Task_04_Financial_Data_Analysis—Model_Evaluation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 模型评估\n",
    "\n",
    "- **模型评估（Model Evaluation）**\n",
    "\n",
    "- **模型评估主要分为离散评估和在线评估（针对分类、回归、排序、序列预测等问题）**\n",
    "\n",
    "- **评估指标的局限性**\n",
    "\n",
    "> **精准率的局限性：当不同类别的样本比例不均衡时，占比大的类别往往将影响其精准率**\n",
    "\n",
    "> **精准率与召回率的权衡：**\n",
    "\n",
    "> **精准率指分类正确的正样本个数占分类器判定为正样本的样本个数比例**\n",
    "\n",
    "> **召回率指分类正确的正样本个数占真正的正样本个数的比例**\n",
    "\n",
    "> **为了评估一个排序模型的好坏，不仅要看模型在不同的 Top N下的 Precision N 和 Recall N 同时还要绘制出模型的 P-R 曲线，通过 P-R 曲线的整体表现，对模型进行全面的评估**\n",
    "\n",
    "> **ROC曲线（二值分类问题）：ROC曲线的横坐标为 假阳性率 FPR 纵坐标为真阳性率 TPR 其中 FPR = FP/N TPR = TP/P 其中，P为真实正样本数量 N为真实负样本数量 TP 是P个样本中被分类器预测为正样本的个数，FP 是N个负样本中被分类器预测为正样本的个数**\n",
    "\n",
    "> **AUC的计算：AUC 指 ROC 曲线下的面积大小，其值能够量化的反映基于ROC曲线衡量出的模型性能**\n",
    "\n",
    "> **ROC曲线与P-R曲线的特点：相比 P-R 曲线，当正负样本的分布发生变化时，ROC曲线的形状能够基本保持不变，而 P-R 曲线的形状一般会发生剧烈变化；如果选择不同的测试集， ROC 曲线能更加稳定的反映模型本身的好坏， 而 P-R 曲线能够直观的反映在特定数据集上的表现性能**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 任务说明：\n",
    "\n",
    "- 利用 **金融数据** 来预测贷款用户是否会逾期\n",
    "- 其中，**“status”** 是结果标签：**0** 表示未逾期，**1** 表示逾期"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 任务要求：\n",
    "\n",
    "> **模型评估（Model Evaluation）**\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 任务实现（模型评估）：（2 Days）\n",
    "\n",
    "\n",
    "- **记录5个模型（逻辑回归、SVM、决策树、随机森林、XGBoost）关于精确率Accuracy、准确率Precision，召回率Recall和F1-score、AUC值的评分表格，并画出ROC曲线**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.1 数据加载"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4754, 90)\n"
     ]
    },
    {
     "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>Unnamed: 0</th>\n",
       "      <th>custid</th>\n",
       "      <th>trade_no</th>\n",
       "      <th>bank_card_no</th>\n",
       "      <th>low_volume_percent</th>\n",
       "      <th>middle_volume_percent</th>\n",
       "      <th>take_amount_in_later_12_month_highest</th>\n",
       "      <th>trans_amount_increase_rate_lately</th>\n",
       "      <th>trans_activity_month</th>\n",
       "      <th>trans_activity_day</th>\n",
       "      <th>...</th>\n",
       "      <th>loans_max_limit</th>\n",
       "      <th>loans_avg_limit</th>\n",
       "      <th>consfin_credit_limit</th>\n",
       "      <th>consfin_credibility</th>\n",
       "      <th>consfin_org_count_current</th>\n",
       "      <th>consfin_product_count</th>\n",
       "      <th>consfin_max_limit</th>\n",
       "      <th>consfin_avg_limit</th>\n",
       "      <th>latest_query_day</th>\n",
       "      <th>loans_latest_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5</td>\n",
       "      <td>2791858</td>\n",
       "      <td>20180507115231274000000023057383</td>\n",
       "      <td>卡号1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.99</td>\n",
       "      <td>0</td>\n",
       "      <td>0.90</td>\n",
       "      <td>0.55</td>\n",
       "      <td>0.313</td>\n",
       "      <td>...</td>\n",
       "      <td>2900.0</td>\n",
       "      <td>1688.0</td>\n",
       "      <td>1200.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1200.0</td>\n",
       "      <td>1200.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10</td>\n",
       "      <td>534047</td>\n",
       "      <td>20180507121002192000000023073000</td>\n",
       "      <td>卡号1</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.94</td>\n",
       "      <td>2000</td>\n",
       "      <td>1.28</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.458</td>\n",
       "      <td>...</td>\n",
       "      <td>3500.0</td>\n",
       "      <td>1758.0</td>\n",
       "      <td>15100.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>22800.0</td>\n",
       "      <td>9360.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>2849787</td>\n",
       "      <td>20180507125159718000000023114911</td>\n",
       "      <td>卡号1</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.96</td>\n",
       "      <td>0</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.114</td>\n",
       "      <td>...</td>\n",
       "      <td>1600.0</td>\n",
       "      <td>1250.0</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>13</td>\n",
       "      <td>1809708</td>\n",
       "      <td>20180507121358683000000388283484</td>\n",
       "      <td>卡号1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.96</td>\n",
       "      <td>2000</td>\n",
       "      <td>0.13</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.777</td>\n",
       "      <td>...</td>\n",
       "      <td>3200.0</td>\n",
       "      <td>1541.0</td>\n",
       "      <td>16300.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>30000.0</td>\n",
       "      <td>12180.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>14</td>\n",
       "      <td>2499829</td>\n",
       "      <td>20180507115448545000000388205844</td>\n",
       "      <td>卡号1</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.99</td>\n",
       "      <td>0</td>\n",
       "      <td>0.46</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.175</td>\n",
       "      <td>...</td>\n",
       "      <td>2300.0</td>\n",
       "      <td>1630.0</td>\n",
       "      <td>8300.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8400.0</td>\n",
       "      <td>8250.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>120.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 90 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0   custid                          trade_no bank_card_no  \\\n",
       "0           5  2791858  20180507115231274000000023057383          卡号1   \n",
       "1          10   534047  20180507121002192000000023073000          卡号1   \n",
       "2          12  2849787  20180507125159718000000023114911          卡号1   \n",
       "3          13  1809708  20180507121358683000000388283484          卡号1   \n",
       "4          14  2499829  20180507115448545000000388205844          卡号1   \n",
       "\n",
       "   low_volume_percent  middle_volume_percent  \\\n",
       "0                0.01                   0.99   \n",
       "1                0.02                   0.94   \n",
       "2                0.04                   0.96   \n",
       "3                0.00                   0.96   \n",
       "4                0.01                   0.99   \n",
       "\n",
       "   take_amount_in_later_12_month_highest  trans_amount_increase_rate_lately  \\\n",
       "0                                      0                               0.90   \n",
       "1                                   2000                               1.28   \n",
       "2                                      0                               1.00   \n",
       "3                                   2000                               0.13   \n",
       "4                                      0                               0.46   \n",
       "\n",
       "   trans_activity_month  trans_activity_day        ...         \\\n",
       "0                  0.55               0.313        ...          \n",
       "1                  1.00               0.458        ...          \n",
       "2                  1.00               0.114        ...          \n",
       "3                  0.57               0.777        ...          \n",
       "4                  1.00               0.175        ...          \n",
       "\n",
       "   loans_max_limit  loans_avg_limit  consfin_credit_limit  \\\n",
       "0           2900.0           1688.0                1200.0   \n",
       "1           3500.0           1758.0               15100.0   \n",
       "2           1600.0           1250.0                4200.0   \n",
       "3           3200.0           1541.0               16300.0   \n",
       "4           2300.0           1630.0                8300.0   \n",
       "\n",
       "   consfin_credibility  consfin_org_count_current  consfin_product_count  \\\n",
       "0                 75.0                        1.0                    2.0   \n",
       "1                 80.0                        5.0                    6.0   \n",
       "2                 87.0                        1.0                    1.0   \n",
       "3                 80.0                        5.0                    5.0   \n",
       "4                 79.0                        2.0                    2.0   \n",
       "\n",
       "   consfin_max_limit  consfin_avg_limit  latest_query_day  loans_latest_day  \n",
       "0             1200.0             1200.0              12.0              18.0  \n",
       "1            22800.0             9360.0               4.0               2.0  \n",
       "2             4200.0             4200.0               2.0               6.0  \n",
       "3            30000.0            12180.0               2.0               4.0  \n",
       "4             8400.0             8250.0              22.0             120.0  \n",
       "\n",
       "[5 rows x 90 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 数据集的加载\n",
    "data = pd.read_csv('Dataset/data.csv')\n",
    "# 剔除重复特征\n",
    "data.drop_duplicates(inplace=True)\n",
    "print(data.shape)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    3561\n",
       "1    1193\n",
       "Name: status, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pickle\n",
    "\n",
    "# 载入特征\n",
    "with open('feature.pkl', 'rb') as f:\n",
    "    X = pickle.load(f)\n",
    "\n",
    "# 观测正负样本是否均衡\n",
    "y = data.status\n",
    "y.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>regional_mobility</th>\n",
       "      <th>student_feature</th>\n",
       "      <th>is_high_user</th>\n",
       "      <th>avg_consume_less_12_valid_month</th>\n",
       "      <th>reg_preference_for_trad</th>\n",
       "      <th>first_transaction_time_year</th>\n",
       "      <th>first_transaction_time_month</th>\n",
       "      <th>first_transaction_time_day</th>\n",
       "      <th>latest_query_time_year</th>\n",
       "      <th>latest_query_time_month</th>\n",
       "      <th>...</th>\n",
       "      <th>loans_max_limit</th>\n",
       "      <th>loans_avg_limit</th>\n",
       "      <th>consfin_credit_limit</th>\n",
       "      <th>consfin_credibility</th>\n",
       "      <th>consfin_org_count_current</th>\n",
       "      <th>consfin_product_count</th>\n",
       "      <th>consfin_max_limit</th>\n",
       "      <th>consfin_avg_limit</th>\n",
       "      <th>latest_query_day</th>\n",
       "      <th>loans_latest_day</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2013</td>\n",
       "      <td>8</td>\n",
       "      <td>17</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2900.0</td>\n",
       "      <td>1688.0</td>\n",
       "      <td>1200.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1200.0</td>\n",
       "      <td>1200.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3500.0</td>\n",
       "      <td>1758.0</td>\n",
       "      <td>15100.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>22800.0</td>\n",
       "      <td>9360.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>6</td>\n",
       "      <td>17</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1600.0</td>\n",
       "      <td>1250.0</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2013</td>\n",
       "      <td>5</td>\n",
       "      <td>16</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3200.0</td>\n",
       "      <td>1541.0</td>\n",
       "      <td>16300.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>30000.0</td>\n",
       "      <td>12180.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>3</td>\n",
       "      <td>12</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2300.0</td>\n",
       "      <td>1630.0</td>\n",
       "      <td>8300.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8400.0</td>\n",
       "      <td>8250.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>120.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2016</td>\n",
       "      <td>9</td>\n",
       "      <td>20</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5300.0</td>\n",
       "      <td>1941.0</td>\n",
       "      <td>11200.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>20400.0</td>\n",
       "      <td>8130.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>12</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2200.0</td>\n",
       "      <td>2200.0</td>\n",
       "      <td>7600.0</td>\n",
       "      <td>73.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>16800.0</td>\n",
       "      <td>8900.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>5</td>\n",
       "      <td>12</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>1810.0</td>\n",
       "      <td>7700.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>13800.0</td>\n",
       "      <td>7050.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>36.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>2</td>\n",
