vq6HbQGsmaANeuLK0mSJEmSJEmSJEmSJG3SLLQauW8BWyc5pWdsRu+C1qHq9cDnk8xdT6x7gNlJdm73fwrckGQ7YGZVfZ3u8X0DxXgUePoA47cDL0uyfZJpwPHADcN4r6HcDPwJQJI9gDnD2TRnlh2tJEmSJEmSJEmSJEmSNPUN1tFIg6iqSnI03Y5Vfwn8FPg18P5+6xYnOQm4coDOVuvWPN7WXJZkOrAYOA/4PeBrSbam20nqPQNsvxA4L8lqYP+emD9K8l+A69ver1fV1zbopbvOBT6XZBlwJ92jA1eOQVxJkiRJkiRJkiRJkiRp0stvT5KTBte6Y23ZisNeCHwT2LWqnljfvk6nU319feOSoyRJkiRJkiRJkiRJkjRSSZZUVWeodXa00nDNAK5PsiXdTlmnDFVkJUmSJEmSJEmSJEmSJG0qLLTSsFTVo8CQlXuSJEmSJEmSJEmSJEnSpmiLiU5Am7blD62c6BQkSZIkSZIkSZIkSZKkDWah1ThJsjbJ0iQrklyWZMYoYpyfZI92fXq/uVvHKtcWb2GSTr+xQ5OsbO+xNMk/j+UzJUmSJEmSJEmSJEmSpMnKQqvxs7qq5lbVXsATwMkjDVBVb62q77Tb0/vNHTAGOQ7HTe095lbV4eP0TEmSJEmSJEmSJEmSJGlCWWg1MW4CdgZI8t7W5WpFklPb2LZJrk5yVxuf18YXJukkOQvYpnWVurjNrWq/SXJO27e8Z++hbf/lSe5JcnGSbIyXSzI/SV+SvrWPeXSgJEmSJEmSJEmSJEmSpr7pE53A5ibJdOC1wDVJ9gVOAl4KBLg9yQ3ATsDDVXVk2zOzN0ZVnZbknVU1d4BHvB6YC+wNbA8sTnJjm3sxsCfwMHALcCBw8whf4eAkS9v1ZVX1kf4LqmoBsABgqx13qRHGlyRJkiRJkiRJkiRJkiYdO1qNn21agVIf8H3gH4CDgCuq6tdVtQr4CnAwsBw4PMnZSQ6uqpG0hToIuLSq1lbVI8ANwH5tblFV/bCqngKWArNH8R69Rwf+TpGVJEmSJEmSJEmSJEmStCmyo9X4Wd2/A9VgR/dV1X2t29URwEeTXFdVHxzmc9Z3HOCanuu1jMO//5xZM4deJEmSJEmSJEmSJEmSJE1ydrSaWDcCRyeZkWRb4BjgpiTPBR6rqs8DHwP2GWDvk0m2HCTmvCTTkuwAHAIs2kj5S5IkSZIkSZIkSZIkSZsFO1pNoKq6I8mF/LYQ6vyqujPJq4FzkjwFPAmcMsD2BcCyJHdU1Qk941cA+wN3AQX8ZVX9OMnuo0jx6iRPtuvbgE+PIoYkSZIkSZIkSZIkSZI05aWqJjoHbcI6nU719fVNdBqSJEmSJEmSJEmSJEnSgJIsqarOUOs8OlCSJEmSJEmSJEmSJEmShmCh1RhK8pwklyS5P8mSJLclOWYUcWYnWdGuD01yVc/ch5Ncm2SrMcr5iiT3JvlpkqXt79VjERtg+UMrxyqUJEmSJEmSJEmSJEmSNGGmT3QCm4okAb4KfK6q3tjGXgC8bgyfcQZwIHBEVa0Zi5hVNeJCMEmSJEmSJEmSJEmSJGlzY0ersfNy4ImqOm/dQFU9WFWfSrJ1ks8mWZ7kziSHwb91rropyR3t74DBgif5C+AI4A+ranUb2zfJDa171rVJdmzjL0xyTRu/KcnubfzCJOe1sfuSHNXG/61rVpIzk1yQZGHrzPXunhz+W5J7knwjyaVJ3jfmX1GSJEmSJEmSJEmSJEmahOxoNXb2BO4YZO4dAFU1pxU9XZdkV+AnwCur6vEkuwCXAp0B9h8I7AbsW1WrAJJsCXwK+KOq+mmSecBHgDcDC4CTq+q7SV4KnEu3EAxgNvAy4IXA9Ul2HuB5uwOHAU8H7k3yv4C9gTcAL6b7/+YOYMlAL5tkPjAfYNozdhjkk0iSJEmSJEmSJEmSJElTh4VWG0mSTwMHAU8AP6RbFEVV3ZPkQWBX4EHg75LMBda2sYH8C/AfgFcBl7ex3YC9gG90Ty1kGvCjJNsBBwCXtXGArXpifamqngK+m+R+ukVV/V3djiZck+QnwHPau3ytp5vW/x7s3atqAd1iL7bacZcabJ0kSZIkSZIkSZIkSZI0VVhoNXbuptvxCYCqekeS7YE+4KFB9rwHeIRut6gtgMcHWfcIcALwzSQ/r6rrgQB3V9X+vQuTPAP4v1U1d5BY/QufBiqEWtNzvZbu/5MMsE6SJEmSJEmSJEmSJEnaLGwx0QlsQr4FbJ3klJ6xGe33RrqFUrQjA58P3AvMBH7UOkz9Kd2uVAOqqvuA1wOfbx2w7gV2SLJ/i7tlkj2r6lfAA0n+uI0nyd49of44yRZJXgjs1OIMx83AHybZunXNOnI4m+bMmjnM8JIkSZIkSZIkSZIkSdLkZaHVGKmqAo4GXpbkgSSLgM8B7wfOBaYlWQ58ETixHc13LvBnSb5N99jAXw/xjMXAScCVwPOAY4Gzk9wFLKV7ZCB0i7re0sbvBv6oJ8y9wA3APwEnV9VgXbQGevaVwF3AV+h26lo5nL2SJEmSJEmSJEmSJEnSVJdufZA2B0kuBK6qqstHuX+7qlqVZAbdLl3zq+qO9e3pdDrV19c3msdJkiRJkiRJkiRJkiRJG12SJVXVGWrd9PFIRpuMBUn2ALYGPjdUkZUkSZIkSZIkSZIkSZK0qbDQajNSVSdu4P43jlEqkiRJkiRJkiRJkiRJ0pSyxUQnoImR5IwkdydZlmRpkpcm+V6S7QdYe2v7nZ1kRbs+NMlVQz1n+UMrxz55SZIkSZIkSZIkSZIkaZzZ0WozlGR/4Chgn6pa04qrnjbY+qo6YNySkyRJkiRJkiRJkiRJkiYhO1ptnnYEflZVawCq6mdV9fC6ySTbJLkmydva/aoJylOSJEmSJEmSJEmSJEmaFCy02jxdBzwvyX1Jzk3ysp657YD/DVxSVX8/muBJ5ifpS9K39jGPDpQkSZIkSZIkSZIkSdLUZ6HVZqiqVgH7AvOBnwJfTHJim/4a8NmqumgD4i+oqk5VdabNmLnB+UqSJEmSJEmSJEmSJEkTbfpEJ6CJUVVrgYXAwiTLgT9rU7cAr01ySVXVhj5nziwLrSRJkiRJkiRJkiRJkjT12dFqM5RktyS79AzNBR5s1x8Afg6cO+6JSZIkSZIkSZIkSZIkSZOUhVabp+2AzyX5TpJlwB7AmT3zpwJbJ/nbiUhOkiRJkiRJkiRJkiRJmmwyBqfDSYPqdDrV19c30WlIkiRJkiRJkiRJkiRJA0qypKo6Q62zo5UkSZIkSZIkSZIkSZIkDcFCq3GS5P9J8oUk/9qO7Pt6kl0nOi9JkiRJkiRJkiRJkiRJQ7PQahwkCXAFsLCqXlhVewCnA8+Z2MyGL8n00exb/tDKsU5FkiRJkiRJkiRJkiRJGncWWo2Pw4Anq+q8dQNVtRS4Ock5SVYkWZ5kHkCSQ5PckORLSe5LclaSE5Isaute2NZdmOS8JDe1dUe18a2TfLatvTPJYW18zxZjaZJlSXZJMjvJinV5JXlfkjPb9cIkf5PkBuA/JdkhyZeTLG5/B47XB5QkSZIkSZIkSZIkSZIm0qi6FGnE9gKWDDD+emAusDewPbA4yY1tbm/g94FfAPcD51fVS5L8J+BdwKlt3WzgZcALgeuT7Ay8A6Cq5iTZHbiuHVN4MvCJqro4ydOAaQzdVeuZVfUygCSXAB+vqpuTPB+4tuUoSZIkSZIkSZIkSZIkbdIstJpYBwGXVtVa4JHWOWo/4FfA4qr6EUCSfwWua3uW0+2Qtc6Xquop4LtJ7gd2b3E/BVBV9yR5ENgVuA04I8n/C3ylqr7bPdVwvb7Yc304sEfPnmckeXpVPdq7Icl8YD7AtGfsMLwvIUmSJEmSJEmSJEmSJE1iHh04Pu4G9h1gfH1VTmt6rp/quX+Kf18gV/321WBxq+oS4HXAauDaJC8HfsO//3+wdb9tv+653gLYv6rmtr9Z/Yus2nMWVFWnqjrTZswc+O0kSZIkSZIkSZIkSZKkKcRCq/HxLWCrJG9bN5BkP+CXwLwk05LsABwCLBph7D9OskWSFwI7AfcCNwIntOfsCjwfuDfJTsD9VfVJ4ErgRcAjwLOTPCvJVsBR63nWdcA7e95h7lDJzZlloZUkSZIkSZIkSZIkSZKmPo8OHAdVVUmOAf5nktOAx4HvAacC2wF30e1E9ZdV9eMku48g/L3ADcBzgJOr6vEk5wLnJVlOt2PViVW1Jsk84P9L8iTwY+CDVfVkkg8CtwMPAPes51nvBj6dZBnd/zs3AiePIFdJkiRJkiRJkiRJkiRpSkpV/5PnNFUkuRC4qqoun+hcBtPpdKqvr2+i05AkSZIkSZIkSZIkSZIGlGRJVXWGWufRgZIkSZIkSZIkSZIkSZI0BI8OnMKq6sSJzkGSJEmSJEmSJEmSJEnaHNjRahJLckaSu5MsS7I0yUuTnJpkxjg9//Se69lJVozHcyVJkiRJkiRJkiRJkqTJxkKrSSrJ/sBRwD5V9SLgcOAHwKnAgIVWSaaNcRqnD71k/ZY/tHIs8pAkSZIkSZIkSZIkSZImlIVWk9eOwM+qag1AVf0MOBZ4LnB9kusBkqxK8sEktwP7J9k3yQ1JliS5NsmObd3CJGcnWZTkviQHt/EZSb7UumZ9McntSTpJzgK2aZ20Lm45TUvy963L1nVJthnnbyJJkiRJkiRJkiRJkiRNCAutJq/rgOe1oqhzk7ysqj4JPAwcVlWHtXXbAiuq6qXA7cCngGOral/gAuAjPTGnV9VL6HbF+qs29nbgl61r1oeAfQGq6jRgdVXNraoT2tpdgE9X1Z7A/wXeMFDiSeYn6UvSt/YxO1pJkiRJkiRJkiRJkiRp6ps+0QloYFW1Ksm+wMHAYcAXk5w2wNK1wJfb9W7AXsA3kgBMA37Us/Yr7XcJMLtdHwR8oj1zRZJl60nrgapaOkCM/rkvABYAbLXjLrWeeJIkSZIkSZIkSZIkSdKUYKHVJFZVa4GFwMIky4E/G2DZ420dQIC7q2r/QUKuab9r+e2/fUaQ0pqe67WARwdKkiRJkiRJkiRJkiRps+DRgZNUkt2S7NIzNBd4EHgUePog2+4Fdkiyf4uxZZI9h3jUzcCftPV7AHN65p5MsuVo8l9nzqyZG7JdkiRJkiRJkiRJkiRJmhQstJq8tgM+l+Q77Ti/PYAz6R7J909Jru+/oaqeAI4Fzk5yF7AUOGCI55xLtzhrGfB+YBmwss0tAJYluXgM3keSJEmSJEmSJEmSJEmaslJVE52DJlCSacCWVfV4khcC3wR2bUVbG6zT6VRfX99YhJIkSZIkSZIkSZIkSZLGXJIlVdUZat308UhGk9oM4Pp2RGCAU8aqyEqSJEmSJEmSJEmSJEnaVFhotZmrqkeBISvyJEmSJEmSJEmSJEmSpM3ZFhOdgMZOkoVJXt1v7NQkFyS5fIi9t27c7CRJkiRJkiRJkiRJkqSpy0KrTculwHH9xo4DPltVx65vY1UdsDESWv7Qyo0RVpIkSZIkSZIkSZIkSRpXFlptWi4HjkqyFUCS2cBzgR8mWdHG9kyyKMnSJMuS7NLGV7XfJDknyYoky5PMa+OHto5Zlye5J8nFSTIB7yhJkiRJkiRJkiRJkiSNu+kTnYDGTlX9PMki4DXA1+h2s/oiUD3LTgY+UVUXJ3kaMK1fmNcDc4G9ge2BxUlubHMvBvYEHgZuAQ4Ebu6fR5L5wHyAac/YYWxeTpIkSZIkSZIkSZIkSZpAdrTa9PQeH3hcu+91G3B6kvcDL6iq1f3mDwIuraq1VfUIcAOwX5tbVFU/rKqngKXA7IESqKoFVdWpqs60GTM3/I0kSZIkSZIkSZIkSZKkCWah1abnq8ArkuwDbFNVd/ROVtUlwOuA1cC1SV7eb//6jgNc03O9FjuiSZIkSZIkSZIkSZIkaTNhodUmpqpWAQuBC/jdblYk2Qm4v6o+CVwJvKjfkhuBeUmmJdkBOARYNNp85syyo5UkSZIkSZIkSZIkSZKmPgutNk2XAnsDXxhgbh6wIslSYHfgon7zVwDLgLuAbwF/WVU/3oi5SpIkSZIkSZIkSZIkSZNeqmqic9AmrNPpVF9f30SnIUmSJEmSJEmSJEmSJA0oyZKq6gy1zo5WkiRJkiRJkiRJkiRJkjQEC60kSZIkSZIkSZIkSZIkaQgWWk0iSSrJP/bcT0/y0yRXbUDM04e57ntJtm/Xt7bf2UneONpnAyx/aCWzT7t6Q0JIkiRJkiRJkiRJkiRJE85Cq8nl18BeSbZp968EHtrAmMMqtOpVVQe0y9nABhVaSZIkSZIkSZIkSZIkSZsCC60mn38CjmzXxwOXrptIsm2SC5IsTnJnkj9q4ycm+UqSa5J8N8nftvGzgG2SLE1ycRv7apIlSe5OMn+gBJKsapdnAQe3/e9JclOSuT3rbknyorH+AJIkSZIkSZIkSZIkSdJkY6HV5PMF4LgkWwMvAm7vmTsD+FZV7QccBpyTZNs2NxeYB8wB5iV5XlWdBqyuqrlVdUJb9+aq2hfoAO9O8qz15HIacFPb/3HgfOBEgCS7AltV1bL+m5LMT9KXpG/tYytH9REkSZIkSZIkSZIkSZKkycRCq0mmFS7NptvN6uv9pl8FnJZkKbAQ2Bp4fpv7ZlWtrKrHge8ALxjkEe9OchfwbeB5wC4jSO8y4KgkWwJvBi4c5B0WVFWnqjrTZswcQXhJkiRJkiRJkiRJkiRpcpo+0QloQFcCHwMOBXo7TgV4Q1Xd27s4yUuBNT1Daxng3zbJocDhwP5V9ViShXSLtYal7fkG8EfAn9DtiiVJkiRJkiRJkiRJkiRt8uxoNTldAHywqpb3G78WeFeSACR58TBiPdk6UAHMBH7ZCqZ2B/5giL2PAk/vN3Y+8ElgcVX9YqiHz5k1k++ddeQw0pQkSZIkSZIkSZIkSZImLwutJqGq+mFVfWKAqQ8BWwLLkqxo90NZ0NZfDFwDTE+yrO399hB7lwG/SXJXkve03JYAvwI+O7y3kSRJkiRJkiRJkiRJkqa+VNVE56ApJMlzgYXA7lX11FDrO51O9fX1bfS8JEmSJEmSJEmSJEmSpNFIsqSqOkOts6OVhi3Jm4DbgTOGU2QlSZIkSZIkSZIkSZIkbSqmT3QCmjqq6iLgoonOQ5IkSZIkSZIkSZIkSRpvdrQaR0mOSVJJdp/oXHol+WCSwzdG7OUPrWT2aVdvjNCSJEmSJEmSJEmSJEnSuLHQanwdD9wMHDcWwZKMSUeyqvpAVf3zWMSSJEmSJEmSJEmSJEmSNkUWWo2TJNsBBwJvoRVaJdkiyblJ7k5yVZKvJzm2zR2R5J4kNyf5ZJKr2viZSRYkuQ64KMm0JOckWZxkWZI/b+t2THJjkqVJViQ5uK29sN0vT/KetvbCnue+Ismdbf6CJFu18e8l+eskd7S5SdWVS5IkSZIkSZIkSZIkSdqYxqQjkoblaOCaqrovyS+S7APsBMwG5gDPBv4PcEGSrYHPAIdU1QNJLu0Xa1/goKpanWQ+sLKq9mtFUbe0IqzXA9dW1UeSTANmAHOBWVW1F0CSZ/YGbc+9EHhFy/Mi4BTgf7YlP6uqfZK8HXgf8NaBXrTlNB9g2jN2GN3XkiRJkiRJkiRJkiRJkiYRO1qNn+OBL7TrL7T7g4DLquqpqvoxcH2b3x24v6oeaPf9C62urKrV7fpVwJuSLAVuB54F7AIsBk5KciYwp6oeBe4HdkryqSSvAX7VL+5uwANVdV+7/xxwSM/8V9rvEroFYgOqqgVV1amqzrQZMwdbJkmSJEmSJEmSJEmSJE0ZdrQaB0meBbwc2CtJAdOAAq4YbMsQIX/db+27quraAZ57CHAk8I9Jzqmqi5LsDbwaeAfwJ8CbR/DcNe13Lf7fkSRJkiRJkiRJkiRJ0mbEjlbj41jgoqp6QVXNrqrnAQ8APwPekGSLJM8BDm3r76HbeWp2u5+3ntjXAqck2RIgya5Jtk3yAuAnVfX3wD8A+yTZHtiiqr4M/Ddgn36x7gFmJ9m53f8pcMOo3xqYM2sm3zvryA0JIUmSJEmSJEmSJEmSJE04uxKNj+OBs/qNfRn4feCHwArgPrpH/62sqtVJ3g5ck+RnwKL1xD6f7jF+dyQJ8FPgaLpFW/85yZPAKuBNwCzgs0nWFdj9l95AVfV4kpOAy5JMp3v84HmjemNJkiRJkiRJkiRJkiRpE5KqmugcNmtJtquqVe14wUXAgVX1457xAJ8GvltVH5/YbEeu0+lUX1/fRKchSZIkSZIkSZIkSZIkDSjJkqrqDLXOjlYT76okzwSeBnyoqn7cxt+W5M/a+J3AZyYqQUmSJEmSJEmSJEmSJGlzZ6HVBKuqQwcZ/zgw5TpYSZIkSZIkSZIkSZIkSZuiLSY6AW3alj+0ktmnXT3RaUiSJEmSJEmSJEmSJEkbxEKrMZRkbZKlSe5OcleS9yYZ8hsnOaftOWc88uz37L2TLO25Pz7JY0m2bPdzkixr1wuT3NvebXGSueOdryRJkiRJkiRJkiRJkjQRPDpwbK2uqrkASZ4NXALMBP5qiH1/DuxQVWs2cn4DWQ68IMnTq+pR4ADgHuDFwKJ2f0vP+hOqqi/JScA5wCvHO2FJkiRJkiRJkiRJkiRpvNnRaiOpqp8A84F3pmta61y1OMmyJH8OkORKYFvg9iTzkuyQ5Mtt3eIkB7Z1Zya5oHWVuj/Ju9v4tkmubl2mViSZ18b3TXJDkiVJrk2y4yB5PgUsBl7ahvYFPk23wIr2e+sAW28DZg0UM8n8JH1J+tY+tnLE306SJEmSJEmSJEmSJEmabCy02oiq6n663/jZwFuAlVW1H7Af8LYk/7GqXkfrhFVVXwQ+AXy8rXsDcH5PyN2BVwMvAf6qHe/3GuDhqtq7qvYCrmnjnwKOrap9gQuAj6wn1VuBA5JsCzwFLOTfF1rdMsCe1wBfHeS9F1RVp6o602bMXN8nkiRJkiRJkiRJkiRJkqYEjw7c+NJ+XwW8KMmx7X4msAvwQL/1hwN7JOu28YwkT2/XV7fjBdck+QnwHLpH/30sydnAVVV1U5K9gL2Ab7Q404AfrSfHW4C/AG4CFlfVvybZOckOwHatYGydi1tB1jRgn+F/BkmSJEmSJEmSJEmSJGnqstBqI0qyE7AW+Andgqt3VdW1Q2zbAti/qlb3iwWwpmdoLTC9qu5Lsi9wBPDRJNcBVwB3V9X+w0z123S7bB1E90hAgB8Cx/G7xwaeANwFnEX3iMHXry/wnFkz6TvryGGmIUmSJEmSJEmSJEmSJE1OHh24kbRuUOcBf1dVBVwLnNKO9SPJrq0zVH/XAe/siTN3iOc8F3isqj4PfIxul6l7gR2S7N/WbJlkz8FiVNWjwA+AE/ltodVtwKn8bqEVVfUk8F+BP0jy++vLT5IkSZIkSZIkSZIkSdoUWGg1trZJsjTJ3cA/0y2a+us2dz7wHeCOJCuAzzBwR7F3A50ky5J8Bzh5iGfOARYlWQqcAXy4qp4AjgXOTnIXsBQ4YIg4twBbVdUP2v1twE4MUGgF0Dpu/XfgfUPElSRJkiRJkiRJkiRJkqa8dJstSRtHp9Opvr6+iU5DkiRJkiRJkiRJkiRJGlCSJVXVGWqdHa0kSZIkSZIkSZIkSZIkaQgWWm1Gkny6HW3Y+3fSxnzm8odWMvu0qzfmIyRJkiRJkiRJkiRJkqSNbtwKrZKsbYU9K5JclmTGKGKcn2SPdn16v7lbxyrXFm96kr9J8t2eoqQzxvIZI8jl0CQre/L459HEqap3VNXcfn+f3YC8zkzyvtHulyRJkiRJkiRJkiRJkqaK8exotboV9uwFPAGcPNIAVfXWqvpOuz2939wBY5Bjrw8DzwXmVNVc4GBgy/6L0jUe3/GmnuKow4e7Kcn0sXh4kmljEUeSJEmSJEmSJEmSJEmaiibq6MCbgJ0Bkry3dblakeTUNrZtkquT3NXG57XxhUk6Sc4CtmndnS5uc6vab5Kc0/Yt79l7aNt/eZJ7klycJAMl17ptvQ14V1U9DlBVj1bVmW1+dpL/k+Rc4A7geUmOb89bkeTsnlireq6PTXJhu74wyXlJbkpyX5KjRvoRk7wgyTeTLGu/z++J/T+SXA+c3fJ6Zvs2P0/yprbuH5Mc3t7npiR3tL8Der7Z9UkuAZa3sTOS3Nu6au02SF7zk/Ql6Vv72MqRvpYkSZIkSZIkSZIkSZI06YxJt6ORaB2WXgtck2Rf4CTgpUCA25PcAOwEPFxVR7Y9M3tjVNVpSd7ZOk3193pgLrA3sD2wOMmNbe7FwJ7Aw8AtwIHAzQPE2Bn4flU9up5X2Q04qarenuS5wNnAvsAvgeuSHF1VXx3ic8wGXga8ELg+yc7rCrsGcHCSpe36sqr6CPB3wEVV9bkkbwY+CRzd1uwKHF5Va5Oc1971QeB+ut25LgL+ADgFeAp4ZVU9nmQX4FKg0+K8BNirqh5o/17H0f2O0+kWmS3pn2hVLQAWAGy14y41xDeQJEmSJEmSJEmSJEmSJr3x7Gi1TSsU6gO+D/wDcBBwRVX9uqpWAV+hWwS0HDg8ydlJDq6qkbRFOgi4tKrWVtUjwA3Afm1uUVX9sKqeApbSLXQaUpKTWvesHyR5Xht+sKq+3a73AxZW1U+r6jfAxcAhwwj9pap6qqq+S7cAavf1rO09OvAjbWx/4JJ2/Y90332dy6pq7bq9LZ9DgP8FzEkyC/hF++5bAn+fZDlwGbBHT5xFVfVAuz6Y7r/XY1X1K+DKYbyjJEmSJEmSJEmSJEmSNOWNZ6HV6p5CoXdV1RN0u1j9jqq6j253qOXAR5OW788mAAAgAElEQVR8YATPGTBms6bnei2Dd/T6F+D5SZ7e8vls6561EpjW1vx6mM/s7ei09XrmBrofqd79vfndSLdI6mBgIfBT4Fi6BVgA7wEeodsFrAM8bZA4I85xzqyZfO+sI0eyRZIkSZIkSZIkSZIkSZp0xrPQaiA3AkcnmZFkW+AY4KZ2FN9jVfV54GPAPgPsfTLJloPEnJdkWpId6HZxWjSSpKrqMbodt/4uydbw/7N352F6luXdx7+/JBiIgeASMeZVRymiQiDAgCKLoBSt0SKLBaRW3FKrVVGRpsvbUtco7isGi2hBVJRYJIhYZRUEJ5ANBH2VUPaKSjAsEZLz/eO5Rx7GmcxMMpNnJnw/xzHHc9/XfV7ndd538ud5XBckmcgjG5DaXQm8MMkTm7ijae2kBXBnkuckmdC8X7tXJZmQZHtaxyXeMJw6gctpHeUHcAz9H4NIVd1M6xjFHarqV03c8TzcaDUNuL3Z6es1PNxM1tclwKFJtmqa0F4xzHolSZIkSZIkSZIkSZKkcWmgHZ02iaq6OslpPNwI9aWquibJS4CTkqwDHgT+rp/pC4BlSa6uqmPaxhfSOlJvKa3dl06oqjuSrO9Yvv78M/A+YEWS3wP3A18BbgOe0uc9bk/yj8CFtHa3Oq+q/qt5PA84F7gZWAFMbZt6A62GrO2AN1fVA8Os8e3AqUneQ2uXqtetJ/ZKHm6guhT4EA83Zn0e+HaSVzXv0HcXK+CP/17foHXs4k083KglSZIkSZIkSZIkSZIkbdZStbGn1WlDNA1m51bVtzpdy2jq7u6unp6eTpchSZIkSZIkSZIkSZIk9SvJ4qrqHiyu00cHSpIkSZIkSZIkSZIkSdKY96hvtEqyMMmSPn8vGaW11jb5VwCPBc7r8/wl/dSysE/Ml5I8t7n+pz7PLt+I2v6YdyQtv3UVXfMWjXRaSZIkSZIkSZIkSZIkaZPy6MBNKMnqqpraXJ8BLK6qj49EvrFq8owdasZrP8nK+XM6XYokSZIkSZIkSZIkSZL0Jzw6cOy7FPgzgCTvSrKi+TuuGXtskkVJljbjRzbjFyXpTjIf2KrZ9eqM5tnq5jdJTmrmLW+be0Az/1tJrk9yRpK05+3Nk+QDzdo/SbJdM759c//TJO/tXU+SJEmSJEmSJEmSJEna3Nlo1QFJJgF/ASxPsgfwOuB5wPOBNyXZDXgpcFtV7VpVOwPnt+eoqnnA/VU1u6qO6bPEYcBsYFfgIOCkJDOaZ7sBxwHPBZ4J7NNPiY8FflJVuwKXAG9qxj8FfKqq9gRuW8/7zU3Sk6Rn7X2rhvBFJEmSJEmSJEmSJEmSpLHNRqtNa6skS4Ae4H+A/wD2BRZW1b1VtRo4G9gPWA4clOTDSfarquF0LO0LnFlVa6vqTuBiYM/m2VVVdUtVrQOWAF39zP8DcG5zvbgtZm/grOb6awMtXlULqqq7qronTpk2jLIlSZIkSZIkSZIkSZKksWlSpwt4lLm/qma3D/Qe3ddXVf282e3qZcCHklxQVe8d4jr95mysabteS///Bx6sqhokZkhmzZxGz/w5GzpdkiRJkiRJkiRJkiRJGhPc0arzLgFemWRKkscChwKXJnkKcF9VnQ58FNi9n7kPJtligJxHJpmYZDqwP3DVCNT6E+Dw5vqoEcgnSZIkSZIkSZIkSZIkjQvuaNVhVXV1ktN4uBHqS1V1TZKXACclWQc8CPxdP9MXAMuSXF1Vx7SNL6R1zN9SoIATquqOJM/eyHKPA05P8m5gETCc4wwlSZIkSZIkSZIkSZKkcSsPnxAnrV+SKbSOP6wkRwFHV9Uh65vT3d1dPT09m6ZASZIkSZIkSZIkSZIkaZiSLK6q7sHi3NFKw7EH8NkkAe4GXt/heiRJkiRJkiRJkiRJkqRNwkarUZLkn4FXA2uBdcDfVtWVQ5z7l8Bzq2r+emK6gBdU1dea+ynAKcAuQG8j1EuravVGvMYjVNWlwK5tNawGpo5UfkmSJEmSJEmSJEmSJGmsmtDpAjZHSfYGXg7sXlW7AAcBNw9x7qSqOmd9TVaNLlqNXL3eAdxZVbOqamfgDcCDwy5+hC2/dRVd8xZ1ugxJkiRJkiRJkiRJkiRpo7ij1eiYAdxVVWsAquougCQrgW8ABzZxr66q/5fkNOC3wG7A1UmWA91V9ffNs3uAbuDJwAlV9S1gPvCcJEuArzRr3tRbQFXd0KzZBZwPXNnk/znwN1V1X5I9gI/T2pXqLuDYqro9yfbA54DpwH3Am6rq+iTPAL5G6//N+SP5wSRJkiRJkiRJkiRJkqSxzB2tRscFwFOT/DzJ55O8sO3ZPVW1F/BZ4JNt488CDqqqd/eTbwawL61dsnp3upoHXFpVs6vqE8CpwD8kuSLJ+5Ps0DZ/R2BBs7vWPcBbkmwBfAY4oqr2aOZ/oIlfALytGT8e+Hwz/ingC1W1J3DHsL+KJEmSJEmSJEmSJEmSNE65o9UoqKrVzW5R+9HaveobSeY1j89s+/1E27SzqmrtACm/U1XrgOuSbDfAmkuSPBM4mNZRhT9tjjC8H7i5qn7chJ4OvJ3WjlQ7Az9IAjARuD3JVOAFwFnNOMDk5ncf4PDm+j+BD/dXS5K5wFyAidtMH+CVJEmSJEmSJEmSJEmSpPHDRqtR0jRNXQRc1BwF+NreR+1hbdf3rifdmrbrDBRUVauBs4Gzk6wDXgZ8u886vesGuLaq9m5/kGQb4O6qmj3QMuups7eOBbR2xWLyjB0GjZckSZIkSZIkSZIkSZLGOo8OHAVJduxzdN9s4Kbm+si23ys2YpnfA1u3rblPksc1148Bntu25tOa3a0AjgYuA24ApveOJ9kiyU5VdQ9wY5JXNeNJsmsz98fAUc31MUMpctbMaaycP2dD31GSJEmSJEmSJEmSJEkaE2y0Gh1Tga8kuS7JMlpNTyc2zyYnuRJ4B/DOjVhjGfBQkqVJ3glsD1zc7J51DdBDazcrgJ8Br21qeTzwhar6A3AE8OEkS4EltI4MhFYT1Rua8WuBQ5rxdwBvTfJTYNpG1C5JkiRJkiRJkiRJkiSNK6nyZLdNJclKoLuq7tqEa3YB51bVzptqzXbd3d3V09PTiaUlSZIkSZIkSZIkSZKkQSVZXFXdg8W5o5UkSZIkSZIkSZIkSZIkDWJSpwt4NKmqrg6suRLoyG5WkiRJkiRJkiRJkiRJ0ubCHa0kSZIkSZIkSZIkSZIkaRAj2miV5MlJvp7kl0muS3JekmdtYK7jkkxpuz8vybYjVy0k6Ury6kFipiQ5I8nyJCuSXJZk6gaud2ySp7Tdr0zyxCHO/VSSW5OMWnNckjcn+ZuRzLn81lV0zVs0kiklSZIkSZIkSZIkSZKkTW7EmnaSBFgIXFRV21fVc4F/ArbbwJTHAX9stKqql1XV3Rtf6SN0AetttALeAdxZVbOqamfgDcCDG7jescBTBgvqq2muOhS4Gdh/A9cebI1JVXVyVX11NPJLkiRJkiRJkiRJkiRJ49lI7o50IPBgVZ3cO1BVS6rq0rSc1OwItTzJkQBJDkhyUZJvJbm+2TkqSd5OqyHpwiQXNrErkzyx2YXqZ0lOSXJtkguSbNXEbJ/k/CSLk1ya5NnN+GlJPp3k8iS/SnJEU+J8YL8kS5K8c4D3mgHc2vZON1TVmibvu5p3WpHkuGasK8mK3vgkxyc5sVmzGzijWW+rJuRtSa5uvsuz1/NtVwBfAI5uy31ikq8032BlksOSfKTJdX6SLZq4PZJc3HyX7yeZ0YxflOSDSS4G3tHkO7559mdJ/jvJ0qa+7ZNMTfLDtnoP6a/YJHOT9CTpWXvfqgFeSZIkSZIkSZIkSZIkSRo/RrLRamdg8QDPDgNmA7sCBwEn9Tb7ALvR2r3qucAzgX2q6tPAbcCBVXVgP/l2AD5XVTsBdwOHN+MLgLdV1R7A8cDn2+bMAPYFXk6rwQpgHnBpVc2uqk8MUPupwD8kuSLJ+5PsAK3mJeB1wPOA5wNvSrLbADmoqm8BPcAxzXr3N4/uqqrdaTVRHT/A9KOBM2ntGPby3gaqxvbAHOAQ4HTgwqqaBdwPzGliPwMc0XyXU4EPtM3ftqpeWFUf67PmGbS+8a7AC4DbgQeAQ5t6DwQ+1uxk1vddF1RVd1V1T5wybaBPIkmSJEmSJEmSJEmSJI0bkzbROvsCZ1bVWuDOZgelPYF7gKuq6haAJEtoHed32SD5bqyqJc31YqAryVRaDUFntfX+TG6b852qWgdcl2TIxxlW1ZIkzwQOptUk9tMkezfvtLCq7m1qPxvYDzhnqLkbZ7e9x2F9HyZ5DPAy4J1V9fskVza1LGpCvldVDyZZDkwEzm/Gl9P6ljvSaoL7QfNdJtJqmur1jX7W3BqYWVULm2/wQDO+BfDBJPsD64CZtI6GvGOY7yxJkiRJkiRJkiRJkiSNKyPZaHUtcMQAz/5k16M2a9qu1zK0mvrO2YrW7lx3V9XsIcxZXz1/oqpW02qIOjvJOlqNT2sHCH+IR+4UtuUg6XvrGujdXwpMA5Y3jVJTgPt4uNFqTVPjuiQPVlU14+uafAGuraq9B1j/3n7GBvo+xwDTgT2a5q6VDPJ+s2ZOo2f+nPWFSJIkSZIkSZIkSZIkSWPeSB4d+CNgcpI39Q4k2TPJC4FLgCOTTEwyHdgfuGqQfL8Hth7q4lV1D3Bjklc1ayfJrhu7RpJ9kjyuuX4MrSMOb6L1Tq9MMiXJY4FDgUuBO4EnJXlCksm0jircoHdqHA28saq6qqoLeAZwcJIpQ5x/AzC92YWLJFsk2Wl9E5pveUuSVzZzJjfrTQP+t2myOhB4+jDfRZIkSZIkSZIkSZIkSRqXRqzRqtlJ6VDgz5P8Msm1wInAbcBCYBmwlFZD1glVNdhxcwuA7yW5cBhlHAO8IclSWjtsHTJI/DLgoSRLk7xzgJjtgYubo/muAXqAb1fV1cBptBrGrgS+VFXXVNWDwHubsXOB69tynQacnGRJkq0Ge5mmueklPLx7Fc1RhZcBrxhsfhP/B1o7jX24+S5LaB2xOJjXAG9Psgy4HHgycAbQnaSH1re+fj3zJUmSJEmSJEmSJEmSpM1GHj5pThp53d3d1dPT0+kyJEmSJEmSJEmSJEmSpH4lWVxV3YPFjeTRgZIkSZIkSZIkSZIkSZK0WZrU6QLGiiQvAT7cZ/jGqjq0E/V0UpK1wHJa/z9uBF5TVXcnmQB8EngRUMADwF9V1Y0dK1aSJEmSJEmSJEmSJEnaBNzRqlFV36+q2X3+HnVNVo37m/ffGfgt8NZm/EjgKcAuVTULOBS4e32Jlt+6iq55i0a1WEmSJEmSJEmSJEmSJGm0uaOVBnMFsEtzPQO4varWAVTVLR2rSpIkSZIkSZIkSZIkSdqE3NFKA0oyEXgxcE4z9E3gFUmWJPlYkt0GmDc3SU+SnrX3rdpU5UqSJEmSJEmSJEmSJEmjxkYr9WerJEuA3wCPB34Af9zBakfgH4F1wA+TvLjv5KpaUFXdVdU9ccq0TVi2JEmSJEmSJEmSJEmSNDpstFJ/7q+q2cDTgccAb+19UFVrqup7VfUe4IPAKztUoyRJkiRJkiRJkiRJkrTJ2GilAVXVKuDtwPFJtkiye5KnACSZAOwC3LS+HLNmTmPl/DmjX6wkSZIkSZIkSZIkSZI0iiZ1ugCNbVV1TZKlwFHAr4FTkkxuHl8FfLZjxUmSJEmSJEmSJEmSJEmbiI1W+hNVNbXP/Svabs/fxOVIkiRJkiRJkiRJkiRJHefRgZIkSZIkSZIkSZIkSZI0CButJEmSJEmSJEmSJEmSJGkQNlptpCT/nOTaJMuSLEnyvCQXJelunp+XZNt+5p2Y5Pjm+r1JDtqAtZ/QrLkkyR1Jbm27f8wwc70/yXH9jJ+e5JXDra3X8ltX0TVv0YZOlyRJkiRJkiRJkiRJksaESZ0uYDxLsjfwcmD3qlqT5InAIxqcquplg+Wpqn/dkPWr6jfA7KaWE4HVVfXRDcklSZIkSZIkSZIkSZIkaWDuaLVxZgB3VdUagKq6q6puaw9IsrJpwOrd/eqGJP8N7NgWc1qSI9ri/z3J1UmWJ3l2Mz49yQ+a8S8muak370CSfDfJ4mbHrTe2jc9p8ixNckE/8/4uyaIkW/YZ3zPJxU3O7yXZbrgfTJIkSZIkSZIkSZIkSRqPbLTaOBcAT03y8ySfT/LCgQKT7AEcBewGHAbsuZ68d1XV7sAXgOObsX8DftSMLwSeNoT6XltVezRrvSvJ45I8ucl7aFXt2tTUXudxwMHN8wfaxicDnwIOb3KeDrxvgHedm6QnSc/a+1YNoUxJkiRJkiRJkiRJkiRpbPPowI1QVaubBqr9gAOBbySZN0D4fsDCqroPIMk560l9dvO7mFZTFsC+wKHNuucn+d0QSnxnkr9srv8PsD3wVODCqrqpyfXbtvjXATcBh1XVQ31yPQfYCfjvJAATgVv6W7SqFgALACbP2KGGUKckSZIkSZIkSZIkSZI0ptlotZGqai1wEXBRkuXAa9cXPsS0a5rftTz8b5Th1JXkIGB/4PlVdX+Sy4AtmzwD1bEcmA3MpNVw9YiUwLKq2m84dUiSJEmSJEmSJEmSJEmbA48O3AhJdkyyQ9vQbP60QanXJcChSbZKsjXwimEudxnwV826BwOPGyR+GvDbpslqJx4+qvDHwIuSPL3J9fi2OT3AW4HvNkcMtrsOmJlkr2beY5q86zVr5jRWzp8zWJgkSZIkSZIkSZIkSZI0ptlotXGmAl9Jcl2SZcBzgRP7C6yqq4FvAEuAbwOXDnOtfwcOTnI18BfA7cDv1xO/CJiSZCnwr8CVTR13An8H/Ffz7Iw+dV4MzAMWtTdhVdUa4Ajg4828a4DnDfMdJEmSJEmSJEmSJEmSpHEpVUM9zU6dlGQysLaqHkqyN/CFqprd6boG093dXT09PZ0uQ5IkSZIkSZIkSZIkSepXksVV1T1Y3KRNUYxGxNOAbyaZAPwBeFOH65EkSZIkSZIkSZIkSZIeNWy0Gieq6hfAbp2uQ5IkSZIkSZIkSZIkSXo0mtDpAsaDJKs7vP6JSe5L8qSRrilJV5IVI5GrP8tvXUXXvEWjlV6SJEmSJEmSJEmSJEnaJGy0GkFpGa1vehfw7lHKvcGSuCuaJEmSJEmSJEmSJEmSNns2Wg1Dkvck+WmSZUn+vRnrSvKzJJ8HrgaemuS0JCuSLE/yziZu+yTnJ1mc5NIkz06ydZIbk2zRxGyTZGXvfR+nAkcmeXyfmh6xI1WS45Oc2FxflOQTSS5patwzydlJfpHk/W1pJiX5SvNe30oypZm/R5KLm5q/n2RGW94PJrkYeMcIfV5JkiRJkiRJkiRJkiRpzLLRaoiSHAzsAOwFzAb2SLJ/83hH4KtVtRvwRGBmVe1cVbOALzcxC4C3VdUewPHA56vq98BFwJwm5ijg21X1YD8lrKbVbDXcxqY/VNX+wMnAfwFvBXYGjk3yhLb6F1TVLsA9wFuaZq/PAEc0NZ8KfKAt77ZV9cKq+ljfBZPMTdKTpGftfauGWa4kSZIkSZIkSZIkSZI09njs29Ad3Pxd09xPpdV49T/ATVX1k2b8V8Azk3wGWARckGQq8ALgrCS9+SY3v18CTgC+A7wOeNN6avg0sCTJnzQ3rcc5ze9y4Nqquh0gya+ApwJ3AzdX1Y+buNOBtwPn02rI+kFT80Tg9ra83xhowapaQKuxjMkzdqhh1CpJkiRJkiRJkiRJkiSNSTZaDV2AD1XVFx8xmHQB9/beV9XvkuwKvITW7lF/BRwH3F1Vs/smraofN8f/vRCYWFUr+sa0xd6d5GvAW9qGH+KRO5Nt2WfamuZ3Xdt1733vv3/fZqii9b7XVtXeA5Rz7wDjkiRJkiRJkiRJkiRJ0mbHowOH7vvA65vdqUgyM8mT+gYleSIwoaq+DfxfYPequge4Mcmrmpg0zVi9vgqcycPHDK7Px4G/5eEmqTuBJyV5QpLJwMs34N2elqS3oepo4DLgBmB673iSLZLsNNzEs2ZOY+X8OYMHSpIkSZIkSZIkSZIkSWOYjVaDSDIJWFNVFwBfA65Ishz4FrB1P1NmAhclWQKcBvxjM34M8IYkS4FrgUPa5pwBPI5Ws9V6VdVdwEKaower6kHgvcCVwLnA9cN8RYCfAa9Nsgx4PPCFqvoDcATw4abmJbSOP5QkSZIkSZIkSZIkSZIedVLV99Q4tWt2njqlqvYaxTWOAA6pqteM1hqd0t3dXT09PZ0uQ5IkSZIkSZIkSZIkSepXksVV1T1Y3KTBAh7NkrwZeDtw3Ciu8RngL4CXjdYakiRJkiRJkiRJkiRJkjaOjVbrUVUnAyeP8hpvG838kiRJkiRJkiRJkiRJkjbehE4XoKFJsnqU85+XZNuRzrv81lV0zVs00mklSZIkSZIkSZIkSZKkTcodrQRAVXl0oSRJkiRJkiRJkiRJkjQAd7QaR5IckOTctvvPJjm2uV6Z5INJrkjSk2T3JN9P8sskb26bf0mShUmuS3Jykglt85+YpCvJz5KckuTaJBck2aqJ2TPJsmaNk5Ks6MBnkCRJkiRJkiRJkiRJkjY5G602LzdX1d7ApcBpwBHA84H3tsXsBbwbmAVsDxzWT54dgM9V1U7A3cDhzfiXgTc3a6wdqIgkc5tmr561963auDeSJEmSJEmSJEmSJEmSxgAbrTYv5zS/y4Erq+r3VfVr4IEk2zbPrqqqX1XVWuBMYN9+8txYVUua68VAVzN/66q6vBn/2kBFVNWCquququ6JU6Zt9EtJkiRJkiRJkiRJkiRJnWaj1fjyEI/8N9uyz/M1ze+6tuve+0nNdfWZ0/e+PQ+0dq6aBGRYlUqSJEmSJEmSJEmSJEmbERutxpebgOcmmZxkGvDiDcixV5JnJJkAHAlcNpRJVfU74PdJnt8MHTWUebNmTmPl/DkbUKYkSZIkSZIkSZIkSZI0dkwaPESdlmQSsKaqbk7yTWAZ8Avgmg1IdwUwH5gFXAIsHMbcNwCnJLkXuAhYtQHrS5IkSZIkSZIkSZIkSeOOjVbjw07ALwGq6gTghL4BVdXVdn0acFrfZ0kA7quqI9cz/y5g57bxj7aFXVtVuzS55gE9G/IykiRJkiRJkiRJkiRJ0nhjo9UYl+TNwNuB4zpdCzAnyT/S+n9zE3BsZ8uRJEmSJEmSJEmSJEmSNo1UVadr0Gasu7u7enrc+EqSJEmSJEmSJEmSJEljU5LFVdU9WNyETVGM/lSS1aOc/+VJrkmyNMl1Sf62GT8tyRHDyNOV5NUbWsfyW1fRNW/Rhk6XJEmSJEmSJEmSJEmSxgSPDtwMJdkCWADsVVW3JJkMdG1gui7g1cDXRqY6SZIkSZIkSZIkSZIkafxxR6sxJMnTk/wwybLm92nN+GlJPp3k8iS/6t2RKsmEJJ9Pcm2Sc5Oc1zzbmlYT3W8AqmpNVd3QttT+/eRKkpOSrEiyPMmRTex8YL8kS5K8M8lOSa5q7pcl2WFTfR9JkiRJkiRJkiRJkiSpU2y0Gls+C3y1qnYBzgA+3fZsBrAv8HJazU8Ah9HacWoW8EZgb4Cq+i1wDnBTkjOTHJNkwhByzQZ2BQ4CTkoyA5gHXFpVs6vqE8CbgU9V1WygG7il70skmZukJ0nP2vtWbcz3kCRJkiRJkiRJkiRJksYEG63Glr15+Ii+/6TVDNXrO1W1rqquA7ZrxvYFzmrG7wAu7A2uqjcCLwauAo4HTh1CrjOram1V3QlcDOzZT41XAP+U5B+Ap1fV/X0DqmpBVXVXVffEKdOG9QEkSZIkSZIkSZIkSZKkschGq7Gt2q7XtF2nz2//k6uWN7tQ/Tlw+Mbkasv5NeAvgfuB7yd50friZ82cxsr5c4aSWpIkSZIkSZIkSZIkSRqzbLQaWy4HjmqujwEuGyT+MuDwJBOSbAccAJBkapID2uJmAzcNkusS4MgkE5NMB/antRvW74Gte4OSPBP4VVV9mtbxhLsM4b0kSZIkSZIkSZIkSZKkcW1Spwt4FJuS5Ja2+48DbwdOTfIe4NfA6wbJ8W1axwOuAH4OXAmsorU71QlJvkhr56l7gWMHybWQ1tGFS2ntpHVCVd2R5DfAQ0mWAqcBWwJ/neRB4A7gvUN6W0mSJEmSJEmSJEmSJGkcS1UNHqUxK8nUqlqd5Am0dqDap6ru6HRdvbq7u6unp6fTZUiSJEmSJEmSJEmSJEn9SrK4qroHi3NHq/Hv3CTbAo8B3jeWmqwkSZIkSZIkSZIkSZKkzYWNVuNcVR3Q6RokSZIkSZIkSZIkSZKkzd2EThegoUlSSf6z7X5Skl8nOXeYeQ7ob06Sv0wybyRqbbf81lV0zVs00mklSZIkSZIkSZIkSZKkTcodrcaPe4Gdk2xVVfcDfw7cOpwESQb8966qc4BzNq5ESZIkSZIkSZIkSZIkafPkjlbjy/eAOc310cCZvQ+S7JXk8iTXNL87NuPHJjkryXeBC9qTJdmziX9mE/fZZvy0JJ9u8vwqyRHN+IQkn09ybZJzk5zX+0ySJEmSJEmSJEmSJEnanNloNb58HTgqyZbALsCVbc+uB/avqt2AfwU+2PZsb+C1VfWi3oEkLwBOBg6pql/1s9YMYF/g5cD8ZuwwoAuYBbyxyfsnksxN0pOkZ+19q4b9kpIkSZIkSZIkSZIkSdJY49GB40hVLUvSRWs3q/P6PJ4GfCXJDkABW7Q9+0FV/bbt/jnAAuDgqrptgOW+U1XrgOuSbNeM7Quc1YzfkeTCAepc0ORn8owdaqjvJ0mSJEmSJEmSJEmSJI1V7mg1/pwDfJS2YwMb7wMurKqdgVcAW7Y9u7dP7O3AA8Bu61lnTdt1+vwO2ayZ01g5f87ggZIkSZIkSZIkSZIkSdIYZqPV+HMq8N6qWt5nfBpwa3N97CA57gbmAB9McsAw1r4MODzJhGaXq+HMlSRJkiRJkiRJkiRJksYtG63Gmaq6pao+1d21B/oAACAASURBVM+jjwAfSvJjYOIQ8txJa+erzyV53hCX/zZwC7AC+CJwJbBqiHMlSZIkSZIkSZIkSZKkcStV1ekaNI4kmVpVq5M8AbgK2Keq7hgovru7u3p6ejZdgZIkSZIkSZIkSZIkSdIwJFlcVd2DxU3aFMVos3Jukm2BxwDvW1+TlSRJkiRJkiRJkiRJkrS5sNFKw1JVB3S6BkmSJEmSJEmSJEmSJGlTm9DpAjotyZOTfD3JL5Ncl+S8JM9KsmIIc1cPY53TktyYZEmSpUlevHGV97vGtkneMty4JE9J8q2Rrgdg+a2r6Jq3aDRSS5IkSZIkSZIkSZIkSZvMo7rRKkmAhcBFVbV9VT0X+Cdgu1Fa8j1VNRs4Djh5FPJvCwzaaNU3rqpuq6ojRqEeSZIkSZIkSZIkSZIkabPwqG60Ag4EHqyqPzY9VdUS4Obe+yTHJvls2/25SQ5ou/9YkquT/DDJ9CGuewUwsy3HHkkuTrI4yfeTzGjGL0ryySSXJ1mRZK9m/MQkx7fNX5GkC5gPbN/smnVSkqlNXVcnWZ7kkGZK37iu3h28kmyZ5MtN/DVJDmz7DmcnOT/JL5J8ZIjvKkmSJEmSJEmSJEmSJI17kzpdQIftDCzeiPmPBa6uqncn+Vfg34C/H8K8lwLfAUiyBfAZ4JCq+nWSI4EPAK/vXaOqXpBkf+DUpuaBzAN2bnbNIskk4NCquifJE4GfJDmnn7iuthxvBaiqWUmeDVyQ5FnNs9nAbsAa4IYkn6mqm+kjyVxgLsDEbYbaeyZJkiRJkiRJkiRJkiSNXY/2RquNtQ74RnN9OnD2IPEnNTtBPQl4fjO2I63mqR+0TjJkInB725wzAarqkiTbJNl2GPUF+GDTpLWO1i5agx2LuC+txi+q6vokNwG9jVY/rKpVAEmuA55O2+5fvapqAbAAYPKMHWoY9UqSJEmSJEmSJEmSJElj0qO90epa4IhBYh7ikUcsbrme2MGait5Dqxnr7cBXgD1oNUNdW1V7DzFnDaOmY4DpwB5V9WCSleuJ7ZX1PFvTdr2WIfz/mTVzGj3z5wwWJkmSJEmSJEmSJEmSJI1pEwYP2az9CJic5E29A0n2pLVTU6+VwOwkE5I8Fdir7dkEHm7UejVw2WALVtU64FPAhCQvAW4ApifZu1l/iyQ7tU05shnfF1jV7Ci1Eti9Gd8deEYT+3tg67a504D/bZqsDmx7r75x7S6h1aBFc2Tg05oaJUmSJEmSJEmSJEmSpEetR/WOVlVVSQ4FPplkHvAArSam49rCfgzcCCwHVgBXtz27F9gpyWJgFU1T1BDXfT9wQlV9P8kRwKeTTKP1b/JJWrttAfwuyeXANsDrm7FvA3+TZAnwU+DnTd7fJPlxkhXA94APA99N0gMsAa4fIO5zbeV9Hjg5yXJaO2cdW1VrmmMNJUmSJEmSJEmSJEmSpEelVA122p06JclFwPFV1dPpWjZUd3d39fSM2/IlSZIkSZIkSZIkSZK0mUuyuKq6B4t7tB8dKEmSJEmSJEmSJEmSJEmDelQfHTgaknwO2KfP8Keq6svDzVVVB7TlXUvr+MIAa4G/r6rLB6lldVVNHe66kiRJkiRJkiRJkiRJkh7JowPHifamqSQvAf6pql441DmdMnnGDjXjtZ9k5fw5nSxDkiRJkiRJkiRJkiRJ6pdHB27etgF+B5BkapIfJrk6yfIkh/QNHigmSVeSnyU5Jcm1SS5IslXz7M+S/HeSpc287Zvx9yT5aZJlSf59E76zJEmSJEmSJEmSJEmS1DEeHTh+bJVkCbAlMAN4UTP+AHBoVd2T5InAT5KcU4/cqqzfmObZDsDRVfWmJN8EDgdOB84A5lfVwiRbAhOSHNzE70XrCMNzkuxfVZe0F5pkLjAXYOI200f8Q0iSJEmSJEmSJEmSJEmbmo1W48f9VTUbIMnewFeT7Eyr4emDSfYH1gEzge2AO9rmDhQDcGNVLWmuFwNdSbYGZlbVQoCqeqBZ92DgYOCaJn4qrcarRzRaVdUCYAG0jg4cmdeXJEmSJEmSJEmSJEmSOsdGq3Goqq5odqaaDrys+d2jqh5MspLWrlftjllPzJq2uLXAVrQas/oT4ENV9cUReRFJkiRJkiRJkiRJkiRpnJjQ6QI0fEmeDUwEfgNMA/63aaA6EHh6P1OGEvNHVXUPcEuSVzbrTU4yBfg+8PokU5vxmUmetL5cs2ZOY+X8OcN8Q0mSJEmSJEmSJEmSJGlscUer8WOrJL1H/AV4bVWtTXIG8N0kPcAS4Pp+5g4lpq/XAF9M8l7gQeBVVXVBkucAVyQBWA38NfC/G/NikiRJkiRJkiRJkiRJ0liXqup0DdqMdXd3V09PT6fLkCRJkiRJkiRJkiRJkvqVZHFVdQ8W59GBkiRJkiRJkiRJkiRJkjQIG60kSZIkSZIkSZIkSZIkaRA2Wo2yJGuTLGn760pyQJJz22Len+T7SSY399OTPJjkbztXuSRJkiRJkiRJkiRJkqReNlqNvvuranbb38r2h0n+GdgHeGVVrWmGXwX8BDh6NAtLMmk08wMsv3UVXfMWjfYykiRJkiRJkiRJkiRJ0qiy0aqDkrwbeBnwiqq6v+3R0cC7gf+TZGZb/EuTXJ1kaZIfNmNTk3w5yfIky5Ic3oyvbpt3RJLTmuvTknw8yYXAh5PsleTyJNc0vzs2cROTfLQt79uSvDjJwra8f57k7FH7QJIkSZIkSZIkSZIkSdIYMeo7Gomtkixprm+sqkOb632AHYE9qqq9KeqpwJOr6qok3wSOBD6eZDpwCrB/Vd2Y5PHNlP8LrKqqWc38xw2hpmcBB1XV2iTbNDkfSnIQ8EHgcGAu8Axgt+bZ44HfAZ9LMr2qfg28Dvhy3+RJ5jbzmbjN9KF9JUmSJEmSJEmSJEmSJGkMc0er0dd+dOChbeP/DwhwcJ/4o4BvNtdf5+HjA58PXFJVNwJU1W+b8YOAz/VOrqrfDaGms6pqbXM9DTgryQrgE8BObXlPrqqHeterqgL+E/jrJNsCewPf65u8qhZUVXdVdU+cMm0I5UiSJEmSJEmSJEmSJEljmztadc6dwDHAD5P8pqoubMaPBrZLckxz/5QkO9Bqyqp+8gw03j62ZZ9n97Zdvw+4sKoOTdIFXDRI3i8D3wUeoNWw9VA/MZIkSZIkSZIkSZIkSdJmxR2tOqiqfg4cBpyeZHaSHYHHVtXMquqqqi7gQ7R2uboCeGGSZwC0HR14AfD3vTnbjg68M8lzkkwA2nfS6msacGtzfWzb+AXAm5NMal+vqm4DbgP+BThtsHecNXMaK+fPGSxMkiRJkiRJkiRJkiRJGtNstOqwqvop8DrgHFrNSwv7hHwbOLqqfg3MBc5OshT4RvP8/cDjkqxoxg9sxucB5wI/Am5fTwkfAT6U5MfAxLbxLwH/Ayxr8r667dkZwM1Vdd2wXlaSJEmSJEmSJEmSJEkap1LV3+lw0sCSfBa4pqr+Y7DY7u7u6unp2QRVSZIkSZIkSZIkSZIkScOXZHFVdQ8WN2lTFKPNR5LFwL3AuztdiyRJkiRJkiRJkiRJkrSp2GilYamqPTpdgyRJkiRJkiRJkiRJkrSpTeh0ARpZSVav59lTknxrhNY5IMm5g8Utv3UVXfMWjcSSkiRJkiRJkiRJkiRJUse4o9WjSFXdBhzR6TokSZIkSZIkSZIkSZKk8cYdrTZDaTkpyYoky5Mc2Yx3JVnRXB+b5Owk5yf5RZKPtM0/OMkVSa5OclaSqc34S5Ncn+Qy4LCOvJwkSZIkSZIkSZIkSZLUATZabZ4OA2YDuwIHASclmdFP3GzgSGAWcGSSpyZ5IvAvwEFVtTvQA7wryZbAKcArgP2AJw+0eJK5SXqS9Ky9b9VIvpckSZIkSZIkSZIkSZLUETZabZ72Bc6sqrVVdSdwMbBnP3E/rKpVVfUAcB3wdOD5wHOBHydZAry2GX82cGNV/aKqCjh9oMWrakFVdVdV98Qp00b2zSRJkiRJkiRJkiRJkqQOmNTpAjQqMsS4NW3Xa2n9fwjwg6o6+hEJk9lAjUx5kiRJkiRJkiRJkiRJ0vjijlabp0toHQU4Mcl0YH/gqiHO/QmwT5I/A0gyJcmzgOuBZyTZvok7eqAE7WbNnMbK+XOGV70kSZIkSZIkSZIkSZI0xthotRlJMonWLlULgWXAUuBHwAlVdcdQclTVr4FjgTOTLKPVePXs5njBucCiJJcBN438G0iSJEmSJEmSJEmSJEljU6o8DW5zkWRX4JSq2qvTtfTq7u6unp6eTpchSZIkSZIkSZIkSZIk9SvJ4qrqHizOHa02E0neDJwJ/Euna5EkSZIkSZIkSZIkSZI2N5M6XYBGRlWdDJzc6TokSZIkSZIkSZIkSZKkzZE7Wo1hSZ6c5OtJfpnkuiTnJXnWKKzz5iR/M9J5AZbfuoqueYvomrdoNNJLkiRJkiRJkiRJkiRJm4Q7Wo1RSQIsBL5SVUc1Y7OB7YCfN/cTq2rtxq7V7IYlSZIkSZIkSZIkSZIkaQDuaDV2HQg82N4EVVVLgIlJLkzyNWA5QJJ3JVnR/B3XjD02yaIkS5vxI5vx+c3uWMuSfLQZOzHJ8c31RUk+nOSqJD9Psl8zPiXJN5t530hyZZLuTfpFJEmSJEmSJEmSJEmSpA5xR6uxa2dg8QDP9gJ2rqobk+wBvA54HhDgyiQXA88EbquqOQBJpiV5PHAo8OyqqiTbDpB/UlXtleRlwL8BBwFvAX5XVbsk2RlYMlDhSeYCcwEmbjN9eG8tSZIkSZIkSZIkSZIkjUHuaDU+XVVVNzbX+/5/9u493M6yvPP490cSDgGMrSADVN2KCALBAAuVkwZLqZrWAaUiRQecSoodpWjRpuIBba1xYMQRixipghUFUbBKVEDkEE7CzoEkgKBiUAErjBoJhwDhnj/WG1hu987aO9mnJN/PdeVa73oO93O/a+fP+7of4OKqeqiqVgAXAQfR7nZ1SNOd6qCqWg78DngUODvJ64GHB4h/UfM5H+jpOOd8gKpaCiweKLmqmlNVrapqTZg8Za1fUpIkSZIkSZIkSZIkSRovLLQav24F9hlg7qGO5/S3oKrubPYvAT6W5INV9QTtblhfBw4DvjtA/JXN5yqe7nrW7zmSJEmSJEmSJEmSJEnSxsBCq/Hr+8BmSY5bPZBkX+CVfdZdAxyWZHKSLWlfDTgvyQ7Aw1X1JeA0YO8kWwFTqurbwInAtCHkcy3wxiaP3YCpg9k0dccpLJs9g2WzZwzhKEmSJEmSJEmSJEmSJGl8mdh9icZCVVWSw4FPJplF+8q/ZcA3+qxbkOQc4KZm6OyqWpjkz4FTkzwJPA68Hdga+M8km9PuUPWuIaR0JnBuksXAQtpXBy5f2/eTJEmSJEmSJEmSJEmS1iepqrHOQeuBJBOASVX1aJKdgCuAF1XVY2va12q1qre3d1RylCRJkiRJkiRJkiRJkoYqyfyqanVbZ0crDdZk4Mokk2h3w3p7tyIrSZIkSZIkSZIkSZIkaUNhoZUGpaoeBLpW7kmSJEmSJEmSJEmSJEkbok3GOgGNniT/Lcn5SX6S5LYk307yoiSPJFmY5PYkNyU5pmPPsUkqyZ92jB3ejB3R7cwl9yynZ9ZcembNHanXkiRJkiRJkiRJkiRJkkacHa02EkkCXAycW1VvasamAdsBP6mqvZqxFwAXJdmkqr7QbF8CHAVc0Xx/E3DLaOYvSZIkSZIkSZIkSZIkjSU7Wm08DgYer6qzVg9U1SLg552Lquou4N3ACR3D84CXJpmUZCvghcCikU9ZkiRJkiRJkiRJkiRJGh/saLXx2AOYP8i1C4BdO74X8D3gz4EpwDeB5w+0OclMYCbAhGdsuza5SpIkSZIkSZIkSZIkSeOKHa3Un/Qzdj7tKwPfBHxlTZurak5VtaqqNWHylJHIT5IkSZIkSZIkSZIkSRpVFlptPG4F9hnk2r2A2zsHquom2l2xtqmqO4c5N0mSJEmSJEmSJEmSJGlc8+rAjcf3gX9NclxVfQ4gyb7A5M5FSXqA04Az+onxT8CjQzl06o5T6J09Y23ylSRJkiRJkiRJkiRJksYNC602ElVVSQ4HPplkFu2CqWXAicBOSRYCmwMPAmdU1Rf6ifGdUUxZkiRJkiRJkiRJkiRJGjcstNqIVNW9wBv7mdpiDXvOAc7pZ/zY4cpLkiRJkiRJkiRJkiRJGu82GesEJEmSJEmSJEmSJEmSJGm8s9BKkiRJkiRJkiRJkiRJkrqw0GoDlOS/JTk/yU+S3Jbk20leNBa5LLlnOT2z5tIza+5YHC9JkiRJkiRJkiRJkiQNCwutNjBJAlwMXFVVO1XVbsD7gO0GuX/CSOYnSZIkSZIkSZIkSZIkrY8stNrwHAw8XlVnrR6oqkXAhCSXrB5L8ukkxzbPy5J8MMm1wHuT3NSxrifJ4uZ5nyRXJ5mf5NIk24/WS0mSJEmSJEmSJEmSJEljyUKrDc8ewPy12PdoVR1YVR8DNk3ygmb8SOCrSSYBZwBHVNU+wOeBj/YXKMnMJL1Jelc9vHwtUpEkSZIkSZIkSZIkSZLGl4ljnYDGjQs6nr8KvBGYTbvQ6khgF9pFXJe3bydkAnBff4Gqag4wB2Cz7XeukUtZkiRJkiRJkiRJkiRJGh0WWm14bgWO6Gf8CX6/g9nmfeYf6ni+ALgwyUVAVdWPkkwFbq2q/YaSzNQdp9A7e8ZQtkiSJEmSJEmSJEmSJEnjjlcHbni+D2yW5LjVA0n2pd2BarckmyWZAvzpQAGq6ifAKuADPN3p6g5g2yT7NTEnJdl9hN5BkiRJkiRJkiRJkiRJGlcstNrAVFUBhwN/luQnSW4FTgHupX0l4GLgPGBhl1AXAG9u9lBVj9HulPXxJLcAi4D9R+IdJEmSJEmSJEmSJEmSpPEm7bocaWS0Wq3q7e0d6zQkSZIkSZIkSZIkSZKkfiWZX1WtbuvsaCVJkiRJkiRJkiRJkiRJXVhoJUmSJEmSJEmSJEmSJEldWGg1SpKcnOTWJIuTLErysmZ8WZJt+ll/ffPZk2Rp8zw9ySX9rJ2epJL8TcfYXs3YSSP3Vt0tuWc5PbPm0jNr7limIUmSJEmSJEmSJEmSJK2TiWOdwMYgyX7AXwB7V9XKprBq0zXtqar9h3jMEuBI4N+b728CbhlqrpIkSZIkSZIkSZIkSZL+kB2tRsf2wANVtRKgqh6oqns7FyTZIsl3kxzXfF8xxDN+BmyeZLskAV4NfKcj/rQkNzYdtS5O8kdJXpzkpo41PUkWN8/7JLk6yfwklybZvhk/IcltTZzz1+K3kCRJkiRJkiRJkiRJktY7FlqNjsuA5yS5M8mZSV7ZZ34r4FvAl6vqc+twzteAvwL2BxYAKzvmvgj8Y1XtSbv71Yeq6nZg0yQvaNYcCXw1ySTgDOCIqtoH+Dzw0WbNLGCvJs7x/SWRZGaS3iS9qx5evg6vI0mSJEmSJEmSJEmSJI0PFlqNgqpaAewDzATuBy5IcmzHkv8EvlBVX1zHo75Ku9DqKOArqweTTAGeWVVXN0PnAq/o2PPG5vlI4AJgF2AP4PIki4D3A3/SrFkMnJfkzcAT/SVRVXOqqlVVrQmTp6zjK0mSJEmSJEmSJEmSJEljb+JYJ7CxqKpVwFXAVUmWAMcA5zTT1wGvSfLlqqp1OOOXSR4H/gz4e9qdrbq5ALgwyUXtEPWjJFOBW6tqv37Wz6BdpPU64ANJdq+qfguuAKbuOIXe2TOG/C6SJEmSJEmSJEmSJEnSeGJHq1GQZJckO3cMTQPu7vj+QeD/AWcOw3EfpH1F4KrVA1W1HPhNkoOaobcAVzdzPwFWAR+gXXQFcAewbZL9mvwnJdk9ySbAc6rqSuC9wDNpX3soSZIkSZIkSZIkSZIkbdDsaDU6tgLOSPJM2tft/Zj2NYKdTgQ+n+R/V9V71/agqrp+gKljgLOSTAbuAt7aMXcBcCrw/CbGY0mOAD7VXDs4EfgkcCfwpWYswOlV9du1zVWSJEmSJEmSJEmSJElaX2QdbqqTumq1WtXb2zvWaUiSJEmSJEmSJEmSJEn9SjK/qlrd1nl1oCRJkiRJkiRJkiRJkiR1YaHVeirJ6UlO7Ph+aZKzO77/nyTvXsP+ga4Y7FyzLMk2/YxPT7L/2uQtSZIkSZIkSZIkSZIkrY8stFp/XQ/sD5BkE2AbYPeO+f2B6wbaXFXrUig1ffXZ3Sy5Zzk9s+bSM2vuOhwnSZIkSZIkSZIkSZIkjS0LrdZf1/F0sdPuwFLgwSR/lGQz4MXAwiTvSXJzksVJPrx6c5IVzecmSc5McmuSS5J8O8kRHee8M8mCJEuS7JqkBzgeeFeSRUkOGoV3lSRJkiRJkiRJkiRJksbUxLFOQGunqu5N8kSS59IuuLoB2BHYD1gOLKbdeWpn4KVAgG8meUVVXdMR6vVADzAVeDZwO/D5jvkHqmrvJH8HnFRVb0tyFrCiqk7rL7ckM4GZABOese0wvbEkSZIkSZIkSZIkSZI0duxotX5b3dVqdaHVDR3frwcObf4tBBYAu9IuvOp0IHBhVT1ZVb8Eruwzf1HzOZ92QVZXVTWnqlpV1ZowecpQ30mSJEmSJEmSJEmSJEkad+xotX67nnZR1VTaVwf+HPgH4He0u1JNBz5WVZ9dQ4x0OWNl87kK/79IkiRJkiRJkiRJkiRpI2VHq/XbdcBfAL+uqlVV9WvgmbSvD7wBuBT4n0m2AkiyY5Jn94lxLfCGJJsk2Y52cVY3DwJbDybBqTtOYdnsGSybPWNQLyRJkiRJkiRJkiRJkiSNRxZard+WANsAN/YZW15VD1TVZcCXgRuSLAG+xh8WSH0d+AXtjlifBX4ALO9y7reAw5MsSnLQur+GJEmSJEmSJEmSJEmSNL6lqsY6B42xJFtV1YokzwJuAg6oql8OR+xWq1W9vb3DEUqSJEmSJEmSJEmSJEkadknmV1Wr27qJo5GMxr1LkjwT2BT45+EqspIkSZIkSZIkSZIkSZI2FBZaiaqaPtY5SJIkSZIkSZIkSZIkSePZJmOdwMYqyclJbk2yOMmiJC9LsizJNkOIMT3J/h3fT0lyTxNvaZLXjUz2kiRJkiRJkiRJkiRJ0sbFjlZjIMl+wF8Ae1fVyqa4atO1CDUdWAFc3zF2elWdluTFwLwkz66qJ9c56QEkmVhVTww0v+Se5fTMmgvAstkzRioNSZIkSZIkSZIkSZIkaUTZ0WpsbA88UFUrAarqgaq6t5l7Z5IFSZYk2RUgyR8n+UbT/erGJHsm6QGOB97VdLA6qPOAqrodeALYJsnzklzR7L8iyXOTTEhyV9qemeTJJK9ozpuX5IVJtkzy+SQ3J1mY5L8388cmuTDJt4DLRuH3kiRJkiRJkiRJkiRJksaUhVZj4zLgOUnuTHJmkld2zD1QVXsDnwFOasY+DCysqj2B9wFfrKplwFm0O1hNq6p5nQckeRnwJHA/8Olmz57AecCnqmoVcCewG3AgMB84KMlmwJ9U1Y+Bk4HvV9W+wMHAqUm2bI7YDzimql7V9+WSzEzSm6R31cPL1+mHkiRJkiRJkiRJkiRJksYDC63GQFWtAPYBZtIuhLogybHN9EXN53ygp3k+EPiPZu/3gWclmTJA+HclWQScBhxZVUW7KOrLzfx/NPEA5gGvaP59rBnfF7i5mT8UmNXEuwrYHHhuM3d5Vf16gPebU1WtqmpNmDxQmpIkSZIkSZIkSZIkSdL6Y+JYJ7CxajpKXQVclWQJcEwztbL5XMXTf5/0F2KA0KdX1Wndjm8+59G+fnAH4IPAe4DpwDUd576hqu7o3Nx0y3qoyxmSJEmSJEmSJEmSJEnSBsOOVmMgyS5Jdu4YmgbcvYYt1wBHN3un075e8HfAg8DWgzjyeuBNzfPRwLXN8w+A/YEnq+pRYBHwt7QLsAAuBd6ZJM3Zew3irN8zdccpLJs9g2WzZwx1qyRJkiRJkiRJkiRJkjRuWGg1NrYCzk1yW5LFwG7AKWtYfwrQatbO5unuV98CDk+yKMlBa9h/AvDWZv9bgL8HqKqVwM+BG5t182gXbi1pvv8zMAlYnGRp812SJEmSJEmSJEmSJEna6KRqoBvopHXXarWqt7d3rNOQJEmSJEmSJEmSJEmS+pVkflW1uq2zo5UkSZIkSZIkSZIkSZIkdWGhlSRJkiRJkiRJkiRJkiR1YaHVCEhyeJJKsuswx10xjLFOTDK54/u3kzxzuOJLkiRJkiRJkiRJkiRJGxILrUbGUcC1wJvGKoG0renveyLwVKFVVb22qn473HksuWc5PbPm0jNr7nCHliRJkiRJkiRJkiRJkkaNhVbDLMlWwAHA39BRaJXkvUmWJLklyexm7IVJvteMLUiyUzP+niQ3J1mc5MMDnPMHa5L0JLk9yZnAAuA5ST6TpDfJrR3rTgB2AK5McmUztizJNs3zu5Msbf6d2Cf255pYlyXZYkR+REmSJEmSJEmSJEmSJGmcmTjWCWyADgO+W1V3Jvl1kr2B7Zrxl1XVw0n+uFl7HjC7qi5OsjmwSZJDgZ2BlwIBvpnkFVV1zeoDBloD/AzYBXhrVf1ds/bkqvp1kgnAFUn2rKpPJXk3cHBVPdCZfJJ9gLcCL2ti/yDJ1cBvmjOPqqrjknwVeAPwpb4/QJKZwEyACc/Ydt1+TUmSJEmSJEmSJEmSJGkcsKPV8DsKOL95Pr/5fgjwhap6GKApfNoa2LGqLm7GHm3mD23+LaTdlWpX2gVOnda05u6qurFj7RuTLGjW7g7s1iX/A4GLq+qhqloBXAQc1Mz9tKoWNc/zgZ7+AlTVnKpqVVVrwuQpXY6TJEmSJEmSJEmSJEmSxj87Wg2jJM8CXgXskaSACUABX28+f2/5QGGAj1XVZ9d0VH9rkvQAD3V8fz5wErBvVf0myTnA5t1eYw1zKzueVwFeHShJkiRJkiRJkiRJkqSNgh2thtcRwBerwfV8DwAAIABJREFU6nlV1VNVzwF+Cvwa+J9JJgMk+eOq+h3wiySHNWObNfOXNmu3asZ3TPLsPucMZg3AM2gXXi1Psh3wmo65B4Gt+9lzDXBYkslJtgQOB+atxW8BwNQdp7Bs9gyWzZ6xtiEkSZIkSZIkSZIkSZKkMWdHq+F1FDC7z9jXgRcD3wR6kzwGfBt4H/AW4LNJPgI8DvxVVV2W5MXADUkAVgBvBn61OuAa1qzqPLiqbkmyELgVuAu4rmN6DvCdJPdV1cEdexY0na9uaobOrqqFTbcsSZIkSZIkSZIkSZIkaaOUqr432knDp9VqVW9v71inIUmSJEmSJEmSJEmSJPUryfyqanVb59WBkiRJkiRJkiRJkiRJktSFhVaSJEmSJEmSJEmSJEmS1IWFVmMkyaoki5LcmuSWJO9O0vXvkeTUZs+pw5DDKUlOap4/kuSQdY3Z15J7ltMzay49s+YOd2hJkiRJkiRJkiRJkiRp1Ewc6wQ2Yo9U1TSAJM8GvgxMAT7UZd/fAttW1crhTKaqPjic8SRJkiRJkiRJkiRJkqQNiR2txoGq+hUwE3hH2iY0natuTrI4yd8CJPkmsCXwgyRHJtk2ydebdTcnOaBZd0qSzye5KsldSU5YfVaSk5PckeR7wC4d4+ckOaJ5Xpbkw0kWJFmSZNdmfNsklzfjn01yd5JtRu2HkiRJkiRJkiRJkiRJksaIHa3Giaq6q7k68NnAfweWV9W+STYDrktyWVW9LsmKjk5YXwZOr6prkzwXuBR4cRNyV+BgYGvgjiSfAfYE3gTsRftvvwCYP0BKD1TV3kn+DjgJeBvtblvfr6qPJXk17eKwP5Bk5uq5Cc/Ydl1+FkmSJEmSJEmSJEmSJGlcsNBqfEnzeSiw5+oOU7SvFNwZ+Gmf9YcAuyWrt/GMJFs3z3Ob6wVXJvkVsB1wEHBxVT0MT3XIGshFzed84PXN84HA4QBV9d0kv+lvY1XNAeYAbLb9zrWGMyRJkiRJkiRJkiRJkqT1goVW40SSFwCrgF/RLrh6Z1Vd2mXbJsB+VfVIn1gAKzuGVvH033qwhU+r93fuzQBrJUmSJEmSJEmSJEmSpA3aJmOdgCDJtsBZwKerqmhfAfj2JJOa+Rcl2bKfrZcB7+iIM63LUdcAhyfZoul89ZdDTPVa4I3NWYcCf9Rtw9Qdp7Bs9gyWzZ4xxKMkSZIkSZIkSZIkSZKk8cOOVmNniySLgEnAE8B/AJ9o5s4GeoAFabenuh84rJ8YJwD/lmQx7b/lNcDxAx1YVQuSXAAsAu4G5g0x5w8DX0lyJHA1cB/w4BBjSJIkSZIkSZIkSZIkSeudtBsoSd0l2QxYVVVPJNkP+ExVrbGLVqvVqt7e3tFJUJIkSZIkSZIkSZIkSRqiJPOrqtVtnR2tNBTPBb6aZBPgMeC4Mc5HkiRJkiRJkiRJkiRJGhUWWmnQqupHwF5jnYckSZIkSZIkSZIkSZI02jYZ6wS0YVtyz3J6Zs2lZ9bcsU5FkiRJkiRJkiRJkiRJWmvDXmiVZFWSRUmWJrkwyeS1iHF2kt2a5/f1mbt+uHJt4k1M8q9JftTkvSjJycN5xhBymZ5keZKFSe5Ick2SvxjEvs2SfK/J/chhyOOqJF3vnZQkSZIkSZIkSZIkSZI2FiPR0eqRqppWVXsAjwHHDzVAVb2tqm5rvr6vz9z+w5Bjp38BdgCmVtU04CBgUt9FaRuNDmDzqmqvqtoFOAH4dJI/7bJnL2BS87tfMPIp/r4kE0b7TEmSJEmSJEmSJEmSJGk0jXTh0DzghQBJ3t10uVqa5MRmbMskc5Pc0owf2YxflaSVZDawRdOp6bxmbkXzmSSnNvuWdOyd3uz/WpIfJjkvSfpLrum2dRzwzqp6FKCqHqyqU5r5niS3JzkTWAA8J8lRzXlLk3y8I9aKjucjkpzTPJ+T5Kwk85LcOZgOVatV1SLgI8A7mljbJvl6kpubfwckeTbwJWBa8zvtlGSfJFcnmZ/k0iTbd/yuH09yU5PLQc34FknOT7I4yQXAFh3vcmiSG5IsaDqUbdWML0vywSTXAn/V53edmaQ3Se+qh5cP9nUlSZIkSZIkSZIkSZKkcWvECq2STAReAyxJsg/wVuBlwMuB45LsBbwauLeqXtJ0wPpuZ4yqmsXTHbKO7nPE64FpwEuAQ4BTVxcU0e7wdCKwG/AC4IAB0nwh8LOqenANr7IL8MWq2gt4HPg48Krm7H2THNblpwDoAV4JzADOSrL5IPastgDYtXn+v8DpVbUv8Abg7Kr6FfA22p2wpgE/A84AjqiqfYDPAx/tiDexql5K+/f5UDP2duDhqtqzWbsPQJJtgPcDh1TV3kAv8O6OWI9W1YFVdX5nwlU1p6paVdWaMHnKEF5VkiRJkiRJkiRJkiRJGp8mjkDMLZIsap7nAf9Ou5Dn4qp6CCDJRbSv6PsucFrTGeqSqpo3hHMOBL5SVauA/0pyNbAv8Dvgpqr6RXPWItqFTtd2C5jkrcDfA88CVl9ReHdV3dg87wtcVVX3N+vPA14BfKNL6K9W1ZPAj5LcRbtwalGXPU+l1fF8CLBbR4OuZyTZus/6XYA9gMubdROA+zrmL2o+59P+XWje4VMAVbU4yeJm/OW0i9Wua2JtCtzQEWvUrymUJEmSJEmSJEmSJEmSxsJIFFo90nRWespAV/dV1Z1Nt6vXAh9LcllVfWSQ5/Qbs7Gy43kVA7/nj4HnJtm6uTLwC8AXkiylXaAE8NAgz6yO574dq6rL9zXZC7i9ed4E2K+qHulc0OfnDXBrVe03QLzVv03f36W/nAJcXlVHDRDroQHGnzJ1xyn0zp7RbZkkSZIkSZIkSZIkSZI0ro3Y1YF9XAMclmRyki2Bw4F5SXagfWXdl4DTgL372ft4kkkDxDwyyYQk29LuynTTUJKqqodpd9z69Orr/JJMoN25qT8/AF6ZZJtm3VHA1c3cfyV5cZJNmvfr9FdJNkmyE+2rDO8YTH5J9gQ+APxbM3QZ8I6O+Wn9bLsD2DbJfs2aSUl273LUNcDRzfo9gD2b8RuBA5K8sJmbnORFg8ldkiRJkiRJkiRJkiRJ2pCMREerP1BVC5Kcw9OFUGdX1cIkfw6cmuRJ4HHaVwz2NQdYnGRBVR3dMX4xsB9wC+1uTO+tql8m2XWI6Z0M/DOwNMmDwCPAucC9wA593uO+JP8EXEm729O3q+o/m+lZwCXAz4GlwFYdW++gXZC1HXB8VT26hnwOSrIQmAz8Cjihqq5o5k4A/q252m8i7QKp4/vk+FiSI4BPJZnSrPskcOsazvwM7U5ei2lfaXhTE+v+JMcCX0myWbP2/cCda4glSZIkSZIkSZIkSZIkbXBSNZRb7DRUTYHZJVX1tbHOZSy0Wq3q7e0d6zQkSZIkSZIkSZIkSZKkfiWZX1WtbutG6+pASZIkSZIkSZIkSZIkSVpvjcrVgeNBkouB5/cZ/sequnQkz62qY/vJ5c+Bj/cZ/mlVHT6SuQxFkhVVtVWfsVOA44D7gc1pX6H4v6rqyYHiLLlnOT2z5gKwbPaMEctXkiRJkiRJkiRJkiRJGkkbTaHVeCpiaoq7RrTAawSdXlWnJdkEuAZ4Je2CK0mSJEmSJEmSJEmSJGmD5dWBWlub0u5q9ZuxTkSSJEmSJEmSJEmSJEkaaRZaaajelWQRcB9wZ1Ut6rsgycwkvUl6Vz28fPQzlCRJkiRJkiRJkiRJkoaZhVYaqtOrahrwbGDLJG/qu6Cq5lRVq6paEyZPGf0MJUmSJEmSJEmSJEmSpGFmoZXWSlU9DnwXeMVY5yJJkiRJkiRJkiRJkiSNtIljnYDWT0kC7A/8wdWBnabuOIXe2TNGJylJkiRJkiRJkiRJkiRphFhopYFMTvKLju+faD7fleTNwCRgMXDmqGcmSZIkSZIkSZIkSZIkjTILrdSvqhroWslTRjMPSZIkSZIkSZIkSZIkaTwYqJhGkiRJkiRJkiRJkiRJktSw0EqSJEmSJEmSJEmSJEmSurDQagQl2S7Jl5PclWR+khuSHJ5kepJLOtb9S5JLk2yW5KokrY65niRLhymfp85NcmySTzfPxyf5H8NxRl9L7llOz6y5T/2TJEmSJEmSJEmSJEmS1kcTxzqBDVWSAN8Azq2qv27Gnge8DvhNx7qTgQOA11bVyva20VVVZ436oZIkSZIkSZIkSZIkSdJ6xI5WI+dVwGOdRUxVdXdVnbH6e5J/AF4L/GVVPdItYNPdal6SBc2//Zvx6U0nrK8l+WGS85pCL5K8uhm7Fnj9AHFPSXJS83xVko8nuSnJnUkOasYnJ/lqksVJLkjyg87OW5IkSZIkSZIkSZIkSdKGzI5WI2d3YMEa5g8AdgH2qaoVfebOS7K68GpT4Mnm+VfAn1XVo0l2Br4CrC522qs5817gOuCAJL3A52gXff0YuGCQuU+sqpcmeS3wIeAQ4O+A31TVnkn2ABYNtDnJTGAmwIRnbDvIIyVJkiRJkiRJkiRJkqTxy45WoyTJvyW5JcnNzdCPgQCH9rP86KqaVlXTaHe8Wm0S8LkkS4ALgd065m6qql9U1ZO0i6B6gF2Bn1bVj6qqgC8NMt2Lms/5TRyAA4HzAapqKbB4oM1VNaeqWlXVmjB5yiCPlCRJkiRJkiRJkiRJksYvO1qNnFuBN6z+UlX/K8k2QG8z9F/A0cAVSf5fVV05iJjvava9hHaR3KMdcys7nlfx9N+21iL31bE642Qt4jB1xyn0zp6xNlslSZIkSZIkSZIkSZKkccOOViPn+8DmSd7eMTa5c0FV3Qm8HvhSkmmDiDkFuK/pWvUWYEKX9T8Enp9kp+b7UYPKvH/XAm8ESLIbMHUdYkmSJEmSJEmSJEmSJEnrFQutRkhzVd9hwCuT/DTJTcC5wD/2WXcz8Fbgmx0FUQM5EzgmyY3Ai4CHuuTwKDATmJvkWuDutXqZp8/eNsli2u+wGFi+DvEkSZIkSZIkSZIkSZKk9Uba9UDSmiWZAEyqqkebgrArgBdV1WNr2tdqtaq3t3dNSyRJkiRJkiRJkiRJkqQxk2R+VbW6rZs4GslogzAZuDLJJCDA27sVWUmSJEmSJEmSJEmSJEkbCq8O3MgkWdHP2ClJ7kmyKMmPklyUZLeO+auAW4F9q+olVbUn8Lf9xZIkSZIkSZIkSZIkSZI2RHa00mqnV9VpAEmOBL6fZGpV3d/M/xY4ALg2yTOB7QcTdMk9y+mZNfep78tmzxjerCVJkiRJkiRJkiRJkqRRYEcr/YGqugC4DPjrjuHzgTc1z68HLhrtvCRJkiRJkiRJkiRJkqSxYqGVBrIA2LXj+xXAK5JMoF1wdcGYZCVJkiRJkiRJkiRJkiSNAQutNJD0+b4KuBY4EtiiqpYNuDGZmaQ3Se+qh5ePYIqSJEmSJEmSJEmSJEnS6LDQSgPZC7i9z9j5wBnAV9e0sarmVFWrqloTJk8ZqfwkSZIkSZIkSZIkSZKkUTNxrBPQ+JPkDcChwD/0mZoHfAz4ymBjTd1xCr2zZwxjdpIkSZIkSZIkSZIkSdLos9Bq4zM5yS86vn+i+XxXkjcDWwJLgVdV1f2dG6uqgNNGJ01JkiRJkiRJkiRJkiRp/LDQaiNTVQNdF3nKGvZMH2B8q2FISZIkSZIkSZIkSZIkSRr3Biq6kSRJkiRJkiRJkiRJkiQ1LLSSJEmSJEmSJEmSJEmSpC4stFoPJDk5ya1JFidZlORlSU5MMnmY4u+Q5GvDEUuSJEmSJEmSJEmSJEnaEKWqxjoHrUGS/YBPANOramWSbYBNgeuBVlU9MKYJdrHZ9jvX9sd88vfGls2eMUbZSJIkSZIkSZIkSZIkSb8vyfyqanVbZ0er8W974IGqWgnQFFYdAewAXJnkSoAkn0nS23S++vDqzUmWJfnXJDc083snuTTJT5Ic36zpSbK0eT42yUVJvpvkR0n+d0esfs+QJEmSJEmSJEmSJEmSNnQWWo1/lwHPSXJnkjOTvLKqPgXcCxxcVQc3605uKuv2BF6ZZM+OGD+vqv2AecA5tAu1Xg58ZIAzpwFHAlOBI5M8ZxBnPCXJzKYgq3fVw8vX9r0lSZIkSZIkSZIkSZKkccNCq3GuqlYA+wAzgfuBC5Ic28/SNyZZACwEdgd265j7ZvO5BPhBVT1YVfcDjyZ5Zj+xrqiq5VX1KHAb8LxBnNGZ85yqalVVa8LkKUN5XUmSJEmSJEmSJEmSJGlcmjjWCai7qloFXAVclWQJcEznfJLnAycB+1bVb5KcA2zesWRl8/lkx/Pq7/39H+hcswqYOIgzJEmSJEmSJEmSJEmSpA2WhVbjXJJdgCer6kfN0DTgbqAH2Bp4AHgG8BCwPMl2wGtoF2YNp7U6Y+qOU+idPWOYU5EkSZIkSZIkSZIkSZJGl4VW499WwBnNFX9PAD+mfY3gUcB3ktxXVQcnWQjcCtwFXDfcSVTVLSN9hiRJkiRJkiRJkiRJkjReparGOgdtwFqtVvX29o51GpIkSZIkSZIkSZIkSVK/ksyvqla3dZuMRjKSJEmSJEmSJEmSJEmStD6z0EqSJEmSJEmSJEmSJEmSurDQapxJsirJoiRLk1yYZPIwxz82yaeb58OS7NYx95EkhwzneZIkSZIkSZIkSZIkSdKGYOJYJ6A/8EhVTQNIch5wPPCJETrrMOAS4DaAqvrgcB+w5J7l9Mya+3tjy2bPGO5jJEmSJEmSJEmSJEmSpBFlR6vxbR7wQoAkb05yU9Pt6rNJJjTjK5J8NMktSW5Msl0z/pdJfpBkYZLvrR5fLcn+wOuAU5uYOyU5J8kRzfw+Sa5OMj/JpUm2b8ZPSHJbksVJzh/F30KSJEmSJEmSJEmSJEkaMxZajVNJJgKvAZYkeTFwJHBA0+1qFXB0s3RL4MaqeglwDXBcM34t8PKq2gs4H3hvZ/yquh74JvCeqppWVT/pOHsScAZwRFXtA3we+GgzPQvYq6r2pN1tq7/cZybpTdK76uHl6/Q7SJIkSZIkSZIkSZIkSeOBVweOP1skWdQ8zwP+HZgJ7APcnARgC+BXzZrHaF//BzAf+LPm+U+AC5pOVJsCPx1CDrsAewCXN+dNAO5r5hYD5yX5BvCN/jZX1RxgDsBm2+9cQzhXkiRJkiRJkiRJkiRJGpcstBp/Hmm6Vj0l7Wqnc6vqn/pZ/3hVrS5mWsXTf9MzgE9U1TeTTAdOGUIOAW6tqv36mZsBvIL2tYMfSLJ7VT0xhNiSJEmSJEmSJEmSJEnSesdCq/XDFcB/Jjm9qn6V5I+Bravq7jXsmQLc0zwfM8CaB4Gt+xm/A9g2yX5VdUNzleCLgNuB51TVlUmuBf4a2Ar47UBJTN1xCr2zZ6zx5SRJkiRJkiRJkiRJkqTxbpOxTkDdVdVtwPuBy5IsBi4Htu+y7RTgwiTzgAcGWHM+8J4kC5Ps1HHeY8ARwMeT3AIsAvanfYXgl5IsARYCp1fVgEVWkiRJkiRJkiRJkiRJ0oYiT986Jw2/VqtVvb29Y52GJEmSJEmSJEmSJEmS1K8k86uq1W2dHa0kSZIkSZIkSZIkSZIkqQsLrSRJkiRJkiRJkiRJkiSpi4ljnYA2bEvuWU7PrLm/N7Zs9owxykaSJEmSJEmSJEmSJElaOxtNR6skq5IsSrI0yYVJJq9FjLOT7NY8v6/P3PXDlWsT76okdyS5Jcl1SXbpsv7EtXmnjv3Tk+zf8f34JP+jy55Tkpy0tmdKkiRJkiRJkiRJkiRJ64uNptAKeKSqplXVHsBjwPFDDVBVb6uq25qv7+szt38/W9bV0VX1EuBc4NQua08E1rrQCpgOPPUOVXVWVX1xHeJJkiRJkiRJkiRJkiRJG4yNqdCq0zzghQBJ3t10uVqa5MRmbMskc5tuUkuTHNmMX5WklWQ2sEXTIeu8Zm5F85kkpzb7lnTsnd7s/1qSHyY5L0kGme81Hfn+aZKFTezPJ9ksyQnADsCVSa5s1h2a5IYkC5oOXls148uSfLgZX5Jk1yQ9tAvP3tW800Gd3aqSHJfk5ub3+Hq3zllJZibpTdK76uHlg3xFSZIkSZIkSZIkSZIkafza6AqtkkwEXgMsSbIP8FbgZcDLgeOS7AW8Gri3ql7SdMD6bmeMqprF0x2yju5zxOuBacBLgEOAU5Ns38ztRbvz1G7AC4ADBpn2Xzb5bg6cAxxZVVOBicDbq+pTwL3AwVV1cJJtgPcDh1TV3kAv8O6OeA80458BTqqqZcBZwOnNO83rc/5FVbVv013rduBv1pRsVc2pqlZVtSZMnjLIV5QkSZIkSZIkSZIkSZLGr42p0GqLJItoFx39DPh34EDg4qp6qKpWABcBBwFLgEOSfDzJQVU1lLZMBwJfqapVVfVfwNXAvs3cTVX1i6p6ElgE9HSJdV6T8wHAScAuwE+r6s5m/lzgFf3sezntYq7rmv3HAM/rmL+o+Zw/iBwA9kgyL8kS4Ghg90HskSRJkiRJkiRJkiRJkjYYE8c6gVH0SFVN6xwY6Oq+qrqz6Xb1WuBjSS6rqo8M8pw1XQe4suN5Fd1//6Orqrcj32cNIYfLq+qoLnkMJgdod9E6rKpuSXIsMH2QeTB1xyn0zp4x2OWSJEmSJEmSJEmSJEnSuLQxdbTqzzXAYUkmJ9kSOByYl2QH4OGq+hJwGrB3P3sfTzJpgJhHJpmQZFvaHaduGqZ8fwj0JHlh8/0ttDtmATwIbN083wgcsHpd834v6hK7c39fWwP3Ne/b96pESZIkSZIkSZIkSZIkaYO3URdaVdUC2t2abgJ+AJxdVQuBqcBNzbV7JwP/0s/2OcDiJOf1Gb8YWAzcAnwfeG9V/XKY8n0UeCtwYXON35PAWR35fCfJlVV1P3As8JUki2kXXu3aJfy3gMOTLEpyUJ+5D9D+fS6nXewlSZIkSZIkSZIkSZIkbVRSVWOdgzZgrVarent7uy+UJEmSJEmSJEmSJEmSxkCS+VXV6rZuo+5oJUmSJEmSJEmSJEmSJEmDMXGsE9jYJbkYeH7H0EuAu4BHgSeAc4FPVtWT/ezdAfhUVR2R5FigVVXv6LOmB7ikqvboZ/9VwElV1Zvk28BfV9Vvh+O9Vltyz3J6Zs39g/Fls2cM5zGSJEmSJEmSJEmSJEnSiLLQaoxV1eGd35OsqKqdmudnA18GpgAf6rNuYlXdCxwxTHn8f/buPdzTsqob+Hcxg8KIjpmIOG85iXgEHGXjAdEg06yxkiQBTaXSKfON0LR3LC8PWTmprxrmachjKpgGZvKWoILimT04MIBKKaMBGXloEjmIw3r/2M/kz83es/cMM3vvmfl8rotrP899r3vd6/kxf67rvn9pR+QBAAAAAAAAAIDdkasDF7DuvjbJqiT/uyacVFXvr6p/THJOVS2vqktHlvxUVf1zVX2lqkYbsxZX1Tur6pKq+kBVLZm8V1VtrKq7Djm/VFWnVdVlVXVOVe07xBwx5PhsVb1q0t4AAAAAAAAAALDb0mi1wHX31zLx/+luw9Ajkjyju39uivCHJnlqkhVJfr2qxobx+yZZ292HJfnvJL83w7YHJ3lDdz8wyX8ledIw/vYkv9vdj0iyebrFVbWqqsaranzz9Ztm/EYAAAAAAAAAAFjoNFrtGmrk+dzu/s40ced297e7+4YkZyY5ahj/t+7+9PD87pHx6VzZ3euH53VJllfVnZPcsbs/M4y/d7rF3b22u8e6e2zRkqUzbAUAAAAAAAAAAAufRqsFrqrulYnTo64dhr6/lfCe5n268encNPK8Ocni/HizFwAAAAAAAAAA7FEWz3cBTK+q9k/y5iR/3d1dNWOv02Or6i5JbkjyxCS/NYz/dFU9ors/m+TEJJ/a1lq6+7tV9b2qenh3fy7JCbNZd+iypRlfs3JbtwMAAAAAAAAAgAXFiVYLz75Vtb6qLkvy0STnJHnZLNd+KsnfJlmf5O+7e3wY/1KSZ1TVJUnukuRN21nbbydZW1WfzcQJV5u2Mw8AAAAAAAAAAOxSqnumW+RgQlXt193XDc+rkxzY3X+wtTVjY2M9Pj6+tRAAAAAAAAAAAJg3VbWuu8dminN1INtiZVW9MBP/br6e5KT5LQcAAAAAAAAAAOaGRitmrbvfl+R9810HAAAAAAAAAADMNY1We6Cquq6795tm7uIkl3f3iSNj70jy5CQHdPf3hrG/SnJykv27+1vT7bXh6k1ZvvrsKec2rlm53d8AAAAAAAAAAABzaa/5LoCFo6run4l/E4+uqjtMmv7XJL86xO2V5JgkV89thQAAAAAAAAAAMD80WjHqKUn+Nsk5SX5l0tzpSY4fno9O8ukkP5yzygAAAAAAAAAAYB5ptGLU8Unel4mmqhMnzf1Lkv2r6ieGuTOmS1JVq6pqvKrGN1+/aacVCwAAAAAAAAAAc0WjFUmSqjoiyX9299eTfCzJQ4amqlFnJjkhycOSXDBdru5e291j3T22aMnSnVYzAAAAAAAAAADMFY1WbHFikvtV1cYkX01ypyRPmhRzRpKXJzm3u2+Z2/IAAAAAAAAAAGD+LJ7vAph/VbVXkl9Pclh3Xz2MHZPkRUn+Zktcd3+jqv4kyUdnm/vQZUszvmblDq4YAAAAAAAAAADmlkarPdOSqrpq5P01Sa7e0mQ1+GSSB1TVgaMLu/stc1EgAAAAAAAAAAAsJBqt9kDdPdWVka+ZFLM5yZYmq5OmybN8hxYGAAAAAAAAAAAL1FQNNwAAAAAAAAAAAIzQaAUAAAAAAAAAADADVwfuYarq7klel+SIJDcl2ZjklO6+oqqem+QVSQ7o7k1D/NFJzkvyzO5+6zD24CQXJXlBd796a/ttuHpTlq8+e8q5jWtW7ohPAgAAAAAAAACAnc6JVnuQqqokZyU5v7sP6u4HJPnjJAcMIScmuTDJsZOWbkhy/Mj7CUku3snlAgAAAAAAAADAgqHRas9yTJKbu/vNWwa6e313X1BVByXZL8mLMtFwNeobSfY7DvgRAAAgAElEQVSpqgOGZq3HJ/mnuSoaAAAAAAAAAADmm0arPcshSdZNM3diktOTXJDkvlV1t0nzH0jy60mOzMS1gTdNt0lVraqq8aoa33z9ptteNQAAAAAAAAAAzDONVmxxQpIzuvuWJGdmoqlq1N8NY1sasqbV3Wu7e6y7xxYtWbpTigUAAAAAAAAAgLmk0WrPclmSwycPVtVhSQ5Ocm5VbcxE09WPXR/Y3d9McnOSxyb52E6vFAAAAAAAAAAAFpDF810Ac+rjSf6iqp7V3aclSVUdkeSVSV7a3a/YElhVV1bVPSetf3GSu3X35qqa1YaHLlua8TUrd0z1AAAAAAAAAAAwTzRa7UG6u6vq2CSvq6rVSW5MsjHJ0UmePSn8rEycbPX5kfWfmZtKAQAAAAAAAABgYanunu8a2I2NjY31+Pj4fJcBAAAAAAAAAABTqqp13T02U9xec1EMAAAAAAAAAADArkyjFQAAAAAAAAAAwAwWz3cBe4KqOiDJa5M8PMl3k/wgySuH5+d39xOGuD9LckSSX0nykSQHJrlxiH9Wd6+fYZ9HJXlzkpuTnJjk/d19yM74ptnacPWmLF999pRzG9esnONqAAAAAAAAAABg+zjRaierqkrywSSf7O57dffhSU5I8r8mxf1JkkcmeWJ33zQMP7W7H5TkjUleNYvtnprk1d29IskNO+obAAAAAAAAAABgT6fRauf7uSQ/6O43bxno7q939+u3vFfVHyb5pSS/3N1TNUh9NsmykfjHVdVnq+qiqnp/Ve1XVc9M8uQkL66q94wurqpFVfWqqrqwqi6pqt8Zxo+tqo/WhAOr6oqquvtW4g+sqk9W1fqqunQ4QQsAAAAAAAAAAHZ7rg7c+R6Y5KKtzD8yyX2THN7d100T8/hMnIqVqrprkhcl+fnu/n5V/Z8kz+vuP62qo5J8uLs/UFXLR9b/dpJN3X1EVd0+yaer6pzuPquqnpTkOcMeL+nub1bVqqnik/xako90959X1aIkS6Yqdli/KkkW3Wn/mX4fAAAAAAAAAABY8DRazbGqekOSo5L8IMkLkvxrkp9I8rgkH5gU/p6qukOSRUkeMow9PMkDMtH8lCS3y8SJV1vzuCSHVdVxw/vSJAcnuTLJ7ye5NMnnuvv0GeIvTPK2qto7yQe7e/1Um3X32iRrk+T2Bx7cM9QGAAAAAAAAAAALnkarne+yJE/a8tLdzxlOpRofhv4jyVOTfKyqvt3d542sfWqSi5OsSfKGTJwoVUnO7e4Tt6GGSvL73f2RKeaWJbklyQFVtVd337K1+Kp6dJKVSf62ql7V3e/a2saHLlua8TUrt6FUAAAAAAAAAABYePaa7wL2AB9Psk9VPXtk7Meu3OvuKzLRRPXuqloxae7mTFwV+PCqun+SzyV5ZFXdO0mqaklV3WeGGj6S5NnDSVSpqvtU1R2qanGStyd5SpIvJXneDPH3THJtd5+W5K350SlbAAAAAAAAAACwW3Oi1U7W3V1VT0zy2qr6oyT/meT7Sf7PpLgLq+o3k3yoqo6ZNHdDVf3fJM/v7t+uqpOSnF5Vtx9CXpTkiq2U8TdJlie5qCbuG/zPJE9M8odJLujuC6pqfZILq+rsrcQfneQFVXVzkuuSPH17fhMAAAAAAAAAANjVVHfPdw3sxsbGxnp8fHzmQAAAAAAAAAAAmAdVta67x2aKc3UgAAAAAAAAAADADDRa7YGq6rppxp9eVZdW1WVVdXlVPX8Yf0dVXV9VdxyJ/auq6qq661zVDQAAAAAAAAAA82XxfBfAwlBVv5jklCSP6+5rqmqfJE8bCfnXJL+a5N1VtVeSY5JcPVPeDVdvyvLVZ281ZuOaldtdNwAAAAAAAAAAzAUnWrHFC5M8v7uvSZLuvrG7TxuZPz3J8cPz0Uk+neSHc1ohAAAAAAAAAADME41WbHFIknVbmf+XJPtX1U8kOTHJGXNSFQAAAAAAAAAALAAardgWZyY5IcnDklwwXVBVraqq8aoa33z9pjkrDgAAAAAAAAAAdhaNVmxxWZLDZ4g5I8nLk5zb3bdMF9Tda7t7rLvHFi1ZuiNrBAAAAAAAAACAebF4vgtgwXhFkldW1RO6+5tVdfskv9Pdp24J6O5vVNWfJPnobJMeumxpxtes3AnlAgAAAAAAAADA3NFotWdaUlVXjby/prtfU1UHJPloVVWSTvK2yQu7+y1zVSQAAAAAAAAAACwUGq32QN095ZWR3f32JG+fYvykaeKX79DCAAAAAAAAAABggZqy4QYAAAAAAAAAAIAf0WgFAAAAAAAAAAAwA1cHMqWq+pMkT0myOcktSb6b5CeS7Jdk/yRXDqG/192fmZciAQAAAAAAAABgjmi04laq6hFJnpDkId19U1XdNcntuvuaqjo6yfO7+wmzybXh6k1ZvvrsGeM2rll5W0oGAAAAAAAAAICdSqMVUzkwybe6+6Yk6e5vzXM9AAAAAAAAAAAwr/aa7wJYkM5J8lNVdUVVvbGqfnZbFlfVqqoar6rxzddv2kklAgAAAAAAAADA3NFoxa1093VJDk+yKsl/JnlfVZ20DevXdvdYd48tWrJ0J1UJAAAAAAAAAABzx9WBTKm7Nyc5P8n5VbUhyTOSvGM+awIAAAAAAAAAgPmi0Ypbqar7Jrmlu/9lGFqR5Ovbk+vQZUszvmblDqsNAAAAAAAAAADmg0YrprJfktdX1Z2T/DDJv2biGkEAAAAAAAAAANgjabTiVrp7XZIjp5k7PxNXCgIAAAAAAAAAwB5jr/kuAAAAAAAAAAAAYKHTaAUAAAAAAAAAADADjVYLUFVtrqr1I/8tr6qxqjp1Fms/M8P8ddOM/25VPX17awYAAAAAAAAAgN1Zdfd818AkVXVdd++3q+Weyu0PPLgPfMbrZhW7cc3KnVwNAAAAAAAAAAD8uKpa191jM8U50WoXUVVHV9WHh+eXVtXbqur8qvpaVZ08Enfd8PfAqvrkcCLWpVX1qJGYP6+qi6vqc1V1wEjO5w/P51fVX1bVF6rqii1rq2pJVf1dVV1SVe+rqs9X1Yz/yAAAAAAAAAAAYFen0Wph2nfk2sCzpom5X5JfSPLQJC+pqr0nzT8lyUe6e0WSByVZP4zfIcnnuvtBST6Z5FnT5F/c3Q9NckqSlwxjv5fku919WJKXJzl8qoVVtaqqxqtqfPP1m2b8WAAAAAAAAAAAWOgWz3cBTOmGoUFqa87u7puS3FRV1yY5IMlVI/MXJnnb0ID1we7e0mj1gyQfHp7XJXnsNPnPHIlZPjwfleSvkqS7L62qS6Za2N1rk6xNJq4OnOE7AAAAAAAAAABgwXOi1a7rppHnzZnUNNfdn0zy6CRXJ/nbqnr6MHVzd/d066bIPxpTt7VoAAAAAAAAAADYFTnRajdVVfdMcnV3n1ZVd0jykCTvuo1pP5XkyUnOq6oHJDl0pgWHLlua8TUrb+O2AAAAAAAAAAAwvzRa7b6OTvKCqro5yXVJnr718Fl5Y5J3DlcGfjHJJUk27YC8AAAAAAAAAACwoNWPbpGDrauqRUn27u4bq+qgJB9Lcp/u/sF0a8bGxnp8fHzOagQAAAAAAAAAgG1RVeu6e2ymOCdasS2WZOLawL2TVJJnb63JCgAAAAAAAAAAdhcarZi17v5ekhm79wAAAAAAAAAAYHej0Wo3UlXHJjkzyf27+8szxP5Nktd09+VVtTHJWHd/q6o+091HVtXyJEd293tvS00brt6U5avPnnX8xjUrb8t2AAAAAAAAAACwU+w13wWwQ52Y5FNJTpgpsLuf2d2XTzF+5PC4PMlTdmh1AAAAAAAAAACwi9JotZuoqv2SPDLJb2dotKqqvarqjVV1WVV9uKr+X1UdN8ydX1W3ugawqq4bHtckeVRVra+q51bVBVW1YiTu01V12E7/MAAAAAAAAAAAWABcHbj7eGKSf+7uK6rqO1X1kCT3ysTJVIcmuVuSLyV52yzzrU7y/O5+QpJU1XeSnJTklKq6T5Lbd/clUy2sqlVJViXJojvtv90fBAAAAAAAAAAAC4UTrXYfJyY5Y3g+Y3g/Ksn7u/uW7v5mkvNuQ/73J3lCVe2d5LeSvGO6wO5e291j3T22aMnS27AlAAAAAAAAAAAsDE602g1U1U8m+bkkh1RVJ1mUpJOctaP26O7rq+rcJL+a5MlJbnXtIAAAAAAAAAAA7K40Wu0ejkvyru7+nS0DVfWJJN9K8qSqemeS/ZMcneS9s8z5vSR3nDT2N0n+MckF3f2d2SQ5dNnSjK9ZOcstAQAAAAAAAABgYXJ14O7hxNz69Kq/T3KPJFcluTTJW5J8PsmmWea8JMkPq+riqnpuknT3uiT/neTtO6JoAAAAAAAAAADYVTjRajfQ3UdPMXZqklTVft193XC94BeSbJi8pruXjzzvN/y9OcljRnNW1T0y0Zx3zo7+BgAAAAAAAAAAWMg0Wu3+PlxVd05yuyQv7+5vbk+Sqnp6kj9P8rzuvmVHFggAAAAAAAAAAAudRqvd3FSnXW1nnncledeOyAUAAAAAAAAAALsajVbcSlXdPcnrkhyR5KYkG5OckuTiJF9JUkm+n+Q3u/srW8u14epNWb767O2uZeOaldu9FgAAAAAAAAAAdpS95rsAFpaqqiRnJTm/uw/q7gck+eMkByT5anev6O4HJXnnMA4AAAAAAAAAALs9jVZMdkySm7v7zVsGunt9kn+bFHenJN+dy8IAAAAAAAAAAGC+uDqQyQ5Jsm6auYOqan2SOyZZkuRhUwVV1aokq5Jk0Z323xk1AgAAAAAAAADAnHKiFdtiy9WBByU5JcnaqYK6e213j3X32KIlS+e2QgAAAAAAAAAA2Ak0WjHZZUkOn0Xch5I8eifXAgAAAAAAAAAAC4KrA5ns40n+oqqe1d2nJUlVHZGJqwJHHZXkqzMlO3TZ0oyvWbnjqwQAAAAAAAAAgDmk0Yof091dVccmeV1VrU5yY5KNmbgq8KCqWp+kkvwgyTPnrVAAAAAAAAAAAJhDGq24le6+JsmTp5jad65rAQAAAAAAAACAhWCv+S4AAAAAAAAAAABgodNoBQAAAAAAAAAAMINd/urAqrp7ktclOSLJTUk2Jjmlu6/YjlynJFnb3dfvgLqWJzmyu987xdyVSR7f3V8ZGXtdkmu6+5Vbyffh7j6kqsaSPL27T54ibmOSse7+1qTx67p7v5H3k4a4/z3bbxjddzbrk2TD1ZuyfPXZWwvZqo1rVm73WgAAAAAAAAAA2FF26ROtqqqSnJXk/O4+qLsfkOSPkxywnSlPSbJkmr0WbWOu5UmeMs3cGUlOGMm9V5LjkrxvNom7e3yqJqudYHlGvmEO9wUAAAAAAAAAgAVll260SnJMkpu7+81bBrp7fXdfUBNeVVWXVtWGqjo+Sarq6Ko6v6o+UFVfrqr3DLEnJ7lHkvOq6rwh9rqq+tOq+nySR1TVi6vqwiHn2qHRK1V176r6aFVdXFUXVdVBSdYkeVRVra+q506q+/SMNFoleXSSjd399apaXlUXDHkuqqojJ3/08A0fHp5/sqrOqaovVtVbktS2/ohV9Y6qOrWqPlNVX6uq44apH/uG0X0BAAAAAAAAAGBPsqs3Wh2SZN00c7+WZEWSByX5+SSvqqoDh7kHZ+L0qgckuVeSR3b3qUmuSXJMdx8zxN0hyaXd/bDu/lSSv+7uI7r7kCT7JnnCEPeeJG/o7gclOTLJvydZneSC7l7R3a8dLay7L0lyS1U9aBg6IRPNV0lybZLHdvdDkhyf5NQZfoOXJPlUdz84yYeS/PQM8dM5MMlRwzetGcam/YatqapVVTVeVeObr9+0neUAAAAAAAAAAMDCsas3Wm3NUUlO7+7N3f0fST6R5Ihh7gvdfVV335JkfSauyJvK5iR/P/J+TFV9vqo2JPm5JA+sqjsmWdbdZyVJd9/Y3dfPor7Tk5xQVYuT/GqS9w/jeyc5bdjj/ZloBtuaRyd597D32Um+O4u9t+iR5w929y3dfXm2/+rFDHWs7e6x7h5btGTpbUkFAAAAAAAAAAALwuL5LuA2uizJcdPMbe0KvZtGnjdn+t/hxu7enCRVtU+SNyYZ6+5/q6qXJtlnhn225vQk52SiAeyS7r52GH9ukv/IxElceyW5cRa5euaQ3FBVt+vuHwzvd0nyrZH50d9ke78JAAAAAAAAAAB2S7t6o9XHk/xFVT2ru09Lkqo6IsmSJJ9M8jtV9c5MNBU9OskLktxvK/m+l+SO+fEGpC32Gf5+q6r2y0SD1we6+7+r6qqqemJ3f7Cqbp9k0UiuKXX3V6vq25m4pu91I1NLk1zV3bdU1TOGXFvzySRPTfJnVfWLSX5imrhPJPmNJG+rqn2TPDnJH82Qe6vfMBuHLlua8TUrb0sKAAAAAAAAAACYd7v01YHd3UmOTfLYqvpqVV2W5KVJrklyVpJLklyciYasP+rub86Qcm2Sf6qq86bY67+SnJZkQ5IPJrlwZPppSU6uqkuSfCbJ3Ye9f1hVF1fVc6fZ7/RMNH6dNTL2xiTPqKrPJblPku/PUPPLkjy6qi5K8rgk35gm7g+S/FpVrU/yuSTv7+5PzpB7Nt8AAAAAAAAAAAC7vZroVYKdY2xsrMfHx+e7DAAAAAAAAAAAmFJVrevusZnidukTrQAAAAAAAAAAAOaCRisAAAAAAAAAAIAZLJ7vAphbVbU5yYZM/L+/MsnTuvu/Ruafm+QVSQ7o7k3D2NFJzkvyzO5+6zD24CQXJXlBd796uv02XL0py1efvd31blyzcrvXAgAAAAAAAADAjuJEqz3PDd29orsPSfKdJM+ZNH9ikguTHDtpfEOS40feT0hy8U6rEgAAAAAAAAAAFhCNVnu2zyZZtuWlqg5Ksl+SF2Wi4WrUN5LsU1UHVFUleXySf5qrQgEAAAAAAAAAYD5ptNpDVdWiJI9J8qGR4ROTnJ7kgiT3raq7TVr2gSS/nuTITFwbeNM0uVdV1XhVjW++ftMOrx0AAAAAAAAAAOaaRqs9z75VtT7Jt5PcJcm5I3MnJDmju29JcmYmmqpG/d0wtqUha0rdvba7x7p7bNGSpTu0eAAAAAAAAAAAmA8arfY8N3T3iiT3THK7JM9Jkqo6LMnBSc6tqo2ZaLr6sesDu/ubSW5O8tgkH5vDmgEAAAAAAAAAYF4tnu8CmB/dvamqTk7yD1X1pkw0Vb20u1+xJaaqrqyqe05a+uIkd+vuzVU14z6HLlua8TUrd2TpAAAAAAAAAAAw5zRa7cG6+4tVdXEmTq86IckvTgo5axj//Miaz8xdhQAAAAAAAAAAsDBUd893DezGxsbGenx8fL7LAAAAAAAAAACAKVXVuu4emylur7koBgAAAAAAAAAAYFem0QoAAAAAAAAAAGAGi+e7AOZeVW1OsiET//+vTPK07v6vYe4+SV6X5D5Jbh7ifj/J/ZP8wxC/V5Jrkzylu6/d2l4brt6U5avPvs01b1yz8jbnAAAAAAAAAACA7eVEqz3TDd29orsPSfKdJM9JkqraJ8nZSd7U3ffu7vsneVOS/Yd1FwzrDkty4ZZ1AAAAAAAAAACwu9NoxWeTLBuen5Lks939j1smu/u87r50dEFVVZI7JvnunFUJAAAAAAAAAADzyNWBe7CqWpTkMUneOgwdkmTdVpY8qqrWJ/nJJN9P8sfT5F2VZFWSLLrT/lOFAAAAAAAAAADALsWJVnumfYeGqW8nuUuSc2e5bsvVgT+V5O1JXjlVUHev7e6x7h5btGTpjqkYAAAAAAAAAADmkROt9kw3dPeKqlqa5MNJnpPk1CSXJfnZWeb4UJK/nyno0GVLM75m5XYXCgAAAAAAAAAAC4ETrfZg3b0pyclJnl9Veyd5b5Ijq+p/OqOq6vFVdegUy49K8tW5qRQAAAAAAAAAAOaXRqs9XHd/McnFSU7o7huSPCHJ71fVv1TV5UlOSnLtEP6oqlpfVRcneVqSP5yPmgEAAAAAAAAAYK65OnAP1N37TXr/5ZHnLyd5/BTL/iPJ0p1cGgAAAAAAAAAALEhOtAIAAAAAAAAAAJjBTm+0qqrrZhFzSlUtmYNallfVU0bex6rq1J2wz8aquus04xuq6pKq+kRV3XMn7L2iqn5pW+Oq6leqavWOrgcAAAAAAAAAAHYHC+XqwFOSvDvJ9bNdUFWLunvzNu6zPMlTkrw3Sbp7PMn4Nua4rY7p7m9V1cuSvCjJs3Zw/hVJxpL8v22J6+4PJfnQDq4lG67elOWrz94huTauWblD8gAAAAAAAAAAwLaas6sDq+roqjq/qj5QVV+uqvfUhJOT3CPJeVV13hD7uKr6bFVdVFXvr6r9hvGNVfXiqvpUkl8f8v1lVX2hqq6oqkcNccur6oJh/UVVdeRQxpokj6qq9VX13KGmDw9r7lJVHxxOm/pcVR02jL+0qt427PW1od4t3/TBqlpXVZdV1apt/Ek+m2TZSK7fGL5jfVW9paoWDePXVdX/Hb7jY1W1/zB+flWNDc93HX6b2yX50yTHD3mOr6qHVtVnquqLw9/7ThN3UlX99ZDvnsNelwx/f3oYf0dVnTrk+VpVHbeN3wwAAAAAAAAAALukOWu0Gjw4E6dXPSDJvZI8srtPTXJNJk56Oma4cu9FSX6+ux+SiROnnjeS48buPqq7zxjeF3f3Q4e8LxnGrk3y2GH98Um2XA+4OskF3b2iu187qbaXJflidx+W5I+TvGtk7n5JfiHJQ5O8pKr2HsZ/q7sPz8TJUCdX1U9uw2/x+CQfTJKquv9Q5yO7e0WSzUmeOsTdIclFw7d8YuQbb6W7f5DkxUneN3zj+5J8Ocmju/vBw9xfTBM36q+TvGv4Ld6TH/1+SXJgkqOSPCETjWsAAAAAAAAAALDbm+urA7/Q3VclSVWtz8RVfp+aFPPwTDRifbqqkuR2mTj9aYvJTUFnDn/XDfmSZO8kf11VW5qW7jOL2o5K8qQk6e6PV9VPVtXSYe7s7r4pyU1VdW2SA5JclYnmqmOHmJ9KcnCSb8+wz3lVdUAmmsFeNIw9JsnhSS4cvnnfYT5Jbhn55nePfO9sLU3yzqo6OEln4reZySOS/Nrw/LdJXjky98HuviXJ5cN33MpwuteqJFl0p/23sVwAAAAAAAAAAFh45rrR6qaR583T7F9Jzu3uE6fJ8f1pco7me26S/0jyoEyc2nXjLGqrKcZ60h7/s09VHZ3k55M8oruvr6rzk+wzi32OycQ3vCMT1/c9b9j7nd39wlms31LTD/OjE8m2tu/Lk5zX3cdW1fIk589ij+n2TH78t5jqN0t3r02yNkluf+DBPVUMAAAAAAAAAADsSua60Wo630tyxyTfSvK5JG+oqnt3979W1ZIk/6u7r9iGfEuTXNXdt1TVM5IsmrTPVD6Ziev6Xj40UX2ru/97OGFquj2+OzRZ3S8TJ3HNSnffUFWnJNlQVX+W5GNJ/qGqXtvd11bVXZLcsbu/nolmquOSnJHkKfnRCWAbM3EK1heG+S0mf+PSJFcPzydtJW7UZ5KckInTrJ6aW586NmuHLlua8TUrt3c5AAAAAAAAAAAsCHvNHDIn1ib5p6o6r7v/MxMNQadX1SWZaLy63zbme2OSZ1TV5zJxbeCWU7AuSfLDqrq4qp47ac1Lk4wNe65J8owZ9vjnTJxsdUkmTo363LYU2N3/nuT0JM/p7sszcY3gOUO+c5McOIR+P8kDq2pdkp/LxClYSfLqJM+uqs8kuetI6vOSPKCq1lfV8Zm49u8VVfXp/KjhbKq4UScn+c2hlqcl+YNt+TYAAAAAAAAAANjdVLeb3Rayqrquu/eb7zq219jYWI+Pj893GQAAAAAAAAAAMKWqWtfdYzPFLZQTrQAAAAAAAAAAABYsjVYL3K58mhUAAAAAAAAAAOwuNFrtwarqumnGn15Vl1bVZVV1eVU9fxh/R1VdWVXrq+rLVfWSua0YAAAAAAAAAADmx+L5LoCFpap+MckpSR7X3ddU1T5JnjYS8oLu/sAwfnlVvau7r5wu34arN2X56rN3eJ0b16zc4TkBAAAAAAAAAGA6Gq2Y7IVJnt/d1yRJd9+Y5LQp4vYZ/n5/rgoDAAAAAAAAAID54upAJjskybqtzL+qqtYnuSrJGd197eSAqlpVVeNVNb75+k07q04AAAAAAAAAAJgzGq3YVi/o7hVJ7p7kMVV15OSA7l7b3WPdPbZoydK5rxAAAAAAAAAAAHYwjVZMdlmSw2cK6u7rkpyf5KidXRAAAAAAAAAAAMy3xfNdAAvOK5K8sqqe0N3frKrbJ/md7j51NKiqFid5WJLXby3ZocuWZnzNyp1XLQAAAAAAAAAAzAGNVnu2JVV11cj7a7r7NVV1QJKPVlUl6SRvG4l5VVW9KMntknwsyZlzVy4AAAAAAAAAAMwPjVZ7sO6e8urI7n57krdPMX7Szq4JAAAAAAAAAAAWoikbbQAAAAAAAAAAAPgRjVYAAAAAAAAAAAAz0Gi1B6mqzVW1vqourap/rKo7D+PLq+qGqvpiVX2pqr5QVc8YWXdSVXVVPWZk7Nhh7Lj5+BYAAAAAAAAAAJhLi+e7AObUDd29Ikmq6p1JnpPkz4e5r3b3g4e5eyU5s6r26u63D/MbkpyY5GPD+wlJLp5pww1Xb8ry1WfvwE+YsHHNyh2eEwAAAAAAAAAApuNEqz3XZ5Msm2qiu7+W5HlJTh4ZviDJQ6tq76raL8m9k6zf6VUCAAAAAAAAAMAC4ESrPVBVLUrymCRv3UrYRUnuN/LeST6a5BeSLE3yoSQ/M03+VUlWJcmiO+2/AyoGAAAAAAAAAID55USrPcu+VbU+ybeT3CXJuVuJrSnGzsjElYEnJDl9uoXdvba7x7p7bNGSpbelXgAAAAAAAAAAWBA0Wu1ZbujuFUnumeR2SZ6zldgHJ/nS6EB3fyHJIUnu2t1X7LQqAQAAAAAAAABggXF14B6ouzdV1clJ/hIX4wwAACAASURBVKGq3jR5vqqWJ3l1ktdPsfyFSW6c7V6HLlua8TUrt7NSAAAAAAAAAABYGDRa7aG6+4tVdXEmrgG8IMlBVfXFJPsk+V6S13f326dY909zWykAAAAAAAAAAMw/jVZ7kO7eb9L7L4+87ruVde9I8o4pxk/aQaUBAAAAAAAAAMCCttd8FwAAAAAAAAAAALDQabQCAAAAAAAAAACYgasDmVJV3T3J65IckeSmJBuTnJLk4iRfGQl9aHf/YLo8G67elOWrz96JlW7dxjUr521vAAAAAAAAAAB2HxqtuJWqqiRnJXlnd58wjK1IckCSr3b3ivmsDwAAAAAAAAAA5pqrA5nKMUlu7u43bxno7vVJ/m3+SgIAAAAAAAAAgPmj0YqpHJJk3TRzB1XV+uG/N0wVUFWrqmq8qsY3X79p51UJAAAAAAAAAABzxNWBbKsZrw7s7rVJ1ibJ7Q88uOekKgAAAAAAAAAA2ImcaMVULkty+HwXAQAAAAAAAAAAC4UTrZjKx5P8RVU9q7tPS5KqOiLJkm1NdOiypRlfs3JH1wcAAAAAAAAAAHPKiVbcSnd3kmOTPLaqvlpVlyV5aZJr5rUwAAAAAAAAAACYJ060YkrdfU2SJ08xdchc1wIAAAAAAAAAAPPNiVYAAAAAAAAAAAAz0GgFAAAAAAAAAAAwA1cHslNtuHpTlq8+e77LmNHGNSvnuwQAAAAAAAAAABawXfJEq6raXFXrq+rSqnp/VS3ZSuydq+r3buN+J1XVPUbe/6aqHjDDmvOramwr8xurasPw3+VV9WdVdfth7h5V9YGR2NOr6pKqem5V3W/49i9W1UG35bsAAAAAAAAAAIDZ2SUbrZLc0N0ruvuQJD9I8rtbib1zktvUaJXkpCT/02jV3c/s7stvY84kOaa7D03y0CT3SrJ2yH9Ndx+XJFV19yRHdvdh3f3aJE9M8g/d/eDu/upsNqmqRTugVgAAAAAAAAAA2GPtqo1Woy5Icu8kqarnDadcXVpVpwzza5IcNJwC9aoh7gVVdeFwStTLhrHlVfWlqjqtqi6rqnOqat+qOi7JWJL3DDn2HT2tqqreVFXjw5qXbc8HdPd1mWgWe2JV3WWo5dJh+pwkdxv2fkmSU5I8s6rOG/b/jar6wjD/li1NVVV1XVX9aVV9PskjqurwqvpEVa2rqo9U1YFD3PlV9ZdDjiuq6lHD+KKqevVw4tYlVfX7w/iUeUZV1arhNxnffP2m7flJAAAAAAAAAABgQdmlG62qanGSX0yyoaoOT/KbSR6W5OFJnlVVD06yOslXhxOwXlBVj0tycCZOkVqR5PCqevSQ8uAkb+juByb5ryRP6u4PJBlP8tQhxw2TyviT7h5LcliSn62qw7bnW7r7v5NcOdQw6ldG6n9ZkjcneW13H1NV909yfJJHdveKJJuTPHVYd4ckl3b3w5J8PsnrkxzX3YcneVuSPx/ZY3F3PzQTTVwvGcZWJfmZJA/u7sMy0Wi29wx5tnzL2u4e6+6xRUuWbs/PAQAAAAAAAAAAC8ri+S5gO+1bVeuH5wuSvDXJs5Oc1d3fT5KqOjPJo5J8aNLaxw3/fXF43y8TzU3fSHJld2/Juy7J8lnU8uSqWpWJ3/LAJA9Icsl2fFOS1DbGPybJ4UkurKok2TfJtcPc5iR/PzzfN8khSc4d4hYl+feRPGcOf0e/+eeTvLm7f5gk3f2dqjpkhjwAAAAAAAAAALBb2lUbrW4YTnD6HzV0/sxCJXlFd79l0vrlSW4aGdqcical6RNV/UyS5yc5oru/W1XvSLLPLOuYnOuOmWhyuiLJbI+BqiTv7O4XTjF3Y3dvHom7rLsfMU2eLd+9OT/6N1FJeor9tpbnVg5dtjTja1bONhwAAAAAAAAAABakXfrqwEk+meSJVbWkqu6Q5NhMnHb1vSR3HIn7SJLfqqr9kqSqllXV3WbIPTnHFndK8v0km6rqgExcY7jNhlremOSD3f3dbVj6sSTHbam/qu5SVfecIu4rSfavqkcMcXtX1QNnyH1Okt8drmdMVd1lO/MAAAAAAMD/b+/ew/Q86zqBf39NOLQWhlNlMWADWqSSYrAjAgpWUWSviMhSpCgseIosha6wsEbdXTl4iIuCsioQEQteKAJyKETAChS1RyZN2rRoObRZbRHRFgZLApTw2z/mKTsOmcwkk3nfmcznc13v9b7P/dz3/fye6fyuJ33zy30DAACseqt1Rauv0d1XDitKXTE0vba7dydJVV1cVdckeU93v7CqTk9y6bAI1q1JnpaZ1Zzmc36SV1fVgSRfXc2pu6+qqt1Jrk1yfZKLjzDsDw4rcZ2Q5O1JXnokg7v7I1X1P5L8ZVWdkOS2JOcm+b9z+n2pqs5O8sqqmsjMf/ffHuKez2uTPDDJ1VV1W5I/6O7fPYp5AAAAAAAAAABg1avuubvDwbEzOTnZU1NT4w4DAAAAAAAAAAAOqap2dffkQv2Op60DAQAAAAAAAAAAlsVxs3XgSlVVlye505zmp3f33nHEM5+qOphkb2Z+J27ITIyfraqNw/GvdPf/HPreK8k/JXlNdz/ncPPuvWk6G7ftXM7QR2Lf9i3jDgEAAAAAAAAAgDGyotUy6+7v7O7Nc14rqshqcGCIbVOSW5KcO+vc9Ul+aNbxk5NcO8rgAAAAAAAAAABgnBRacSiXJtkw6/hAkr+rqtv3onxKkjePPCoAAAAAAAAAABgThVb8O1W1Lsljklww59SbkpxTVfdNcjDJJw8zx9aqmqqqqYP7p5cvWAAAAAAAAAAAGBGFVtzuxKrak+TmJPdIcuGc8+9N8gNJnprkzw43UXfv6O7J7p5cd9LEsgQLAAAAAAAAAACjpNCK2x3o7s1JTk1yxyTnzj7Z3V9KsivJf0vy56MPDwAAAAAAAAAAxmf9uANgZenu6ao6L8k7q+pVc07/VpIPdffNVbWo+c7YMJGp7VuOdZgAAAAAAAAAADBSVrTia3T37iRXJTlnTvu13f368UQFAAAAAAAAAADjY0UrkiTdffKc48fPOtx0iP7nJzl/eaMCAAAAAAAAAICVwYpWAAAAAAAAAAAAC1BoBQAAAAAAAAAAsABbB5KqOphkb2Z+H25I8vTu/mxVvT3J67v7HUO/65L8cXf/ynD850ne2N1vm2/uvTdNZ+O2nct+D6O2b/uWcYcAAAAAAAAAAMAIWdGKJDnQ3Zu7e1OSW5KcO7RfkuSRSVJV90xya5JHzBr3iKEPAAAAAAAAAAAc1xRaMdelSTYMny/OUGg1vL87ySk14/6ZKdD61BhiBAAAAAAAAACAkVJoxVdV1bokj0lywdC0K8mmqrpjZgqtLk1yXZLTh+OL55lna1VNVdXUwf3Tyx84AAAAAAAAAAAsM4VWJMmJVbUnyc1J7pHkwiTp7i8muTbJtyd5eJLLM1Ns9cjhdchtA7t7R3dPdvfkupMmRhA+AAAAAAAAAAAsr/XjDoAV4UB3b66qicxsD3huklcO5y5J8ugkd+nuz1TVZUmek+ShSV690MRnbJjI1PYtyxQ2AAAAAAAAAACMhhWt+Krunk5yXpIXVNUdhuaLk/xskquG46szs7rVN2ZmtSsAAAAAAAAAADjuKbTi3+nu3ZkpqjpnaLokyQMys2VguvvLST6dZKq7vzKWIAEAAAAAAAAAYMRsHUi6++Q5x4+f9fnTSWrO+bNGExkAAAAAAAAAAKwMVrQCAAAAAAAAAABYgEIrAAAAAAAAAACABdg6cI2rqicmeVuS07v774/1/Htvms7GbTuP9bTHlX3bt4w7BAAAAAAAAAAAFmBFK56a5G+TnDPuQAAAAAAAAAAAYKVSaLWGVdXJSb4ryU9lKLSqqhOq6ver6tqqendV/UVVnT2cO7OqPlRVu6rqfVV1nzGGDwAAAAAAAAAAI6PQam37kSTv7e6PJrmlqr49yX9KsjHJGUl+OskjkqSq7pDk/yQ5u7vPTPK6JL96qEmramtVTVXV1MH908t/FwAAAAAAAAAAsMzWjzsAxuqpSX57+Pym4fgOSd7S3V9J8qmq+uBw/luSbEpyYVUlybok/3SoSbt7R5IdSXKn+5zWyxY9AAAAAAAAAACMiEKrNaqq7pnk+5JsqqrOTOFUJ3n7fEOSXNvdjziS65yxYSJT27csKVYAAAAAAAAAABg3WweuXWcneUN3n9rdG7v7fkluSPKvSZ5UVSdU1b2TnDX0vy7JKVX11a0Eq+rB4wgcAAAAAAAAAABGTaHV2vXUfO3qVX+e5BuS3JjkmiSvSXJ5kunu/lJmirN+o6quSrInySNHFy4AAAAAAAAAAIyPrQPXqO4+6xBtr0ySqjq5u28dthe8Isne4fyeJI8eZZwAAAAAAAAAALASKLTiUN5dVXdLcsckL+3uT407IAAAAAAAAAAAGCdbB/JVVXVrMrPaVXdvzszWgL9VVTWcf0RVdVXddzieqKpbqsrvEQAAAAAAAAAAxzUrWjGv7v5sVX0qyelJPpKZwqvdw/ubkzw8yeXd/ZX55th703Q2bts5inCPO/u2bxl3CAAAAAAAAAAADKxExEIuzkxhVYb3V8w5vmQcQQEAAAAAAAAAwCgptGIhl+T/F1Y9IMlbkkwOx4/MTCEWAAAAAAAAAAAc1xRasZCLkzyyqu6fZF93fyFJVdXJSc5McsXcAVW1taqmqmrq4P7pEYcLAAAAAAAAAADHnkIrDqu7P5bk7kken+TSoXlXkp9IckN333qIMTu6e7K7J9edNDG6YAEAAAAAAAAAYJmsH3cArAqXJvmvSZ456/hXkvzFQgPP2DCRqe1bli8yAAAAAAAAAAAYAStaMdtJVXXjrNfzh/aLk9wvydRwfGmSByS5ZBxBAgAAAAAAAADAqFnRiq/q7kMW3nX3y5K8bNbxviQ1orAAAAAAAAAAAGDsrGgFAAAAAAAAAACwAIVWAAAAAAAAAAAAC7B1IF+jqg4m2ZuZ348bkjy9uz9bVRuT/F2S62Z1f1h3f2nkQQIAAAAAAAAAwAgptOJQDnT35iSpqtcnOTfJrw7nPnH7ucXYe9N0Nm7buQwhciT2bd8y7hAAAAAAAAAAAFY1WweykEuTbBh3EAAAAAAAAAAAME4KrZhXVa1L8pgkF8xq/qaq2jO8fm+ecVuraqqqpg7unx5JrAAAAAAAAAAAsJxsHcihnFhVe5JsTLIryYWzzi24dWB370iyI0nudJ/TermCBAAAAAAAAACAUbGiFYdyYCimOjXJHZOcO+Z4AAAAAAAAAABgrKxoxby6e7qqzkvyzqp61dHMccaGiUxt33KMIwMAAAAAAAAAgNGyohWH1d27k1yV5JxxxwIAAAAAAAAAAONiRSu+RnefPOf48bMON404HAAAAAAAAAAAGDsrWgEAAAAAAAAAACxAoRUAAAAAAAAAAMACFFpxSFV16yHavqWqLqqqPVX1d1W1YxyxAQAAAAAAAADAqK0fdwCsKq9M8orufmeSVNUZCw3Ye9N0Nm7bueyBcXzbt33LuEMAAAAAAAAAANY4K1pxJO6T5MbbD7p77xhjAQAAAAAAAACAkVFoxZF4RZIPVNV7qup5VXW3Q3Wqqq1VNVVVUwf3T484RAAAAAAAAAAAOPYUWrFo3f1HSU5P8pYkZyW5rKrudIh+O7p7srsn1500MeIoAQAAAAAAAADg2FNoxRHp7k929+u6+wlJvpxk07hjAgAAAAAAAACA5bZ+3AGwelTV45K8v7tvq6r/kOSeSW463JgzNkxkavuWkcQHAAAAAAAAAADLRaEV8zmpqm6cdfzyJPdN8jtV9YWh7YXd/anRhwYAAAAAAAAAAKOl0IpD6u75tpV8/kgDAQAAAAAAAACAFWC+YhoAAAAAAAAAAAAGCq0AAAAAAAAAAAAWYOvANaaqDibZm5n/9jckeXp3f3Y498Akv53kgUluG/o9N8npST6Y5Ke7+w+Hvg9NcmWSF3b3b853vb03TWfjtp3Ld0OwSPu2bxl3CAAAAAAAAADAKmZFq7XnQHdv7u5NSW5Jcm6SVNWdk+xM8qru/ubuPj3Jq5KcMozbm+Qps+Y5J8lVowsbAAAAAAAAAADGR6HV2nZpkg3D5x9Lcml3v+v2k939we6+Zjj8hyR3rqp7V1UleVyS94w0WgAAAAAAAAAAGBOFVmtUVa1L8pgkFwxNm5LsWmDYW5M8OckjM7Nt4BfnmXtrVU1V1dTB/dPHKGIAAAAAAAAAABgfhVZrz4lVtSfJzUnukeTCIxj75swUWj01yZ/O16m7d3T3ZHdPrjtpYknBAgAAAAAAAADASqDQau050N2bk5ya5I5Jzh3ar01y5uEGdvenktyW5AeSvH85gwQAAAAAAAAAgJVk/bgDYDy6e7qqzkvyzqp6VZI/SfILVbWlu3cmSVU9LslNc4b+ryRf390Hq2rB65yxYSJT27cc4+gBAAAAAAAAAGC0rGi1hnX37iRXJTmnuw8k+aEkz62qj1XVR5I8M8mn54y5pLvfMfJgAQAAAAAAAABgjKq7xx0Dx7HJycmempoadxgAAAAAAAAAAHBIVbWruycX6mdFKwAAAAAAAAAAgAUotAIAAAAAAAAAAFjA+nEHwMpVVbd298lz2l6U5GeS/MvQ9N7u3jbfHHtvms7GbTuXL0g4xvZt3zLuEAAAAAAAAACAFUihFUfjFd39m+MOAgAAAAAAAAAARsXWgQAAAAAAAAAAAAtQaMXReF5V7RlePzj3ZFVtraqpqpo6uH96HPEBAAAAAAAAAMAxZetAjsZhtw7s7h1JdiTJne5zWo8sKgAAAAAAAAAAWCZWtAIAAAAAAAAAAFiAFa1YVmdsmMjU9i3jDgMAAAAAAAAAAJZEoRWHc1JV3Tjr+OVjiwQAAAAAAAAAAMZIoRXz6m5bSwIAAAAAAAAAQBKFNAAAAAAAAAAAAAtQaAUAAAAAAAAAALAAWweOUFUdTLI3SSU5mOQ53X3JAmNu7e6Tq2pjknd396ah/WFJfjPJvZN0kr9Ncl537z/CmO6W5Me6+/eP8Ha+Gtvh+uy9aTobt+080qlhzdm3fcu4QwAAAAAAAAAADsOKVqN1oLs3d/e3JfmFJL9+NJNU1b2TvCXJz3f3tyQ5Pcl7k9zlKKa7W5Jnz3OddUcTHwAAAAAAAAAAHG8UWo3PXZN8Jkmq6uSqen9VXVlVe6vqCQuMPTfJ67v70iTpGW/t7n+uqntU1Tuq6uqquqyqHjJc40VV9bqquqiqrq+q84a5tif5pqraU1Uvq6qzquqDVfUnmVl9K1X1/Kq6Znj93HL8MAAAAAAAAAAAYCWzdeBonVhVe5LcOcl9knzf0P6FJE/s7s9V1b2SXFZVF3R3zzPPpiSvn+fci5Ps7u4fqarvS/KGJJuHcw9K8r2ZWfnquqp6VZJtSTZ19+YkqaqzkjxsaLuhqs5M8hNJvjMzWx5eXlUf6u7d891kVW1NsjVJ1t31lMP/RAAAAAAAAAAAYBWwotVo3b514IOSPC7JG6qqMlPA9GtVdXWSv0qyIcm9j/Ia353kj5Okuz+Q5J5VNTGc29ndX+zuf03y6cNc44ruvmHWfG/v7s93961J3pbkUYcLoLt3dPdkd0+uO2nicF0BAAAAAAAAAGBVUGg1JsO2f/dKckqSHx/ezxxWlvrnzKx6NZ9rk5w5z7k61OWG9y/OajuY+Vc0+/wC8wEAAAAAAAAAwJpi68AxqaoHJVmX5OYkE0k+3d23VdX3Jjl1geG/m+SKqtrZ3ZcP8z0tM6th/XVmCrdeOmwD+K/DloTzzfVvmdlKcD5/neT8qtqemaKrJyZ5+iJuMUlyxoaJTG3fstjuAAAAAAAAAACwIim0Gq0Tq2rP