       "      <td>18</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5300.0</td>\n",
       "      <td>4750.0</td>\n",
       "      <td>5500.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>19200.0</td>\n",
       "      <td>7987.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2800.0</td>\n",
       "      <td>1520.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>142.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>8</td>\n",
       "      <td>14</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1800.0</td>\n",
       "      <td>1325.0</td>\n",
       "      <td>9900.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>20400.0</td>\n",
       "      <td>7757.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2014</td>\n",
       "      <td>10</td>\n",
       "      <td>26</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3900.0</td>\n",
       "      <td>2450.0</td>\n",
       "      <td>8300.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>13200.0</td>\n",
       "      <td>9400.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>35.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2013</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3500.0</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>19300.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>42000.0</td>\n",
       "      <td>16985.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>8</td>\n",
       "      <td>7</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>9300.0</td>\n",
       "      <td>2240.0</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>83.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>11400.0</td>\n",
       "      <td>5460.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>4.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>12</td>\n",
       "      <td>20</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>1600.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>193.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>4.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>1</td>\n",
       "      <td>20</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4100.0</td>\n",
       "      <td>2633.0</td>\n",
       "      <td>18400.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>33600.0</td>\n",
       "      <td>17633.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>27.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2015</td>\n",
       "      <td>12</td>\n",
       "      <td>3</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4500.0</td>\n",
       "      <td>2480.0</td>\n",
       "      <td>21300.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>27600.0</td>\n",
       "      <td>9900.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2013</td>\n",
       "      <td>12</td>\n",
       "      <td>22</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>1575.0</td>\n",
       "      <td>4700.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>6600.0</td>\n",
       "      <td>4062.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2016</td>\n",
       "      <td>10</td>\n",
       "      <td>18</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>2533.0</td>\n",
       "      <td>10100.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>15600.0</td>\n",
       "      <td>10275.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>77.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2013</td>\n",
       "      <td>6</td>\n",
       "      <td>13</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>6200.0</td>\n",
       "      <td>1957.0</td>\n",
       "      <td>10100.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20400.0</td>\n",
       "      <td>8187.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>37.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>7</td>\n",
       "      <td>21</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>500.0</td>\n",
       "      <td>500.0</td>\n",
       "      <td>8400.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8400.0</td>\n",
       "      <td>8400.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>71.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2014</td>\n",
       "      <td>7</td>\n",
       "      <td>6</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2400.0</td>\n",
       "      <td>1633.0</td>\n",
       "      <td>16500.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>32400.0</td>\n",
       "      <td>18600.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>23.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>4.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2014</td>\n",
       "      <td>12</td>\n",
       "      <td>23</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>9000.0</td>\n",
       "      <td>3080.0</td>\n",
       "      <td>21900.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>39600.0</td>\n",
       "      <td>20454.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2014</td>\n",
       "      <td>9</td>\n",
       "      <td>16</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3500.0</td>\n",
       "      <td>1900.0</td>\n",
       "      <td>8100.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>14400.0</td>\n",
       "      <td>6100.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2011</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1000.0</td>\n",
       "      <td>750.0</td>\n",
       "      <td>14000.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>31200.0</td>\n",
       "      <td>15500.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>104.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2014</td>\n",
       "      <td>3</td>\n",
       "      <td>15</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>1560.0</td>\n",
       "      <td>10400.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>42600.0</td>\n",
       "      <td>9841.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>82.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>11</td>\n",
       "      <td>20</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2500.0</td>\n",
       "      <td>1671.0</td>\n",
       "      <td>6500.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8100.0</td>\n",
       "      <td>7050.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>80.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>7</td>\n",
       "      <td>9</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5700.0</td>\n",
       "      <td>3160.0</td>\n",
       "      <td>7000.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>13200.0</td>\n",
       "      <td>6461.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2700.0</td>\n",
       "      <td>1825.0</td>\n",
       "      <td>21600.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>21600.0</td>\n",
       "      <td>21600.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2012</td>\n",
       "      <td>5</td>\n",
       "      <td>20</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>2000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>34.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4724</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>8</td>\n",
       "      <td>26</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3800.0</td>\n",
       "      <td>2271.0</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9000.0</td>\n",
       "      <td>3460.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4725</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>6</td>\n",
       "      <td>16</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2100.0</td>\n",
       "      <td>1166.0</td>\n",
       "      <td>17200.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>21600.0</td>\n",
       "      <td>11850.0</td>\n",
       "      <td>142.0</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4726</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2016</td>\n",
       "      <td>7</td>\n",
       "      <td>15</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>1960.0</td>\n",
       "      <td>4800.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4800.0</td>\n",
       "      <td>4800.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>53.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4727</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>6</td>\n",
       "      <td>10</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>1812.0</td>\n",
       "      <td>4800.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>12000.0</td>\n",
       "      <td>5125.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>57.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4728</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>3</td>\n",
       "      <td>14</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2500.0</td>\n",
       "      <td>1662.0</td>\n",
       "      <td>10800.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>20400.0</td>\n",
       "      <td>8933.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4729</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2500.0</td>\n",
       "      <td>1518.0</td>\n",
       "      <td>6200.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>16800.0</td>\n",
       "      <td>7200.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>104.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4730</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2014</td>\n",
       "      <td>8</td>\n",
       "      <td>26</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>1771.0</td>\n",
       "      <td>15600.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>15600.0</td>\n",
       "      <td>15600.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>128.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4731</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2013</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>1818.0</td>\n",
       "      <td>60200.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>124200.0</td>\n",
       "      <td>25361.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>11.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4732</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>4</td>\n",
       "      <td>21</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4300.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>7200.0</td>\n",
       "      <td>4950.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>250.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4733</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2016</td>\n",
       "      <td>10</td>\n",
       "      <td>20</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4700.0</td>\n",
       "      <td>2320.0</td>\n",
       "      <td>9000.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>13200.0</td>\n",
       "      <td>6900.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4734</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2016</td>\n",
       "      <td>7</td>\n",
       "      <td>19</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>1950.0</td>\n",
       "      <td>1500.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1500.0</td>\n",
       "      <td>1500.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>131.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4735</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>11</td>\n",
       "      <td>10</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4100.0</td>\n",
       "      <td>1750.0</td>\n",
       "      <td>20300.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>34800.0</td>\n",
       "      <td>21100.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4736</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>10</td>\n",
       "      <td>15</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1200.0</td>\n",
       "      <td>850.0</td>\n",
       "      <td>10200.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>10200.0</td>\n",
       "      <td>10200.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>156.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4737</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2014</td>\n",
       "      <td>3</td>\n",
       "      <td>13</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3700.0</td>\n",
       "      <td>1757.0</td>\n",
       "      <td>14800.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>20400.0</td>\n",
       "      <td>11714.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4738</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>8</td>\n",
       "      <td>12</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>900.0</td>\n",
       "      <td>900.0</td>\n",
       "      <td>6800.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>7200.0</td>\n",
       "      <td>5700.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>89.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4739</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2015</td>\n",
       "      <td>3</td>\n",
       "      <td>29</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2300.0</td>\n",
       "      <td>1766.0</td>\n",
       "      <td>3800.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3900.0</td>\n",
       "      <td>3600.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4740</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>9</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3900.0</td>\n",
       "      <td>2200.0</td>\n",
       "      <td>11300.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>12000.0</td>\n",
       "      <td>9200.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>163.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4741</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2300.0</td>\n",
       "      <td>2300.0</td>\n",
       "      <td>13700.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>19800.0</td>\n",
       "      <td>13900.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>140.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4742</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>30</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4200.0</td>\n",
       "      <td>1928.0</td>\n",
       "      <td>9200.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>21600.0</td>\n",
       "      <td>7557.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4743</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>17</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3500.0</td>\n",
       "      <td>1825.0</td>\n",
       "      <td>1700.0</td>\n",
       "      <td>72.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3600.0</td>\n",
       "      <td>2300.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4744</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2011</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3000.0</td>\n",
       "      <td>1691.0</td>\n",
       "      <td>5200.0</td>\n",
       "      <td>77.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>8400.0</td>\n",
       "      <td>3344.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>88.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4745</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "      <td>2016</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2100.0</td>\n",
       "      <td>1280.0</td>\n",
       "      <td>6800.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>7200.0</td>\n",
       "      <td>5850.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4746</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>11</td>\n",
       "      <td>15</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>1810.0</td>\n",
       "      <td>7700.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>13800.0</td>\n",
       "      <td>7050.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>36.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4747</th>\n",
       "      <td>4.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2016</td>\n",
       "      <td>9</td>\n",
       "      <td>25</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1100.0</td>\n",
       "      <td>1100.0</td>\n",
       "      <td>4500.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6000.0</td>\n",
       "      <td>3260.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>22.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4748</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2013</td>\n",
       "      <td>9</td>\n",
       "      <td>14</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3100.0</td>\n",
       "      <td>2450.0</td>\n",
       "      <td>17300.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>19800.0</td>\n",
       "      <td>18300.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>54.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4749</th>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>23</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>49200.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>49200.0</td>\n",
       "      <td>49200.0</td>\n",
       "      <td>44.0</td>\n",
       "      <td>58.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4750</th>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2014</td>\n",
       "      <td>5</td>\n",
       "      <td>27</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>1952.0</td>\n",
       "      <td>10100.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>19200.0</td>\n",
       "      <td>7500.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>55.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4751</th>\n",
       "      <td>3.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2013</td>\n",
       "      <td>2</td>\n",
       "      <td>24</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3300.0</td>\n",
       "      <td>1540.0</td>\n",
       "      <td>7600.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>21000.0</td>\n",
       "      <td>7250.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>55.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4752</th>\n",
       "      <td>4.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>3</td>\n",
       "      <td>23</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1900.0</td>\n",
       "      <td>1216.0</td>\n",
       "      <td>7100.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>12000.0</td>\n",
       "      <td>3950.0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>16.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4753</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2015</td>\n",
       "      <td>7</td>\n",
       "      <td>7</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4000.0</td>\n",
       "      <td>2180.0</td>\n",
       "      <td>4500.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>7800.0</td>\n",
       "      <td>4360.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>54.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4754 rows × 87 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      regional_mobility  student_feature  is_high_user  \\\n",
       "0                   3.0             -1.0             0   \n",
       "1                   4.0              1.0             0   \n",
       "2                   1.0             -1.0             0   \n",
       "3                   3.0             -1.0             0   \n",
       "4                   1.0             -1.0             0   \n",
       "5                   3.0              1.0             0   \n",
       "6                   2.0             -1.0             0   \n",
       "7                   2.0             -1.0             0   \n",
       "8                   2.0             -1.0             0   \n",
       "9                   2.0              1.0             0   \n",
       "10                  2.0             -1.0             0   \n",
       "11                  1.0              1.0             0   \n",
       "12                  4.0              1.0             0   \n",
       "13                  1.0             -1.0             0   \n",
       "14                  4.0             -1.0             0   \n",
       "15                  4.0             -1.0             0   \n",
       "16                  3.0             -1.0             0   \n",
       "17                  4.0              1.0             0   \n",
       "18                  3.0              1.0             0   \n",
       "19                  3.0             -1.0             0   \n",
       "20                  3.0              1.0             0   \n",
       "21                  3.0             -1.0             0   \n",
       "22                  4.0             -1.0             0   \n",
       "23                  3.0             -1.0             0   \n",
       "24                  3.0              1.0             0   \n",
       "25                  3.0             -1.0             0   \n",
       "26                  3.0              1.0             0   \n",
       "27                  3.0              1.0             0   \n",
       "28                  3.0              1.0             0   \n",
       "29                  4.0              1.0             0   \n",
       "...                 ...              ...           ...   \n",
       "4724                3.0              1.0             0   \n",
       "4725                2.0             -1.0             0   \n",
       "4726                2.0             -1.0             0   \n",
       "4727                3.0             -1.0             0   \n",
       "4728                3.0             -1.0             1   \n",
       "4729                2.0             -1.0             0   \n",
       "4730                4.0              1.0             0   \n",
       "4731                4.0              1.0             0   \n",
       "4732                1.0             -1.0             0   \n",
       "4733                3.0             -1.0             0   \n",
       "4734                3.0             -1.0             0   \n",
       "4735                3.0              1.0             0   \n",
       "4736                2.0              1.0             0   \n",
       "4737                3.0              1.0             0   \n",
       "4738                3.0             -1.0             0   \n",
       "4739                2.0             -1.0             0   \n",
       "4740                2.0             -1.0             0   \n",
       "4741                2.0              1.0             0   \n",
       "4742                2.0              1.0             0   \n",
       "4743                2.0             -1.0             0   \n",
       "4744                3.0             -1.0             0   \n",
       "4745                3.0             -1.0             0   \n",
       "4746                3.0              1.0             0   \n",
       "4747                4.0             -1.0             0   \n",
       "4748                2.0             -1.0             0   \n",
       "4749                4.0              1.0             0   \n",
       "4750                3.0              1.0             0   \n",
       "4751                3.0             -1.0             0   \n",
       "4752                4.0             -1.0             0   \n",
       "4753                2.0              1.0             0   \n",
       "\n",
       "      avg_consume_less_12_valid_month  reg_preference_for_trad  \\\n",
       "0                                 7.0                        0   \n",
       "1                                 5.0                        0   \n",
       "2                                 5.0                        0   \n",
       "3                                 7.0                        1   \n",
       "4                                 6.0                        0   \n",
       "5                                 6.0                        1   \n",
       "6                                 6.0                        0   \n",
       "7                                 8.0                        0   \n",
       "8                                 8.0                        1   \n",
       "9                                 7.0                        0   \n",
       "10                                5.0                        0   \n",
       "11                                7.0                        0   \n",
       "12                                7.0                        1   \n",
       "13                                7.0                        0   \n",
       "14                                9.0                        0   \n",
       "15                                6.0                        0   \n",
       "16                                8.0                        2   \n",
       "17                                7.0                        0   \n",
       "18                                6.0                        2   \n",
       "19                                8.0                        0   \n",
       "20                                5.0                        0   \n",
       "21                                7.0                        0   \n",
       "22                                9.0                        2   \n",
       "23                                8.0                        1   \n",
       "24                                7.0                        0   \n",
       "25                                7.0                        2   \n",
       "26                                5.0                        0   \n",
       "27                                8.0                        1   \n",
       "28                                7.0                        0   \n",
       "29                                8.0                        0   \n",