8LmSPKO7D1bVG5O8q6qmkuxJ8veHm6S7/7mqzknym1X19Um+kpmCqLcleVGSPxq2Idyf5BkLzHVzVV1cVdckeU+SnXPOX1lV5ye5Ymh6bXfvXvQdAwAAAAAAAADAcaC6e+FecJQmJyd7ampq3GEAAAAAAAAAAMAhVdWu7p5cqN8JowgGAAAAAAAAAABgNVNoBQAAAAAAAAAAsID14w6ApKoOJtmbmf8eNyR5end/9hjOf1aSL3X3JcPxs5Ls7+43HKtrzGfvTdPZuG3ncl8GALJv+5ZxhwAAAAAAAAAcx6xotTIc6O7N3b0pyS1Jzj3G85+V5JG3H3T3q0dRZAUAAAAAAAAAAMcLhVYrz6VJNtx+UFUvrKoPV9XVVfXiWe3vqKpdVXVtVW2d1f64qrqyqq6qqvdX1cYkz0ryvKraU1WPqqoXVdULhv6bq+qyYf63V9Xdh/aLquo3quqKqvpoVT1qaH/w0LZnGHPaSH4qAAAAAAAAAAAwRgqtVpCqWpfkMUkuGI4fm+S0JA9LsjnJmVX16KH7T3b3mUkmk5xXVfesqlOS/EGSJ3X3tyV5cnfvS/LqJK8YVs36mzmXfUOSn+/uh2Rm+8JfnnVufXc/LMnPzWp/VpLf6e7Nw7VvPMR9bK2qqaqaOrh/eik/EgAAAAAAAAAAWBEUWq0MJ1bVniQ3J7lHkguH9scOr91JrkzyoMwUXiUzxVVXJbksyf2G9ocn+evuviFJuvuWw120qiaS3K27PzQ0vT7Jo2d1edvwvivJxuHzpUl+sap+Psmp3X1g7rzdvaO7J7t7ct1JE4u4fQAAAAAAAAAAWNnWjzsAkiQHunvzUPj07iTnJnllkkry6939mtmdq+qsJN+f5BHdvb+qLkpy56F/H8O4vji8H8zwu9Ldf1JVlyfZkuR9VfXT3f2B+SY4Y8NEprZvOYYhAQAAAAAAAADA6FnRagXp7ukk5yV5QVXdIcn7kvxkVZ2cJFW1oaq+PslEks8MRVYPysxKVsnMalPfU1X3H/rfY2j/tyR3med6n6mqRw1NT0/yobn9ZquqByS5vrtfmZktDh9y1DcMAAAAAAAAAACrhBWtVpju3j1sCXhOd/9xVZ2e5NKqSpJbkzwtyXuTPKuqrk5yXWa2D0x3/0tVbU3ytqo6Icmnk/xAkncleWtVPSHJc+dc8hlJXl1VJyW5PslPLBDiU5I8rapuS/KpJC9Z8k0DAAAAAAAAAMAKV93Hcqc5+PcmJyd7ampq3GEAAAAAAAAAAMAhVdWu7p5cqJ+tAwEAAAAAAAAAABag0AoAAAAAAAAAAGAB68cdACtLVd3a3SfPaXtRkl9Oclp3f3xoe16Slyf5ju6ed2/AvTdNZ+O2ncsYMQAsbN/2LeMOAQAAAAAAAFjlrGjFYu1Ncs6s47OTfGRMsQAAAAAAAAAAwEgptGKx3pHkCUlSVQ9IMp3kX8YaEQAAAAAAAAAAjIhCKxbrc0n+sao2JXlqkj+br2NVba2qqaqaOrh/emQBAgAAAAAAAADAclFoxZF4U2a2D/yRJG+fr1N37+juye6eXHfSxMiCAwAAAAAAAACA5bJ+3AGwqrwrycuSTHX356pqwQFnbJjI1PYtyx4YAAAAAAAAAAAsJ4VWLFp3H6iqn0/y0XHHAgAAAAAAAAAAo6TQirlOqqobZx2/fPbJ7n7TiOMBAAAAAAAAAICxU2jFv9PdJxxB37OWMRQAAAAAAAAAAFgxFl1UAwAAAAAAAAAAsFYptFqhqupgVe2pqmuq6i1VddJh+t6tqp69xOs9s6q+Ydbxa6vqWxcYc1FVTS7lugAAAAAAAAAAsBrYOnDlOtDdm5Okqt6Y5FlJXj5P37sleXaS31/C9Z6Z5Jokn0yS7v7pJcz1VXtvms7GbTuPxVQAACyzfdu3jDsEAAAAAACAFcuKVqvD3yT55iSpqucPq1xdU1U/N5zfnuSbhhWwXjb0e2FVfbiqrq6qFw9tG6vq76rqD6rq2qr6y6o6sarOTjKZ5I3DHCfOXq2qql5VVVPDmBeP/O4BAAAAAAAAAGDMFFqtcFW1Psl/TLK3qs5M8hNJvjPJw5P8TFU9NMm2JJ/o7s3d/cKqemyS05I8LMnmJGdW1aOHKU9L8nvd/eAkn03ypO5+a5KpJD8+zHFgThi/1N2TSR6S5Huq6iELxLx1KMyaOrh/+hj8FAAAAAAAAAAAYLwUWq1cJ1bVnswUQP1Dkj9M8t1J3t7dn+/uW5O8LcmjDjH2scNrd5IrkzwoMwVWSXJDd+8ZPu9KsnERsfxoVV05zPfgJN96uM7dvaO7J7t7ct1JE4uYHgAAAAAAAAAAVrb14w6AeR3o7s2zG6qqFjm2kvx6d79mzviNSb44q+lgkhMPO1HV/ZO8IMl3dPdnqur8JHdeZBwAAAAAAAAAAHBcUGi1uvx1kvOrantmiqmemOTpSf4tyV1m9XtfkpdW1Ru7+9aq2pDktgXmnjvH7e6a5PNJpqvq3pnZxvCixQZ8xoaJTG3fstjuAAAAAAAAAACwIim0WkW6+8phRakrhqbXdvfuJKmqi6vqmiTv6e4XVtXpSS4dFsG6NcnTMrOC1XzOT/LqqjqQ5BGzrnlVVe1Ocm2S65NcfGzvCgAAAAAAAAAAVr7q7nHHwHFscnKyp6amxh0GAAAAAAAAAAAcUlXt6u7JhfqdMIpgAAAAAAAAAAAAVjOFVgAAAAAAAAAAAAtQaLWKVdV9q+qdVfWxqvpEVf1OVd1xma956/C+saquWc5rAQAAAAAAAADASrF+3AFwdKqqkrwtyau6+wlVtS7JjiS/muSFS5h3fXd/+RiFmb03TWfjtp3HajoAAAAAAAAAgFVj3/Yt4w6BY8iKVqvX9yX5Qnf/UZJ098Ekz0vyk1X14ap68O0dq+qiqjqzqr6uql43nN9dVU8Yzj+zqt5SVe9K8pdVdXJVvb+qrqyqvbf3AwAAAAAAAACAtcqKVqvXg5Psmt3Q3Z+rqn9I8u4kP5rkl6vqPkm+obt3VdWvJflAd/9kVd0tyRVV9VfD8EckeUh331JV65M8cZjvXkkuq6oLursXE1hVbU2yNUnW3fWUY3GvAAAAAAAAAAAwVla0Wr0qdEvHeAAAB95JREFUyaEKnyrJRUmePBz/aJK3DJ8fm2RbVe0Z+tw5yTcO5y7s7ltmzfFrVXV1kr9KsiHJvRcbWHfv6O7J7p5cd9LEom8IAAAAAAAAAABWKitarV7XJnnS7IaqumuS+yX5cJKbq+ohSZ6S5Gdv75LkSd193Zxx35nk87OafjzJKUnO7O7bqmpfZoqyAAAAAAAAAABgTVJotXq9P8n2qvrP3f2GqlqX5LeSnN/d+6vqTUn+e5KJ7t47jHlfkudW1XO7u6vqod29+xBzTyT59FBk9b1JTj3aIM/YMJGp7VuOdjgAAAAAAAAAAKwItg5cpbq7kzwxyZOr6mNJPprkC0l+cejy1iTnJHnzrGEvTXKHJFdX1TXD8aG8MclkVU1lZnWrvz/2dwAAAAAAAAAAAKtHzdTrwPKYnJzsqampcYcBAAAAAAAAAACHVFW7untywX4KrVhOVfVvSa4bdxzAsrtXkn8ddxDAspPrsDbIdVgb5DqsDXId1ga5DmuDXIfjnzyH8Tq1u09ZqNP6UUTCmnbdYir+gNWtqqbkOhz/5DqsDXId1ga5DmuDXIe1Qa7D2iDX4fgnz2F1OGHcAQAAAAAAAAAAAKx0Cq0AAAAAAAAAAAAWoNCK5bZj3AEAIyHXYW2Q67A2yHVYG+Q6rA1yHdYGuQ5rg1yH4588h1WgunvcMQAAAAAAAAAAAKxoVrQCAAAAAAAAAABYgEIrAAAAAAAAAACABSi04qhV1eOq6rqq+nhVbTvE+TtV1Z8N5y+vqo2zzv3C0H5dVf3gKOMGFu9o87yqNlbVgaraM7xePerYgcVbRK4/uqqurKovV9XZc849o6o+NryeMbqogSO1xFw/OOu5fsHoogaO1CJy/flV9ZGqurqq3l9Vp84657kOq8QSc91zHVaJReT6s6pq75DPf1tV3zrrnO/gYZU42lz3PTysLgvl+qx+Z1dVV9XkrDbPdVhBqrvHHQOrUFWtS/LRJD+Q5MYkH07y1O7+yKw+z07ykO5+VlWdk+SJ3f2U4Q+Af5rkYUm+IclfJXlgdx8c9X0A81tinm9M8u7u3jT6yIEjschc35jkrklekOSC7n7r0H6PJFNJJpN0kl1Jzuzuz4zwFoBFWEquD+du7e6TRxkzcOQWmevfm+Ty7t5fVf8lyVnDn+E912GVWEquD+c812EVWGSu37W7Pzd8/uEkz+7ux/kOHlaPJeb6xvgeHlaFxeT60O8uSXYmuWOS53T3lOc6rDxWtOJoPSzJx7v7+u7+UpI3JXnCnD5PSPL64fNbkzymqmpof1N3f7G7b0jy8WE+YGVZSp4Dq8eCud7d+7r76iRfmTP2B5Nc2N23DH8Je2GSx40iaOCILSXXgdVjMbn+we7ePxxeluS+w2fPdVg9lpLrwOqxmFz/3KzDr8tMsXTiO3hYTZaS68DqsZi/c0uSlyb530m+MKvNcx1WGIVWHK0NSf5x1vGNQ9sh+3T3l5NMJ7nnIscC47eUPE+S+1fV7qr6UFU9armDBY7aUp7Lnumweiw1X+9cVVNVdVlV/cixDQ04ho40138qyXuOciwwPkvJ9cRzHVaLReV6VZ1bVZ/IzF/KnnckY4EVYSm5nvgeHlaLBXO9qh6a5H7d/e4jHQuM1vpxB8CqdagVa+ZW0M/XZzFjgfFbSp7/U5Jv7O6bq+rMJO+oqgfP+Zc3wMqwlOeyZzqsHkvN12/s7k9W1QOSfKCq9nb3J45RbMCxs+hcr6qnZWabwO850rHA2C0l1xPPdVgtFpXr3f17SX6vqn4syf9I8ozFjgVWhKXkuu/hYfU4bK5X1QlJXpHkmUc6Fhg9K1pxtG5Mcr9Zx/dN8sn5+lTV+iQTSW5Z5Fhg/I46z4flS29Oku7eleQTSR647BEDR2Mpz2XPdFg9lpSv3f3J4f36JBcleeixDA44ZhaV61X1/Ul+KckPd/cXj2QssCIsJdc912H1ONJn85uS3L5Knec6rB5Hneu+h4dVZaFcv0uSTUkuqqp9SR6e5IKqmlzEWGDEFFpxtD6c5LSqun9V3THJOUkumNPngsxU1CfJ2Uk+0N09tJ9TVXeqqvsnOS3JFSOKG1i8o87zqjqlqtYlyfAvZE9Lcv2I4gaOzGJyfT7vS/LYqrp7Vd09yWOHNmDlOepcH3L8TsPneyX5riQfWbZIgaVYMNeHrQhek5nCi0/POuW5DqvHUee65zqsKovJ9dNmHW5J8rHhs+/gYfU46lz3PTysKofN9e6e7u57dffG7t6Y5LLM/Fl+Kp7rsOLYOpCj0t1frqrnZOZL13VJXtfd11bVS5JMdfcFSf4wyR9X1cczs5LVOcPYa6vqzZn5EufLSc7t7oNjuRFgXkvJ8ySPTvKSqvpykoNJntXdt4z+LoCFLCbXq+o7krw9yd2TPL6qXtzdD+7uW6rqpZn5n8QkeYlch5VpKbme5PQkr6mqr2TmH+ts725/IQsr0CL/DP+yJCcneUtVJck/dPcPe67D6rGUXI/nOqwai8z15wyr192W5DMZ/kGk7+Bh9VhKrsf38LBqLDLX5xvruQ4rTM0sMAQAAAAAAAAAAMB8bB0IAAAAAAAAAACwAIVWAAAAAAAAAAAAC1BoBQAAAAAAAAAAsACFVgAAAAAAAAAAAAtQaAUAAAAAAAAAALAAhVYAAAAAAAAAAAALUGgFAAAAAAAAAACwgP8H2SUgQrEHp2MAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 2880x1800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.ensemble import ExtraTreesRegressor\n",
    "\n",
    "model = ExtraTreesRegressor()\n",
    "model.fit(X,y)\n",
    "\n",
    "#plot graph of feature importances for better visualization\n",
    "fig, ax = plt.subplots(figsize=(40,25))\n",
    "feat_importances = pd.Series(model.feature_importances_, index=X.columns)\n",
    "feat_importances.nlargest(125).plot(kind='barh')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Based on the ExtraTreeRegressor scores, the following features are the ten most important:<br>\n",
    "- Overall\n",
    "- Potential\n",
    "- LAM\n",
    "- RF\n",
    "- International Reputation\n",
    "- Ball Control\n",
    "- Position Potential\n",
    "- Reactions\n",
    "- LF\n",
    "- Age\n",
    "\n",
    "Some of the features are in line with the F-scores, however, some features vary vastly in their importance between the two methods. E.g., Real Face is very important according to F-score, but not at all according to ExtraTreeRegressor. Same goes for Special. Reversely, Age is very important according to ExtraTreeRegressor, but not si much based on its F-value."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Additionally, we are also looking at the RMSE when using each individual feature as a stand alone predictor for the player's value. To do so, we are running each model using only the repsective feature. We then calculate the RMSE and compare the results. The features with a higher RMSE relative to the others are considered to have less explanatory power (WARNING: This is not taking effects between features into consideration)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_feats=predictors_df.shape[1]\n",
    "col_names=pd.DataFrame(np.array(predictors_df.columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "cv_score_features=pd.DataFrame()\n",
    "best_models=[dtr,rfr,xbr,lre,svr]\n",
    "index=0\n",
    "\n",
    "for mod in best_models:\n",
    "    score_list=[]\n",
    "    index+=1\n",
    "    for i in range(n_feats):\n",
    "        X_feature = pd.DataFrame(np.array(predictors_df.iloc[:, i]).reshape(-1,1))\n",
    "        scores = -cross_val_score(mod, X_feature, target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "        score_list.append(scores.mean())\n",
    "    cv_score_features[str(index)]=score_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature</th>\n",
       "      <th>DTR</th>\n",
       "      <th>RFR</th>\n",
       "      <th>XBR</th>\n",
       "      <th>LRE</th>\n",
       "      <th>SVC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>Overall</td>\n",
       "      <td>3.291131e+13</td>\n",
       "      <td>3.294635e+13</td>\n",
       "      <td>3.291657e+13</td>\n",
       "      <td>6.343037e+13</td>\n",
       "      <td>3.559205e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50</td>\n",
       "      <td>BallControl</td>\n",
       "      <td>3.363363e+13</td>\n",
       "      <td>3.431617e+13</td>\n",
       "      <td>3.359435e+13</td>\n",
       "      <td>3.976522e+13</td>\n",
       "      <td>3.559360e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>22</td>\n",
       "      <td>RW</td>\n",
       "      <td>3.404289e+13</td>\n",
       "      <td>3.425635e+13</td>\n",
       "      <td>3.403353e+13</td>\n",
       "      <td>4.016000e+13</td>\n",
       "      <td>3.559329e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>LW</td>\n",
       "      <td>3.404289e+13</td>\n",
       "      <td>3.418134e+13</td>\n",
       "      <td>3.403353e+13</td>\n",
       "      <td>4.016000e+13</td>\n",
       "      <td>3.559329e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>23</td>\n",
       "      <td>LAM</td>\n",
       "      <td>3.503860e+13</td>\n",
       "      <td>3.495623e+13</td>\n",
       "      <td>3.501742e+13</td>\n",
       "      <td>4.009771e+13</td>\n",
       "      <td>3.559356e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25</td>\n",
       "      <td>RAM</td>\n",
       "      <td>3.503860e+13</td>\n",
       "      <td>3.493453e+13</td>\n",
       "      <td>3.501742e+13</td>\n",
       "      <td>4.009771e+13</td>\n",
       "      <td>3.559356e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>24</td>\n",
       "      <td>CAM</td>\n",
       "      <td>3.503860e+13</td>\n",
       "      <td>3.500816e+13</td>\n",
       "      <td>3.501742e+13</td>\n",
       "      <td>4.009771e+13</td>\n",
       "      <td>3.559356e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>21</td>\n",
       "      <td>RF</td>\n",
       "      <td>3.510001e+13</td>\n",
       "      <td>3.507777e+13</td>\n",
       "      <td>3.507277e+13</td>\n",
       "      <td>4.010982e+13</td>\n",
       "      <td>3.559353e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>LF</td>\n",
       "      <td>3.510001e+13</td>\n",
       "      <td>3.516358e+13</td>\n",
       "      <td>3.507277e+13</td>\n",
       "      <td>4.010982e+13</td>\n",
       "      <td>3.559353e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>CF</td>\n",
       "      <td>3.510001e+13</td>\n",
       "      <td>3.506381e+13</td>\n",
       "      <td>3.507277e+13</td>\n",
       "      <td>4.010982e+13</td>\n",
       "      <td>3.559353e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>30</td>\n",
       "      <td>RM</td>\n",
       "      <td>3.523238e+13</td>\n",
       "      <td>3.496112e+13</td>\n",
       "      <td>3.521924e+13</td>\n",
       "      <td>4.015433e+13</td>\n",
       "      <td>3.559332e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>26</td>\n",
       "      <td>LM</td>\n",
       "      <td>3.523238e+13</td>\n",
       "      <td>3.561174e+13</td>\n",
       "      <td>3.521924e+13</td>\n",
       "      <td>4.015433e+13</td>\n",
       "      <td>3.559332e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>44</td>\n",
       "      <td>ShortPassing</td>\n",
       "      <td>3.530417e+13</td>\n",
       "      <td>3.542422e+13</td>\n",
       "      <td>3.527164e+13</td>\n",
       "      <td>3.969333e+13</td>\n",
       "      <td>3.559412e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>Potential</td>\n",
       "      <td>3.532868e+13</td>\n",
       "      <td>3.537813e+13</td>\n",
       "      <td>3.533525e+13</td>\n",
       "      <td>4.168772e+13</td>\n",
       "      <td>3.559093e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>46</td>\n",
       "      <td>Dribbling</td>\n",
       "      <td>3.547305e+13</td>\n",
       "      <td>3.545947e+13</td>\n",
       "      <td>3.541716e+13</td>\n",
       "      <td>4.004378e+13</td>\n",
       "      <td>3.559388e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>54</td>\n",
       "      <td>Reactions</td>\n",
       "      <td>3.568605e+13</td>\n",
       "      <td>3.564557e+13</td>\n",
       "      <td>3.574151e+13</td>\n",
       "      <td>4.535629e+13</td>\n",
       "      <td>3.559331e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>LS</td>\n",
       "      <td>3.587456e+13</td>\n",
       "      <td>3.587886e+13</td>\n",
       "      <td>3.586294e+13</td>\n",
       "      <td>4.012114e+13</td>\n",
       "      <td>3.559382e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>ST</td>\n",
       "      <td>3.587456e+13</td>\n",
       "      <td>3.581031e+13</td>\n",
       "      <td>3.586294e+13</td>\n",
       "      <td>4.012114e+13</td>\n",
       "      <td>3.559382e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>RS</td>\n",
       "      <td>3.587456e+13</td>\n",
       "      <td>3.576295e+13</td>\n",
       "      <td>3.586294e+13</td>\n",
       "      <td>4.012114e+13</td>\n",
       "      <td>3.559382e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>64</td>\n",
       "      <td>Vision</td>\n",
       "      <td>3.646856e+13</td>\n",
       "      <td>3.644209e+13</td>\n",
       "      <td>3.641733e+13</td>\n",
       "      <td>3.969074e+13</td>\n",
       "      <td>3.559435e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>27</td>\n",
       "      <td>LCM</td>\n",
       "      <td>3.646782e+13</td>\n",
       "      <td>3.605517e+13</td>\n",
       "      <td>3.643896e+13</td>\n",
       "      <td>4.001880e+13</td>\n",
       "      <td>3.559367e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>29</td>\n",
       "      <td>RCM</td>\n",
       "      <td>3.646782e+13</td>\n",
       "      <td>3.604131e+13</td>\n",
       "      <td>3.643896e+13</td>\n",
       "      <td>4.001880e+13</td>\n",
       "      <td>3.559367e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>28</td>\n",
       "      <td>CM</td>\n",
       "      <td>3.646782e+13</td>\n",
       "      <td>3.606953e+13</td>\n",
       "      <td>3.643896e+13</td>\n",
       "      <td>4.001880e+13</td>\n",
       "      <td>3.559367e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>66</td>\n",
       "      <td>Composure</td>\n",
       "      <td>3.667634e+13</td>\n",
       "      <td>3.665991e+13</td>\n",
       "      <td>3.661637e+13</td>\n",
       "      <td>4.021195e+13</td>\n",
       "      <td>3.559377e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>77</td>\n",
       "      <td>Attacking</td>\n",
       "      <td>3.718240e+13</td>\n",
       "      <td>3.699858e+13</td>\n",
       "      <td>3.665597e+13</td>\n",
       "      <td>3.964200e+13</td>\n",
       "      <td>3.559391e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>Special</td>\n",
       "      <td>4.039301e+13</td>\n",
       "      <td>3.967300e+13</td>\n",
       "      <td>3.677384e+13</td>\n",
       "      <td>3.936163e+13</td>\n",
       "      <td>3.559508e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>78</td>\n",
       "      <td>Skill</td>\n",
       "      <td>3.761462e+13</td>\n",
       "      <td>3.773848e+13</td>\n",
       "      <td>3.691074e+13</td>\n",
       "      <td>3.960158e+13</td>\n",
       "      <td>3.559421e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>63</td>\n",
       "      <td>Positioning</td>\n",
       "      <td>3.710834e+13</td>\n",
       "      <td>3.702116e+13</td>\n",
       "      <td>3.705036e+13</td>\n",
       "      <td>4.008411e+13</td>\n",
       "      <td>3.559426e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>49</td>\n",
       "      <td>LongPassing</td>\n",
       "      <td>3.703052e+13</td>\n",
       "      <td>3.725048e+13</td>\n",
       "      <td>3.707486e+13</td>\n",
       "      <td>3.972126e+13</td>\n",
       "      <td>3.559443e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>International Reputation</td>\n",
       "      <td>3.712385e+13</td>\n",
       "      <td>3.708087e+13</td>\n",
       "      <td>3.712412e+13</td>\n",
       "      <td>3.702092e+13</td>\n",
       "      <td>3.559517e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>34</td>\n",
       "      <td>RDM</td>\n",
       "      <td>3.725316e+13</td>\n",
       "      <td>3.732581e+13</td>\n",
       "      <td>3.726221e+13</td>\n",
       "      <td>4.018408e+13</td>\n",
       "      <td>3.559431e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>33</td>\n",
       "      <td>CDM</td>\n",
       "      <td>3.725316e+13</td>\n",
       "      <td>3.751836e+13</td>\n",
       "      <td>3.726221e+13</td>\n",
       "      <td>4.018408e+13</td>\n",
       "      <td>3.559431e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32</td>\n",
       "      <td>LDM</td>\n",
       "      <td>3.725316e+13</td>\n",
       "      <td>3.754583e+13</td>\n",
       "      <td>3.726221e+13</td>\n",
       "      <td>4.018408e+13</td>\n",
       "      <td>3.559431e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>60</td>\n",
       "      <td>LongShots</td>\n",
       "      <td>3.730820e+13</td>\n",
       "      <td>3.732562e+13</td>\n",
       "      <td>3.729669e+13</td>\n",
       "      <td>3.983257e+13</td>\n",
       "      <td>3.559450e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>56</td>\n",
       "      <td>ShotPower</td>\n",
       "      <td>3.774815e+13</td>\n",
       "      <td>3.765473e+13</td>\n",
       "      <td>3.769899e+13</td>\n",
       "      <td>3.973684e+13</td>\n",
       "      <td>3.559433e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>42</td>\n",
       "      <td>Finishing</td>\n",
       "      <td>3.785361e+13</td>\n",
       "      <td>3.764233e+13</td>\n",
       "      <td>3.779046e+13</td>\n",
       "      <td>4.018861e+13</td>\n",
       "      <td>3.559469e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>41</td>\n",
       "      <td>Crossing</td>\n",
       "      <td>3.797931e+13</td>\n",
       "      <td>3.794522e+13</td>\n",
       "      <td>3.788268e+13</td>\n",
       "      <td>4.002732e+13</td>\n",
       "      <td>3.559454e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>45</td>\n",
       "      <td>Volleys</td>\n",
       "      <td>3.795030e+13</td>\n",
       "      <td>3.795453e+13</td>\n",
       "      <td>3.792699e+13</td>\n",
       "      <td>3.986659e+13</td>\n",
       "      <td>3.559466e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>81</td>\n",
       "      <td>Mentality</td>\n",
       "      <td>3.901254e+13</td>\n",
       "      <td>3.914657e+13</td>\n",
       "      <td>3.798656e+13</td>\n",
       "      <td>3.928514e+13</td>\n",
       "      <td>3.559373e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>36</td>\n",
       "      <td>LB</td>\n",
       "      <td>3.803766e+13</td>\n",
       "      <td>3.821657e+13</td>\n",
       "      <td>3.803114e+13</td>\n",
       "      <td>4.026994e+13</td>\n",
       "      <td>3.559417e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>40</td>\n",
       "      <td>RB</td>\n",
       "      <td>3.803766e+13</td>\n",
       "      <td>3.808355e+13</td>\n",
       "      <td>3.803114e+13</td>\n",
       "      <td>4.026994e+13</td>\n",
       "      <td>3.559417e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>47</td>\n",
       "      <td>Curve</td>\n",
       "      <td>3.807209e+13</td>\n",
       "      <td>3.814493e+13</td>\n",
       "      <td>3.804161e+13</td>\n",
       "      <td>3.979783e+13</td>\n",
       "      <td>3.559453e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>31</td>\n",
       "      <td>LWB</td>\n",
       "      <td>3.823767e+13</td>\n",
       "      <td>3.829537e+13</td>\n",
       "      <td>3.821822e+13</td>\n",
       "      <td>4.022300e+13</td>\n",
       "      <td>3.559454e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>35</td>\n",
       "      <td>RWB</td>\n",
       "      <td>3.823767e+13</td>\n",
       "      <td>3.817528e+13</td>\n",
       "      <td>3.821822e+13</td>\n",
       "      <td>4.022300e+13</td>\n",
       "      <td>3.559454e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>Skill Moves</td>\n",
       "      <td>3.846869e+13</td>\n",
       "      <td>3.847564e+13</td>\n",
       "      <td>3.846849e+13</td>\n",
       "      <td>3.957370e+13</td>\n",
       "      <td>3.558797e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>80</td>\n",
       "      <td>Power</td>\n",
       "      <td>3.926227e+13</td>\n",
       "      <td>3.944756e+13</td>\n",
       "      <td>3.847206e+13</td>\n",
       "      <td>3.972625e+13</td>\n",
       "      <td>3.559380e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>76</td>\n",
       "      <td>Position_Potential</td>\n",
       "      <td>3.859366e+13</td>\n",
       "      <td>3.853693e+13</td>\n",
       "      <td>3.859348e+13</td>\n",
       "      <td>4.097735e+13</td>\n",
       "      <td>3.559578e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>Real Face</td>\n",
       "      <td>3.868142e+13</td>\n",
       "      <td>3.869908e+13</td>\n",
       "      <td>3.868123e+13</td>\n",
       "      <td>3.868142e+13</td>\n",
       "      <td>3.559496e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>62</td>\n",
       "      <td>Interceptions</td>\n",
       "      <td>3.894271e+13</td>\n",
       "      <td>3.900229e+13</td>\n",
       "      <td>3.881148e+13</td>\n",
       "      <td>4.046144e+13</td>\n",
       "      <td>3.559447e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>43</td>\n",
       "      <td>HeadingAccuracy</td>\n",
       "      <td>3.919032e+13</td>\n",
       "      <td>3.921825e+13</td>\n",
       "      <td>3.906376e+13</td>\n",
       "      <td>4.018653e+13</td>\n",
       "      <td>3.559467e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>79</td>\n",
       "      <td>Movement</td>\n",
       "      <td>3.950834e+13</td>\n",
       "      <td>3.948714e+13</td>\n",
       "      <td>3.911419e+13</td>\n",
       "      <td>4.045258e+13</td>\n",
       "      <td>3.559410e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>48</td>\n",
       "      <td>FKAccuracy</td>\n",
       "      <td>3.920840e+13</td>\n",
       "      <td>3.917831e+13</td>\n",
       "      <td>3.911572e+13</td>\n",
       "      <td>3.991706e+13</td>\n",
       "      <td>3.559464e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>68</td>\n",
       "      <td>StandingTackle</td>\n",
       "      <td>3.923663e+13</td>\n",
       "      <td>3.938180e+13</td>\n",
       "      <td>3.916141e+13</td>\n",
       "      <td>4.051417e+13</td>\n",
       "      <td>3.559436e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>69</td>\n",
       "      <td>SlidingTackle</td>\n",
       "      <td>3.935045e+13</td>\n",
       "      <td>3.930441e+13</td>\n",
       "      <td>3.919273e+13</td>\n",
       "      <td>4.054869e+13</td>\n",
       "      <td>3.559440e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>65</td>\n",
       "      <td>Penalties</td>\n",
       "      <td>3.948346e+13</td>\n",
       "      <td>3.943302e+13</td>\n",
       "      <td>3.938622e+13</td>\n",
       "      <td>3.999467e+13</td>\n",
       "      <td>3.559464e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>67</td>\n",
       "      <td>Marking</td>\n",
       "      <td>3.951506e+13</td>\n",
       "      <td>3.962020e+13</td>\n",
       "      <td>3.941244e+13</td>\n",
       "      <td>4.046294e+13</td>\n",
       "      <td>3.559463e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>82</td>\n",
       "      <td>Defending</td>\n",
       "      <td>4.000452e+13</td>\n",
       "      <td>4.023640e+13</td>\n",
       "      <td>3.941836e+13</td>\n",
       "      <td>4.050480e+13</td>\n",
       "      <td>3.559433e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>39</td>\n",
       "      <td>RCB</td>\n",
       "      <td>3.962185e+13</td>\n",
       "      <td>3.966552e+13</td>\n",
       "      <td>3.953114e+13</td>\n",
       "      <td>4.035521e+13</td>\n",
       "      <td>3.559418e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>38</td>\n",
       "      <td>CB</td>\n",
       "      <td>3.962185e+13</td>\n",
       "      <td>3.965921e+13</td>\n",
       "      <td>3.953114e+13</td>\n",
       "      <td>4.035521e+13</td>\n",
       "      <td>3.559418e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>37</td>\n",
       "      <td>LCB</td>\n",
       "      <td>3.962185e+13</td>\n",
       "      <td>3.965597e+13</td>\n",
       "      <td>3.953114e+13</td>\n",
       "      <td>4.035521e+13</td>\n",
       "      <td>3.559418e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>61</td>\n",
       "      <td>Aggression</td>\n",
       "      <td>4.030169e+13</td>\n",
       "      <td>4.035343e+13</td>\n",
       "      <td>4.016983e+13</td>\n",
       "      <td>4.028812e+13</td>\n",
       "      <td>3.559470e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>71</td>\n",
       "      <td>GKHandling</td>\n",
       "      <td>4.017113e+13</td>\n",
       "      <td>4.024809e+13</td>\n",
       "      <td>4.016995e+13</td>\n",
       "      <td>4.058494e+13</td>\n",
       "      <td>3.559492e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>75</td>\n",
       "      <td>Club_Reputation</td>\n",
       "      <td>4.043025e+13</td>\n",
       "      <td>4.039635e+13</td>\n",
       "      <td>4.019410e+13</td>\n",
       "      <td>3.949069e+13</td>\n",
       "      <td>3.559478e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>83</td>\n",
       "      <td>GoalKeeping</td>\n",
       "      <td>4.070534e+13</td>\n",
       "      <td>4.059779e+13</td>\n",
       "      <td>4.029215e+13</td>\n",
       "      <td>4.058338e+13</td>\n",
       "      <td>3.559438e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>70</td>\n",
       "      <td>GKDiving</td>\n",
       "      <td>4.061509e+13</td>\n",
       "      <td>4.050402e+13</td>\n",
       "      <td>4.042991e+13</td>\n",
       "      <td>4.058557e+13</td>\n",
       "      <td>3.559484e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>116</td>\n",
       "      <td>Continent_Asia</td>\n",
       "      <td>4.049309e+13</td>\n",
       "      <td>4.046750e+13</td>\n",
       "      <td>4.049280e+13</td>\n",
       "      <td>4.049309e+13</td>\n",
       "      <td>3.559517e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>51</td>\n",
       "      <td>Acceleration</td>\n",
       "      <td>4.054570e+13</td>\n",
       "      <td>4.052450e+13</td>\n",
       "      <td>4.050439e+13</td>\n",
       "      <td>4.063825e+13</td>\n",
       "      <td>3.559464e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>58</td>\n",
       "      <td>Stamina</td>\n",
       "      <td>4.061329e+13</td>\n",
       "      <td>4.065371e+13</td>\n",
       "      <td>4.053882e+13</td>\n",
       "      <td>4.044311e+13</td>\n",
       "      <td>3.559439e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>120</td>\n",
       "      <td>Continent_South America</td>\n",
       "      <td>4.055348e+13</td>\n",
       "      <td>4.055314e+13</td>\n",
       "      <td>4.055318e+13</td>\n",
       "      <td>4.055348e+13</td>\n",
       "      <td>3.559474e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>123</td>\n",
       "      <td>Position_Group_GoalKeeper</td>\n",
       "      <td>4.056619e+13</td>\n",
       "      <td>4.058352e+13</td>\n",
       "      <td>4.056590e+13</td>\n",
       "      <td>4.056619e+13</td>\n",
       "      <td>3.559532e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>93</td>\n",
       "      <td>Position_GK</td>\n",
       "      <td>4.056619e+13</td>\n",
       "      <td>4.042336e+13</td>\n",
       "      <td>4.056590e+13</td>\n",
       "      <td>4.056619e+13</td>\n",
       "      <td>3.559532e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>101</td>\n",
       "      <td>Position_LS</td>\n",
       "      <td>4.056743e+13</td>\n",
       "      <td>4.059826e+13</td>\n",
       "      <td>4.056713e+13</td>\n",
       "      <td>4.056743e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>124</td>\n",
       "      <td>Position_Group_Midfielder</td>\n",
       "      <td>4.057505e+13</td>\n",
       "      <td>4.065053e+13</td>\n",
       "      <td>4.057475e+13</td>\n",
       "      <td>4.057505e+13</td>\n",
       "      <td>3.559397e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>104</td>\n",
       "      <td>Position_RAM</td>\n",
       "      <td>4.057549e+13</td>\n",
       "      <td>4.064459e+13</td>\n",
       "      <td>4.057520e+13</td>\n",
       "      <td>4.057549e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>94</td>\n",
       "      <td>Position_LAM</td>\n",
       "      <td>4.057999e+13</td>\n",
       "      <td>4.062341e+13</td>\n",
       "      <td>4.057970e+13</td>\n",
       "      <td>4.057999e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>89</td>\n",
       "      <td>Position_CB</td>\n",
       "      <td>4.058014e+13</td>\n",
       "      <td>4.051565e+13</td>\n",
       "      <td>4.057984e+13</td>\n",
       "      <td>4.058014e+13</td>\n",
       "      <td>3.559516e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>111</td>\n",
       "      <td>Position_RS</td>\n",
       "      <td>4.058024e+13</td>\n",
       "      <td>4.054201e+13</td>\n",
       "      <td>4.057994e+13</td>\n",
       "      <td>4.058024e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>119</td>\n",
       "      <td>Continent_Oceania</td>\n",
       "      <td>4.058075e+13</td>\n",
       "      <td>4.059081e+13</td>\n",
       "      <td>4.058046e+13</td>\n",
       "      <td>4.058075e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>88</td>\n",
       "      <td>Position_CAM</td>\n",
       "      <td>4.058364e+13</td>\n",
       "      <td>4.053184e+13</td>\n",
       "      <td>4.058335e+13</td>\n",
       "      <td>4.058364e+13</td>\n",
       "      <td>3.559511e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>100</td>\n",
       "      <td>Position_LM</td>\n",
       "      <td>4.058558e+13</td>\n",
       "      <td>4.061478e+13</td>\n",
       "      <td>4.058528e+13</td>\n",
       "      <td>4.058558e+13</td>\n",
       "      <td>3.559509e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>110</td>\n",
       "      <td>Position_RM</td>\n",
       "      <td>4.058744e+13</td>\n",