       "...                               ...                      ...   \n",
       "4724                              5.0                        0   \n",
       "4725                              5.0                        0   \n",
       "4726                              8.0                        2   \n",
       "4727                              8.0                        0   \n",
       "4728                              8.0                        0   \n",
       "4729                              7.0                        0   \n",
       "4730                              5.0                        1   \n",
       "4731                              8.0                        1   \n",
       "4732                              6.0                        0   \n",
       "4733                              8.0                        3   \n",
       "4734                              5.0                        2   \n",
       "4735                              7.0                        0   \n",
       "4736                              5.0                        0   \n",
       "4737                              5.0                        0   \n",
       "4738                              5.0                        0   \n",
       "4739                              6.0                        1   \n",
       "4740                              5.0                        0   \n",
       "4741                              8.0                        0   \n",
       "4742                              6.0                        0   \n",
       "4743                              8.0                        1   \n",
       "4744                              7.0                        0   \n",
       "4745                              5.0                        1   \n",
       "4746                              5.0                        0   \n",
       "4747                              8.0                        2   \n",
       "4748                              6.0                        0   \n",
       "4749                              7.0                        0   \n",
       "4750                              7.0                        0   \n",
       "4751                              8.0                        0   \n",
       "4752                              7.0                        0   \n",
       "4753                              7.0                        0   \n",
       "\n",
       "      first_transaction_time_year  first_transaction_time_month  \\\n",
       "0                            2013                             8   \n",
       "1                            2016                             4   \n",
       "2                            2017                             6   \n",
       "3                            2013                             5   \n",
       "4                            2017                             3   \n",
       "5                            2016                             9   \n",
       "6                            2016                            11   \n",
       "7                            2015                             5   \n",
       "8                            2015                             2   \n",
       "9                            2017                             1   \n",
       "10                           2015                             8   \n",
       "11                           2014                            10   \n",
       "12                           2013                             2   \n",
       "13                           2015                             8   \n",
       "14                           2016                            12   \n",
       "15                           2015                             1   \n",
       "16                           2015                            12   \n",
       "17                           2013                            12   \n",
       "18                           2016                            10   \n",
       "19                           2013                             6   \n",
       "20                           2017                             7   \n",
       "21                           2014                             7   \n",
       "22                           2014                            12   \n",
       "23                           2014                             9   \n",
       "24                           2011                             1   \n",
       "25                           2014                             3   \n",
       "26                           2015                            11   \n",
       "27                           2015                             7   \n",
       "28                           2017                             1   \n",
       "29                           2012                             5   \n",
       "...                           ...                           ...   \n",
       "4724                         2016                             8   \n",
       "4725                         2015                             6   \n",
       "4726                         2016                             7   \n",
       "4727                         2015                             6   \n",
       "4728                         2017                             3   \n",
       "4729                         2016                             6   \n",
       "4730                         2014                             8   \n",
       "4731                         2013                             3   \n",
       "4732                         2017                             4   \n",
       "4733                         2016                            10   \n",
       "4734                         2016                             7   \n",
       "4735                         2015                            11   \n",
       "4736                         2016                            10   \n",
       "4737                         2014                             3   \n",
       "4738                         2016                             8   \n",
       "4739                         2015                             3   \n",
       "4740                         2017                             2   \n",
       "4741                         2016                            10   \n",
       "4742                         2016                             1   \n",
       "4743                         2017                             2   \n",
       "4744                         2011                             1   \n",
       "4745                         2016                             9   \n",
       "4746                         2015                            11   \n",
       "4747                         2016                             9   \n",
       "4748                         2013                             9   \n",
       "4749                         2016                            11   \n",
       "4750                         2014                             5   \n",
       "4751                         2013                             2   \n",
       "4752                         2016                             3   \n",
       "4753                         2015                             7   \n",
       "\n",
       "      first_transaction_time_day  latest_query_time_year  \\\n",
       "0                             17                  2018.0   \n",
       "1                              2                  2018.0   \n",
       "2                             17                  2018.0   \n",
       "3                             16                  2018.0   \n",
       "4                             12                  2018.0   \n",
       "5                             20                  2018.0   \n",
       "6                             12                  2018.0   \n",
       "7                             12                  2018.0   \n",
       "8                             18                  2018.0   \n",
       "9                              5                  2018.0   \n",
       "10                            14                  2018.0   \n",
       "11                            26                  2018.0   \n",
       "12                             6                  2018.0   \n",
       "13                             7                  2018.0   \n",
       "14                            20                  2018.0   \n",
       "15                            20                  2018.0   \n",
       "16                             3                  2018.0   \n",
       "17                            22                  2018.0   \n",
       "18                            18                  2018.0   \n",
       "19                            13                  2018.0   \n",
       "20                            21                  2018.0   \n",
       "21                             6                  2018.0   \n",
       "22                            23                  2018.0   \n",
       "23                            16                  2018.0   \n",
       "24                            16                  2018.0   \n",
       "25                            15                  2018.0   \n",
       "26                            20                  2018.0   \n",
       "27                             9                  2018.0   \n",
       "28                            11                  2018.0   \n",
       "29                            20                  2018.0   \n",
       "...                          ...                     ...   \n",
       "4724                          26                  2018.0   \n",
       "4725                          16                  2017.0   \n",
       "4726                          15                  2018.0   \n",
       "4727                          10                  2018.0   \n",
       "4728                          14                  2018.0   \n",
       "4729                           7                  2018.0   \n",
       "4730                          26                  2018.0   \n",
       "4731                           3                  2018.0   \n",
       "4732                          21                  2018.0   \n",
       "4733                          20                  2018.0   \n",
       "4734                          19                  2018.0   \n",
       "4735                          10                  2018.0   \n",
       "4736                          15                  2018.0   \n",
       "4737                          13                  2018.0   \n",
       "4738                          12                  2018.0   \n",
       "4739                          29                  2018.0   \n",
       "4740                           9                  2018.0   \n",
       "4741                          10                  2018.0   \n",
       "4742                          30                  2018.0   \n",
       "4743                          17                  2018.0   \n",
       "4744                           3                  2018.0   \n",
       "4745                           9                  2018.0   \n",
       "4746                          15                  2018.0   \n",
       "4747                          25                  2018.0   \n",
       "4748                          14                  2018.0   \n",
       "4749                          23                  2018.0   \n",
       "4750                          27                  2018.0   \n",
       "4751                          24                  2018.0   \n",
       "4752                          23                  2018.0   \n",
       "4753                           7                  2018.0   \n",
       "\n",
       "      latest_query_time_month        ...         loans_max_limit  \\\n",
       "0                         4.0        ...                  2900.0   \n",
       "1                         5.0        ...                  3500.0   \n",
       "2                         5.0        ...                  1600.0   \n",
       "3                         5.0        ...                  3200.0   \n",
       "4                         4.0        ...                  2300.0   \n",
       "5                         5.0        ...                  5300.0   \n",
       "6                         5.0        ...                  2200.0   \n",
       "7                         4.0        ...                  3100.0   \n",
       "8                         4.0        ...                  5300.0   \n",
       "9                         4.0        ...                  2800.0   \n",
       "10                        4.0        ...                  1800.0   \n",
       "11                        4.0        ...                  3900.0   \n",
       "12                        3.0        ...                  3500.0   \n",
       "13                        4.0        ...                  9300.0   \n",