       "      <td>4.060298e+13</td>\n",
       "      <td>4.058715e+13</td>\n",
       "      <td>4.058744e+13</td>\n",
       "      <td>3.559511e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>108</td>\n",
       "      <td>Position_RDM</td>\n",
       "      <td>4.058798e+13</td>\n",
       "      <td>4.064590e+13</td>\n",
       "      <td>4.058769e+13</td>\n",
       "      <td>4.058798e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>103</td>\n",
       "      <td>Position_LWB</td>\n",
       "      <td>4.058846e+13</td>\n",
       "      <td>4.056909e+13</td>\n",
       "      <td>4.058816e+13</td>\n",
       "      <td>4.058846e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>113</td>\n",
       "      <td>Position_RWB</td>\n",
       "      <td>4.059027e+13</td>\n",
       "      <td>4.059812e+13</td>\n",
       "      <td>4.058998e+13</td>\n",
       "      <td>4.059027e+13</td>\n",
       "      <td>3.559515e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>87</td>\n",
       "      <td>Defense_WR</td>\n",
       "      <td>4.059084e+13</td>\n",
       "      <td>4.061972e+13</td>\n",
       "      <td>4.059055e+13</td>\n",
       "      <td>4.062020e+13</td>\n",
       "      <td>3.559449e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>91</td>\n",
       "      <td>Position_CF</td>\n",
       "      <td>4.059545e+13</td>\n",
       "      <td>4.060072e+13</td>\n",
       "      <td>4.059516e+13</td>\n",
       "      <td>4.059545e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>114</td>\n",
       "      <td>Position_ST</td>\n",
       "      <td>4.059688e+13</td>\n",
       "      <td>4.065603e+13</td>\n",
       "      <td>4.059659e+13</td>\n",
       "      <td>4.059688e+13</td>\n",
       "      <td>3.559509e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>86</td>\n",
       "      <td>Attack_WR</td>\n",
       "      <td>4.059741e+13</td>\n",
       "      <td>4.052668e+13</td>\n",
       "      <td>4.059711e+13</td>\n",
       "      <td>4.061665e+13</td>\n",
       "      <td>3.559312e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>98</td>\n",
       "      <td>Position_LDM</td>\n",
       "      <td>4.059742e+13</td>\n",
       "      <td>4.053577e+13</td>\n",
       "      <td>4.059713e+13</td>\n",
       "      <td>4.059742e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>90</td>\n",
       "      <td>Position_CDM</td>\n",
       "      <td>4.059848e+13</td>\n",
       "      <td>4.062884e+13</td>\n",
       "      <td>4.059818e+13</td>\n",
       "      <td>4.059848e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>107</td>\n",
       "      <td>Position_RCM</td>\n",
       "      <td>4.059857e+13</td>\n",
       "      <td>4.051201e+13</td>\n",
       "      <td>4.059827e+13</td>\n",
       "      <td>4.059857e+13</td>\n",
       "      <td>3.559513e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>115</td>\n",
       "      <td>Continent_Africa</td>\n",
       "      <td>4.060050e+13</td>\n",
       "      <td>4.065196e+13</td>\n",
       "      <td>4.060021e+13</td>\n",
       "      <td>4.060050e+13</td>\n",
       "      <td>3.559506e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>95</td>\n",
       "      <td>Position_LB</td>\n",
       "      <td>4.060203e+13</td>\n",
       "      <td>4.061604e+13</td>\n",
       "      <td>4.060174e+13</td>\n",
       "      <td>4.060203e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>97</td>\n",
       "      <td>Position_LCM</td>\n",
       "      <td>4.060453e+13</td>\n",
       "      <td>4.067682e+13</td>\n",
       "      <td>4.060423e+13</td>\n",
       "      <td>4.060453e+13</td>\n",
       "      <td>3.559513e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>Preferred Foot</td>\n",
       "      <td>4.060673e+13</td>\n",
       "      <td>4.065177e+13</td>\n",
       "      <td>4.060644e+13</td>\n",
       "      <td>4.060673e+13</td>\n",
       "      <td>3.559479e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>112</td>\n",
       "      <td>Position_RW</td>\n",
       "      <td>4.060816e+13</td>\n",
       "      <td>4.058124e+13</td>\n",
       "      <td>4.060786e+13</td>\n",
       "      <td>4.060816e+13</td>\n",
       "      <td>3.559513e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>Weak Foot</td>\n",
       "      <td>4.060848e+13</td>\n",
       "      <td>4.065150e+13</td>\n",
       "      <td>4.060818e+13</td>\n",
       "      <td>4.062002e+13</td>\n",
       "      <td>3.559302e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>Loaned From</td>\n",
       "      <td>4.061036e+13</td>\n",
       "      <td>4.064751e+13</td>\n",
       "      <td>4.061006e+13</td>\n",
       "      <td>4.061036e+13</td>\n",
       "      <td>3.559503e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>106</td>\n",
       "      <td>Position_RCB</td>\n",
       "      <td>4.061562e+13</td>\n",
       "      <td>4.061802e+13</td>\n",
       "      <td>4.061532e+13</td>\n",
       "      <td>4.061562e+13</td>\n",
       "      <td>3.559512e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>118</td>\n",
       "      <td>Continent_North America</td>\n",
       "      <td>4.061745e+13</td>\n",
       "      <td>4.061235e+13</td>\n",
       "      <td>4.061716e+13</td>\n",
       "      <td>4.061745e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>105</td>\n",
       "      <td>Position_RB</td>\n",
       "      <td>4.062090e+13</td>\n",
       "      <td>4.057000e+13</td>\n",
       "      <td>4.062060e+13</td>\n",
       "      <td>4.062090e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>102</td>\n",
       "      <td>Position_LW</td>\n",
       "      <td>4.062130e+13</td>\n",
       "      <td>4.062293e+13</td>\n",
       "      <td>4.062100e+13</td>\n",
       "      <td>4.062130e+13</td>\n",
       "      <td>3.559513e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>96</td>\n",
       "      <td>Position_LCB</td>\n",
       "      <td>4.062844e+13</td>\n",
       "      <td>4.066746e+13</td>\n",
       "      <td>4.062815e+13</td>\n",
       "      <td>4.062844e+13</td>\n",
       "      <td>3.559513e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>99</td>\n",
       "      <td>Position_LF</td>\n",
       "      <td>4.063674e+13</td>\n",
       "      <td>4.067840e+13</td>\n",
       "      <td>4.063643e+13</td>\n",
       "      <td>4.063674e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>122</td>\n",
       "      <td>Position_Group_Forward</td>\n",
       "      <td>4.063847e+13</td>\n",
       "      <td>4.062271e+13</td>\n",
       "      <td>4.063817e+13</td>\n",
       "      <td>4.063847e+13</td>\n",
       "      <td>3.559463e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>Height</td>\n",
       "      <td>4.064589e+13</td>\n",
       "      <td>4.067548e+13</td>\n",
       "      <td>4.064278e+13</td>\n",
       "      <td>4.059698e+13</td>\n",
       "      <td>3.559501e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>53</td>\n",
       "      <td>Agility</td>\n",
       "      <td>4.072630e+13</td>\n",
       "      <td>4.062334e+13</td>\n",
       "      <td>4.064309e+13</td>\n",
       "      <td>4.055930e+13</td>\n",
       "      <td>3.559455e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>92</td>\n",
       "      <td>Position_CM</td>\n",
       "      <td>4.065203e+13</td>\n",
       "      <td>4.072975e+13</td>\n",
       "      <td>4.065174e+13</td>\n",
       "      <td>4.065203e+13</td>\n",
       "      <td>3.559518e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>72</td>\n",
       "      <td>GKKicking</td>\n",
       "      <td>4.135352e+13</td>\n",
       "      <td>4.100044e+13</td>\n",
       "      <td>4.065761e+13</td>\n",
       "      <td>4.057962e+13</td>\n",
       "      <td>3.559474e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>117</td>\n",
       "      <td>Continent_Europe</td>\n",
       "      <td>4.066388e+13</td>\n",
       "      <td>4.062134e+13</td>\n",
       "      <td>4.066358e+13</td>\n",
       "      <td>4.066388e+13</td>\n",
       "      <td>3.559486e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>109</td>\n",
       "      <td>Position_RF</td>\n",
       "      <td>4.069445e+13</td>\n",
       "      <td>4.068104e+13</td>\n",
       "      <td>4.069413e+13</td>\n",
       "      <td>4.069445e+13</td>\n",
       "      <td>3.559514e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>121</td>\n",
       "      <td>Position_Group_Defender</td>\n",
       "      <td>4.070186e+13</td>\n",
       "      <td>4.070087e+13</td>\n",
       "      <td>4.070156e+13</td>\n",
       "      <td>4.070186e+13</td>\n",
       "      <td>3.559457e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>57</td>\n",
       "      <td>Jumping</td>\n",
       "      <td>4.079931e+13</td>\n",
       "      <td>4.081470e+13</td>\n",
       "      <td>4.071576e+13</td>\n",
       "      <td>4.062103e+13</td>\n",
       "      <td>3.559474e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>52</td>\n",
       "      <td>SprintSpeed</td>\n",
       "      <td>4.079643e+13</td>\n",
       "      <td>4.088547e+13</td>\n",
       "      <td>4.073411e+13</td>\n",
       "      <td>4.063930e+13</td>\n",
       "      <td>3.559450e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>85</td>\n",
       "      <td>Body Type</td>\n",
       "      <td>4.073744e+13</td>\n",
       "      <td>4.078065e+13</td>\n",
       "      <td>4.073714e+13</td>\n",
       "      <td>4.072909e+13</td>\n",
       "      <td>3.559507e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>Weight</td>\n",
       "      <td>4.084161e+13</td>\n",
       "      <td>4.085690e+13</td>\n",
       "      <td>4.078900e+13</td>\n",
       "      <td>4.067301e+13</td>\n",
       "      <td>3.559496e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>55</td>\n",
       "      <td>Balance</td>\n",
       "      <td>4.091976e+13</td>\n",
       "      <td>4.095201e+13</td>\n",
       "      <td>4.083913e+13</td>\n",
       "      <td>4.064931e+13</td>\n",
       "      <td>3.559493e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>Jersey Number</td>\n",
       "      <td>4.092715e+13</td>\n",
       "      <td>4.093653e+13</td>\n",
       "      <td>4.085776e+13</td>\n",
       "      <td>4.071989e+13</td>\n",
       "      <td>3.559471e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>59</td>\n",
       "      <td>Strength</td>\n",
       "      <td>4.103734e+13</td>\n",
       "      <td>4.108102e+13</td>\n",
       "      <td>4.095919e+13</td>\n",
       "      <td>4.075654e+13</td>\n",
       "      <td>3.559478e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>Contract Valid Until</td>\n",
       "      <td>4.097543e+13</td>\n",
       "      <td>4.084194e+13</td>\n",
       "      <td>4.097354e+13</td>\n",
       "      <td>4.115312e+13</td>\n",
       "      <td>3.559390e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>84</td>\n",
       "      <td>Potential_Difference</td>\n",
       "      <td>4.134340e+13</td>\n",
       "      <td>4.133834e+13</td>\n",
       "      <td>4.134291e+13</td>\n",
       "      <td>4.143626e+13</td>\n",
       "      <td>3.559468e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>Age</td>\n",
       "      <td>4.170572e+13</td>\n",
       "      <td>4.173941e+13</td>\n",
       "      <td>4.170205e+13</td>\n",
       "      <td>4.218390e+13</td>\n",
       "      <td>3.559363e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>74</td>\n",
       "      <td>GKReflexes</td>\n",
       "      <td>4.405168e+13</td>\n",
       "      <td>4.063388e+13</td>\n",
       "      <td>4.172554e+13</td>\n",
       "      <td>4.058282e+13</td>\n",
       "      <td>3.559485e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>Joined</td>\n",
       "      <td>4.730089e+13</td>\n",
       "      <td>4.530670e+13</td>\n",
       "      <td>4.236384e+13</td>\n",
       "      <td>4.082795e+13</td>\n",
       "      <td>3.559494e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>73</td>\n",
       "      <td>GKPositioning</td>\n",
       "      <td>4.607656e+13</td>\n",
       "      <td>4.412526e+13</td>\n",
       "      <td>4.294646e+13</td>\n",
       "      <td>4.058505e+13</td>\n",
       "      <td>3.559484e+13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       Feature           DTR           RFR           XBR  \\\n",
       "1                      Overall  3.291131e+13  3.294635e+13  3.291657e+13   \n",
       "50                 BallControl  3.363363e+13  3.431617e+13  3.359435e+13   \n",
       "22                          RW  3.404289e+13  3.425635e+13  3.403353e+13   \n",
       "18                          LW  3.404289e+13  3.418134e+13  3.403353e+13   \n",
       "23                         LAM  3.503860e+13  3.495623e+13  3.501742e+13   \n",
       "25                         RAM  3.503860e+13  3.493453e+13  3.501742e+13   \n",
       "24                         CAM  3.503860e+13  3.500816e+13  3.501742e+13   \n",
       "21                          RF  3.510001e+13  3.507777e+13  3.507277e+13   \n",
       "19                          LF  3.510001e+13  3.516358e+13  3.507277e+13   \n",
       "20                          CF  3.510001e+13  3.506381e+13  3.507277e+13   \n",
       "30                          RM  3.523238e+13  3.496112e+13  3.521924e+13   \n",
       "26                          LM  3.523238e+13  3.561174e+13  3.521924e+13   \n",
       "44                ShortPassing  3.530417e+13  3.542422e+13  3.527164e+13   \n",
       "2                    Potential  3.532868e+13  3.537813e+13  3.533525e+13   \n",
       "46                   Dribbling  3.547305e+13  3.545947e+13  3.541716e+13   \n",
       "54                   Reactions  3.568605e+13  3.564557e+13  3.574151e+13   \n",
       "15                          LS  3.587456e+13  3.587886e+13  3.586294e+13   \n",
       "16                          ST  3.587456e+13  3.581031e+13  3.586294e+13   \n",
       "17                          RS  3.587456e+13  3.576295e+13  3.586294e+13   \n",
       "64                      Vision  3.646856e+13  3.644209e+13  3.641733e+13   \n",
       "27                         LCM  3.646782e+13  3.605517e+13  3.643896e+13   \n",
       "29                         RCM  3.646782e+13  3.604131e+13  3.643896e+13   \n",
       "28                          CM  3.646782e+13  3.606953e+13  3.643896e+13   \n",
       "66                   Composure  3.667634e+13  3.665991e+13  3.661637e+13   \n",
       "77                   Attacking  3.718240e+13  3.699858e+13  3.665597e+13   \n",
       "3                      Special  4.039301e+13  3.967300e+13  3.677384e+13   \n",
       "78                       Skill  3.761462e+13  3.773848e+13  3.691074e+13   \n",
       "63                 Positioning  3.710834e+13  3.702116e+13  3.705036e+13   \n",
       "49                 LongPassing  3.703052e+13  3.725048e+13  3.707486e+13   \n",
       "5     International Reputation  3.712385e+13  3.708087e+13  3.712412e+13   \n",
       "34                         RDM  3.725316e+13  3.732581e+13  3.726221e+13   \n",
       "33                         CDM  3.725316e+13  3.751836e+13  3.726221e+13   \n",
       "32                         LDM  3.725316e+13  3.754583e+13  3.726221e+13   \n",
       "60                   LongShots  3.730820e+13  3.732562e+13  3.729669e+13   \n",
       "56                   ShotPower  3.774815e+13  3.765473e+13  3.769899e+13   \n",
       "42                   Finishing  3.785361e+13  3.764233e+13  3.779046e+13   \n",
       "41                    Crossing  3.797931e+13  3.794522e+13  3.788268e+13   \n",
       "45                     Volleys  3.795030e+13  3.795453e+13  3.792699e+13   \n",
       "81                   Mentality  3.901254e+13  3.914657e+13  3.798656e+13   \n",
       "36                          LB  3.803766e+13  3.821657e+13  3.803114e+13   \n",
       "40                          RB  3.803766e+13  3.808355e+13  3.803114e+13   \n",
       "47                       Curve  3.807209e+13  3.814493e+13  3.804161e+13   \n",
       "31                         LWB  3.823767e+13  3.829537e+13  3.821822e+13   \n",
       "35                         RWB  3.823767e+13  3.817528e+13  3.821822e+13   \n",
       "7                  Skill Moves  3.846869e+13  3.847564e+13  3.846849e+13   \n",
       "80                       Power  3.926227e+13  3.944756e+13  3.847206e+13   \n",
       "76          Position_Potential  3.859366e+13  3.853693e+13  3.859348e+13   \n",
       "8                    Real Face  3.868142e+13  3.869908e+13  3.868123e+13   \n",
       "62               Interceptions  3.894271e+13  3.900229e+13  3.881148e+13   \n",
       "43             HeadingAccuracy  3.919032e+13  3.921825e+13  3.906376e+13   \n",
       "79                    Movement  3.950834e+13  3.948714e+13  3.911419e+13   \n",
       "48                  FKAccuracy  3.920840e+13  3.917831e+13  3.911572e+13   \n",
       "68              StandingTackle  3.923663e+13  3.938180e+13  3.916141e+13   \n",
       "69               SlidingTackle  3.935045e+13  3.930441e+13  3.919273e+13   \n",
       "65                   Penalties  3.948346e+13  3.943302e+13  3.938622e+13   \n",
       "67                     Marking  3.951506e+13  3.962020e+13  3.941244e+13   \n",
       "82                   Defending  4.000452e+13  4.023640e+13  3.941836e+13   \n",
       "39                         RCB  3.962185e+13  3.966552e+13  3.953114e+13   \n",
       "38                          CB  3.962185e+13  3.965921e+13  3.953114e+13   \n",
       "37                         LCB  3.962185e+13  3.965597e+13  3.953114e+13   \n",
       "61                  Aggression  4.030169e+13  4.035343e+13  4.016983e+13   \n",
       "71                  GKHandling  4.017113e+13  4.024809e+13  4.016995e+13   \n",
       "75             Club_Reputation  4.043025e+13  4.039635e+13  4.019410e+13   \n",
       "83                 GoalKeeping  4.070534e+13  4.059779e+13  4.029215e+13   \n",
       "70                    GKDiving  4.061509e+13  4.050402e+13  4.042991e+13   \n",
       "116             Continent_Asia  4.049309e+13  4.046750e+13  4.049280e+13   \n",
       "51                Acceleration  4.054570e+13  4.052450e+13  4.050439e+13   \n",
       "58                     Stamina  4.061329e+13  4.065371e+13  4.053882e+13   \n",
       "120    Continent_South America  4.055348e+13  4.055314e+13  4.055318e+13   \n",
       "123  Position_Group_GoalKeeper  4.056619e+13  4.058352e+13  4.056590e+13   \n",
       "93                 Position_GK  4.056619e+13  4.042336e+13  4.056590e+13   \n",
       "101                Position_LS  4.056743e+13  4.059826e+13  4.056713e+13   \n",
       "124  Position_Group_Midfielder  4.057505e+13  4.065053e+13  4.057475e+13   \n",
       "104               Position_RAM  4.057549e+13  4.064459e+13  4.057520e+13   \n",
       "94                Position_LAM  4.057999e+13  4.062341e+13  4.057970e+13   \n",
       "89                 Position_CB  4.058014e+13  4.051565e+13  4.057984e+13   \n",
       "111                Position_RS  4.058024e+13  4.054201e+13  4.057994e+13   \n",
       "119          Continent_Oceania  4.058075e+13  4.059081e+13  4.058046e+13   \n",
       "88                Position_CAM  4.058364e+13  4.053184e+13  4.058335e+13   \n",
       "100                Position_LM  4.058558e+13  4.061478e+13  4.058528e+13   \n",
       "110                Position_RM  4.058744e+13  4.060298e+13  4.058715e+13   \n",
       "108               Position_RDM  4.058798e+13  4.064590e+13  4.058769e+13   \n",
       "103               Position_LWB  4.058846e+13  4.056909e+13  4.058816e+13   \n",
       "113               Position_RWB  4.059027e+13  4.059812e+13  4.058998e+13   \n",
       "87                  Defense_WR  4.059084e+13  4.061972e+13  4.059055e+13   \n",
       "91                 Position_CF  4.059545e+13  4.060072e+13  4.059516e+13   \n",
       "114                Position_ST  4.059688e+13  4.065603e+13  4.059659e+13   \n",
       "86                   Attack_WR  4.059741e+13  4.052668e+13  4.059711e+13   \n",
       "98                Position_LDM  4.059742e+13  4.053577e+13  4.059713e+13   \n",
       "90                Position_CDM  4.059848e+13  4.062884e+13  4.059818e+13   \n",
       "107               Position_RCM  4.059857e+13  4.051201e+13  4.059827e+13   \n",
       "115           Continent_Africa  4.060050e+13  4.065196e+13  4.060021e+13   \n",
       "95                 Position_LB  4.060203e+13  4.061604e+13  4.060174e+13   \n",
       "97                Position_LCM  4.060453e+13  4.067682e+13  4.060423e+13   \n",
       "4               Preferred Foot  4.060673e+13  4.065177e+13  4.060644e+13   \n",
       "112                Position_RW  4.060816e+13  4.058124e+13  4.060786e+13   \n",
       "6                    Weak Foot  4.060848e+13  4.065150e+13  4.060818e+13   \n",
       "11                 Loaned From  4.061036e+13  4.064751e+13  4.061006e+13   \n",
       "106               Position_RCB  4.061562e+13  4.061802e+13  4.061532e+13   \n",
       "118    Continent_North America  4.061745e+13  4.061235e+13  4.061716e+13   \n",
       "105                Position_RB  4.062090e+13  4.057000e+13  4.062060e+13   \n",
       "102                Position_LW  4.062130e+13  4.062293e+13  4.062100e+13   \n",
       "96                Position_LCB  4.062844e+13  4.066746e+13  4.062815e+13   \n",
       "99                 Position_LF  4.063674e+13  4.067840e+13  4.063643e+13   \n",
       "122     Position_Group_Forward  4.063847e+13  4.062271e+13  4.063817e+13   \n",
       "13                      Height  4.064589e+13  4.067548e+13  4.064278e+13   \n",
       "53                     Agility  4.072630e+13  4.062334e+13  4.064309e+13   \n",
       "92                 Position_CM  4.065203e+13  4.072975e+13  4.065174e+13   \n",
       "72                   GKKicking  4.135352e+13  4.100044e+13  4.065761e+13   \n",
       "117           Continent_Europe  4.066388e+13  4.062134e+13  4.066358e+13   \n",
       "109                Position_RF  4.069445e+13  4.068104e+13  4.069413e+13   \n",
       "121    Position_Group_Defender  4.070186e+13  4.070087e+13  4.070156e+13   \n",
       "57                     Jumping  4.079931e+13  4.081470e+13  4.071576e+13   \n",
       "52                 SprintSpeed  4.079643e+13  4.088547e+13  4.073411e+13   \n",
       "85                   Body Type  4.073744e+13  4.078065e+13  4.073714e+13   \n",
       "14                      Weight  4.084161e+13  4.085690e+13  4.078900e+13   \n",
       "55                     Balance  4.091976e+13  4.095201e+13  4.083913e+13   \n",
       "9                Jersey Number  4.092715e+13  4.093653e+13  4.085776e+13   \n",
       "59                    Strength  4.103734e+13  4.108102e+13  4.095919e+13   \n",
       "12        Contract Valid Until  4.097543e+13  4.084194e+13  4.097354e+13   \n",
       "84        Potential_Difference  4.134340e+13  4.133834e+13  4.134291e+13   \n",
       "0                          Age  4.170572e+13  4.173941e+13  4.170205e+13   \n",
       "74                  GKReflexes  4.405168e+13  4.063388e+13  4.172554e+13   \n",
       "10                      Joined  4.730089e+13  4.530670e+13  4.236384e+13   \n",
       "73               GKPositioning  4.607656e+13  4.412526e+13  4.294646e+13   \n",
       "\n",
       "              LRE           SVC  \n",
       "1    6.343037e+13  3.559205e+13  \n",
       "50   3.976522e+13  3.559360e+13  \n",
       "22   4.016000e+13  3.559329e+13  \n",
       "18   4.016000e+13  3.559329e+13  \n",
       "23   4.009771e+13  3.559356e+13  \n",
       "25   4.009771e+13  3.559356e+13  \n",
       "24   4.009771e+13  3.559356e+13  \n",
       "21   4.010982e+13  3.559353e+13  \n",
       "19   4.010982e+13  3.559353e+13  \n",
       "20   4.010982e+13  3.559353e+13  \n",
       "30   4.015433e+13  3.559332e+13  \n",
       "26   4.015433e+13  3.559332e+13  \n",
       "44   3.969333e+13  3.559412e+13  \n",
       "2    4.168772e+13  3.559093e+13  \n",
       "46   4.004378e+13  3.559388e+13  \n",
       "54   4.535629e+13  3.559331e+13  \n",
       "15   4.012114e+13  3.559382e+13  \n",
       "16   4.012114e+13  3.559382e+13  \n",
       "17   4.012114e+13  3.559382e+13  \n",
       "64   3.969074e+13  3.559435e+13  \n",
       "27   4.001880e+13  3.559367e+13  \n",
       "29   4.001880e+13  3.559367e+13  \n",
       "28   4.001880e+13  3.559367e+13  \n",
       "66   4.021195e+13  3.559377e+13  \n",
       "77   3.964200e+13  3.559391e+13  \n",
       "3    3.936163e+13  3.559508e+13  \n",
       "78   3.960158e+13  3.559421e+13  \n",
       "63   4.008411e+13  3.559426e+13  \n",
       "49   3.972126e+13  3.559443e+13  \n",
       "5    3.702092e+13  3.559517e+13  \n",
       "34   4.018408e+13  3.559431e+13  \n",
       "33   4.018408e+13  3.559431e+13  \n",
       "32   4.018408e+13  3.559431e+13  \n",
       "60   3.983257e+13  3.559450e+13  \n",
       "56   3.973684e+13  3.559433e+13  \n",
       "42   4.018861e+13  3.559469e+13  \n",
       "41   4.002732e+13  3.559454e+13  \n",
       "45   3.986659e+13  3.559466e+13  \n",
       "81   3.928514e+13  3.559373e+13  \n",
       "36   4.026994e+13  3.559417e+13  \n",
       "40   4.026994e+13  3.559417e+13  \n",
       "47   3.979783e+13  3.559453e+13  \n",
       "31   4.022300e+13  3.559454e+13  \n",
       "35   4.022300e+13  3.559454e+13  \n",
       "7    3.957370e+13  3.558797e+13  \n",
       "80   3.972625e+13  3.559380e+13  \n",
       "76   4.097735e+13  3.559578e+13  \n",
       "8    3.868142e+13  3.559496e+13  \n",
       "62   4.046144e+13  3.559447e+13  \n",
       "43   4.018653e+13  3.559467e+13  \n",
       "79   4.045258e+13  3.559410e+13  \n",
       "48   3.991706e+13  3.559464e+13  \n",
       "68   4.051417e+13  3.559436e+13  \n",
       "69   4.054869e+13  3.559440e+13  \n",
       "65   3.999467e+13  3.559464e+13  \n",
       "67   4.046294e+13  3.559463e+13  \n",
       "82   4.050480e+13  3.559433e+13  \n",
       "39   4.035521e+13  3.559418e+13  \n",
       "38   4.035521e+13  3.559418e+13  \n",
       "37   4.035521e+13  3.559418e+13  \n",
       "61   4.028812e+13  3.559470e+13  \n",
       "71   4.058494e+13  3.559492e+13  \n",
       "75   3.949069e+13  3.559478e+13  \n",
       "83   4.058338e+13  3.559438e+13  \n",
       "70   4.058557e+13  3.559484e+13  \n",
       "116  4.049309e+13  3.559517e+13  \n",
       "51   4.063825e+13  3.559464e+13  \n",
       "58   4.044311e+13  3.559439e+13  \n",
       "120  4.055348e+13  3.559474e+13  \n",
       "123  4.056619e+13  3.559532e+13  \n",
       "93   4.056619e+13  3.559532e+13  \n",
       "101  4.056743e+13  3.559514e+13  \n",
       "124  4.057505e+13  3.559397e+13  \n",
       "104  4.057549e+13  3.559514e+13  \n",
       "94   4.057999e+13  3.559514e+13  \n",
       "89   4.058014e+13  3.559516e+13  \n",
       "111  4.058024e+13  3.559514e+13  \n",
       "119  4.058075e+13  3.559514e+13  \n",
       "88   4.058364e+13  3.559511e+13  \n",
       "100  4.058558e+13  3.559509e+13  \n",
       "110  4.058744e+13  3.559511e+13  \n",
       "108  4.058798e+13  3.559514e+13  \n",
       "103  4.058846e+13  3.559514e+13  \n",
       "113  4.059027e+13  3.559515e+13  \n",
       "87   4.062020e+13  3.559449e+13  \n",
       "91   4.059545e+13  3.559514e+13  \n",
       "114  4.059688e+13  3.559509e+13  \n",
       "86   4.061665e+13  3.559312e+13  \n",
       "98   4.059742e+13  3.559514e+13  \n",
       "90   4.059848e+13  3.559514e+13  \n",
       "107  4.059857e+13  3.559513e+13  \n",
       "115  4.060050e+13  3.559506e+13  \n",
       "95   4.060203e+13  3.559514e+13  \n",
       "97   4.060453e+13  3.559513e+13  \n",
       "4    4.060673e+13  3.559479e+13  \n",
       "112  4.060816e+13  3.559513e+13  \n",
       "6    4.062002e+13  3.559302e+13  \n",
       "11   4.061036e+13  3.559503e+13  \n",
       "106  4.061562e+13  3.559512e+13  \n",
       "118  4.061745e+13  3.559514e+13  \n",
       "105  4.062090e+13  3.559514e+13  \n",
       "102  4.062130e+13  3.559513e+13  \n",
       "96   4.062844e+13  3.559513e+13  \n",
       "99   4.063674e+13  3.559514e+13  \n",
       "122  4.063847e+13  3.559463e+13  \n",
       "13   4.059698e+13  3.559501e+13  \n",
       "53   4.055930e+13  3.559455e+13  \n",
       "92   4.065203e+13  3.559518e+13  \n",
       "72   4.057962e+13  3.559474e+13  \n",
       "117  4.066388e+13  3.559486e+13  \n",
       "109  4.069445e+13  3.559514e+13  \n",
       "121  4.070186e+13  3.559457e+13  \n",
       "57   4.062103e+13  3.559474e+13  \n",
       "52   4.063930e+13  3.559450e+13  \n",
       "85   4.072909e+13  3.559507e+13  \n",
       "14   4.067301e+13  3.559496e+13  \n",
       "55   4.064931e+13  3.559493e+13  \n",
       "9    4.071989e+13  3.559471e+13  \n",
       "59   4.075654e+13  3.559478e+13  \n",
       "12   4.115312e+13  3.559390e+13  \n",
       "84   4.143626e+13  3.559468e+13  \n",
       "0    4.218390e+13  3.559363e+13  \n",
       "74   4.058282e+13  3.559485e+13  \n",
       "10   4.082795e+13  3.559494e+13  \n",
       "73   4.058505e+13  3.559484e+13  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s = {'Feature': col_names[0],\n",
    "     'DTR': cv_score_features['1'],\n",
    "     'RFR': cv_score_features['2'],\n",
    "     'XBR': cv_score_features['3'],\n",
    "     'LRE': cv_score_features['4'],\n",
    "     'SVC': cv_score_features['5']}\n",
    "feature_scores=pd.DataFrame(data=s)\n",
    "feature_scores=feature_scores.sort_values(by=['XBR'])\n",
    "feature_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2160x2880 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importance = pd.DataFrame(dict(graph=feature_scores['Feature'],\n",
    "                               DTR=feature_scores['DTR'],\n",
    "                               RFR=feature_scores['RFR'],\n",
    "                               XBR=feature_scores['XBR'],\n",
    "                               LRE=feature_scores['LRE'],\n",
    "                               SVC=feature_scores['SVC'])) \n",
    "\n",
    "ind = np.arange(len(importance))\n",
    "width = 0.15\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(30,40))\n",
    "ax.barh(ind, importance.DTR, width, color='maroon', label='DTC')\n",
    "ax.barh(ind + width, importance.RFR, width, color='forestgreen', label='RFR')\n",
    "ax.barh(ind + 2*width, importance.XBR, width, color='slategrey', label='XBR')\n",
    "ax.barh(ind + 3*width, importance.LRE, width, color='lightsalmon', label='LRE')\n",
    "ax.set_xlim(0, max(importance.DTR)+min(importance.DTR))\n",
    "plt.xlabel('Importance', fontsize=16)\n",
    "plt.ylabel('Features', fontsize=16)\n",
    "plt.title('Feature Importance (DTR,RFR,XGB,LRE,SVC)',fontsize=20)\n",
    "\n",
    "ax.set(yticks=ind + width, yticklabels=importance.graph, ylim=[2*width - 1, len(importance)])\n",