       "14                        4.0        ...                  2000.0   \n",
       "15                        5.0        ...                  4100.0   \n",
       "16                        5.0        ...                  4500.0   \n",
       "17                        4.0        ...                  2000.0   \n",
       "18                        4.0        ...                  5000.0   \n",
       "19                        4.0        ...                  6200.0   \n",
       "20                        5.0        ...                   500.0   \n",
       "21                        4.0        ...                  2400.0   \n",
       "22                        4.0        ...                  9000.0   \n",
       "23                        4.0        ...                  3500.0   \n",
       "24                        4.0        ...                  1000.0   \n",
       "25                        4.0        ...                  5000.0   \n",
       "26                        4.0        ...                  2500.0   \n",
       "27                        4.0        ...                  5700.0   \n",
       "28                        5.0        ...                  2700.0   \n",
       "29                        4.0        ...                  2000.0   \n",
       "...                       ...        ...                     ...   \n",
       "4724                      4.0        ...                  3800.0   \n",
       "4725                     12.0        ...                  2100.0   \n",
       "4726                      4.0        ...                  3000.0   \n",
       "4727                      4.0        ...                  3100.0   \n",
       "4728                      4.0        ...                  2500.0   \n",
       "4729                      5.0        ...                  2500.0   \n",
       "4730                      4.0        ...                  3100.0   \n",
       "4731                      5.0        ...                  3100.0   \n",
       "4732                      4.0        ...                     0.0   \n",
       "4733                      4.0        ...                  4700.0   \n",
       "4734                      4.0        ...                  3100.0   \n",
       "4735                      5.0        ...                  4100.0   \n",
       "4736                      4.0        ...                  1200.0   \n",
       "4737                      5.0        ...                  3700.0   \n",
       "4738                      3.0        ...                   900.0   \n",
       "4739                      4.0        ...                  2300.0   \n",
       "4740                      4.0        ...                  3900.0   \n",
       "4741                      4.0        ...                  2300.0   \n",
       "4742                      5.0        ...                  4200.0   \n",
       "4743                      4.0        ...                  3500.0   \n",
       "4744                      5.0        ...                  3000.0   \n",
       "4745                      5.0        ...                  2100.0   \n",
       "4746                      4.0        ...                  3100.0   \n",
       "4747                      5.0        ...                  1100.0   \n",
       "4748                      4.0        ...                  3100.0   \n",
       "4749                      3.0        ...                     0.0   \n",
       "4750                      4.0        ...                  5000.0   \n",
       "4751                      4.0        ...                  3300.0   \n",
       "4752                      3.0        ...                  1900.0   \n",
       "4753                      4.0        ...                  4000.0   \n",
       "\n",
       "      loans_avg_limit  consfin_credit_limit  consfin_credibility  \\\n",
       "0              1688.0                1200.0                 75.0   \n",
       "1              1758.0               15100.0                 80.0   \n",
       "2              1250.0                4200.0                 87.0   \n",
       "3              1541.0               16300.0                 80.0   \n",
       "4              1630.0                8300.0                 79.0   \n",
       "5              1941.0               11200.0                 80.0   \n",
       "6              2200.0                7600.0                 73.0   \n",
       "7              1810.0                7700.0                 79.0   \n",
       "8              4750.0                5500.0                 79.0   \n",
       "9              1520.0                   0.0                  0.0   \n",
       "10             1325.0                9900.0                 80.0   \n",
       "11             2450.0                8300.0                 85.0   \n",
       "12             2000.0               19300.0                 79.0   \n",
       "13             2240.0                5000.0                 83.0   \n",
       "14             1600.0                   0.0                  0.0   \n",
       "15             2633.0               18400.0                 79.0   \n",
       "16             2480.0               21300.0                 77.0   \n",
       "17             1575.0                4700.0                 80.0   \n",
       "18             2533.0               10100.0                 78.0   \n",
       "19             1957.0               10100.0                 79.0   \n",
       "20              500.0                8400.0                 74.0   \n",
       "21             1633.0               16500.0                 76.0   \n",
       "22             3080.0               21900.0                 79.0   \n",
       "23             1900.0                8100.0                 80.0   \n",
       "24              750.0               14000.0                 75.0   \n",
       "25             1560.0               10400.0                 78.0   \n",
       "26             1671.0                6500.0                 77.0   \n",
       "27             3160.0                7000.0                 78.0   \n",
       "28             1825.0               21600.0                 78.0   \n",
       "29             2000.0                3000.0                 79.0   \n",
       "...               ...                   ...                  ...   \n",
       "4724           2271.0                5000.0                 79.0   \n",
       "4725           1166.0               17200.0                 78.0   \n",
       "4726           1960.0                4800.0                 75.0   \n",
       "4727           1812.0                4800.0                 76.0   \n",
       "4728           1662.0               10800.0                 80.0   \n",
       "4729           1518.0                6200.0                 80.0   \n",
       "4730           1771.0               15600.0                 78.0   \n",
       "4731           1818.0               60200.0                 78.0   \n",
       "4732              0.0                4300.0                 80.0   \n",
       "4733           2320.0                9000.0                 81.0   \n",
       "4734           1950.0                1500.0                 78.0   \n",
       "4735           1750.0               20300.0                 79.0   \n",
       "4736            850.0               10200.0                 85.0   \n",
       "4737           1757.0               14800.0                 78.0   \n",
       "4738            900.0                6800.0                 79.0   \n",
       "4739           1766.0                3800.0                 72.0   \n",
       "4740           2200.0               11300.0                 79.0   \n",
       "4741           2300.0               13700.0                 80.0   \n",
       "4742           1928.0                9200.0                 77.0   \n",
       "4743           1825.0                1700.0                 72.0   \n",
       "4744           1691.0                5200.0                 77.0   \n",
       "4745           1280.0                6800.0                 82.0   \n",
       "4746           1810.0                7700.0                 79.0   \n",
       "4747           1100.0                4500.0                 76.0   \n",
       "4748           2450.0               17300.0                 75.0   \n",
       "4749              0.0               49200.0                 78.0   \n",
       "4750           1952.0               10100.0                 78.0   \n",
       "4751           1540.0                7600.0                 79.0   \n",
       "4752           1216.0                7100.0                 79.0   \n",
       "4753           2180.0                4500.0                 74.0   \n",
       "\n",
       "      consfin_org_count_current  consfin_product_count  consfin_max_limit  \\\n",
       "0                           1.0                    2.0             1200.0   \n",
       "1                           5.0                    6.0            22800.0   \n",
       "2                           1.0                    1.0             4200.0   \n",
       "3                           5.0                    5.0            30000.0   \n",
       "4                           2.0                    2.0             8400.0   \n",
       "5                          10.0                   12.0            20400.0   \n",
       "6                           2.0                    2.0            16800.0   \n",
       "7                           4.0                    5.0            13800.0   \n",
       "8                           8.0                   11.0            19200.0   \n",
       "9                           0.0                    0.0                0.0   \n",
       "10                          7.0                    8.0            20400.0   \n",
       "11                          3.0                    3.0            13200.0   \n",
       "12                          7.0                    8.0            42000.0   \n",
       "13                          5.0                    7.0            11400.0   \n",
       "14                          0.0                    0.0                0.0   \n",
       "15                          9.0                   11.0            33600.0   \n",
       "16                          4.0                    4.0            27600.0   \n",
       "17                          8.0                   10.0             6600.0   \n",
       "18                          4.0                    4.0            15600.0   \n",
       "19                          8.0                    9.0            20400.0   \n",
       "20                          1.0                    1.0             8400.0   \n",
       "21                          2.0                    3.0            32400.0   \n",
       "22                         11.0                   11.0            39600.0   \n",
       "23                          6.0                    7.0            14400.0   \n",
       "24                          3.0                    3.0            31200.0   \n",
       "25                         12.0                   14.0            42600.0   \n",
       "26                          2.0                    2.0             8100.0   \n",
       "27                         13.0                   15.0            13200.0   \n",
       "28                          1.0                    1.0            21600.0   \n",
       "29                          1.0                    1.0             3000.0   \n",
       "...                         ...                    ...                ...   \n",
       "4724                        5.0                    6.0             9000.0   \n",
       "4725                        4.0                    4.0            21600.0   \n",
       "4726                        1.0                    1.0             4800.0   \n",
       "4727                        4.0                    4.0            12000.0   \n",
       "4728                        6.0                    6.0            20400.0   \n",
       "4729                        4.0                    4.0            16800.0   \n",
       "4730                        1.0                    1.0            15600.0   \n",
       "4731                       13.0                   15.0           124200.0   \n",
       "4732                        2.0                    2.0             7200.0   \n",
       "4733                        3.0                    3.0            13200.0   \n",
       "4734                        1.0                    1.0             1500.0   \n",
       "4735                        4.0                    4.0            34800.0   \n",
       "4736                        1.0                    1.0            10200.0   \n",
       "4737                        7.0                    8.0            20400.0   \n",
       "4738                        2.0                    3.0             7200.0   \n",