    "ax.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "According to this method, some position skills, as well as Ball Control and the Overall score are the ten most important features for a low RMSE. \n",
    "\n",
    "Based on our feature importance analysis, we are making the following modifications:\n",
    "- Group position skills into 'Forward Skills', 'Midfield Skills' and 'Defence Skills' and drop single position skills to avoid overemphasizing these features\n",
    "- Drop individual position dummies and only keep the grouped position dummies to avoid duplication of features\n",
    "- Drop all features for that the F-value makes up less than one percent of the summ of all F-values, EXCEPT for when one of these features is within the top ten features for either the ExtraTreeRegressor or the single feature RMSE."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Grouping position skills and drop single position skills:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_df['Forward_Skill'] = predictors_df.loc[:,['LS',  'ST', 'RS',  'LW', 'LF', 'CF', 'RF', 'RW']].mean(axis=1)\n",
    "\n",
    "predictors_df['Midfield_Skill'] = predictors_df.loc[:,['LAM','CAM','RAM', 'LM', 'LCM', 'CM' ,'RCM', 'RM','LDM', 'CDM', 'RDM']].mean(axis=1)\n",
    "\n",
    "predictors_df['Defence_Skill'] = predictors_df.loc[:,['LWB','RWB', 'LB','LCB','CB','RCB','RB']].mean(axis=1)\n",
    "\n",
    "predictors_df = predictors_df.drop(['LS',  'ST', 'RS',  'LW', 'LF', 'CF', 'RF', 'RW',\n",
    "          'LAM','CAM','RAM', 'LM', 'LCM', 'CM' ,'RCM', 'RM','LDM', 'CDM', 'RDM',\n",
    "          'LWB','RWB', 'LB','LCB','CB','RCB','RB'], axis = 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop individual position dummies and only keep grouped position dummies:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_df=predictors_df.drop(predictors_df.iloc[:,predictors_df.columns.get_loc('Position_CAM'):predictors_df.columns.get_loc('Position_ST')+1], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we are dropping all features for which the F-value makes up less than one percent of the total sum of F-values, and which are not among the top ten features neither for RMSE nor ExtraTreeRegressor. To do so, we are indexing the most important features first, extract all values that we do not want to keep and use these as our index in the drop function. Also, we are keeping the Position Groups (Goalkeeper, Defender, Midfielder, Forward), since Goalkeepers have a significantly different value distribution than the other groups. We, therefore, think it will be important to take this into account."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Age</th>\n",
       "      <th>Overall</th>\n",
       "      <th>Potential</th>\n",
       "      <th>Special</th>\n",
       "      <th>Preferred Foot</th>\n",
       "      <th>International Reputation</th>\n",
       "      <th>Weak Foot</th>\n",
       "      <th>Skill Moves</th>\n",
       "      <th>Real Face</th>\n",
       "      <th>Jersey Number</th>\n",
       "      <th>Joined</th>\n",
       "      <th>Loaned From</th>\n",
       "      <th>Contract Valid Until</th>\n",
       "      <th>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Crossing</th>\n",
       "      <th>Finishing</th>\n",
       "      <th>HeadingAccuracy</th>\n",
       "      <th>ShortPassing</th>\n",
       "      <th>Volleys</th>\n",
       "      <th>Dribbling</th>\n",
       "      <th>Curve</th>\n",
       "      <th>FKAccuracy</th>\n",
       "      <th>LongPassing</th>\n",
       "      <th>BallControl</th>\n",
       "      <th>Acceleration</th>\n",
       "      <th>SprintSpeed</th>\n",
       "      <th>Agility</th>\n",
       "      <th>Reactions</th>\n",
       "      <th>Balance</th>\n",
       "      <th>ShotPower</th>\n",
       "      <th>Jumping</th>\n",
       "      <th>Stamina</th>\n",
       "      <th>Strength</th>\n",
       "      <th>LongShots</th>\n",
       "      <th>Aggression</th>\n",
       "      <th>Interceptions</th>\n",
       "      <th>Positioning</th>\n",
       "      <th>Vision</th>\n",
       "      <th>Penalties</th>\n",
       "      <th>Composure</th>\n",
       "      <th>Marking</th>\n",
       "      <th>StandingTackle</th>\n",
       "      <th>SlidingTackle</th>\n",
       "      <th>GKDiving</th>\n",
       "      <th>GKHandling</th>\n",
       "      <th>GKKicking</th>\n",
       "      <th>GKPositioning</th>\n",
       "      <th>GKReflexes</th>\n",
       "      <th>Club_Reputation</th>\n",
       "      <th>Position_Potential</th>\n",
       "      <th>Attacking</th>\n",
       "      <th>Skill</th>\n",
       "      <th>Movement</th>\n",
       "      <th>Power</th>\n",
       "      <th>Mentality</th>\n",
       "      <th>Defending</th>\n",
       "      <th>GoalKeeping</th>\n",
       "      <th>Potential_Difference</th>\n",
       "      <th>Body Type</th>\n",
       "      <th>Attack_WR</th>\n",
       "      <th>Defense_WR</th>\n",
       "      <th>Continent_Africa</th>\n",
       "      <th>Continent_Asia</th>\n",
       "      <th>Continent_Europe</th>\n",
       "      <th>Continent_North America</th>\n",
       "      <th>Continent_Oceania</th>\n",
       "      <th>Continent_South America</th>\n",
       "      <th>Position_Group_Defender</th>\n",
       "      <th>Position_Group_Forward</th>\n",
       "      <th>Position_Group_GoalKeeper</th>\n",
       "      <th>Position_Group_Midfielder</th>\n",
       "      <th>Forward_Skill</th>\n",
       "      <th>Midfield_Skill</th>\n",
       "      <th>Defence_Skill</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>94</td>\n",
       "      <td>94</td>\n",
       "      <td>2202</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>10.0</td>\n",
       "      <td>5297</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>170</td>\n",
       "      <td>159</td>\n",
       "      <td>84.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>97.0</td>\n",
       "      <td>93.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>59.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>33.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2.242424</td>\n",
       "      <td>2</td>\n",
       "      <td>85.0</td>\n",
       "      <td>93.4</td>\n",
       "      <td>91.6</td>\n",
       "      <td>75.6</td>\n",
       "      <td>71.500000</td>\n",
       "      <td>29.000000</td>\n",
       "      <td>10.8</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>90.875</td>\n",
       "      <td>81.454545</td>\n",
       "      <td>55.285714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>33</td>\n",
       "      <td>94</td>\n",
       "      <td>94</td>\n",
       "      <td>2228</td>\n",
       "      <td>1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>175</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>188</td>\n",
       "      <td>183</td>\n",
       "      <td>84.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>89.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>88.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>89.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>88.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>93.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>29.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>23.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>2.520000</td>\n",
       "      <td>3</td>\n",
       "      <td>87.0</td>\n",
       "      <td>83.2</td>\n",
       "      <td>86.6</td>\n",
       "      <td>90.0</td>\n",
       "      <td>74.833333</td>\n",
       "      <td>27.333333</td>\n",
       "      <td>11.6</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>90.125</td>\n",
       "      <td>78.727273</td>\n",
       "      <td>58.714286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>26</td>\n",
       "      <td>92</td>\n",
       "      <td>93</td>\n",
       "      <td>2143</td>\n",
       "      <td>1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "      <td>10.0</td>\n",
       "      <td>516</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>175</td>\n",
       "      <td>150</td>\n",
       "      <td>79.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>88.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>49.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>89.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>33.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>2.233333</td>\n",
       "      <td>3</td>\n",
       "      <td>79.2</td>\n",
       "      <td>88.8</td>\n",
       "      <td>91.6</td>\n",
       "      <td>70.6</td>\n",
       "      <td>73.833333</td>\n",
       "      <td>28.000000</td>\n",
       "      <td>11.8</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>87.125</td>\n",
       "      <td>78.727273</td>\n",
       "      <td>55.857143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>27</td>\n",
       "      <td>91</td>\n",
       "      <td>93</td>\n",
       "      <td>1471</td>\n",
       "      <td>1</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2741</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>193</td>\n",
       "      <td>168</td>\n",
       "      <td>17.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>57.0</td>\n",
       "      <td>58.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>43.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>88.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>2.090909</td>\n",
       "      <td>0</td>\n",
       "      <td>22.8</td>\n",
       "      <td>30.2</td>\n",
       "      <td>61.6</td>\n",
       "      <td>43.4</td>\n",
       "      <td>42.666667</td>\n",
       "      <td>16.333333</td>\n",
       "      <td>88.8</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>27</td>\n",
       "      <td>91</td>\n",
       "      <td>92</td>\n",
       "      <td>2281</td>\n",
       "      <td>1</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1220</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>180</td>\n",
       "      <td>154</td>\n",
       "      <td>93.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>92.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>83.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>88.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>58.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>2.090909</td>\n",
       "      <td>3</td>\n",
       "      <td>80.8</td>\n",
       "      <td>87.2</td>\n",
       "      <td>80.2</td>\n",
       "      <td>82.0</td>\n",
       "      <td>80.833333</td>\n",
       "      <td>59.000000</td>\n",
       "      <td>11.2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>85.125</td>\n",
       "      <td>84.727273</td>\n",
       "      <td>71.142857</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Age  Overall  Potential  Special Preferred Foot International Reputation  \\\n",
       "0   31       94         94     2202              0                      5.0   \n",
       "1   33       94         94     2228              1                      5.0   \n",
       "2   26       92         93     2143              1                      5.0   \n",
       "3   27       91         93     1471              1                      4.0   \n",
       "4   27       91         92     2281              1                      4.0   \n",
       "\n",
       "  Weak Foot Skill Moves Real Face  Jersey Number  Joined  Loaned From  \\\n",
       "0       4.0         4.0         1           10.0    5297            0   \n",
       "1       4.0         5.0         1            7.0     175            0   \n",
       "2       5.0         5.0         1           10.0     516            0   \n",
       "3       3.0         1.0         1            1.0    2741            0   \n",
       "4       5.0         4.0         1            7.0    1220            0   \n",
       "\n",
       "   Contract Valid Until  Height  Weight  Crossing  Finishing  HeadingAccuracy  \\\n",
       "0                     3     170     159      84.0       95.0             70.0   \n",
       "1                     4     188     183      84.0       94.0             89.0   \n",
       "2                     4     175     150      79.0       87.0             62.0   \n",
       "3                     2     193     168      17.0       13.0             21.0   \n",
       "4                     5     180     154      93.0       82.0             55.0   \n",
       "\n",
       "   ShortPassing  Volleys  Dribbling  Curve  FKAccuracy  LongPassing  \\\n",
       "0          90.0     86.0       97.0   93.0        94.0         87.0   \n",
       "1          81.0     87.0       88.0   81.0        76.0         77.0   \n",
       "2          84.0     84.0       96.0   88.0        87.0         78.0   \n",
       "3          50.0     13.0       18.0   21.0        19.0         51.0   \n",
       "4          92.0     82.0       86.0   85.0        83.0         91.0   \n",
       "\n",
       "   BallControl  Acceleration  SprintSpeed  Agility  Reactions  Balance  \\\n",
       "0         96.0          91.0         86.0     91.0       95.0     95.0   \n",
       "1         94.0          89.0         91.0     87.0       96.0     70.0   \n",
       "2         95.0          94.0         90.0     96.0       94.0     84.0   \n",
       "3         42.0          57.0         58.0     60.0       90.0     43.0   \n",
       "4         91.0          78.0         76.0     79.0       91.0     77.0   \n",
       "\n",
       "   ShotPower  Jumping  Stamina  Strength  LongShots  Aggression  \\\n",
       "0       85.0     68.0     72.0      59.0       94.0        48.0   \n",
       "1       95.0     95.0     88.0      79.0       93.0        63.0   \n",
       "2       80.0     61.0     81.0      49.0       82.0        56.0   \n",
       "3       31.0     67.0     43.0      64.0       12.0        38.0   \n",
       "4       91.0     63.0     90.0      75.0       91.0        76.0   \n",
       "\n",
       "   Interceptions  Positioning  Vision  Penalties  Composure  Marking  \\\n",
       "0           22.0         94.0    94.0       75.0       96.0     33.0   \n",
       "1           29.0         95.0    82.0       85.0       95.0     28.0   \n",
       "2           36.0         89.0    87.0       81.0       94.0     27.0   \n",
       "3           30.0         12.0    68.0       40.0       68.0     15.0   \n",
       "4           61.0         87.0    94.0       79.0       88.0     68.0   \n",
       "\n",
       "   StandingTackle  SlidingTackle  GKDiving  GKHandling  GKKicking  \\\n",
       "0            28.0           26.0       6.0        11.0       15.0   \n",
       "1            31.0           23.0       7.0        11.0       15.0   \n",
       "2            24.0           33.0       9.0         9.0       15.0   \n",
       "3            21.0           13.0      90.0        85.0       87.0   \n",
       "4            58.0           51.0      15.0        13.0        5.0   \n",
       "\n",
       "   GKPositioning  GKReflexes  Club_Reputation  Position_Potential  Attacking  \\\n",
       "0           14.0         8.0         2.242424                   2       85.0   \n",
       "1           14.0        11.0         2.520000                   3       87.0   \n",
       "2           15.0        11.0         2.233333                   3       79.2   \n",
       "3           88.0        94.0         2.090909                   0       22.8   \n",
       "4           10.0        13.0         2.090909                   3       80.8   \n",
       "\n",
       "   Skill  Movement  Power  Mentality  Defending  GoalKeeping  \\\n",
       "0   93.4      91.6   75.6  71.500000  29.000000         10.8   \n",
       "1   83.2      86.6   90.0  74.833333  27.333333         11.6   \n",
       "2   88.8      91.6   70.6  73.833333  28.000000         11.8   \n",
       "3   30.2      61.6   43.4  42.666667  16.333333         88.8   \n",
       "4   87.2      80.2   82.0  80.833333  59.000000         11.2   \n",
       "\n",
       "   Potential_Difference Body Type Attack_WR Defense_WR Continent_Africa  \\\n",
       "0                     0         1         2          2                0   \n",
       "1                     0         2         3          1                0   \n",
       "2                     1         1         3          2                0   \n",
       "3                     2         1         2          2                0   \n",
       "4                     1         2         3          3                0   \n",
       "\n",
       "  Continent_Asia Continent_Europe Continent_North America Continent_Oceania  \\\n",
       "0              0                0                       0                 0   \n",
       "1              0                1                       0                 0   \n",
       "2              0                0                       0                 0   \n",
       "3              0                1                       0                 0   \n",
       "4              0                1                       0                 0   \n",
       "\n",
       "  Continent_South America Position_Group_Defender Position_Group_Forward  \\\n",
       "0                       1                       0                      1   \n",
       "1                       0                       0                      1   \n",
       "2                       1                       0                      1   \n",
       "3                       0                       0                      0   \n",
       "4                       0                       0                      0   \n",
       "\n",
       "  Position_Group_GoalKeeper Position_Group_Midfielder  Forward_Skill  \\\n",
       "0                         0                         0         90.875   \n",
       "1                         0                         0         90.125   \n",
       "2                         0                         0         87.125   \n",
       "3                         1                         0          0.000   \n",
       "4                         0                         1         85.125   \n",
       "\n",
       "   Midfield_Skill  Defence_Skill  \n",
       "0       81.454545      55.285714  \n",
       "1       78.727273      58.714286  \n",
       "2       78.727273      55.857143  \n",
       "3        0.000000       0.000000  \n",
       "4       84.727273      71.142857  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictors_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "featureScores_topten=list(featureScores.sort_values(by='Expl_percent', ascending=False)[0:10]['Specs'])\n",
    "feat_importances_topten=list(feat_importances.sort_values(ascending=False)[0:10].index.values)\n",
    "feature_scores_topten=list(feature_scores[0:10]['Feature'])\n",
    "Position_features=['Position_Group_Defender','Position_Group_Forward','Position_Group_GoalKeeper','Position_Group_Midfielder']\n",
    "features_keep=list(set(featureScores_topten+feat_importances_topten+feature_scores_topten+Position_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Matching features to keep with features left in predictors_df\n",
    "features_keep=set(features_keep)& set(predictors_df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_drop=predictors_df.columns.drop(features_keep)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_df=predictors_df.drop(features_drop,axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Overall</th>\n",
       "      <th>Potential</th>\n",
       "      <th>Special</th>\n",
       "      <th>International Reputation</th>\n",
       "      <th>Real Face</th>\n",
       "      <th>BallControl</th>\n",
       "      <th>Reactions</th>\n",
       "      <th>Vision</th>\n",
       "      <th>Composure</th>\n",
       "      <th>Club_Reputation</th>\n",
       "      <th>Mentality</th>\n",
       "      <th>Position_Group_Defender</th>\n",
       "      <th>Position_Group_Forward</th>\n",
       "      <th>Position_Group_GoalKeeper</th>\n",
       "      <th>Position_Group_Midfielder</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>94</td>\n",
       "      <td>94</td>\n",
       "      <td>2202</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "      <td>96.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>2.242424</td>\n",
       "      <td>71.500000</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>94</td>\n",
       "      <td>94</td>\n",
       "      <td>2228</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "      <td>94.0</td>\n",
       "      <td>96.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>2.520000</td>\n",
       "      <td>74.833333</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>92</td>\n",
       "      <td>93</td>\n",
       "      <td>2143</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "      <td>95.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>2.233333</td>\n",
       "      <td>73.833333</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>91</td>\n",
       "      <td>93</td>\n",
       "      <td>1471</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>42.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>2.090909</td>\n",
       "      <td>42.666667</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>91</td>\n",
       "      <td>92</td>\n",
       "      <td>2281</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>91.0</td>\n",
       "      <td>91.0</td>\n",
       "      <td>94.0</td>\n",
       "      <td>88.0</td>\n",
       "      <td>2.090909</td>\n",
       "      <td>80.833333</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Overall  Potential  Special International Reputation Real Face  \\\n",
       "0       94         94     2202                      5.0         1   \n",
       "1       94         94     2228                      5.0         1   \n",
       "2       92         93     2143                      5.0         1   \n",
       "3       91         93     1471                      4.0         1   \n",
       "4       91         92     2281                      4.0         1   \n",
       "\n",
       "   BallControl  Reactions  Vision  Composure  Club_Reputation  Mentality  \\\n",
       "0         96.0       95.0    94.0       96.0         2.242424  71.500000   \n",
       "1         94.0       96.0    82.0       95.0         2.520000  74.833333   \n",
       "2         95.0       94.0    87.0       94.0         2.233333  73.833333   \n",
       "3         42.0       90.0    68.0       68.0         2.090909  42.666667   \n",
       "4         91.0       91.0    94.0       88.0         2.090909  80.833333   \n",
       "\n",
       "  Position_Group_Defender Position_Group_Forward Position_Group_GoalKeeper  \\\n",
       "0                       0                      1                         0   \n",
       "1                       0                      1                         0   \n",
       "2                       0                      1                         0   \n",
       "3                       0                      0                         1   \n",
       "4                       0                      0                         0   \n",
       "\n",
       "  Position_Group_Midfielder  \n",
       "0                         0  \n",
       "1                         0  \n",
       "2                         0  \n",
       "3                         0  \n",
       "4                         1  "
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictors_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can use the reduced data frame for predicting values again now, and compare the results to our baseline models.\n",
    "\n",
    "### 2.4) Prediction on reduced dataset\n",
    "\n",
    "We first split the data into train and test set. As before, we are using stratified sampling evenly distribute the outliers among samples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_predictors_val, test_predictors_val, train_target_val, test_target_val = train_test_split(predictors_df,\n",
    "                                                                                                target_val,\n",
    "                                                                                                test_size=0.2,\n",
    "                                                                                                stratify=y_binned)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before using the data for prediction, we have to scale them again as we did in the beginning."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "\n",
    "train_predictors_val[train_predictors_val.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']).columns] = scaler.fit_transform(train_predictors_val.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']))\n",
    "\n",
    "test_predictors_val[test_predictors_val.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']).columns] = scaler.fit_transform(test_predictors_val.select_dtypes(include=['float64','float32','int64','int32'],exclude=['category']))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_predictors_val[train_predictors_val.select_dtypes(include='category').columns]=train_predictors_val[train_predictors_val.select_dtypes(include='category').columns].astype('int')\n",
    "test_predictors_val[test_predictors_val.select_dtypes(include='category').columns]=test_predictors_val[test_predictors_val.select_dtypes(include='category').columns].astype('int')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After we have prepared the train and test split, we fit the models and calculate the respective CV, Train and Test errors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py:245: FutureWarning:\n",
      "\n",
      "The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "\n",
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:37:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:193: FutureWarning:\n",
      "\n",
      "The default value of gamma will change from 'auto' to 'scale' in version 0.22 to account better for unscaled features. Set gamma explicitly to 'auto' or 'scale' to avoid this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "dtr=DecisionTreeRegressor()\n",
    "rfr=RandomForestRegressor()\n",
    "xbr=xgb.XGBRegressor()\n",
    "lre=LinearRegression()\n",
    "svr=SVR()\n",
    "\n",
    "models = [dtr,rfr,xbr,lre,svr]\n",
    "\n",
    "for mod in models:\n",
    "    mod.fit(train_predictors_val, train_target_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_error = []\n",
    "\n",
    "# Sequentially fit all models and calculate test score\n",
    "for mod in models:\n",
    "    y_pred = mod.predict(test_predictors_val)\n",
    "    mse_mod = MSE(test_target_val, y_pred)\n",
    "    rmse_mod = round(mse_mod**(1/2))\n",
    "    test_error.append(rmse_mod)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "cv_error=[]\n",
    "\n",
    "for mod in models:\n",
    "    MSE_cv_error = -cross_val_score(mod, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "    RMSE_cv = round((MSE_cv_error.mean())**(1/2))\n",
    "    cv_error.append(RMSE_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_error=[]\n",
    "\n",
    "# Sequentially fit all models and calculate train score\n",
    "for mod in models:\n",
    "    y_pred = mod.predict(train_predictors_val)\n",
    "    mse_mod = MSE(train_target_val, y_pred)\n",
    "    rmse_mod = round(mse_mod**(1/2))\n",
    "    train_error.append(rmse_mod)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_names = ['dtr','rfr','xbr','lre','svc']\n",
    "\n",
    "error = pd.DataFrame(dict(graph=model_names,test_error=test_error, cv_error=cv_error, train_error=train_error)) \n",
    "\n",
    "ind = np.arange(len(error))\n",
    "width = 0.2\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12,6))\n",
    "ax.barh(ind, error.test_error, width, color='maroon', label='test')\n",
    "ax.barh(ind + width, error.cv_error, width, color='forestgreen', label='cv')\n",
    "ax.barh(ind + 2*width, error.train_error, width, color='slategrey', label='train')\n",
    "ax.set_xlim(0, max(error.cv_error)+min(error.cv_error))\n",
    "plt.xlabel('RMSE', fontsize=16)\n",
    "plt.ylabel('Algorithms', fontsize=16)\n",
    "plt.title('Test, Cross Validation and Train Error per Algorithm',fontsize=20)\n",
    "\n",
    "ax.set(yticks=ind + width, yticklabels=error.graph, ylim=[2*width - 1, len(error)])\n",
    "ax.legend()\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>graph</th>\n",
       "      <th>test_error</th>\n",
       "      <th>cv_error</th>\n",
       "      <th>train_error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>dtr</td>\n",
       "      <td>1607480.0</td>\n",
       "      <td>1652020.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>rfr</td>\n",
       "      <td>1109561.0</td>\n",
       "      <td>1252162.0</td>\n",
       "      <td>539757.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>xbr</td>\n",
       "      <td>1084035.0</td>\n",
       "      <td>1185917.0</td>\n",
       "      <td>765875.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>lre</td>\n",
       "      <td>3587924.0</td>\n",
       "      <td>3438748.0</td>\n",
       "      <td>3434508.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>svc</td>\n",
       "      <td>6097043.0</td>\n",
       "      <td>5803173.0</td>\n",
       "      <td>5805194.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  graph  test_error   cv_error  train_error\n",
       "0   dtr   1607480.0  1652020.0          0.0\n",
       "1   rfr   1109561.0  1252162.0     539757.0\n",
       "2   xbr   1084035.0  1185917.0     765875.0\n",
       "3   lre   3587924.0  3438748.0    3434508.0\n",
       "4   svc   6097043.0  5803173.0    5805194.0"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compared to our first baseline model, the linear regression became a lot more stable, as the test error reduced significantly. The other models where almost uneffected by the reduced data frame. This means we did not loose any predictive power by reducing our dataset, and can continue with the parameter tuning. Decision Tree, Random Forest and XGBoost look most promising.\n",
    "\n",
    "Final look at the feature importance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py:245: FutureWarning:\n",
      "\n",
      "The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2880x1800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = ExtraTreesRegressor()\n",
    "\n",
    "X = train_predictors_val  \n",
    "y = train_target_val\n",
    "\n",
    "model.fit(X,y)\n",
    "\n",
    "#plot graph of feature importances for better visualization\n",
    "fig, ax = plt.subplots(figsize=(40,25))\n",
    "feat_importances = pd.Series(model.feature_importances_, index=X.columns)\n",
    "feat_importances.nlargest(125).plot(kind='barh')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.) Parameter tuning\n",
    "\n",
    "We are using GridSearch to iterate over a list of possible parameters for the most important parameters of each model.\n",
    "\n",
    "### 3.1) Decision Tree Regressor"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We are starting by tuning parameters for the Decision Tree Regressor. As most important features we are identifying the maximum depth, minimum sample leafs, minimum sample split and the maximum number of features to be considered when for the best split. As we can see from the baseline model, the Decision Tree Regressor at this point is highly overfit. \n",
    "\n",
    "**Default settings:**<br>\n",
    "DecisionTreeRegressor(criterion=’mse’, splitter=’best’, max_depth=None, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0.0, max_features=None, random_state=None, max_leaf_nodes=None, min_impurity_decrease=0.0, min_impurity_split=None, presort=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 10 folds for each of 800 candidates, totalling 8000 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-2856077542264.771\n",
      "DecisionTreeRegressor(criterion='mse', max_depth=9, max_features=4,\n",
      "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
      "                      min_impurity_split=None, min_samples_leaf=3,\n",
      "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
      "                      presort=False, random_state=None, splitter='best')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done 8000 out of 8000 | elapsed:  2.3min finished\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "# Parameter's to search\n",
    "max_depth = [1,2,3,4,5,6,7,8,9,10]\n",
    "min_samples_split = [0.1,0.3,0.5,0.7,2]\n",
    "min_samples_leaf = [1,2,3,4]\n",
    "max_features = [1,2,3,4]\n",
    "\n",
    "# Setting up parameter grid\n",
    "hyperparams = {'max_depth': max_depth, 'min_samples_split': min_samples_split, 'min_samples_leaf': min_samples_leaf, \n",
    "               'max_features': max_features}\n",
    "\n",
    "# Run GridSearch CV\n",
    "gd=GridSearchCV(estimator = dtr, param_grid = hyperparams, \n",
    "                verbose=True, cv=10, scoring = 'neg_mean_squared_error')\n",
    "\n",
    "# Fitting model and return results\n",
    "gd.fit(train_predictors_val, train_target_val)\n",
    "print(gd.best_score_)\n",
    "print(gd.best_estimator_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After using GridSearch to identify the best parameters for minimizing the mean squared error, we also have to make sure the model is not under- or overfit."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "dtr=DecisionTreeRegressor(criterion='mse', max_depth=9, max_features=4,\n",
    "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
    "                      min_impurity_split=None, min_samples_leaf=3,\n",
    "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
    "                      presort=False, random_state=None, splitter='best')\n",
    "\n",
    "dtr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = dtr.predict(test_predictors_val)\n",
    "test_mse_dtr = MSE(test_target_val, y_pred)\n",
    "test_rmse_dtr = round(test_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_dtr = -cross_val_score(dtr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_dtr = round((MSE_cv_error_dtr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = dtr.predict(train_predictors_val)\n",
    "train_mse_dtr = MSE(train_target_val, y_pred)\n",
    "train_rmse_dtr = round(train_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>DTR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1817524.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1777679.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>1211908.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             DTR errors\n",
       "Test error    1817524.0\n",
       "CV error      1777679.0\n",
       "Train error   1211908.0"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_dtr,RMSE_cv_dtr,train_rmse_dtr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['DTR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The train error is significantly lower than the test and cv error, which is an indicator for overfitting. Therefore, we are manually reducing complexity beginning with the optimized parameters model.\n",
    "\n",
    "We recognized that we the maximum depth is quite high. A high maximum depth is more prone to overfitting. We try to reduce the depth to six."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "dtr=DecisionTreeRegressor(criterion='mse', max_depth=6, max_features=4,\n",
    "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
    "                      min_impurity_split=None, min_samples_leaf=3,\n",
    "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
    "                      presort=False, random_state=None, splitter='best')\n",
    "\n",
    "dtr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = dtr.predict(test_predictors_val)\n",
    "test_mse_dtr = MSE(test_target_val, y_pred)\n",
    "test_rmse_dtr = round(test_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_dtr = -cross_val_score(dtr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_dtr = round((MSE_cv_error_dtr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = dtr.predict(train_predictors_val)\n",
    "train_mse_dtr = MSE(train_target_val, y_pred)\n",
    "train_rmse_dtr = round(train_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>DTR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1990149.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1961092.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>1774496.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             DTR errors\n",
       "Test error    1990149.0\n",
       "CV error      1961092.0\n",
       "Train error   1774496.0"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_dtr,RMSE_cv_dtr,train_rmse_dtr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['DTR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model still shows some overfitting, but at least the CV and training error are closer now. However, we need to further reduce complexity. We do so by increasing the minimum sample leaf from three to five."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "dtr=DecisionTreeRegressor(criterion='mse', max_depth=6, max_features=4,\n",
    "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
    "                      min_impurity_split=None, min_samples_leaf=5,\n",
    "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
    "                      presort=False, random_state=None, splitter='best')\n",
    "\n",
    "dtr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = dtr.predict(test_predictors_val)\n",
    "test_mse_dtr = MSE(test_target_val, y_pred)\n",
    "test_rmse_dtr = round(test_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_dtr = -cross_val_score(dtr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_dtr = round((MSE_cv_error_dtr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = dtr.predict(train_predictors_val)\n",
    "train_mse_dtr = MSE(train_target_val, y_pred)\n",
    "train_rmse_dtr = round(train_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>DTR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>2261811.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1972206.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>1724263.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             DTR errors\n",
       "Test error    2261811.0\n",
       "CV error      1972206.0\n",
       "Train error   1724263.0"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_dtr,RMSE_cv_dtr,train_rmse_dtr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['DTR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "CV and train error are pretty much aligned now. The test error is still too high though. Further complexity reduction by increasing the minimum sample split."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "dtr=DecisionTreeRegressor(criterion='mse', max_depth=6, max_features=4,\n",
    "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
    "                      min_impurity_split=None, min_samples_leaf=5,\n",
    "                      min_samples_split=4, min_weight_fraction_leaf=0.0,\n",
    "                      presort=False, random_state=None, splitter='best')\n",
    "\n",
    "dtr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = dtr.predict(test_predictors_val)\n",
    "test_mse_dtr = MSE(test_target_val, y_pred)\n",
    "test_rmse_dtr = round(test_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_dtr = -cross_val_score(dtr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_dtr = round((MSE_cv_error_dtr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = dtr.predict(train_predictors_val)\n",
    "train_mse_dtr = MSE(train_target_val, y_pred)\n",
    "train_rmse_dtr = round(train_mse_dtr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>DTR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>2315590.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1964846.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>1731071.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             DTR errors\n",
       "Test error    2315590.0\n",
       "CV error      1964846.0\n",
       "Train error   1731071.0"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_dtr,RMSE_cv_dtr,train_rmse_dtr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['DTR errors'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.1785\n",
      "0.881\n"
     ]
    }
   ],
   "source": [
    "print(round(test_rmse_dtr/RMSE_cv_dtr,4))\n",
    "print(round(train_rmse_dtr/RMSE_cv_dtr,4))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The test error and the CV error show difference of less than five percent (difference for the train error is higher, but this is fluctuating a lot and, therefore, not reliable). We, therefore, assume approximately now under- or overfitting."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2) Random Forest Regressor\n",
    "\n",
    "We repeat the same process as before for the Random Forest Regressor. Since Random Forest equals a collection of multiple decision trees, we identify the same parameters as for Decision Tree to be the most important ones. We add only one parameter, namely the number of trees in the forest.\n",
    "\n",
    "**Default Settings:**<br>\n",
    "RandomForestRegressor(n_estimators=’warn’, criterion=’mse’, max_depth=None, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0.0, max_features=’auto’, max_leaf_nodes=None, min_impurity_decrease=0.0, min_impurity_split=None, bootstrap=True, oob_score=False, n_jobs=None, random_state=None, verbose=0, warm_start=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 81 candidates, totalling 243 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 243 out of 243 | elapsed:  1.1min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-1365217956288.885\n",
      "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=8,\n",
      "                      max_features='auto', max_leaf_nodes=None,\n",
      "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                      min_samples_leaf=2, min_samples_split=2,\n",
      "                      min_weight_fraction_leaf=0.0, n_estimators=100,\n",
      "                      n_jobs=None, oob_score=False, random_state=None,\n",
      "                      verbose=0, warm_start=False)\n"
     ]
    }
   ],
   "source": [
    "# Parameter's to search\n",
    "n_estimators = [5,10,100]\n",
    "max_depth = [1,4,8]\n",
    "min_samples_split = [0.1,0.5,2]\n",
    "min_samples_leaf = [1,2,4]\n",
    "\n",
    "# Setting up parameter grid\n",
    "hyperparams = {'n_estimators': n_estimators,\n",
    "               'max_depth': max_depth, \n",
    "               'min_samples_split': min_samples_split,\n",
    "               'min_samples_leaf': min_samples_leaf}\n",
    "\n",
    "# Run GridSearch CV\n",
    "gd=GridSearchCV(estimator = rfr, param_grid = hyperparams, \n",
    "                verbose=True, cv=3, scoring = 'neg_mean_squared_error')\n",
    "\n",
    "# Fitting model and return results\n",
    "gd.fit(train_predictors_val, train_target_val)\n",
    "print(gd.best_score_)\n",
    "print(gd.best_estimator_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After using GridSearch to identify the best parameters for minimizing the mean squared error, we also have to make sure the model is not under- or overfit."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfr=RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=8,\n",
    "                      max_features='auto', max_leaf_nodes=None,\n",
    "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
    "                      min_samples_leaf=2, min_samples_split=2,\n",
    "                      min_weight_fraction_leaf=0.0, n_estimators=100,\n",
    "                      n_jobs=None, oob_score=False, random_state=None,\n",
    "                      verbose=0, warm_start=False)\n",
    "\n",
    "rfr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = rfr.predict(test_predictors_val)\n",
    "test_mse_rfr = MSE(test_target_val, y_pred)\n",
    "test_rmse_rfr = round(test_mse_rfr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_rfr = -cross_val_score(rfr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_rfr = round((MSE_cv_error_rfr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = rfr.predict(train_predictors_val)\n",
    "train_mse_rfr = MSE(train_target_val, y_pred)\n",
    "train_rmse_rfr = round(train_mse_rfr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>RFR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1140582.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1165992.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>748189.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             RFR errors\n",
       "Test error    1140582.0\n",
       "CV error      1165992.0\n",
       "Train error    748189.0"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_rfr,RMSE_cv_rfr,train_rmse_rfr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['RFR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The CV error is higher than the test error, which can be interpreted as a sign for underfitting (even though the train error is lower than the test error, but that measurement is highly fluctuating). Therefore, we are increasing complexity. First, we are decreasing the the minimum sample leaf to 0.5 and increase the minimum sample split to 200 (as GridSearch chose the highest value for this parameter, so higher values might perform even better)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfr=RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=10,\n",
    "                      max_features='auto', max_leaf_nodes=None,\n",
    "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
    "                      min_samples_leaf=0.5, min_samples_split=2,\n",
    "                      min_weight_fraction_leaf=0.0, n_estimators=200,\n",
    "                      n_jobs=None, oob_score=False, random_state=None,\n",
    "                      verbose=0, warm_start=False)\n",
    "\n",
    "rfr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = rfr.predict(test_predictors_val)\n",
    "test_mse_rfr = MSE(test_target_val, y_pred)\n",
    "test_rmse_rfr = round(test_mse_rfr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_rfr = -cross_val_score(rfr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_rfr = round((MSE_cv_error_rfr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = rfr.predict(train_predictors_val)\n",
    "train_mse_rfr = MSE(train_target_val, y_pred)\n",
    "train_rmse_rfr = round(train_mse_rfr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>RFR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>5837430.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>5540566.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>5540533.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             RFR errors\n",
       "Test error    5837430.0\n",
       "CV error      5540566.0\n",
       "Train error   5540533.0"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_rfr,RMSE_cv_rfr,train_rmse_rfr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['RFR errors'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.0536\n",
      "1.0\n"
     ]
    }
   ],
   "source": [
    "print(round(test_rmse_rfr/RMSE_cv_rfr,4))\n",
    "print(round(train_rmse_rfr/RMSE_cv_rfr,4))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train and CV error show virtually no difference, and the difference between CV and test error is below four percent. Therefore, we assume no significant under- or overfitting."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.3) Extreme Gradient Boosting (XGB Regressor)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "Helpful information on XGBoost parameter tuning:\n",
    "https://www.analyticsvidhya.com/blog/2016/03/complete-guide-parameter-tuning-xgboost-with-codes-python/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 108 candidates, totalling 324 fits\n",
      "[14:46:23] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n",
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:23] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:24] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:24] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:25] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:25] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:26] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:29] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:29] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:30] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:30] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:30] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:32] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:33] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:34] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:34] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:35] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:36] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:37] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:38] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:38] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:39] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:40] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:41] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:41] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:42] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:43] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:44] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:44] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:46] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:47] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:51] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:52] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:54] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:55] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:57] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:46:58] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:00] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:01] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:02] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:04] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:05] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:07] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:08] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:09] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:11] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:11] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:12] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:12] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:12] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:13] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:13] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:13] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:14] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:14] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:18] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:19] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:20] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:21] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:22] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:23] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:24] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:24] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:25] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:26] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:29] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:30] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:32] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:33] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:35] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:37] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:39] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:40] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:42] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:43] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:45] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:46] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:48] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:50] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:52] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:53] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:55] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:56] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:58] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:47:59] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:00] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:01] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:01] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:02] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:02] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:02] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:03] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:03] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:04] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:04] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:04] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:05] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:05] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:06] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:06] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:06] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:07] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:07] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:08] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:08] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:09] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:10] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:11] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:12] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:12] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:13] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:14] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:18] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:19] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:20] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:21] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:21] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:22] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:24] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:25] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:29] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:32] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:34] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:35] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:37] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:38] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:39] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:41] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:42] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:44] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:45] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:46] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:48] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:48] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:48] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:50] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:50] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:51] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:51] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:51] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:52] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:52] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:53] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:53] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:53] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:54] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:54] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:54] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:55] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:56] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:57] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:58] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:58] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:48:59] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:00] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:00] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:01] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:02] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:02] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:03] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:04] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:05] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:05] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:06] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:07] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:07] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:09] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:10] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:12] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:13] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:14] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:18] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:20] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:21] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:22] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:23] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:25] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:26] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:29] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:30] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:32] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:32] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:32] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:33] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:33] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:33] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:34] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:34] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:34] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:35] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:35] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:36] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:36] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:36] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:37] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:37] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:37] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:38] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:39] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:40] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:41] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:41] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:42] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:43] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:44] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:44] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:45] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:46] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:46] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:47] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:48] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:50] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:50] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:52] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:53] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:55] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:56] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:57] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:49:59] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:00] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:02] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:03] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:04] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:05] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:06] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:08] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:09] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:11] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:12] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:13] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:14] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:14] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:15] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:18] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:18] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:18] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:19] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:19] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:19] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:20] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:20] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:20] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:21] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:22] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:23] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:23] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:24] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:25] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:26] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:26] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:28] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:29] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:30] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:31] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:32] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:33] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:33] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:35] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:36] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:38] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:39] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:40] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:41] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:43] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:44] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:46] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:47] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:48] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:49] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:51] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:52] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:53] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:55] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:56] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done 324 out of 324 | elapsed:  4.6min finished\n",