       "4739                        3.0                    3.0             3900.0   \n",
       "4740                        3.0                    3.0            12000.0   \n",
       "4741                        3.0                    3.0            19800.0   \n",
       "4742                        7.0                    8.0            21600.0   \n",
       "4743                        2.0                    2.0             3600.0   \n",
       "4744                        9.0                   11.0             8400.0   \n",
       "4745                        4.0                    4.0             7200.0   \n",
       "4746                        4.0                    5.0            13800.0   \n",
       "4747                        5.0                    6.0             6000.0   \n",
       "4748                        2.0                    2.0            19800.0   \n",
       "4749                        1.0                    1.0            49200.0   \n",
       "4750                        7.0                    8.0            19200.0   \n",
       "4751                        8.0                    8.0            21000.0   \n",
       "4752                        8.0                   11.0            12000.0   \n",
       "4753                        5.0                    5.0             7800.0   \n",
       "\n",
       "      consfin_avg_limit  latest_query_day  loans_latest_day  \n",
       "0                1200.0              12.0              18.0  \n",
       "1                9360.0               4.0               2.0  \n",
       "2                4200.0               2.0               6.0  \n",
       "3               12180.0               2.0               4.0  \n",
       "4                8250.0              22.0             120.0  \n",
       "5                8130.0               3.0               4.0  \n",
       "6                8900.0               1.0               3.0  \n",
       "7                7050.0              14.0              36.0  \n",
       "8                7987.0              24.0               7.0  \n",
       "9                   0.0              18.0             142.0  \n",
       "10               7757.0              12.0               9.0  \n",
       "11               9400.0              31.0              35.0  \n",
       "12              16985.0              38.0               6.0  \n",
       "13               5460.0              13.0               9.0  \n",
       "14                  0.0              16.0             193.0  \n",
       "15              17633.0               4.0              27.0  \n",
       "16               9900.0               3.0               1.0  \n",
       "17               4062.0               9.0               7.0  \n",
       "18              10275.0              26.0              77.0  \n",
       "19               8187.0              12.0              37.0  \n",
       "20               8400.0               3.0              71.0  \n",
       "21              18600.0               7.0              23.0  \n",
       "22              20454.0               4.0               1.0  \n",
       "23               6100.0              12.0               9.0  \n",
       "24              15500.0              14.0             104.0  \n",
       "25               9841.0              21.0              82.0  \n",
       "26               7050.0              10.0              80.0  \n",
       "27               6461.0              20.0               6.0  \n",
       "28              21600.0               5.0               9.0  \n",
       "29               3000.0              14.0              34.0  \n",
       "...                 ...               ...               ...  \n",
       "4724             3460.0              21.0               5.0  \n",
       "4725            11850.0             142.0             100.0  \n",
       "4726             4800.0              19.0              53.0  \n",
       "4727             5125.0              28.0              57.0  \n",
       "4728             8933.0              18.0               9.0  \n",
       "4729             7200.0               3.0             104.0  \n",
       "4730            15600.0              28.0             128.0  \n",
       "4731            25361.0               2.0              11.0  \n",
       "4732             4950.0               9.0             250.0  \n",
       "4733             6900.0              14.0               9.0  \n",
       "4734             1500.0              24.0             131.0  \n",
       "4735            21100.0               2.0               6.0  \n",
       "4736            10200.0              18.0             156.0  \n",
       "4737            11714.0               1.0               7.0  \n",
       "4738             5700.0              48.0              89.0  \n",
       "4739             3600.0              12.0               9.0  \n",
       "4740             9200.0              31.0             163.0  \n",
       "4741            13900.0               9.0             140.0  \n",
       "4742             7557.0               4.0               4.0  \n",
       "4743             2300.0              11.0              12.0  \n",
       "4744             3344.0               5.0              88.0  \n",
       "4745             5850.0               1.0               8.0  \n",
       "4746             7050.0              14.0              36.0  \n",
       "4747             3260.0               1.0              22.0  \n",
       "4748            18300.0              35.0              54.0  \n",
       "4749            49200.0              44.0              58.0  \n",
       "4750             7500.0              17.0              55.0  \n",
       "4751             7250.0              18.0              55.0  \n",
       "4752             3950.0              55.0              16.0  \n",
       "4753             4360.0              20.0              54.0  \n",
       "\n",
       "[4754 rows x 87 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.2 划分数据集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 划分数据集（训练集和测试集7:3 无验证集）\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3,random_state=2333)\n",
    "\n",
    "# 特征归一化处理\n",
    "std = StandardScaler()\n",
    "X_train = std.fit_transform(X_train.values)\n",
    "X_test = std.transform(X_test.values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.3 模型评估"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
    "from sklearn.metrics import roc_auc_score,roc_curve, auc\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def model_metrics(clf, X_train, X_test, y_train, y_test):\n",
    "    # 预测训练集和测试集\n",
    "    y_train_pred = clf.predict(X_train)\n",
    "    y_test_pred = clf.predict(X_test)\n",
    "    \n",
    "    y_train_proba = clf.predict_proba(X_train)[:,1]\n",
    "    y_test_proba = clf.predict_proba(X_test)[:,1]\n",
    "    \n",
    "    # 准确率Accuracy\n",
    "    print('【Accuracy准确率】', end = ' ')\n",
    "    print('训练集：', '%.4f'%accuracy_score(y_train, y_train_pred), end = ' ')\n",
    "    print('测试集：', '%.4f'%accuracy_score(y_test, y_test_pred))\n",
    "    \n",
    "    # 精准率Precision\n",
    "    print('【Prediction精准率】', end = ' ')\n",
    "    print('训练集：', '%.4f'%precision_score(y_train, y_train_pred), end = ' ')\n",
    "    print('测试集：', '%.4f'%precision_score(y_test, y_test_pred))\n",
    "\n",
    "    # 召回率Recall\n",
    "    print('【Recall召回率】', end = ' ')\n",
    "    print('训练集：', '%.4f'%recall_score(y_train, y_train_pred), end = ' ')\n",
    "    print('测试集：', '%.4f'%recall_score(y_test, y_test_pred))\n",
    "    \n",
    "    # f1-score\n",
    "    print('【F1-score】', end = ' ')\n",
    "    print('训练集：', '%.4f'%f1_score(y_train, y_train_pred), end = ' ')\n",
    "    print('测试集：', '%.4f'%f1_score(y_test, y_test_pred))\n",
    "    \n",
    "    # AUC取值\n",
    "    print('【AUC取值】', end = ' ')\n",
    "    print('训练集：', '%.4f'%roc_auc_score(y_train, y_train_proba), end = ' ')\n",
    "    print('测试集：', '%.4f'%roc_auc_score(y_test, y_test_proba))\n",
    "    \n",
    "    # ROC曲线\n",
    "    fpr_train, tpr_train, thresholds_train = roc_curve(y_train, y_train_proba, pos_label = 1)\n",
    "    fpr_test, tpr_test, thresholds_test = roc_curve(y_test, y_test_proba, pos_label = 1)\n",
    "    \n",
    "    label = [\"Train - AUC:{:.4f}\".format(auc(fpr_train, tpr_train)), \n",
    "             \"Test - AUC:{:.4f}\".format(auc(fpr_test, tpr_test))]\n",
    "    plt.plot(fpr_train,tpr_train)\n",
    "    plt.plot(fpr_test,tpr_test)\n",
    "    plt.plot([0, 1], [0, 1], 'd--')\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.legend(label, loc = 4)\n",
    "    plt.title(\"ROC Curve\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4.4 导入不同分类器模型\n",
    "\n",
    "> ### 逻辑回归、SVM、决策树、随机森林、XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import svm\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from xgboost.sklearn import XGBClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ###  4.4.1 Logistic Regresssion 逻辑回归"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = LogisticRegression(C = 0.1, penalty = 'l1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.7977 测试集： 0.8094\n",
      "【Prediction精准率】 训练集： 0.7227 测试集： 0.7547\n",
      "【Recall召回率】 训练集： 0.3226 测试集： 0.3399\n",
      "【F1-score】 训练集： 0.4461 测试集： 0.4688\n",
      "【AUC取值】 训练集： 0.8023 测试集： 0.7980\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1b9540f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr.fit(X_train, y_train.values)\n",
    "model_metrics(lr, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### 4.4.2 SVM 支持向量机\n",
    "\n",
    "> #### 线性核函数 svm_linear = svm.SVC(kernel = 'linear', probability=True).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_linear = svm.SVC(C = 0.01, kernel = 'linear', probability=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.7854 测试集： 0.7919\n",
      "【Prediction精准率】 训练集： 0.7763 测试集： 0.8256\n",
      "【Recall召回率】 训练集： 0.2107 测试集： 0.2011\n",
      "【F1-score】 训练集： 0.3315 测试集： 0.3235\n",
      "【AUC取值】 训练集： 0.8042 测试集： 0.8097\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ba40a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_linear.fit(X_train, y_train.values)\n",
    "model_metrics(svm_linear, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> #### 多元线性核函数 svm_poly = svm.SVC(kernel = 'poly', probability=True).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_poly =  svm.SVC(C = 0.01, kernel = 'poly', probability=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.7532 测试集： 0.7554\n",
      "【Prediction精准率】 训练集： 0.9524 测试集： 0.7500\n",
      "【Recall召回率】 训练集： 0.0238 测试集： 0.0170\n",
      "【F1-score】 训练集： 0.0465 测试集： 0.0332\n",
      "【AUC取值】 训练集： 0.8689 测试集： 0.7220\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bbbe208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_poly.fit(X_train, y_train.values)\n",
    "model_metrics(svm_poly, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> #### 径向基函数 svm_rbf = svm.SVC(probability=True).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_rbf =  svm.SVC(gamma = 0.01, C =0.01 , probability=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.7475 测试集： 0.7526\n",
      "【Prediction精准率】 训练集： 0.0000 测试集： 0.0000\n",
      "【Recall召回率】 训练集： 0.0000 测试集： 0.0000\n",
      "【F1-score】 训练集： 0.0000 测试集： 0.0000\n",
      "【AUC取值】 训练集： 0.8603 测试集： 0.7941\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bbbe390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_rbf.fit(X_train, y_train.values)\n",
    "model_metrics(svm_rbf, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> #### 非线性sigmoid核函数 svm_sigmoid = svm.SVC(kernel = 'sigmoid',probability=True).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_sigmoid =  svm.SVC(C = 0.01, kernel = 'sigmoid',probability=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.7475 测试集： 0.7526\n",
      "【Prediction精准率】 训练集： 0.0000 测试集： 0.0000\n",
      "【Recall召回率】 训练集： 0.0000 测试集： 0.0000\n",
      "【F1-score】 训练集： 0.0000 测试集： 0.0000\n",
      "【AUC取值】 训练集： 0.7768 测试集： 0.8011\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ba39710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "svm_sigmoid.fit(X_train, y_train.values)\n",
    "model_metrics(svm_sigmoid, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### 4.4.3 Decision Tree 决策树"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "dt = DecisionTreeClassifier(max_depth=5,min_samples_split=50,min_samples_leaf=60, max_features=9, random_state =2018)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.7869 测试集： 0.7765\n",
      "【Prediction精准率】 训练集： 0.6446 测试集： 0.5885\n",
      "【Recall召回率】 训练集： 0.3476 测试集： 0.3201\n",
      "【F1-score】 训练集： 0.4517 测试集： 0.4147\n",
      "【AUC取值】 训练集： 0.7726 测试集： 0.7432\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ba97cc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dt.fit(X_train, y_train.values)\n",