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:50:57] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
      "-1146679041680.3508\n",
      "XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
      "             colsample_bynode=1, colsample_bytree=1, eta=0.1, gamma=0,\n",
      "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
      "             max_depth=8, min_child_weight=6, missing=None, n_estimators=100,\n",
      "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
      "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
      "             silent=None, subsample=0.5, verbosity=1)\n"
     ]
    }
   ],
   "source": [
    "max_depth = [1,4,8]\n",
    "min_child_weight = [1,4,6]\n",
    "gamma = [0,0.5]\n",
    "eta = [0.1,0.2,0.3]\n",
    "subsample = [0.5,1]\n",
    "\n",
    "# Setting up parameter grid\n",
    "hyperparams = {'max_depth': max_depth,\n",
    "               'min_child_weight': min_child_weight, \n",
    "               'gamma': gamma,\n",
    "               'eta': eta,\n",
    "              'subsample': subsample}\n",
    "\n",
    "# Run GridSearch CV\n",
    "gd=GridSearchCV(estimator = xbr, param_grid = hyperparams, \n",
    "                verbose=True, cv=3, scoring = 'neg_mean_squared_error')\n",
    "\n",
    "# Fitting model and return results\n",
    "gd.fit(train_predictors_val, train_target_val)\n",
    "print(gd.best_score_)\n",
    "print(gd.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:52:39] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "xbr=xgb.XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "             colsample_bynode=1, colsample_bytree=1, eta=0.1, gamma=0,\n",
    "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
    "             max_depth=8, min_child_weight=6, missing=None, n_estimators=100,\n",
    "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
    "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "             silent=None, subsample=0.5, verbosity=1)\n",
    "\n",
    "xbr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = xbr.predict(test_predictors_val)\n",
    "test_mse_xbr = MSE(test_target_val, y_pred)\n",
    "test_rmse_xbr = round(test_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_xbr = -cross_val_score(xbr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_xbr = round((MSE_cv_error_xbr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = xbr.predict(train_predictors_val)\n",
    "train_mse_xbr = MSE(train_target_val, y_pred)\n",
    "train_rmse_xbr = round(train_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>XBR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1068804.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1070831.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>580964.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             XBR errors\n",
       "Test error    1068804.0\n",
       "CV error      1070831.0\n",
       "Train error    580964.0"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_xbr,RMSE_cv_xbr,train_rmse_xbr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['XBR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once again, as for Random Forest, the model shows signs of underfitting (since the test error is lower than the cross-validation error). Also, train error and cv error do not align. We increase complexity by decreasing the subsample (also, setting number of estimates to 140, as GridSearch chose the maximum number given, which indicates that more estimators might be needed; 140 equals one percent of the training dataset)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:57:00] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "xbr=xgb.XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "             colsample_bynode=1, colsample_bytree=1, eta=0.1, gamma=0,\n",
    "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
    "             max_depth=8, min_child_weight=6, missing=None, n_estimators=140,\n",
    "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
    "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "             silent=None, subsample=.15, verbosity=1)\n",
    "\n",
    "xbr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = xbr.predict(test_predictors_val)\n",
    "test_mse_xbr = MSE(test_target_val, y_pred)\n",
    "test_rmse_xbr = round(test_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_xbr = -cross_val_score(xbr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_xbr = round((MSE_cv_error_xbr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = xbr.predict(train_predictors_val)\n",
    "train_mse_xbr = MSE(train_target_val, y_pred)\n",
    "train_rmse_xbr = round(train_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>XBR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1186515.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1273765.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>899196.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             XBR errors\n",
       "Test error    1186515.0\n",
       "CV error      1273765.0\n",
       "Train error    899196.0"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_xbr,RMSE_cv_xbr,train_rmse_xbr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['XBR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train and CV error came a little bit closer toghether. However, the overall error also increased. Next, we are trying to decrease the variance in the model by reducing the maximum depth."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:58:29] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "xbr=xgb.XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "             colsample_bynode=1, colsample_bytree=1, eta=0.1, gamma=0,\n",
    "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
    "             max_depth=6, min_child_weight=6, missing=None, n_estimators=140,\n",
    "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
    "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "             silent=None, subsample=.15, verbosity=1)\n",
    "\n",
    "xbr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = xbr.predict(test_predictors_val)\n",
    "test_mse_xbr = MSE(test_target_val, y_pred)\n",
    "test_rmse_xbr = round(test_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_xbr = -cross_val_score(xbr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_xbr = round((MSE_cv_error_xbr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = xbr.predict(train_predictors_val)\n",
    "train_mse_xbr = MSE(train_target_val, y_pred)\n",
    "train_rmse_xbr = round(train_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>XBR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1186330.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1265565.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>915878.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             XBR errors\n",
       "Test error    1186330.0\n",
       "CV error      1265565.0\n",
       "Train error    915878.0"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_xbr,RMSE_cv_xbr,train_rmse_xbr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['XBR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Decreasing the maximum depth helped stabilizing the model variance. However, it looks like it is still some tuning needed. \n",
    "\n",
    "A node is split only when the resulting split gives a positive reduction in the loss function. Gamma specifies the minimum loss reduction required to make a split. Therefore, we are increasing gamma to see whether that could improve our results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14:59:17] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "xbr=xgb.XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "             colsample_bynode=1, colsample_bytree=1, eta=0.1, gamma=0.5,\n",
    "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
    "             max_depth=6, min_child_weight=6, missing=None, n_estimators=140,\n",
    "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
    "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "             silent=None, subsample=.15, verbosity=1)\n",
    "\n",
    "xbr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = xbr.predict(test_predictors_val)\n",
    "test_mse_xbr = MSE(test_target_val, y_pred)\n",
    "test_rmse_xbr = round(test_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_xbr = -cross_val_score(xbr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_xbr = round((MSE_cv_error_xbr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = xbr.predict(train_predictors_val)\n",
    "train_mse_xbr = MSE(train_target_val, y_pred)\n",
    "train_rmse_xbr = round(train_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>XBR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1186330.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1265565.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>915878.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             XBR errors\n",
       "Test error    1186330.0\n",
       "CV error      1265565.0\n",
       "Train error    915878.0"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_xbr,RMSE_cv_xbr,train_rmse_xbr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['XBR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Changing gamma did not reduce the bias, but increased the variance of the model. Therefore, we are going back to gamma = 0. Instead, we decrease eta (the learning rate per node), to hopefully increase our accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:00:08] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "xbr=xgb.XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "             colsample_bynode=1, colsample_bytree=1, eta=0.01, gamma=0.5,\n",
    "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
    "             max_depth=6, min_child_weight=6, missing=None, n_estimators=140,\n",
    "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
    "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "             silent=None, subsample=.15, verbosity=1)\n",
    "\n",
    "\n",
    "xbr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = xbr.predict(test_predictors_val)\n",
    "test_mse_xbr = MSE(test_target_val, y_pred)\n",
    "test_rmse_xbr = round(test_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_xbr = -cross_val_score(xbr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_xbr = round((MSE_cv_error_xbr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = xbr.predict(train_predictors_val)\n",
    "train_mse_xbr = MSE(train_target_val, y_pred)\n",
    "train_rmse_xbr = round(train_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>XBR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1186330.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1265565.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>915878.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             XBR errors\n",
       "Test error    1186330.0\n",
       "CV error      1265565.0\n",
       "Train error    915878.0"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_xbr,RMSE_cv_xbr,train_rmse_xbr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['XBR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Decreasing eta did not help to significantly decrease our RMSE or decrease the variance in the model. In a final attempt, we decrease the minimum sum of weights of all observations required in a child (min_child_weight) to hopefully increase accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:01:16] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "xbr=xgb.XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "             colsample_bynode=1, colsample_bytree=1, eta=0.1, gamma=0.5,\n",
    "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
    "             max_depth=6, min_child_weight=2, missing=None, n_estimators=140,\n",
    "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
    "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "             silent=None, subsample=.15, verbosity=1)\n",
    "\n",
    "xbr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = xbr.predict(test_predictors_val)\n",
    "test_mse_xbr = MSE(test_target_val, y_pred)\n",
    "test_rmse_xbr = round(test_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_xbr = -cross_val_score(xbr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_xbr = round((MSE_cv_error_xbr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = xbr.predict(train_predictors_val)\n",
    "train_mse_xbr = MSE(train_target_val, y_pred)\n",
    "train_rmse_xbr = round(train_mse_xbr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>XBR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>1182594.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>1142286.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>773277.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             XBR errors\n",
       "Test error    1182594.0\n",
       "CV error      1142286.0\n",
       "Train error    773277.0"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_xbr,RMSE_cv_xbr,train_rmse_xbr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['XBR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Doing so does not have a significant effect either. Instead, it even slightly increases variance. Therefore, we accept the variance in the model (before changing min_child_weight) and move on to Linear Regression."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.4) Linear Regression\n",
    "\n",
    "For Linear Regression, there is not much parameter tuning to do. The only change we are making to the model is to set fit_intercept parameter to False, as we already centered the data before.\n",
    "\n",
    "**Default model:**\n",
    "LinearRegression(fit_intercept=True, normalize=False, copy_X=True, n_jobs=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [],
   "source": [
    "lre=LinearRegression(fit_intercept=False, normalize=False, copy_X=True, n_jobs=None)\n",
    "\n",
    "lre.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = lre.predict(test_predictors_val)\n",
    "test_mse_lre = MSE(test_target_val, y_pred)\n",
    "test_rmse_lre = round(test_mse_lre**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_lre = -cross_val_score(lre, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_lre = round((MSE_cv_error_lre.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = lre.predict(train_predictors_val)\n",
    "train_mse_lre = MSE(train_target_val, y_pred)\n",
    "train_rmse_lre = round(train_mse_lre**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>RFR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>3587924.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>3438748.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>3434508.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             RFR errors\n",
       "Test error    3587924.0\n",
       "CV error      3438748.0\n",
       "Train error   3434508.0"
      ]
     },
     "execution_count": 173,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_lre,RMSE_cv_lre,train_rmse_lre], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['RFR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Excluding the intercept term did not change anything for our prediction compared to the base model. Even though the RMSE is quite high, the model shows very stable results."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.5) Epsilon Support Vector Regression\n",
    "\n",
    "\n",
    "**Default model:**<br>\n",
    "SVR(kernel=’rbf’, degree=3, gamma=’auto_deprecated’, coef0=0.0, tol=0.001, C=1.0, epsilon=0.1, shrinking=True, cache_size=200, verbose=False, max_iter=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 20 candidates, totalling 60 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  60 out of  60 | elapsed:  4.5min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-33642840426335.918\n",
      "SVR(C=10, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma=0.1,\n",
      "    kernel='rbf', max_iter=-1, shrinking=True, tol=0.001, verbose=False)\n"
     ]
    }
   ],
   "source": [
    "C = [0.001, 0.01, 0.1, 1, 10]\n",
    "gamma = [0.001, 0.01, 0.1, 1]\n",
    "\n",
    "# Setting up parameter grid\n",
    "hyperparams = {'C': C,\n",
    "               'gamma': gamma}\n",
    "\n",
    "# Run GridSearch CV\n",
    "gd=GridSearchCV(estimator = svr, param_grid = hyperparams, \n",
    "                verbose=True, cv=3, scoring = 'neg_mean_squared_error')\n",
    "\n",
    "# Fitting model and return results\n",
    "gd.fit(train_predictors_val, train_target_val)\n",
    "print(gd.best_score_)\n",
    "print(gd.best_estimator_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [],
   "source": [
    "svr=SVR(C=10, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma=0.1,\n",
    "    kernel='rbf', max_iter=-1, shrinking=True, tol=0.001, verbose=False)\n",
    "\n",
    "svr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = svr.predict(test_predictors_val)\n",
    "test_mse_svr = MSE(test_target_val, y_pred)\n",
    "test_rmse_svr = round(test_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_svr = -cross_val_score(svr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_svr = round((MSE_cv_error_svr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = svr.predict(train_predictors_val)\n",
    "train_mse_svr = MSE(train_target_val, y_pred)\n",
    "train_rmse_svr = round(train_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SVR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>6091026.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>5800242.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>5798892.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             SVR errors\n",
       "Test error    6091026.0\n",
       "CV error      5800242.0\n",
       "Train error   5798892.0"
      ]
     },
     "execution_count": 178,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_svr,RMSE_cv_svr,train_rmse_svr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['SVR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Even though the RMSE itself is quite high, we get very stable results. Lets see whether using different kernels could improve our RMSE. By default, SVR uses the Gaussian RBF kernel. We try to use a linear kernel instead."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [],
   "source": [
    "svr=SVR(C=10, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma=0.1,\n",
    "    kernel='linear', max_iter=-1, shrinking=True, tol=0.001, verbose=False)\n",
    "\n",
    "svr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = svr.predict(test_predictors_val)\n",
    "test_mse_svr = MSE(test_target_val, y_pred)\n",
    "test_rmse_svr = round(test_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_svr = -cross_val_score(svr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_svr = round((MSE_cv_error_svr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = svr.predict(train_predictors_val)\n",
    "train_mse_svr = MSE(train_target_val, y_pred)\n",
    "train_rmse_svr = round(train_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SVR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>5854694.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>5619461.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>5560873.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             SVR errors\n",
       "Test error    5854694.0\n",
       "CV error      5619461.0\n",
       "Train error   5560873.0"
      ]
     },
     "execution_count": 182,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_svr,RMSE_cv_svr,train_rmse_svr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['SVR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A Linear kernel did not help much. Next we use a polynomil kernel."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [],
   "source": [
    "svr=SVR(C=10, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma=0.1,\n",
    "    kernel='poly', max_iter=-1, shrinking=True, tol=0.001, verbose=False)\n",
    "\n",
    "svr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = svr.predict(test_predictors_val)\n",
    "test_mse_svr = MSE(test_target_val, y_pred)\n",
    "test_rmse_svr = round(test_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_svr = -cross_val_score(svr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_svr = round((MSE_cv_error_svr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = svr.predict(train_predictors_val)\n",
    "train_mse_svr = MSE(train_target_val, y_pred)\n",
    "train_rmse_svr = round(train_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SVR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>5917741.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>5690448.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>5638680.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             SVR errors\n",
       "Test error    5917741.0\n",
       "CV error      5690448.0\n",
       "Train error   5638680.0"
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_svr,RMSE_cv_svr,train_rmse_svr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['SVR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As for the linear kernel, our result did not change significantly. Finally, we are using a sigmoid kernel."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 630,
   "metadata": {},
   "outputs": [],
   "source": [
    "svr=SVR(C=10, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma=0.1,\n",
    "    kernel='sigmoid', max_iter=-1, shrinking=True, tol=0.001, verbose=False)\n",
    "\n",
    "svr.fit(train_predictors_val, train_target_val)\n",
    "\n",
    "# Test error\n",
    "y_pred = svr.predict(test_predictors_val)\n",
    "test_mse_svr = MSE(test_target_val, y_pred)\n",
    "test_rmse_svr = round(test_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 631,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Cross-validation error\n",
    "MSE_cv_error_svr = -cross_val_score(svr, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "RMSE_cv_svr = round((MSE_cv_error_svr.mean())**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 632,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train error\n",
    "y_pred = svr.predict(train_predictors_val)\n",
    "train_mse_svr = MSE(train_target_val, y_pred)\n",
    "train_rmse_svr = round(train_mse_svr**(1/2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 633,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SVR errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>Test error</td>\n",
       "      <td>6009328.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>CV error</td>\n",
       "      <td>5823318.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>Train error</td>\n",
       "      <td>5823789.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             SVR errors\n",
       "Test error    6009328.0\n",
       "CV error      5823318.0\n",
       "Train error   5823789.0"
      ]
     },
     "execution_count": 633,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([test_rmse_svr,RMSE_cv_svr,train_rmse_svr], \n",
    "             index=['Test error','CV error','Train error'], \n",
    "             columns=['SVR errors'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once again, no significant change. In the following, we are using the linear kernel, since that one performed best (even though the differences where marginal)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.6) Final comparison of tuned models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In order to make a decision for a final model, we are comparing the tuned models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning:\n",
      "\n",
      "Series.base is deprecated and will be removed in a future version\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[15:12:27] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n"
     ]
    }
   ],
   "source": [
    "dtr=DecisionTreeRegressor(criterion='mse', max_depth=6, max_features=4,\n",
    "                      max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
    "                      min_impurity_split=None, min_samples_leaf=5,\n",
    "                      min_samples_split=4, min_weight_fraction_leaf=0.0,\n",
    "                      presort=False, random_state=None, splitter='best')\n",
    "\n",
    "rfr=RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=10,\n",
    "                      max_features='auto', max_leaf_nodes=None,\n",
    "                      min_impurity_decrease=0.0, min_impurity_split=None,\n",
    "                      min_samples_leaf=0.5, min_samples_split=2,\n",
    "                      min_weight_fraction_leaf=0.0, n_estimators=200,\n",
    "                      n_jobs=None, oob_score=False, random_state=None,\n",
    "                      verbose=0, warm_start=False)\n",
    "\n",
    "xbr=xgb.XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "             colsample_bynode=1, colsample_bytree=1, eta=0.01, gamma=0.5,\n",
    "             importance_type='gain', learning_rate=0.1, max_delta_step=0,\n",
    "             max_depth=6, min_child_weight=6, missing=None, n_estimators=140,\n",
    "             n_jobs=1, nthread=None, objective='reg:linear', random_state=0,\n",
    "             reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "             silent=None, subsample=.15, verbosity=1)\n",
    "\n",
    "lre=LinearRegression(fit_intercept=False, normalize=False, copy_X=True, n_jobs=None)\n",
    "\n",
    "svr=SVR(C=10, cache_size=200, coef0=0.0, degree=3, epsilon=0.1, gamma=0.1,\n",
    "    kernel='linear', max_iter=-1, shrinking=True, tol=0.001, verbose=False)\n",
    "\n",
    "models = [dtr,rfr,xbr,lre,svr]\n",
    "\n",
    "for mod in models:\n",
    "    mod.fit(train_predictors_val, train_target_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_error = []\n",
    "\n",
    "# Sequentially fit all models and calculate test score\n",
    "for mod in models:\n",
    "    y_pred = mod.predict(test_predictors_val)\n",
    "    mse_mod = MSE(test_target_val, y_pred)\n",
    "    rmse_mod = round(mse_mod**(1/2))\n",
    "    test_error.append(rmse_mod)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wehrm\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:1978: FutureWarning:\n",
      "\n",
      "The default value of cv will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "cv_error=[]\n",
    "\n",
    "for mod in models:\n",
    "    MSE_cv_error = -cross_val_score(mod, train_predictors_val,\n",
    "                                    train_target_val, scoring='neg_mean_squared_error', n_jobs=-1)\n",
    "    RMSE_cv = round((MSE_cv_error.mean())**(1/2))\n",
    "    cv_error.append(RMSE_cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_error=[]\n",
    "\n",
    "# Sequentially fit all models and calculate train score\n",
    "for mod in models:\n",
    "    y_pred = mod.predict(train_predictors_val)\n",
    "    mse_mod = MSE(train_target_val, y_pred)\n",
    "    rmse_mod = round(mse_mod**(1/2))\n",
    "    train_error.append(rmse_mod)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_names = ['dtr','rfr','xbr','lre','svc']\n",
    "\n",
    "error = pd.DataFrame(dict(graph=model_names,test_error=test_error, cv_error=cv_error, train_error=train_error)) \n",
    "\n",
    "ind = np.arange(len(error))\n",
    "width = 0.2\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12,6))\n",
    "ax.barh(ind, error.test_error, width, color='maroon', label='test')\n",
    "ax.barh(ind + width, error.cv_error, width, color='forestgreen', label='cv')\n",
    "ax.barh(ind + 2*width, error.train_error, width, color='slategrey', label='train')\n",
    "ax.set_xlim(0, max(error.cv_error)+min(error.cv_error))\n",
    "plt.xlabel('RMSE', fontsize=16)\n",
    "plt.ylabel('Algorithms', fontsize=16)\n",
    "plt.title('Test, Cross Validation and Train Error per Algorithm',fontsize=20)\n",
    "\n",
    "ax.set(yticks=ind + width, yticklabels=error.graph, ylim=[2*width - 1, len(error)])\n",
    "ax.legend()\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>graph</th>\n",
       "      <th>test_error</th>\n",
       "      <th>cv_error</th>\n",
       "      <th>train_error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>dtr</td>\n",
       "      <td>1769111.0</td>\n",
       "      <td>1933184.0</td>\n",
       "      <td>1601429.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>rfr</td>\n",
       "      <td>5837386.0</td>\n",
       "      <td>5540553.0</td>\n",
       "      <td>5540532.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>xbr</td>\n",
       "      <td>1186330.0</td>\n",
       "      <td>1265565.0</td>\n",
       "      <td>915878.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>lre</td>\n",
       "      <td>3587924.0</td>\n",
       "      <td>3438748.0</td>\n",
       "      <td>3434508.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>svc</td>\n",
       "      <td>5854694.0</td>\n",
       "      <td>5619461.0</td>\n",
       "      <td>5560873.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  graph  test_error   cv_error  train_error\n",
       "0   dtr   1769111.0  1933184.0    1601429.0\n",
       "1   rfr   5837386.0  5540553.0    5540532.0\n",
       "2   xbr   1186330.0  1265565.0     915878.0\n",
       "3   lre   3587924.0  3438748.0    3434508.0\n",
       "4   svc   5854694.0  5619461.0    5560873.0"
      ]
     },
     "execution_count": 192,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Extreme Gradient Boosting delivers by far the lowest RMSE amongst our models, even though it shows some variance due to underfitting (based on CV and test error comparison). The most stable models, but with a significantly higher RMSE than XGBoost, are Linear Regression, Random Forest and Support Vector Regression. Decision Tree has a relatively low RMSE, but shows signs of high variance. \n",
    "\n",
    "Considering the high variance of the Decision Tree and the high bias (RMSE) of Support Vector Regressor, Linear Regression and Random Forest, we decide to accept the variance in XGBoost in favour of the lower overall bias this model seems to have."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4.) Using models for final prediction\n",
    "\n",
    "**Putting the prediction into business presepective:**<br>\n",
    "In this step we are loading the pickle we have derived from the anomaly detection algorithm. In terms of business this means the clubs has\n",
    "\n",
    "1.) Selected a subset of players based either on position and overall skill or based on a similar player and<br>\n",
    "2.) Our anomaly detection model has identified under- and overvalued players amongst the previously defined subset.\n",
    "\n",
    "The subset of undervalued players is now used to predict these players instrinsic value, so the club gets an idea about how high the first bid for this player should be.\n",
    "\n",
    "**Technical perspective:**<br>\n",
    "From a technical perspective, we are loading the subset of undervalued players and apply the same transformations to it as we did for the training set. We then use the transformed subset to predict the players intrinsic value and save the results in a csv file to give to the club as a final report."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('undervalue_players.pkl', 'rb') as f:\n",
    "    undervalued_players = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('predictors_anomaly.pkl', 'rb') as f:\n",
    "    predictors_anomaly = pickle.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Extracting the subset of undervalued players from predictors_anomaly data frame."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_anomaly_processed=predictors_anomaly[predictors_anomaly['ID'].isin(list(undervalued_players['ID']))].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>Age</th>\n",
       "      <th>Overall</th>\n",
       "      <th>Potential</th>\n",
       "      <th>Value</th>\n",
       "      <th>Special</th>\n",
       "      <th>Preferred Foot</th>\n",
       "      <th>International Reputation</th>\n",
       "      <th>Weak Foot</th>\n",
       "      <th>Skill Moves</th>\n",
       "      <th>Real Face</th>\n",
       "      <th>Jersey Number</th>\n",
       "      <th>Joined</th>\n",
       "      <th>Loaned From</th>\n",
       "      <th>Contract Valid Until</th>\n",
       "      <th>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>LS</th>\n",
       "      <th>ST</th>\n",
       "      <th>RS</th>\n",
       "      <th>LW</th>\n",
       "      <th>LF</th>\n",
       "      <th>CF</th>\n",
       "      <th>RF</th>\n",
       "      <th>RW</th>\n",
       "      <th>LAM</th>\n",
       "      <th>CAM</th>\n",
       "      <th>RAM</th>\n",
       "      <th>LM</th>\n",
       "      <th>LCM</th>\n",
       "      <th>CM</th>\n",
       "      <th>RCM</th>\n",
       "      <th>RM</th>\n",
       "      <th>LWB</th>\n",
       "      <th>LDM</th>\n",
       "      <th>CDM</th>\n",
       "      <th>RDM</th>\n",
       "      <th>RWB</th>\n",
       "      <th>LB</th>\n",
       "      <th>LCB</th>\n",
       "      <th>CB</th>\n",
       "      <th>RCB</th>\n",
       "      <th>RB</th>\n",
       "      <th>Crossing</th>\n",
       "      <th>Finishing</th>\n",
       "      <th>HeadingAccuracy</th>\n",
       "      <th>ShortPassing</th>\n",
       "      <th>Volleys</th>\n",
       "      <th>Dribbling</th>\n",
       "      <th>Curve</th>\n",
       "      <th>FKAccuracy</th>\n",
       "      <th>LongPassing</th>\n",
       "      <th>BallControl</th>\n",
       "      <th>Acceleration</th>\n",
       "      <th>SprintSpeed</th>\n",
       "      <th>Agility</th>\n",
       "      <th>Reactions</th>\n",
       "      <th>Balance</th>\n",
       "      <th>ShotPower</th>\n",
       "      <th>Jumping</th>\n",
       "      <th>Stamina</th>\n",
       "      <th>Strength</th>\n",
       "      <th>LongShots</th>\n",
       "      <th>Aggression</th>\n",
       "      <th>Interceptions</th>\n",
       "      <th>Positioning</th>\n",
       "      <th>Vision</th>\n",
       "      <th>Penalties</th>\n",
       "      <th>Composure</th>\n",
       "      <th>Marking</th>\n",
       "      <th>StandingTackle</th>\n",
       "      <th>SlidingTackle</th>\n",
       "      <th>GKDiving</th>\n",
       "      <th>GKHandling</th>\n",
       "      <th>GKKicking</th>\n",
       "      <th>GKPositioning</th>\n",
       "      <th>GKReflexes</th>\n",
       "      <th>Club_Reputation</th>\n",
       "      <th>Position_Potential</th>\n",
       "      <th>Attacking</th>\n",
       "      <th>Skill</th>\n",
       "      <th>Movement</th>\n",
       "      <th>Power</th>\n",
       "      <th>Mentality</th>\n",
       "      <th>Defending</th>\n",
       "      <th>GoalKeeping</th>\n",
       "      <th>Potential_Difference</th>\n",
       "      <th>anomaly</th>\n",
       "      <th>Body Type</th>\n",
       "      <th>Attack_WR</th>\n",
       "      <th>Defense_WR</th>\n",
       "      <th>Position_CAM</th>\n",
       "      <th>Position_CB</th>\n",
       "      <th>Position_CDM</th>\n",
       "      <th>Position_CF</th>\n",
       "      <th>Position_CM</th>\n",
       "      <th>Position_GK</th>\n",
       "      <th>Position_LAM</th>\n",
       "      <th>Position_LB</th>\n",
       "      <th>Position_LCB</th>\n",
       "      <th>Position_LCM</th>\n",
       "      <th>Position_LDM</th>\n",
       "      <th>Position_LF</th>\n",
       "      <th>Position_LM</th>\n",
       "      <th>Position_LS</th>\n",
       "      <th>Position_LW</th>\n",
       "      <th>Position_LWB</th>\n",
       "      <th>Position_RAM</th>\n",
       "      <th>Position_RB</th>\n",
       "      <th>Position_RCB</th>\n",
       "      <th>Position_RCM</th>\n",
       "      <th>Position_RDM</th>\n",
       "      <th>Position_RF</th>\n",
       "      <th>Position_RM</th>\n",
       "      <th>Position_RS</th>\n",
       "      <th>Position_RW</th>\n",
       "      <th>Position_RWB</th>\n",
       "      <th>Position_ST</th>\n",
       "      <th>Continent_Africa</th>\n",
       "      <th>Continent_Asia</th>\n",
       "      <th>Continent_Europe</th>\n",
       "      <th>Continent_North America</th>\n",
       "      <th>Continent_Oceania</th>\n",
       "      <th>Continent_South America</th>\n",
       "      <th>Position_Group_Defender</th>\n",
       "      <th>Position_Group_Forward</th>\n",
       "      <th>Position_Group_GoalKeeper</th>\n",
       "      <th>Position_Group_Midfielder</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1045</td>\n",
       "      <td>1045</td>\n",
       "      <td>31</td>\n",
       "      <td>77</td>\n",
       "      <td>77</td>\n",
       "      <td>7500000.0</td>\n",
       "      <td>2094</td>\n",
       "      <td>1</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>14.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>178</td>\n",
       "      <td>168</td>\n",
       "      <td>73</td>\n",
       "      <td>73</td>\n",
       "      <td>73</td>\n",
       "      <td>76</td>\n",
       "      <td>76</td>\n",
       "      <td>76</td>\n",
       "      <td>76</td>\n",
       "      <td>76</td>\n",
       "      <td>77</td>\n",
       "      <td>77</td>\n",
       "      <td>77</td>\n",
       "      <td>76</td>\n",
       "      <td>75</td>\n",
       "      <td>75</td>\n",
       "      <td>75</td>\n",
       "      <td>76</td>\n",
       "      <td>69</td>\n",
       "      <td>70</td>\n",
       "      <td>70</td>\n",
       "      <td>70</td>\n",
       "      <td>69</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>77.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>59.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>63.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>3</td>\n",
       "      <td>74.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>74.8</td>\n",
       "      <td>71.0</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>64.000000</td>\n",
       "      <td>10.4</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1443</td>\n",
       "      <td>1443</td>\n",
       "      <td>33</td>\n",
       "      <td>76</td>\n",
       "      <td>76</td>\n",
       "      <td>4700000.0</td>\n",
       "      <td>2001</td>\n",
       "      <td>1</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>706</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>178</td>\n",
       "      <td>165</td>\n",
       "      <td>72</td>\n",
       "      <td>72</td>\n",
       "      <td>72</td>\n",
       "      <td>75</td>\n",
       "      <td>74</td>\n",
       "      <td>74</td>\n",
       "      <td>74</td>\n",
       "      <td>75</td>\n",
       "      <td>74</td>\n",
       "      <td>74</td>\n",
       "      <td>74</td>\n",
       "      <td>74</td>\n",
       "      <td>71</td>\n",
       "      <td>71</td>\n",
       "      <td>71</td>\n",
       "      <td>74</td>\n",
       "      <td>65</td>\n",
       "      <td>63</td>\n",
       "      <td>63</td>\n",
       "      <td>63</td>\n",
       "      <td>65</td>\n",
       "      <td>62</td>\n",
       "      <td>57</td>\n",
       "      <td>57</td>\n",
       "      <td>57</td>\n",
       "      <td>62</td>\n",
       "      <td>75.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>59.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>86.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>52.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>1.222222</td>\n",
       "      <td>2</td>\n",
       "      <td>69.6</td>\n",
       "      <td>77.0</td>\n",
       "      <td>75.6</td>\n",
       "      <td>71.4</td>\n",
       "      <td>68.166667</td>\n",
       "      <td>49.666667</td>\n",
       "      <td>9.4</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2802</td>\n",
       "      <td>2802</td>\n",
       "      <td>32</td>\n",
       "      <td>73</td>\n",
       "      <td>73</td>\n",
       "      <td>2300000.0</td>\n",
       "      <td>2050</td>\n",
       "      <td>0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0</td>\n",
       "      <td>44.0</td>\n",
       "      <td>74</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>180</td>\n",
       "      <td>174</td>\n",
       "      <td>69</td>\n",
       "      <td>69</td>\n",
       "      <td>69</td>\n",
       "      <td>72</td>\n",
       "      <td>71</td>\n",
       "      <td>71</td>\n",
       "      <td>71</td>\n",
       "      <td>72</td>\n",
       "      <td>74</td>\n",
       "      <td>74</td>\n",
       "      <td>74</td>\n",
       "      <td>72</td>\n",
       "      <td>76</td>\n",
       "      <td>76</td>\n",
       "      <td>76</td>\n",
       "      <td>72</td>\n",
       "      <td>68</td>\n",
       "      <td>71</td>\n",
       "      <td>71</td>\n",
       "      <td>71</td>\n",
       "      <td>68</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>67</td>\n",
       "      <td>82.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>83.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>83.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>53.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>54.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>71.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>1.444444</td>\n",
       "      <td>3</td>\n",
       "      <td>72.6</td>\n",
       "      <td>80.8</td>\n",
       "      <td>65.2</td>\n",
       "      <td>72.0</td>\n",
       "      <td>73.500000</td>\n",
       "      <td>60.666667</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        ID  Age  Overall  Potential      Value  Special  Preferred Foot  \\\n",
       "1045  1045   31       77         77  7500000.0     2094               1   \n",
       "1443  1443   33       76         76  4700000.0     2001               1   \n",
       "2802  2802   32       73         73  2300000.0     2050               0   \n",
       "\n",
       "      International Reputation  Weak Foot  Skill Moves  Real Face  \\\n",
       "1045                       3.0        4.0          4.0          1   \n",
       "1443                       2.0        3.0          4.0          0   \n",
       "2802                       3.0        3.0          3.0          0   \n",
       "\n",
       "      Jersey Number  Joined  Loaned From  Contract Valid Until  Height  \\\n",
       "1045           14.0       0            1                     2     178   \n",
       "1443           19.0     706            0                     1     178   \n",
       "2802           44.0      74            0                     3     180   \n",
       "\n",
       "      Weight  LS  ST  RS  LW  LF  CF  RF  RW  LAM  CAM  RAM  LM  LCM  CM  RCM  \\\n",
       "1045     168  73  73  73  76  76  76  76  76   77   77   77  76   75  75   75   \n",
       "1443     165  72  72  72  75  74  74  74  75   74   74   74  74   71  71   71   \n",
       "2802     174  69  69  69  72  71  71  71  72   74   74   74  72   76  76   76   \n",
       "\n",
       "      RM  LWB  LDM  CDM  RDM  RWB  LB  LCB  CB  RCB  RB  Crossing  Finishing  \\\n",
       "1045  76   69   70   70   70   69  67   67  67   67  67      77.0       71.0   \n",
       "1443  74   65   63   63   63   65  62   57  57   57  62      75.0       73.0   \n",
       "2802  72   68   71   71   71   68  67   67  67   67  67      82.0       61.0   \n",
       "\n",
       "      HeadingAccuracy  ShortPassing  Volleys  Dribbling  Curve  FKAccuracy  \\\n",
       "1045             65.0          79.0     78.0       79.0   79.0        64.0   \n",
       "1443             59.0          80.0     61.0       74.0   86.0        79.0   \n",
       "2802             67.0          83.0     70.0       73.0   83.0        81.0   \n",
       "\n",
       "      LongPassing  BallControl  Acceleration  SprintSpeed  Agility  Reactions  \\\n",
       "1045         78.0         80.0          75.0         68.0     81.0       74.0   \n",
       "1443         69.0         77.0          76.0         77.0     78.0       74.0   \n",
       "2802         85.0         82.0          60.0         53.0     69.0       69.0   \n",
       "\n",
       "      Balance  ShotPower  Jumping  Stamina  Strength  LongShots  Aggression  \\\n",
       "1045     76.0       75.0     78.0     59.0      68.0       75.0        80.0   \n",
       "1443     73.0       73.0     64.0     78.0      68.0       74.0        71.0   \n",
       "2802     75.0       80.0     74.0     54.0      72.0       80.0        71.0   \n",
       "\n",
       "      Interceptions  Positioning  Vision  Penalties  Composure  Marking  \\\n",
       "1045           61.0         77.0    80.0       73.0       79.0     69.0   \n",
       "1443           42.0         71.0    72.0       81.0       72.0     50.0   \n",
       "2802           71.0         65.0    80.0       73.0       81.0     40.0   \n",
       "\n",
       "      StandingTackle  SlidingTackle  GKDiving  GKHandling  GKKicking  \\\n",
       "1045            63.0           60.0      11.0         7.0       11.0   \n",
       "1443            52.0           47.0       8.0         8.0        8.0   \n",
       "2802            72.0           70.0      16.0         7.0        9.0   \n",
       "\n",
       "      GKPositioning  GKReflexes  Club_Reputation  Position_Potential  \\\n",
       "1045           12.0        11.0         1.500000                   3   \n",
       "1443           11.0        12.0         1.222222                   2   \n",
       "2802            8.0        15.0         1.444444                   3   \n",
       "\n",
       "      Attacking  Skill  Movement  Power  Mentality  Defending  GoalKeeping  \\\n",
       "1045       74.0   76.0      74.8   71.0  75.000000  64.000000         10.4   \n",
       "1443       69.6   77.0      75.6   71.4  68.166667  49.666667          9.4   \n",
       "2802       72.6   80.8      65.2   72.0  73.500000  60.666667         11.0   \n",
       "\n",
       "      Potential_Difference  anomaly  Body Type  Attack_WR  Defense_WR  \\\n",
       "1045                     0        1          2          3           3   \n",
       "1443                     0        1          2          3           2   \n",
       "2802                     0        1          2          2           2   \n",
       "\n",
       "      Position_CAM  Position_CB  Position_CDM  Position_CF  Position_CM  \\\n",
       "1045             0            0             0            0            1   \n",
       "1443             0            0             0            0            0   \n",
       "2802             0            0             1            0            0   \n",
       "\n",
       "      Position_GK  Position_LAM  Position_LB  Position_LCB  Position_LCM  \\\n",
       "1045            0             0            0             0             0   \n",
       "1443            0             0            0             0             0   \n",
       "2802            0             0            0             0             0   \n",
       "\n",
       "      Position_LDM  Position_LF  Position_LM  Position_LS  Position_LW  \\\n",
       "1045             0            0            0            0            0   \n",
       "1443             0            0            0            0            0   \n",
       "2802             0            0            0            0            0   \n",
       "\n",
       "      Position_LWB  Position_RAM  Position_RB  Position_RCB  Position_RCM  \\\n",
       "1045             0             0            0             0             0   \n",
       "1443             0             0            0             0             0   \n",
       "2802             0             0            0             0             0   \n",
       "\n",
       "      Position_RDM  Position_RF  Position_RM  Position_RS  Position_RW  \\\n",
       "1045             0            0            0            0            0   \n",
       "1443             0            0            0            0            1   \n",
       "2802             0            0            0            0            0   \n",
       "\n",
       "      Position_RWB  Position_ST  Continent_Africa  Continent_Asia  \\\n",
       "1045             0            0                 0               1   \n",
       "1443             0            0                 0               0   \n",
       "2802             0            0                 0               0   \n",
       "\n",
       "      Continent_Europe  Continent_North America  Continent_Oceania  \\\n",
       "1045                 0                        0                  0   \n",
       "1443                 1                        0                  0   \n",
       "2802                 1                        0                  0   \n",
       "\n",
       "      Continent_South America  Position_Group_Defender  \\\n",
       "1045                        0                        0   \n",
       "1443                        0                        0   \n",
       "2802                        0                        0   \n",
       "\n",
       "      Position_Group_Forward  Position_Group_GoalKeeper  \\\n",
       "1045                       0                          0   \n",
       "1443                       1                          0   \n",
       "2802                       0                          0   \n",
       "\n",
       "      Position_Group_Midfielder  \n",
       "1045                          1  \n",
       "1443                          0  \n",
       "2802                          1  "
      ]
     },
     "execution_count": 196,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictors_anomaly_processed.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Overall</th>\n",
       "      <th>Potential</th>\n",
       "      <th>Special</th>\n",
       "      <th>International Reputation</th>\n",
       "      <th>Real Face</th>\n",
       "      <th>BallControl</th>\n",
       "      <th>Reactions</th>\n",
       "      <th>Vision</th>\n",
       "      <th>Composure</th>\n",
       "      <th>Club_Reputation</th>\n",
       "      <th>Mentality</th>\n",
       "      <th>Position_Group_Defender</th>\n",
       "      <th>Position_Group_Forward</th>\n",
       "      <th>Position_Group_GoalKeeper</th>\n",
       "      <th>Position_Group_Midfielder</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2540</td>\n",
       "      <td>1.121234</td>\n",
       "      <td>1.736760</td>\n",
       "      <td>1.087923</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.879658</td>\n",
       "      <td>0.683807</td>\n",
       "      <td>1.321442</td>\n",
       "      <td>0.468988</td>\n",
       "      <td>0.309200</td>\n",
       "      <td>0.996850</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14643</td>\n",
       "      <td>-0.904369</td>\n",
       "      <td>-1.845045</td>\n",
       "      <td>-0.259031</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.099739</td>\n",
       "      <td>-0.758910</td>\n",
       "      <td>-1.366165</td>\n",
       "      <td>-1.635112</td>\n",
       "      <td>-0.505285</td>\n",
       "      <td>-0.615435</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6185</td>\n",
       "      <td>0.397804</td>\n",
       "      <td>0.271476</td>\n",
       "      <td>0.548407</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.819664</td>\n",
       "      <td>-0.093041</td>\n",
       "      <td>0.684904</td>\n",
       "      <td>0.907342</td>\n",
       "      <td>-0.505285</td>\n",
       "      <td>0.204371</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>531</td>\n",
       "      <td>1.989350</td>\n",
       "      <td>1.411141</td>\n",
       "      <td>1.983446</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1.299614</td>\n",
       "      <td>2.015546</td>\n",
       "      <td>1.887254</td>\n",
       "      <td>2.047063</td>\n",
       "      <td>0.580695</td>\n",
       "      <td>2.130915</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4458</td>\n",
       "      <td>0.687176</td>\n",
       "      <td>-0.054143</td>\n",
       "      <td>1.069572</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.699677</td>\n",
       "      <td>0.794785</td>\n",
       "      <td>0.967810</td>\n",
       "      <td>0.819671</td>\n",
       "      <td>-0.505285</td>\n",
       "      <td>1.174475</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Overall  Potential   Special  International Reputation  Real Face  \\\n",
       "2540   1.121234   1.736760  1.087923                         1          0   \n",
       "14643 -0.904369  -1.845045 -0.259031                         1          0   \n",
       "6185   0.397804   0.271476  0.548407                         1          0   \n",
       "531    1.989350   1.411141  1.983446                         2          1   \n",
       "4458   0.687176  -0.054143  1.069572                         1          0   \n",
       "\n",
       "       BallControl  Reactions    Vision  Composure  Club_Reputation  \\\n",
       "2540      0.879658   0.683807  1.321442   0.468988         0.309200   \n",
       "14643     0.099739  -0.758910 -1.366165  -1.635112        -0.505285   \n",
       "6185      0.819664  -0.093041  0.684904   0.907342        -0.505285   \n",
       "531       1.299614   2.015546  1.887254   2.047063         0.580695   \n",
       "4458      0.699677   0.794785  0.967810   0.819671        -0.505285   \n",
       "\n",
       "       Mentality  Position_Group_Defender  Position_Group_Forward  \\\n",
       "2540    0.996850                        0                       0   \n",
       "14643  -0.615435                        1                       0   \n",
       "6185    0.204371                        0                       0   \n",
       "531     2.130915                        0                       0   \n",
       "4458    1.174475                        1                       0   \n",
       "\n",
       "       Position_Group_GoalKeeper  Position_Group_Midfielder  \n",
       "2540                           0                          1  \n",
       "14643                          0                          0  \n",
       "6185                           0                          1  \n",
       "531                            0                          1  \n",
       "4458                           0                          0  "
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_predictors_val.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.1) Preparing subset\n",
    "\n",
    "First, we are dropping all features that we did not use to train our final models. Also, we are transforming the categorical features to type categorical (if they were not already of type categorical)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_anomaly_processed=predictors_anomaly_processed[train_predictors_val.columns]\n",
    "predictors_anomaly_processed[['International Reputation','Real Face']]=predictors_anomaly_processed[['International Reputation','Real Face']].astype('category')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we are scaling all continuous features and back-transform categorical features to int (as the model does not accept categorical input)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "\n",
    "predictors_anomaly_processed[predictors_anomaly_processed.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']).columns] = scaler.fit_transform(predictors_anomaly_processed.select_dtypes(include=['float64','float32','int64','int32'], exclude=['category']))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors_anomaly_processed[predictors_anomaly_processed.select_dtypes(include='category').columns]=predictors_anomaly_processed[predictors_anomaly_processed.select_dtypes(include='category').columns].astype('int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Overall</th>\n",
       "      <th>Potential</th>\n",
       "      <th>Special</th>\n",
       "      <th>International Reputation</th>\n",
       "      <th>Real Face</th>\n",
       "      <th>BallControl</th>\n",
       "      <th>Reactions</th>\n",
       "      <th>Vision</th>\n",
       "      <th>Composure</th>\n",
       "      <th>Club_Reputation</th>\n",
       "      <th>Mentality</th>\n",
       "      <th>Position_Group_Defender</th>\n",
       "      <th>Position_Group_Forward</th>\n",
       "      <th>Position_Group_GoalKeeper</th>\n",
       "      <th>Position_Group_Midfielder</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1045</td>\n",
       "      <td>0.980581</td>\n",
       "      <td>0.980581</td>\n",
       "      <td>1.202217</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.162221</td>\n",
       "      <td>0.707107</td>\n",
       "      <td>0.707107</td>\n",
       "      <td>0.431934</td>\n",
       "      <td>0.92582</td>\n",
       "      <td>0.947282</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1443</td>\n",
       "      <td>0.392232</td>\n",
       "      <td>0.392232</td>\n",
       "      <td>-1.246094</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.297771</td>\n",
       "      <td>0.707107</td>\n",
       "      <td>-1.414214</td>\n",
       "      <td>-1.382189</td>\n",
       "      <td>-1.38873</td>\n",
       "      <td>-1.383032</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2802</td>\n",
       "      <td>-1.372813</td>\n",
       "      <td>-1.372813</td>\n",
       "      <td>0.043877</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1.135550</td>\n",
       "      <td>-1.414214</td>\n",
       "      <td>0.707107</td>\n",
       "      <td>0.950255</td>\n",
       "      <td>0.46291</td>\n",
       "      <td>0.435750</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Overall  Potential   Special  International Reputation  Real Face  \\\n",
       "1045  0.980581   0.980581  1.202217                         3          1   \n",
       "1443  0.392232   0.392232 -1.246094                         2          0   \n",
       "2802 -1.372813  -1.372813  0.043877                         3          0   \n",
       "\n",
       "      BallControl  Reactions    Vision  Composure  Club_Reputation  Mentality  \\\n",
       "1045     0.162221   0.707107  0.707107   0.431934          0.92582   0.947282   \n",
       "1443    -1.297771   0.707107 -1.414214  -1.382189         -1.38873  -1.383032   \n",
       "2802     1.135550  -1.414214  0.707107   0.950255          0.46291   0.435750   \n",
       "\n",
       "      Position_Group_Defender  Position_Group_Forward  \\\n",
       "1045                        0                       0   \n",
       "1443                        0                       1   \n",
       "2802                        0                       0   \n",
       "\n",
       "      Position_Group_GoalKeeper  Position_Group_Midfielder  \n",
       "1045                          0                          1  \n",
       "1443                          0                          0  \n",
       "2802                          0                          1  "
      ]
     },
     "execution_count": 201,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictors_anomaly_processed"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have prepared our final data frame, we can move on to predicting the players values."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.2) Final bid recommendation (prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = abs(xbr.predict(predictors_anomaly_processed))\n",
    "\n",
    "recommendation = pd.concat([pd.DataFrame(undervalued_players['ID']), pd.DataFrame(undervalued_players['Name']), pd.DataFrame(predictions)], axis = 'columns')\n",
    "\n",
    "recommendation.columns = ['ID', 'Name', 'Bid Recommendation']\n",
    "recommendation.to_csv('UndervaluedPlayers_BidRecommendation.csv', header = True, index = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [
    "wkjJ5jFt1atv",
    "dlA9qH-wjrs0"
   ],
   "name": "Copy of Data Mining Project.ipynb",
   "provenance": [],
   "toc_visible": true,
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}