    "model_metrics(dt, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### 4.4.4 Random Forests 随机森林"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = RandomForestClassifier(oob_score=True, random_state=2018)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.9823 测试集： 0.7821\n",
      "【Prediction精准率】 训练集： 0.9975 测试集： 0.6479\n",
      "【Recall召回率】 训练集： 0.9321 测试集： 0.2606\n",
      "【F1-score】 训练集： 0.9637 测试集： 0.3717\n",
      "【AUC取值】 训练集： 0.9996 测试集： 0.7090\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ba59898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rf.fit(X_train, y_train.values)\n",
    "model_metrics(rf, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### 4.4.5 XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb = XGBClassifier(learning_rate =0.1, n_estimators=80, max_depth=3, min_child_weight=5,\\\n",
    "                    gamma=0.2, subsample=0.8, colsample_bytree=0.8, reg_alpha=1e-5,\\\n",
    "                    objective= 'binary:logistic', nthread=4,scale_pos_weight=1, seed=27)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "【Accuracy准确率】 训练集： 0.8392 测试集： 0.7947\n",
      "【Prediction精准率】 训练集： 0.8308 测试集： 0.6667\n",
      "【Recall召回率】 训练集： 0.4560 测试集： 0.3399\n",
      "【F1-score】 训练集： 0.5888 测试集： 0.4503\n",
      "【AUC取值】 训练集： 0.8947 测试集： 0.7989\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xd4VGX2wPHvSQKEErrSQi/Sa+goKNhwkSAo4NqwsFhWkWWVXTUU0Z+IFcFVbGADVERQQERWxZUaBASjaOihSG+BkJCc3x8zGYeQMimTaefzPHmcO3Nn7rkh3jNvuecVVcUYY4wBCPN1AMYYY/yHJQVjjDEulhSMMca4WFIwxhjjYknBGGOMiyUFY4wxLpYUjDHGuFhSMEFFRHaIyBkROSUi+0VkhoiUy7JPNxH5r4icFJHjIvK5iDTPsk95EXlJRHY5PyvRuV01h+OKiDwoIptFJFlEkkTkYxFp5c3zNaaoWVIwwaifqpYD2gLtgH9lviAiXYGvgPlATaA+sBH4QUQaOPcpCSwDWgDXAOWBbsBhoFMOx3wZeAh4EKgMNAE+A67Lb/AiEpHf9xhTVMTuaDbBRER2AHer6tfO7WeBFqp6nXP7e2CTqt6X5X2LgYOqepuI3A08BTRU1VMeHLMx8CvQVVXX5LDPt8D7qvqmc/sOZ5w9nNsKPACMBCKAJcApVR3t9hnzge9U9QURqQm8AlwGnAJeVNUpHvyKjMmVtRRM0BKRaOBaING5XQbHN/6Ps9n9I+BK5+M+wJeeJASn3kBSTgkhH2KBzkBz4ENgsIgIgIhUAq4CZotIGPA5jhZOLefxR4rI1YU8vjGWFExQ+kxETgK7gQPAWOfzlXH8ze/L5j37gMzxgio57JOT/O6fk/9T1SOqegb4HlDgUudrg4CVqroX6AhcpKoTVDVVVbcBbwBDiiAGE+IsKZhgFKuqUUAvoCl/XuyPAhlAjWzeUwM45Hx8OId9cpLf/XOyO/OBOvp1ZwNDnU/dDHzgfFwXqCkixzJ/gH8D1YogBhPiLCmYoKWq3wEzgOec28nASuDGbHa/CcfgMsDXwNUiUtbDQy0DokUkJpd9koEybtvVsws5y/YsYJCI1MXRrTTX+fxuYLuqVnT7iVLVvh7Ga0yOLCmYYPcScKWItHVujwFud04fjRKRSiIyEegKjHfu8x6OC+9cEWkqImEiUkVE/i0iF1x4VfV34FVgloj0EpGSIhIpIkNEZIxztw3ADSJSRkQaAXflFbiqrgcOAm8CS1T1mPOlNcAJEXlUREqLSLiItBSRjgX5BRnjzpKCCWqqehB4F3jCuf0/4GrgBhzjADtxTFvt4by4o6pncQw2/wosBU7guBBXBVbncKgHganANOAYsBUYgGNAGOBFIBX4A5jJn11BeZnljOVDt3NKB/rhmHK7HUe315tABQ8/05gc2ZRUY4wxLtZSMMYY42JJwRhjjIslBWOMMS6WFIwxxrgEXOGtqlWrar169XwdhjHGBJR169YdUtWL8tov4JJCvXr1iI+P93UYxhgTUERkpyf7WfeRMcYYF0sKxhhjXCwpGGOMcbGkYIwxxsWSgjHGBIDEo4nEzo8l8WiiV4/jtaQgIm+LyAER2ZzD6yIiU5wLov8kIu29FYsxxgSy02mnuW/ZfWw7to37l93P6bTTXjuWN1sKM3Asep6Ta4HGzp/hwH+8GIsxxgSsuBVxHEk5gqIcPnOYsSvG5v2mAvLafQqqulxE6uWyS3/gXecKU6tEpKKI1FDVoljW0Bhj8vTh6l3M37DH12Gcp/fpRXQ/841r+5syZ/mmQgqpzq/wZzPO8u3ub5n3+zwGNB5Q5Mf35c1rtXBbfhBIcj53QVIQkeE4WhPUqVOnWIIzxvgHb164V28/AsC/q60670LsSy1SNwHwc8lWALznlhAypaSn8NKPLwVdUpBsnst2cQdVnQ5MB4iJibEFIIwpAH/8VuyJzAt35/qVc9wn67drj5WHquVKUe2Is0pC3R4FCbGI9YBWg6jdeiAvxL/A6d/nIgjqdnmMDI/k4Q4Pe+XovkwKSUBtt+1oYK+PYjEmIBTmwu7JxdUfda5fmf5ta3Fz5yy9BPHvwKZPHI/3/c/x34Je1KMcF2JihhU80CKUnpHOrQtuYMeJHQxrOYzdJ3bz/Z7vOZt+llJhpehVuxexjWK9cmxfJoUFwAMiMhvHouTHbTzBhAJfXdhzvLj6A/cLfHYSnD/udrolgrr+dVEvqGMpx6hQqgLhYeE82O5BqpetTouqLTiddprY+bHsT95PldJVGN9tfN4fVkBeW45TRGYBvXCsa/sHMBYoAaCqr4mI4FjT9hrgNDBMVfOsdBcTE6NWEM/4u9wu/IX9xu63F/b8ck8EOwv4TT8IEgGAqvLFti+YtHYSI9uPZFCTQRfsk3g0kdHLR/PcZc/RqFKjfB9DRNapakye+wXaGs2WFIw/yenin9eFP2gu7DnJ65s/XJgIguQCn1/7k/czYeUEvt/zPa0vas2EbhNoWLFhkR/H06QQcKWzjSlOeXX15HTx9+uuGm/KTAaefPMPki6fwli0bRETVk0gQzN4tOOjDG06lPCwcJ/GZEnBGDdZk0Be3/hD9uIP2bcG3JNBiF/wPVG+VHlaVW3F2K5jiY6K9nU4gHUfmRCWXSsguyQQshf9nOTVGrBkkKNzGed4L+E90jLSGN56OOAYT3AMsXqXdR8Z48bTBBDS3/wz5TUeYK2BAtlyZAtxK+JIOJzA1fWudiWD4kgI+WFJwQS1zGRgCSAHeXUBZceSQb6kpqfy+k+v8/amtylfqjzP93yeK+te6XfJIJMlBRMUPJkFFPIJIDubPoH9m6B6qz+fs4t+kdp5Yidvb36bvg368s+Yf1IxsqKvQ8qVJQUT8D5cvYt/z3PUi7FZQFnk1RWUmRCGLSy+mELA6bTT/Hf3f/lLg7/QuFJjFsQuoHZU7bzf6AcsKZiAktvYwNMDWoXuxT8rT6eGVm/laBWYIrNi7womrJzA3lN7aV65OQ0qNgiYhACWFIyfys9NYSHdGsipJWCDwcXu+NnjPB//PPMS51GvfD3eueYdGlRs4Ouw8s2SgvE5T2cGZW6HbALIKv4d+GKk43HWloAlg2KVnpHObYtvY+eJndzd6m5GtBlBqfBSvg6rQCwpGJ+xmUGF4J4Q/vKSXfx95GjK0T8L2LV/kBpla9C8SnNfh1UolhRMsXJvFdjMoHzKroCcJQSfUFU+3/Y5k9ZMYmSHkdzY5EZ61+nt67CKhCUFUyyyaxVYMvBATpVErXvIZ/ae2suElRP4Ye8PtL2oLR2qdfB1SEXKkoLxuqxTRi0ReCC72UOWCHzu862fM3HVRBTlX53+xZCmQwiTsLzfGEAsKRivydo6sCmjHso6gGyJwG9UjqxMu4vbEdc1jprlavo6HK+wpGAKze4mLoTcykzYeIHPpWWkMfPnmZzLOMeINiPoXqs73Wp289sSFUXBkoIpFLubuJCszITf+uXwL4xdMZZfjvzCtfWu9dsCdkXNkoLJl5zWG7CuoUKwMhN+5Wz6WV7b+BrvbH6HiqUq8mKvF+lTt4+vwyo2lhSMx7JrFVhrIAtPlqF0l7WVYHxu14ldzPh5Bv0a9mN0zGgqlKrg65CKlSUF4xH3hGCtglxk1x2UG6s95BdOp51m2a5l9GvYj8aVGvN57Od+sxJacbOkYHJlM4jyIf4dxyBx3R7WHRRAftjzA+NXjmd/8n5aVGlBg4oNQjYhgCUFk4PsbjazbqJsZHdzmX3zDwjHUo4xOX4yC7YuoH6F+sy8dmZAFrArapYUzHksGXjIbi4LaOkZ6dy6+FZ2n9zNPa3u4W9t/hawBeyKmiUFY/WIPJVTyQlLBAHjSMoRKpaqSHhYOA93eJia5WrStHJTX4flVywphDCrR5SLvNYutmQQUFSVzxI/Y3L8ZEa2H8lNl9zEFXWu8HVYfsmSQoiyekRuPF283hJBQNpzag/jV4xn5b6VtL+4PZ2qd/J1SH7NkkIIsumlbnJaqMYSQFD4fOvnPLnqSQTh8c6Pc+MlNwZdAbuiZkkhhNj00ixsoZqgVyWyCh2qdSCuSxw1ytXwdTgBwZJCiLDuIjdZZw5ZQggaaRlpvLP5HdI1nXvb3Eu3Wt3oVqubr8MKKJYUgpy1Dpxs5lDQSzicQNwPcWw5uoW+9fu6CtiZ/LGkEMSsdeCUddzAkkFQSTmXwn82/oeZP8+kUmQlXrr8paBZGtMXvJoUROQa4GUgHHhTVZ/J8nodYCZQ0bnPGFVd5M2YQoG1DtzYuEHQSzqZxLsJ79K/UX9GdRgVcgXsiprXkoKIhAPTgCuBJGCtiCxQ1QS33R4HPlLV/4hIc2ARUM9bMYWCkG0d5FSd1MYNgtKp1FN8vetrYhvF0qhSIxYOWBi0K6EVN2+2FDoBiaq6DUBEZgP9AfekoEB55+MKwF4vxhPUQrZ1kF25CXfWVRR0lict58lVT3Lg9AFaV21Ng4oNLCEUIW8mhVrAbrftJKBzln3GAV+JyN+BskC2K1mIyHBgOECdOiFwocunkGsd2KBxSDqacpRn1z7LF9u+oGGFhrx77btWwM4LvJkUshv21yzbQ4EZqvq8iHQF3hORlqqacd6bVKcD0wFiYmKyfkZIC4kb0bJ2DVm5iZCTnpHObYtvI+lkEiPajOCeVvdQMrykr8MKSt5MCklAbbftaC7sHroLuAZAVVeKSCRQFTjgxbiCRkgkBLhw4RpLBCHj0JlDVI6sTHhYOP+I+Qc1ytbgksqX+DqsoObNpLAWaCwi9YE9wBDg5iz77AJ6AzNEpBkQCRz0YkxBI+gSQm7LWGYmBFu4JmSoKvMS5/Hc2ucY2cFRwK5X7V6+DiskeC0pqOo5EXkAWIJjuunbqvqziEwA4lV1AfAP4A0ReRhH19IdqmrdQ3kIyoSQXf2hTLZkZUjZfXI341eMZ/X+1cRUi6FLjS6+DimkePU+Bec9B4uyPBfn9jgB6O7NGIJNUCUEKzdhspifOJ+nVj9FmITxRJcnGNRkkBWwK2Z2R3MACbiEkFuXENjMIXOBi8pcRKfqnXi8y+NUL1vd1+GEJEsKfi67VdH8PiHkde9AJksGIS8tPY03N7+JqnJf2/voVrMb3WpaATtfsqTg5+Zv2EPCvhM0r1E+MO5ByDo+YBd9k4PNhzbzxA9PkHgskX4N+lkBOz9hScGPfbh6F6u3H6Fz/crM+VtXX4eTN6szZDxw5twZpq2fxnu/vEfV0lV55YpXbGaRH7Gk4Kfcxw/6t63l42g8YAnBeGjPyT18+OuHDGw8kIc7PExUyShfh2TcWFLwU5njCH4/fgCWEEyeTqae5OudXzOg8QAaVWrEohsW2UCyn7Kk4Ifcu40sIZhAtzxpOeNXjufQmUO0ubgNDSo0sITgxywp+JGslU79utvI7jEweTiScoRJayaxaPsiGlVsxEu9XqJBBStg5+8sKfiBrMnA72cZ2Qwjk4f0jHRuX3w7SaeSuK/tfdzd8m5KhJfwdVjGAx4lBREpCdRR1UQvxxOSMqedBkQysNaByYV7AbvRMaOpWa4mjSs19nVYJh/yvH9cRK4DNgFLndttRWSetwMLFZnjB81rlGfO37r6d0L4YqQjIdTtYQnBnCdDM/hoy0f8Zd5f+HjLxwD0rN3TEkIA8qSlMAHH4jjfAKjqBhFp5NWoQoTfTzvNbjEbSwYmi10ndjFu5TjW7l9L5+qd6VbL7kgOZJ4khTRVPZblTkOrZFpIfl3HKLsyFTZ2YLIx7/d5PLX6KUqElWBc13Hc0PgGuys5wHmSFH4RkZuAMOfaCA8Bq7wbVvDy67WUs0sGlghMLmqUq0G3mt14rPNjVCtbzdfhmCLgSVJ4AIgDMoBPcayP8C9vBhXM/HpQOXOFM0sGJgep6am8uelNMjSDB9o9QJcaXWy9gyDjSVK4WlUfBR7NfEJEbsCRIEwBZA4q+yVb4czk4KeDPzF2xVgSjyVyfcPrrYBdkPIkKTzOhQngsWyeM7nI7DbKrHhqTKA4nXaaqRum8n7C+1xc5mKm9Z7GZdGX+Tos4yU5JgURuRq4BqglIi+4vVQeR1eS8ZD7oHJmt5HfiX/nz+mmxrjZl7yPOb/O4aZLbmJk+5GUK1nO1yEZL8qtpXAA2AykAD+7PX8SGOPNoIKN3xa3y27Kqa2FbIATqSdYumMpA5sMpGHFhiy8YaHVKwoROSYFVV0PrBeRD1Q1pRhjChruXUZ+VdzOppyaXPx313+ZuGoiR1KO0K5aOytgF2I8GVOoJSJPAc2ByMwnVbWJ16IKAn7bZWR1i0wODp85zDNrnuHLHV/SpFITXrniFStgF4I8SQozgInAc8C1wDBsTCFXfntjmpW5NjlIz0jntsW3sS95H39v93eGtRxGiTArYBeKPEkKZVR1iYg8p6pbgcdF5HtvBxao/DYhwJ/jB5YQjNOB0weoWroq4WHhPNrpUWqVq0XDig19HZbxoTwL4gFnxTEZeauIjBCRfsDFXo4rYPntoHKmuj0sIRgyNIM5v87h+s+u56MtHwFwWfRllhCMRy2Fh4FywIPAU0AF4E5vBhWI/HZQOZNNOTVOO47vYNzKcaz7Yx1danShRy37mzB/yjMpqOpq58OTwK0AIhLtzaACkfuNaX4zqOwus+vIppyGtE9//5SnVz9NyfCSTOg2gdhGsXZXsjlPrklBRDoCtYD/qeohEWmBo9zFFYAlhiz8snxF5vTTzJpG1nUU0mqWq0mPWj14rPNjXFTmIl+HY/xQbnc0/x8wENiIY3B5Ho4KqZOAEcUTXmDIXCinc/3Kvg7lfNlNPzUhJTU9ldc2vgbAg+0ftAJ2Jk+5tRT6A21U9YyIVAb2Ore3FE9ogSNzcNmvuo1s+mnI23BgA3Er4th+fDsDGg2wAnbGI7klhRRVPQOgqkdE5FdLCBdybyX4dHDZvWQF2EppIex02mmmrJ/Ch798SPWy1Xmtz2t0r9Xd12GZAJFbUmggIpmVUAWo57aNqt6Q14eLyDXAy0A48KaqPpPNPjcB43Cs5rZRVW/2PHzf85tWQua4QfVWjm27Wzlk7Uvex8dbPmZI0yE81P4hypYo6+uQTADJLSkMzLI9NT8fLCLhwDTgSiAJWCsiC1Q1wW2fxjgW7OmuqkdFJCDvf/CLVkLmdFNbCyEkHT97nK92fsWNTW6kYcWGLB64mIvLBOT/TsbHciuIt6yQn90JSFTVbQAiMhvHOEWC2z73ANNU9ajzmAcKecxi5RcDzO5jBzaQHJKW7VzGxNUTOZpylJhqMdSvUN8SgikwT+5oLqhawG637STnc+6aAE1E5AcRWeXsbrqAiAwXkXgRiT948KCXws0f93IWPus6ssHkkHbozCFGfTuKkd+OpGrpqnx43YfUr1Df12GZAOfJHc0Fld00B83m+I2BXjjue/heRFqq6rHz3qQ6HZgOEBMTk/UzfMLn5SwsIYS09Ix0bl98O/uT9/NQ+4e4vcXtVsDOFAmPk4KIlFLVs/n47CSgttt2NI5prVn3WaWqacB2EdmCI0mszcdxfMZnYwmWEELW/uT9XFzmYsLDwhnTaQy1ompZeWtTpPLsPhKRTiKyCfjdud1GRF7x4LPXAo1FpL6IlASGAAuy7PMZcLnzc6vi6E7alo/4fSJzLMEnLCGEpAzN4INfPuD6z65nzpY5AFwafaklBFPkPGkpTAH+guMCjqpuFJHL83qTqp4TkQeAJTimpL6tqj+LyAQgXlUXOF+7SkQSgHTgn6p6uIDnUmx8Ng3VEkJI2nZ8G+NWjGP9gfV0r9mdntE9fR2SCWKeJIUwVd2Z5U7IdE8+XFUXAYuyPBfn9liBUc6fgOCzm9UsIYSkub/N5enVTxMZEclTPZ6iX4N+dley8SpPksJuEekEqPPeg78Dv3k3LP/ls1aCLZATkmpH1aZn7Z78u/O/qVq6qq/DMSHAk6RwL44upDrAH8DXzudCjs9LWliV06B3Nv2sq4DdQ+0folONTnSq0cnHUZlQ4klSOKeqQ7weSQAo9laCez0j9xIWJiitP7CeuB/i2HFiBwMbD7QCdsYnPEkKa51TRecAn6rqSS/H5Jd80kpwr2dUvZXdsRykktOSefnHl5n962xqlqvJ631ep1utbr4Oy4QoT1Zeaygi3XBMKR0vIhuA2ao62+vR+Yliv3vZfWGc6q2snlGQ+yP5Dz79/VNubnYzD7Z7kDIlyvg6JBPCPCpzoaorVPVBoD1wAvjAq1H5mWK9ezlzltHO/1nrIIgdSznGnF8d9xs0qNiAxTcsZkynMZYQjM/l2VIQkXI4CtkNAZoB84GQadsWW7dRZuvA1kEIaqrK0p1LeWr1U5w4e4JONTpRv0J9WxrT+A1PxhQ2A58Dz6rq916Ox+8U2+Cy+zrKtg5CUDp4+iBPrX6KZbuW0bxKc6ZfOd0K2Bm/40lSaKCqGV6PxA8VSyvBxg9CQnpGOrd/eTsHTh9gVIdR3Nr8ViLCvFmP0piCyfGvUkSeV9V/AHNF5ILKpJ6svBbovNpKyNpdlNlCMEHFvYDdY50fo1a5WtSrUM/XYRmTo9y+qsxx/jdfK64FG6+1Eqy7KKilZ6Qze8tsXv7xZR7u8DBDmw61dZJNQMht5bU1zofNVPW8xOAsdFfYldn8mldXVbPlM4PatmPbiFsRx8aDG+lRqwe9onv5OiRjPObJlNQ7s3nurqIOxN94teso8y5l6y4KOh//9jGDPh/EzhM7ebrH07za+1VqlKvh67CM8VhuYwqDcUxDrS8in7q9FAUcy/5dwcUrXUfurQTrMgo6daPq0rtOb8Z0GkOV0lV8HY4x+ZbbmMIa4DCOFdOmuT1/EljvzaCClnv5a2slBIWUcym8uvFVBOHhDg9bATsT8HIbU9gObMdRFTWkeGU8wdZDCDrx++MZt3IcO0/s5KYmN1kBOxMUcus++k5Ve4rIUcB9SqrgWB/HCyOw/qHIxhPcq5zancpB41TqKV768SXmbJlDdLlo3rzqTTrX6OzrsIwpErl1H2UuuRmSK3sUejzBvWVQt4dNPQ0iB84cYH7ifG5rfhv3t73f6hWZoJJb91HmXcy1gb2qmioiPYDWwPs4CuMFnSLpOrKuoqBzNOUoS3YsYUjTITSo0IDFAxfbSmgmKHkyJfUzHEtxNgTexVEU70OvRuVDRdJ1ZEtnBg1V5cvtXxI7P5ZJayex4/gOAEsIJmh5UnwlQ1XTROQG4CVVnSIiQT37qFBdRzblNGgcOH2AJ1c9ybe7v6VFlRa8cdUbVqLCBD2PluMUkRuBW4FY53MlvBdSgLMb04JCekY6d3x5BwdOH2B0zGj+2uyvVsDOhARP/srvBO7DUTp7m4jUB2Z5NyzfKLKpqNZKCFh7T+2lWplqhIeF83jnx4mOiqZO+WJaftUYP5DnmIKqbgYeBOJFpCmwW1Wf8npkPlDo8YTMriMTcNIz0pn580z6f9afOVsctSC71epmCcGEHE9WXrsUeA/Yg+Meheoicquq/uDt4HyhUOMJ1nUUkH4/+jtjV4xl06FN9IzuyRV1rvB1SMb4jCfdRy8CfVU1AUBEmuFIEjHeDKy4FbrryAaYA9JHWz7i/9b8H1Eloph06SSurX+t3ZVsQponSaFkZkIAUNVfRKSkF2PyiUJ3HVkrIaBklqSoX6E+V9W9ikc7PUrlyKC9Sd8Yj3mSFH4UkddxtA4A/kqQFsQrcNeRtRICxplzZ5i2fhphYWGM6jCKjtU70rF6R1+HZYzf8OTmtRHAVuAR4FFgG/A3bwYVcKyVEBDW7l/LwAUDmZkwk9Npp1G9YJVZY0Jeri0FEWkFNATmqeqzxRNS8SuSqajWSvBbJ1NP8sK6F/jkt0+oHVWbt656y8pbG5ODHFsKIvJvHCUu/gosFZHsVmALCoUaT7BpqH7v4JmDLNy2kDta3MHc6+daQjAmF7l1H/0VaK2qNwIdgXvz++Eico2IbBGRRBEZk8t+g0RERcRnM5oKPJ5gXUd+6UjKET745QMAGlRowJcDv+QfMf+gdERpH0dmjH/LrfvorKomA6jqQRHxZPzBRUTCcazYdiWQBKwVkQXuM5mc+0XhuDludb4i97XMtRL2b7KuIz+iqizavohn1jzDqbRTdK/ZnXoV6tnMImM8lFtSaOC2NrMADd3XalbVG/L47E5AoqpuAxCR2UB/ICHLfk8CzwKj8xO4z2UmhOqtrJXgJ/Yn7+fJVU+yPGk5rau2Zny38VbAzph8yi0pDMyyPTWfn10L2O22nQSctzyViLQDaqvqFyKSY1IQkeHAcIA6dYq27EChBpmrt4JhC4s0HlMw5zLOMezLYRxOOcwjHR/h5qY3Ex4W7uuwjAk4uS2ys6yQn53dbaGuOYDO7qgXgTvy+iBVnQ5MB4iJiSnSeYQFGmR2vy/B+NSeU3uoXqY6EWERxHWNIzoqmtpRtX0dljEBK1/jBPmUhGPVtkzRwF637SigJfCtiOwAugALfDHYnO9BZhtc9rlzGeeYsXkG/T/rz+wtswHoWrOrJQRjCsmbBeLXAo2dpbb3AEOAmzNfVNXjuK3/LCLfAqNVNd6LMRWe3b3sc1uObGHsirH8fPhnLq99OVfWvdLXIRkTNDxOCiJSSlXPerq/qp4TkQeAJUA48Laq/iwiE4B4VV2Q/3D9gLUSfGr2r7OZtGYS5UuVZ3LPyVxd92orYGdMEfKkdHYn4C2gAlBHRNoAd6vq3/N6r6ouAhZleS4uh317eRKwX7BWQrHLLGDXqGIjrql/DY90fIRKkZV8HZYxQceTlsIU4C847m5GVTeKyOVejcoYp9Npp3ll/StEhEXwj5h/EFM9hpjqQVW13Ri/4slAc5iq7szyXLo3gjHG3ap9q7hhwQ28/8v7pKanWgE7Y4qBJy2F3c4uJHXepfx34DfvhlU8imxNZlOkTqSe4Pn45/n090+pW74uM66ZQYdqHXwdljEhwZOkcC+OLqQ6wB/A1xSgDpI/KvTCOsYrDp85zOLti7mz5Z3c2+ZeIiMifR2SMSEjz6SgqgdwTCcNSoVak9kUmUNnDvHl9i+5pfkt1K9QnyUDl9hAsjE+4MnsozdwuxM5k6oO90pE/szuZC5yqspEA1MsAAAfo0lEQVQX275g0tpJnE47zaXRl1K3fF1LCMb4iCfdR1+7PY4EBnB+TaPQEP8OfDHS8djuUSgS+07tY8KqCfxvz/9oc1EbJnSbQN3ydX0dljEhzZPuoznu2yLyHrDUaxEVk3wNMrsnhL+8ZPcoFIFzGecYtmQYR1KOMKbTGIZcMsQK2BnjBwpS5qI+EPBf5zweZLaEUKR2n9xNzbI1iQiLYFy3cdSOqk2tcjbQb4y/yPM+BRE5KiJHnD/HcLQS/u390Lwvz0FmSwhF5lzGOd7a9Baxn8W6Cth1qdHFEoIxfibXloI4isq0wVHQDiBDQ+kOosw6R5YQCuXXI78S90Mcvxz5hd51enNV3at8HZIxJge5JgVVVRGZp6pBdedQvsYTrM5RoXz4y4dMXjuZCqUq8EKvF6yiqTF+zpMxhTUi0l5Vf/R6NMXEblrzvswCdk0qNaFvg7480vERKpSq4OuwjDF5yDEpiEiEqp4DegD3iMhWIBnHimqqqu2LKUav8Gg8we5JyLfTaaeZsn4KERLB6I6jrYCdMQEmt5bCGqA9EFtMsfgXWzch31bsWcH4lePZl7yPm5vd7GotGGMCR25JQQBUdWsxxeJ/bDzBI8fPHmfy2snM3zqfeuXrMeOaGbSvFtANSWNCVm5J4SIRGZXTi6r6ghfi8b34dxythP2boHorX0cTEI6kHGHpzqXc3epuRrQZQanwUr4OyRhTQLklhXCgHM4WQ8hwTwjWdZSjQ2cOsWjbIm5rcZurgF3FyIq+DssYU0i5JYV9qjqh2CLxB+6Dy8MW+joav6SqLNi6gGfXPkvKuRR61u5J3fJ1LSEYEyTyHFMIKTa4nKs9p/YwYeUEVuxdQbuL2zGu2zgrYGdMkMktKfQutij8iQ0uZ+tcxjnuWnIXR1OO8ljnx7jpkpsIE09WczXGBJIck4KqHinOQHzO7kvI1q4Tu6hVrhYRYRFM6DaB6Khoapar6euwjDFeEnJf9TJLXFzAuo7Ok5aRxhs/vUHs/D8L2HWq0ckSgjFBriClswNatiUu3FsJ1nVEwuEExq4Yy69HfuWquldxdb2rfR2SMaaYhFxSgGxKXFgrweWDXz5g8trJVIqsxEu9XqJ33dAcWjImVIVkUshWiLcSMktSNK3clH4N+zE6ZrQVsDMmBFlSCHHJacm8tO4lSoaX5J8d/0mHah3oUC2oKqUbY/Ih5AaazZ/+t+d/DJg/gDlb5qAoobR+kjEme9ZSCMGpqMdSjjE5fjILti6gQYUGvHvtu7S9uK2vwzLG+IHQTgruazCH0CDzsbPHWLZrGX9r/TeGtx5OyfCSvg7JGOMnvNp9JCLXiMgWEUkUkTHZvD5KRBJE5CcRWSYiXq2ZcME9CiG0BvPB0weZsXkGqkq9CvVYMnAJD7R7wBKCMeY8XksKIhIOTAOuBZoDQ0WkeZbd1gMxqtoa+AR41lvxQA73KAT5rCNVZd7v8+j/WX+mbpjKrpO7AGxmkTEmW95sKXQCElV1m6qmArOB/u47qOo3qnraubkKiPZiPIAHy3AGkaSTSQxfOpy4FXE0qdyET/p9YgXsjDG58uaYQi1gt9t2EtA5l/3vAhZn94KIDAeGA9SpExoX9MI6l3GOu7+6m2Nnj/FElycY1GSQFbAzxuTJm0khu9Lb2c55FJFbgBigZ3avq+p0YDpATEyMzZvMxc4TO4kuF01EWARPdn+S2lG1qV62uq/DMsYECG9+dUwCarttRwN7s+4kIn2Ax4DrVfWsF+M5X+ZU1CCRlpHG6xtfZ8D8Acz6dRYAHat3tIRgjMkXb7YU1gKNRaQ+sAcYAtzsvoOItANeB65R1QNejOVCQVTv6OdDPxO3Io7fjv7GtfWu5dr61/o6JGNMgPJaUlDVcyLyALAEx3rPb6vqzyIyAYhX1QXAZBzrQH8sIgC7VPV6b8XkEkRVUd9PeJ/J8ZOpGlmVKZdP4fI6l/s6JGNMAPPqzWuqughYlOW5OLfHfbx5/Oz0Pr0Ivpji2AjgVkJmAbsWVVswoNEARsWMonzJ8r4OyxgT4ELujubuZ75xPAjQG9ZOpZ7ixXUvUjK8JI92epR2F7ej3cXtfB2WMSZIhOYcxQDtNlqetJzY+bF88vsnRIRFWAE7Y0yRC7mWQiA6mnKUSWsnsXDbQhpVbMQLvV6g9UWtfR2WMSYIWVIIACdST/Dd7u+4t8293NPqHkqEl/B1SMaYIGVJwU/9kfwHC7cvZFiLYdQtX5clg5bYQLIxxutCakyh9+lFtEjd5OswcqWqfPLbJ8TOj+U/G/7D7pOOSiGWEIwxxSGkWgqumUd+OhV194ndjFs5jjX719CxekfGdR1HnfJW68l4Ji0tjaSkJFJSUnwdivGhyMhIoqOjKVGiYN3MIZUUAH4u2YoWfjjzKLOA3fHU48R1jWNg44FWwM7kS1JSElFRUdSrVw/nzaAmxKgqhw8fJikpifr16xfoM0IuKfib7ce3UzuqNhFhEUzsMdEK2JkCS0lJsYQQ4kSEKlWqcPDgwQJ/hn0V9ZG09DT+s+E/3LDgBmb/OhuwAnam8CwhmML+DVhLwQc2HdxE3Io4Eo8l0rd+X65rcJ2vQzLGGMBaCsXuvYT3uGXxLZxIPcHUK6Yy6bJJVIqs5OuwjCm0w4cP07ZtW9q2bUv16tWpVauWazs1NdWjzxg2bBhbtmwpsphSU1OpXLkyTzzxxHnPR0dHc+zYMdf2119/TWxsrGt74cKFdOjQgebNm9O0aVMeffTRbD9/4sSJNGrUiKZNm/L1119nu89XX31Fu3btaNu2LZdeeinbtm0DYMeOHVxxxRW0bt2ayy+/nL17z19Z4Pjx49SoUYORI0cCcOzYMdfvs23btlSpUoXRo0fn/5eSF1UNqJ8OHTpoQW1+qrtufqp7gd9fGBkZGaqquv6P9Tp+xXg9cfaET+IwwSshIcHXIbiMHTtWJ0+efMHzGRkZmp6eXmxxzJ8/X7t3766NGzc+7/latWrp0aNHXdtLly7V/v37q6rqhg0btFGjRrplyxZVVU1LS9NXX331gs/euHGjtmvXTs+ePauJiYnaqFGjbM+tfv36+ttvv6mq6ssvv6x33XWXqqrGxsbq+++/r6qqS5Ys0TvuuOO899133306dOhQfeihh7I9t9atW+sPP/yQ7WvZ/S3gqE6d5zXWuo+87GTqSV5Y9wKR4ZE82ulR2l7clrYXt/V1WCbIjf/8ZxL2nijSz2xeszxj+7XI9/sSExOJjY2lR48erF69mi+++ILx48fz448/cubMGQYPHkxcnKN4co8ePZg6dSotW7akatWqjBgxgsWLF1OmTBnmz5/PxRdfnK9jz5o1i1GjRvHiiy+ydu1aOnbsmOd7Jk2axBNPPEGTJk0AiIiI4N57771gv/nz5zN06FBKlixJw4YNqVOnDuvWrbvgGCLCiROOf4vjx49Ts2ZNABISEujduzcAvXv35sYbb+Sdd94BYM2aNRw7dowrrriCzZs3X3DsX375hePHj9O1a9d8/DY8Y91HXvTt7m+J/SyWT3//lBLhJayAnQlZCQkJ3HXXXaxfv55atWrxzDPPEB8fz8aNG1m6dCkJCQkXvOf48eP07NmTjRs30rVrV95+++18HTM5OZnvvvuOvn37MnToUGbNmuXR+zZv3kyHDh2yfW3evHlMmDABgD179lC79p+LS0ZHR7Nnz54L3vPWW29x1VVXER0dzZw5c3jkkUcAaNOmDXPnzgVg7ty5nDhxguPHj5Oens4///lPJk+enGOMs2bNYsiQIV6ZWGAtBS84knKEZ9Y8w+Lti2lcqTEvX/EyLau29HVYJoQU5Bu9NzVs2PC8b9CzZs3irbfe4ty5c+zdu5eEhASaN29+3ntKly7Ntdc6VhHs0KED33//fb6OuWDBAq688koiIyO58cYbiYmJ4bnnniMsLCzbi6knF9gBAwYwYMAAgGy/5GX3GS+++CJLliwhJiaG//u//2P06NG89tprvPjiizzwwAO89dZb9OzZk+rVqxMREcErr7xC//79XS2K7MyePZuPP/44z3gLImSSwoerd9Eg5RxRkd4/5VOpp/hf0v+4r+193N3ybitgZ0Je2bJlXY9///13Xn75ZdasWUPFihW55ZZbsr0Lu2TJkq7H4eHhnDt37oJ9+vTpw6FDh+jSpQuvvfbaea/NmjWL1atXU69ePQAOHDjA8uXL6dWrF1WqVOHo0aNUrFgRgCNHjlC1alUAWrRowbp162jRIvfEGh0dze7du13bSUlJF1zI9+3bx6+//kpMTAwAgwcPdg1o16pVi3nz5gFw4sQJ5s6dS9myZVm1ahUrVqxgypQpnDp1itTUVMqWLctTTz0FwLp164iIiKBNmza5xldQIdN9NH+Do1lXtVwpr3z+/uT9vLnpTVSVOuXrsGTQEu5tc68lBGOyOHHiBFFRUZQvX559+/axZMmSAn/W119/zYYNGy5ICEePHmX16tUkJSWxY8cOduzYwZQpU1xdSL169eK9994D4Ny5c3zwwQdcfrljKdtHHnmEiRMnkpiYCEB6ejovvPDCBce+/vrrmTVrFqmpqWzdupWdO3de0O1UpUoVDh065PqspUuX0qxZMwAOHTrkam08/fTT3H333YCjFbBr1y527NjBM888w5133ulKCOBIdkOHDi3w7ywvIZMUAKIiI6gWFVmkn5mhGXy05SNi58cy/afprgJ2USWjivQ4xgSL9u3b07x5c1q2bMk999xD9+7di/wYc+fO5corrzyv/k9sbCzz5s0jLS2NcePGkZCQQJs2bWjfvj3NmjVzXWjbtWvHc889x0033USzZs1o1aqV6w5h9zGFNm3aEBsbS7Nmzejbty+vvvoqYWGOS+rVV1/NgQMHKFmyJNOnTyc2NpY2bdowe/ZsJk2aBMCyZcto0qQJTZo04ciRI4wZMybP81JVPvroI68mBQm0wc+YmBiNj4/P9/sGv76SuMP/pEWNCjBsYZHEsvPETsatGEf8H/F0rtGZsV3HUjuqdt5vNMYLfvnlF9e3UBPasvtbEJF1qhqT13tDZkyhqJ3LOMfwr4ZzMvUkE7pNILZRrJUYMMYEPEsK+bTt2DbqlK9DRFgET1/6NLWjanNxmfzNnTbGGH8VUmMKhZGansq0DdMYuGAgs351DFZ1qNbBEoIxJqhYS8EDGw9uZOwPY9l6fCv9GvSjX4N+vg7JGGO8wpJCHmb+PJPn45+nWtlqvNr7VS6NvtTXIRljjNeETFL4c33mHh7tn6EZhEkYbS5qw02X3MTI9iMpV7Kcd4M0xhgfC5kxBU/XZz6ReoK4H+J4Zs0zALS9uC2Pd3ncEoIxeSiK0tkAb7/9Nvv37y9QDGvXrkVEWLZsmeu5xMRE2rY9vwjl448/zksvvQQ45v4/++yzXHLJJbRs2ZK2bdvywQcfXPDZKSkpDBo0iEaNGtG1a1d27dp1wT4JCQnnlbeOiopi6tSpAKxfv54uXbrQqlUr+vfvz6lTpwBHee9bbrmFVq1a0axZM5599lnX573wwgu0aNGCFi1a8MorrxTod5JfIZMUwLE+M7msz7xs1zJiP4tlwdYFlC1R1grYGZMPVapUYcOGDWzYsIERI0bw8MMPu7bdS1bkpTBJYdasWfTo0cPj4ncA06ZN45tvviE+Pp7Nmzfz7bffkp6efsF+06dPp3r16iQmJnL//ffzr3/964J9mjdv7jrn+Ph4Spcu7Sprceedd/L888+zadMmrrvuOp5//nnAcQczwKZNm4iPj+eVV14hKSmJDRs2MHPmTOLj49mwYQOffvop27dvL8ivJV9CpvsoN4fPHObp1U/z1c6vaFq5KVN7T6V5leZ5v9EYf7V4DOzfVLSfWb0VXPtMgd46c+ZMpk2bRmpqKt26dWPq1KlkZGQwbNgwNmzYgKoyfPhwqlWrxoYNGxg8eDClS5dmzZo1HieUjIwM5s6dyzfffMOll15KamqqR+99+umnWblyJVFRjioEFStW5Lbbbrtgv/nz5/PMM47zv+mmm1yL3+Tkq6++olmzZkRHRwOwdetW193bV155Jf3792fs2LGICMnJyaSnp3PmzBkiIyOJiori+++/p2vXrpQuXRqAyy67jHnz5jFq1CiPfh8FFVIthZwkpyWzct9KHmz3IB9e96ElBGOK0ObNm5k3bx4rVqxgw4YNnDt3jtmzZ7Nu3ToOHTrEpk2b2Lx5M7fddhuDBw+mbdu2zJkzJ98tjOXLl9O0aVMaNGhA9+7d+fLLL/N8z9GjR0lLS6Nu3brZvv7YY4+xaNEi4PxS2SVLlqRs2bLnrd6W1ezZs88rR9G0aVMWLnRUU/j4449dxfSGDBlCREQENWrUoG7duowZM4YKFSrQqlUrvv32W44cOUJycjKLFy8+rwCft4RMS2F3RDovVz7NK0cTaVSpEftO7ePzbZ9zT6t7qFO+DksHLaVsibJ5f5AxgaCA3+i94euvv2bt2rWuSqFnzpyhdu3aXH311WzZsoWHHnqIvn37ctVVVxXqOJlrDIDjQjtr1iyuv/76HCsNiEieXcTuheg8LZUNjvGHhQsXnldIb8aMGTz00EPExcXRv39/V12mlStXEhkZyZ49ezhy5AiXXnopffr0oWXLlowaNYo+ffpQrlw52rVrR0SE9y/ZXm0piMg1IrJFRBJF5IJqTyJSSkTmOF9fLSL1vBHH6bTTTKqSzJ6IDO5bdh/v/vwusfNjeXPTm64CdpYQjPEOVeXOO+909bVv2bKFJ554gipVqvDTTz/Ro0cPpkyZwt/+9rdcP2fHjh2uAdw333zzvNfS0tKYN28ecXFx1KtXj5EjR7Jw4UKSk5NdZbLdZZbKrly5MiVKlMh20Dgr91LZqampJCcnU6FChWz3XbhwIZ07d3aV4wbHeMPSpUtZt26da8Aa4IMPPqBv376UKFGCatWq0aVLF9atWwfA8OHD+fHHH1m+fDkVKlSgcePGecZZWF5LCiISDkwDrgWaA0NFJGu/zF3AUVVtBLwITPJGLHEr4jgepqg4SlxPjp9Mm4vaMK//POqUr+ONQxpjnPr06cNHH33EoUOHAMcspV27dnHw4EFUlRtvvNG1PCdAVFQUJ0+evOBz6tWr50osmWWmM3311Vd07NiR3bt3s2PHDnbt2kW/fv1YsGABFStWpFKlSnz33Xeu43/11Veu/v0xY8Zw3333uY557Ngx3njjjQuOf/311zNz5kwAPvroo1xbNtmVtz5w4ADgGPuYOHEiI0aMAKBOnTr897//BeDUqVOsXr2aSy655Lz37Nixg/nz5zN48OAcj1lUvNkW6QQkquo2ABGZDfQH3Nfd6w+Mcz7+BJgqIqJFOO1n3u/zWJ60nDRn+lOUEmEluKb+NdQqV6uoDmOMyUGrVq0YO3Ysffr0ISMjgxIlSvDaa68RHh7OXXfdhaoiIq6S0sOGDePuu+/O10DzrFmzXCuiZRo4cCDvvPMOQ4cO5f333+f+++/n+PHjAEycONG1+M7f//53kpOT6dChAyVLlqREiRKuJTMfe+wxunfvTt++fRk+fDi33HILjRo1omrVqq5ZQ7t37+b+++9nwYIFgOPC/s0337jWW8703nvvMX36dFcivPXWWwF48MEHueOOO2jRogWqyogRI1wL/MTGxnLs2DFKlizJ66+/nmPLpCh5rXS2iAwCrlHVu53btwKdVfUBt302O/dJcm5vde5zKMtnDQeGA9SpU6fDzp07PY6j55yeHEk5csHzlSMr893g7/J9Xsb4KyudbTIVpnS2N8cUshuByZqBPNkHVZ2uqjGqGnPRRRflK4iR7UdSOqL0ec9FhkfycIeH8/U5xhgTCryZFJIA9xVnooG9Oe0jIhFABeDCr/WFMKDxAC6LvoxS4Y5lOEuFlaJX7V7ENootysMYY0xQ8GZSWAs0FpH6IlISGAIsyLLPAuB25+NBwH+Lcjwh04RuE6gcWRlBqFK6CuO7jS/qQxjjF+wufFPYvwGvJQVVPQc8ACwBfgE+UtWfRWSCiFzv3O0toIqIJAKjgLwXKS2AMiXK8GrvV2lQsQHTek+jTIky3jiMMT4VGRnJ4cOHLTGEMFXl8OHDREYWfC36kFmj2Zhgl5aWRlJSEikpKb4OxfhQZGQk0dHRrpvjMtkazcaEmBIlSlC/fn1fh2ECnNU+MsYY42JJwRhjjIslBWOMMS4BN9AsIgcBz29pPl9V4FCeewUXO+fQYOccGgpzznVVNc+7fwMuKRSGiMR7MvoeTOycQ4Odc2gojnO27iNjjDEulhSMMca4hFpSmO7rAHzAzjk02DmHBq+fc0iNKRhjjMldqLUUjDHG5MKSgjHGGJegTAoico2IbBGRRBG5oPKqiJQSkTnO11eLSL3ij7JoeXDOo0QkQUR+EpFlIlLXF3EWpbzO2W2/QSKiIhLw0xc9OWcRucn5b/2ziHxY3DEWNQ/+tuuIyDcist75993XF3EWFRF5W0QOOFemzO51EZEpzt/HTyLSvkgDUNWg+gHCga1AA6AksBFonmWf+4DXnI+HAHN8HXcxnPPlQBnn43tD4Zyd+0UBy4FVQIyv4y6Gf+fGwHqgknP7Yl/HXQznPB241/m4ObDD13EX8pwvA9oDm3N4vS+wGMfKlV2A1UV5/GBsKXQCElV1m6qmArOB/ln26Q/MdD7+BOgtItktDRoo8jxnVf1GVU87N1fhWAkvkHny7wzwJPAsEAz1pD0553uAaap6FEBVDxRzjEXNk3NWoLzzcQUuXOExoKjqcnJfgbI/8K46rAIqikiNojp+MCaFWsBut+0k53PZ7qOOxYCOA1WKJTrv8OSc3d2F45tGIMvznEWkHVBbVb8ozsC8yJN/5yZAExH5QURWicg1xRadd3hyzuOAW0QkCVgE/L14QvOZ/P7/ni/BuJ5Cdt/4s8679WSfQOLx+YjILUAM0NOrEXlfrucsImHAi8AdxRVQMfDk3zkCRxdSLxytwe9FpKWqHvNybN7iyTkPBWao6vMi0hV4z3nOGd4Pzye8ev0KxpZCElDbbTuaC5uTrn1EJAJHkzO35pq/8+ScEZE+wGPA9ap6tphi85a8zjkKaAl8KyI7cPS9LgjwwWZP/7bnq2qaqm4HtuBIEoHKk3O+C/gIQFVXApE4CscFK4/+fy+oYEwKa4HGIlJfREriGEhekGWfBcDtzseDgP+qcwQnQOV5zs6ulNdxJIRA72eGPM5ZVY+ralVVraeq9XCMo1yvqoG8lqsnf9uf4ZhUgIhUxdGdtK1YoyxanpzzLqA3gIg0w5EUDhZrlMVrAXCbcxZSF+C4qu4rqg8Puu4jVT0nIg8AS3DMXHhbVX8WkQlAvKouAN7C0cRMxNFCGOK7iAvPw3OeDJQDPnaOqe9S1et9FnQheXjOQcXDc14CXCUiCUA68E9VPey7qAvHw3P+B/CGiDyMoxvljkD+kicis3B0/1V1jpOMBUoAqOprOMZN+gKJwGlgWJEeP4B/d8YYY4pYMHYfGWOMKSBLCsYYY1wsKRhjjHGxpGCMMcbFkoIxxhgXSwrG74hIuohscPupl8u+9XKqJpnPY37rrMS50Vki4pICfMYIEbnN+fgOEanp9tqbItK8iONcKyJtPXjPSBEpU9hjm9BgScH4ozOq2tbtZ0cxHfevqtoGR7HEyfl9s6q+pqrvOjfvAGq6vXa3qiYUSZR/xvkqnsU5ErCkYDxiScEEBGeL4HsR+dH50y2bfVqIyBpn6+InEWnsfP4Wt+dfF5HwPA63HGjkfG9vZ53+Tc4696Wczz8jf65P8ZzzuXEiMlpEBuGoL/WB85ilnd/wY0TkXhF51i3mO0TklQLGuRK3Qmgi8h8RiRfHOgrjnc89iCM5fSMi3zifu0pEVjp/jx+LSLk8jmNCiCUF449Ku3UdzXM+dwC4UlXbA4OBKdm8bwTwsqq2xXFRTnKWPRgMdHc+nw78NY/j9wM2iUgkMAMYrKqtcFQAuFdEKgMDgBaq2hqY6P5mVf0EiMfxjb6tqp5xe/kT4Aa37cHAnALGeQ2OshaZHlPVGKA10FNEWqvqFBx1cS5X1cudpS8eB/o4f5fxwKg8jmNCSNCVuTBB4YzzwuiuBDDV2YeejqOmT1YrgcdEJBr4VFV/F5HeQAdgrbO8R2kcCSY7H4jIGWAHjvLLlwDbVfU35+szgfuBqTjWZ3hTRBYCHpfmVtWDIrLNWbPmd+cxfnB+bn7iLIuj7IP7qls3ichwHP9f18Cx4MxPWd7bxfn8D87jlMTxezMGsKRgAsfDwB9AGxwt3AsWzVHVD0VkNXAdsERE7sZRZnimqv7Lg2P81b1gnohku8aGsx5PJxxF2IYADwBX5ONc5gA3Ab8C81RVxXGF9jhOHCuQPQNMA24QkfrAaKCjqh4VkRk4CsNlJcBSVR2aj3hNCLHuIxMoKgD7nDXyb8XxLfk8ItIA2ObsMlmAoxtlGTBIRC527lNZPF+f+legnog0cm7fCnzn7IOvoKqLcAziZjcD6CSO8t3Z+RSIxbEOwBznc/mKU1XTcHQDdXF2PZUHkoHjIlINuDaHWFYB3TPPSUTKiEh2rS4ToiwpmEDxKnC7iKzC0XWUnM0+g4HNIrIBaIpjycIEHBfPr0TkJ2Apjq6VPKlqCo4KlB+LyCYgA3gNxwX2C+fnfYejFZPVDOC1zIHmLJ97FEgA6qrqGudz+Y7TOVbxPDBaVTfiWJv5Z+BtHF1SmaYDi0XkG1U9iGNm1CzncVbh+F0ZA1iVVGOMMW6spWCMMcbFkoIxxhgXSwrGGGNcLCkYY4xxsaRgjDHGxZKCMcYYF0sKxhhjXP4fvIA8PGYWOLQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x18a3a5f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb.fit(X_train, y_train.values)\n",
    "model_metrics(xgb, X_train, X_test, y_train, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion\n",
    "\n",
    "> ### Discussion \n",
    "\n",
    "> #### SVM：RBF vs. Sigmoid中的查准率（精准率）、查全率（召回率）和F1-Score皆为0，原因出在哪里？？\n",
    "\n",
    "|模型|参数|准确率|精确率|召回率|F1-Score|AUC取值|ROC曲线|\n",
    "|:---|:---|:---|:---|:---|:---|:---||:---|\n",
    "|LR逻辑回归|LogisticRegression(C = 0.1, penalty = 'l1')|训练集： 0.7977 测试集： 0.8094|训练集： 0.7227 测试集： 0.7547|训练集： 0.3226 测试集： 0.3399|训练集： 0.4461 测试集： 0.4688|训练集：0.7978 测试集：0.8031|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-LR.png\">|\n",
    "|svm_linear|svm.SVC(C = 0.01, kernel = 'linear', probability=True)|训练集： 0.7839 测试集： 0.7898|训练集： 0.7713 测试集： 0.8046|训练集： 0.2048 测试集： 0.1983|训练集： 0.3236 测试集： 0.3182|训练集： 0.8023 测试集： 0.7980|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-LINEAR.png\">|\n",
    "|svm_poly|svm.SVC(C = 0.01, kernel = 'poly', probability=True)|训练集： 0.7532 测试集： 0.7554|训练集： 0.9524 测试集： 0.7500|训练集： 0.0238 测试集： 0.0170|训练集： 0.0465 测试集： 0.0332|训练集： 0.8689 测试集： 0.7220|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-POLY.png\">|\n",
    "|svm_rbf|svm.SVC(gamma = 0.01, C =0.01 , probability=True)|训练集： 0.7475 测试集： 0.7526|训练集： 0.0000 测试集： 0.0000|训练集： 0.0000 测试集： 0.0000|训练集： 0.0000 测试集： 0.0000|训练集： 0.8603 测试集： 0.7941|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-RBF.png\">|\n",
    "|svm_sigmoid|svm.SVC(C = 0.01, kernel = 'sigmoid',probability=True)|训练集： 0.7475 测试集： 0.7526|训练集： 0.0000 测试集： 0.0000|训练集： 0.0000 测试集： 0.0000|训练集： 0.0000 测试集： 0.0000|训练集： 0.7768 测试集： 0.8011|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-SIGMOID.png\">|\n",
    "|决策树|DecisionTreeClassifier(max_depth=11,min_samples_split=550,min_samples_leaf=80,max_features=19)|训练集： 0.7869 测试集： 0.7765|训练集： 0.6446 测试集： 0.5885|训练集： 0.3476 测试集： 0.3201|训练集： 0.4517 测试集： 0.4147|训练集： 0.7726 测试集： 0.7432|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-DT.png\">|\n",
    "|随机森林|RandomForestClassifier(oob_score=True, random_state=2018)|训练集： 0.9823 测试集： 0.7821|训练集： 0.9975 测试集： 0.6479|训练集： 0.9321 测试集： 0.2606|训练集： 0.9637 测试集： 0.3717|训练集： 0.9996 测试集： 0.7090|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-RF.png\">|\n",
    "|XGBoost|XGBClassifier(learning_rate =0.01, n_estimators=180, max_depth=3, min_child_weight=5,gamma=0.4, subsample=0.5, colsample_bytree=0.9, reg_alpha=1, objective= 'binary:logistic',nthread=4,scale_pos_weight=1, seed=27)|训练集： 0.8392 测试集： 0.7947|训练集： 0.8308 测试集： 0.6667|训练集： 0.4560 测试集： 0.3399|训练集： 0.5888 测试集： 0.4503|训练集： 0.8947 测试集： 0.7989|<img align=\"right\" style=\"padding-right:1px;\" src=\"ROC-XGBOOST.png\">|\n",
    "\n",
    "> #### 利用不同的数据挖掘算法，构建不同的模型，通过对比模型评估之间的差异，可以进一步利用考虑模型调参和模型融合的方法改善模型性能"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}