FeatureEngineering.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "#Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import graphlab\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##Load celtra data (10% random subsample of the large data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2016-05-30 12:10:06,253 [INFO] graphlab.cython.cy_server, 176: GraphLab Create v1.9 started. Logging: C:\\Users\\TimJ\\AppData\\Local\\Temp\\graphlab_server_1464603005.log.0\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre>Read 69661 lines. Lines per second: 93000</pre>"
      ],
      "text/plain": [
       "Read 69661 lines. Lines per second: 93000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Finished parsing file E:\\celtraMiningFrik\\ccdm_mine.csv</pre>"
      ],
      "text/plain": [
       "Finished parsing file E:\\celtraMiningFrik\\ccdm_mine.csv"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Parsing completed. Parsed 245377 lines in 2.36714 secs.</pre>"
      ],
      "text/plain": [
       "Parsing completed. Parsed 245377 lines in 2.36714 secs."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "This non-commercial license of GraphLab Create is assigned to tim.jerman@gmail.com and will expire on December 31, 2016. For commercial licensing options, visit https://dato.com/buy/.\n"
     ]
    }
   ],
   "source": [
    "fileName = loadPath + 'ccdm_large.tsv'\n",
    "sep = '\\t'\n",
    "\n",
    "data = graphlab.SFrame.read_csv(fileName,sep,column_type_hints=str);\n",
    "\n",
    "#perct = 95.25\n",
    "#data = prepareData(train_data, perct)\n",
    "\n",
    "data,_ = data.random_split(0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Convert numeric types from string to int/float"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data['ADLOADINGTIME']=data['ADLOADINGTIME'].astype(float)\n",
    "data['GEOIP_LAT'] = data['GEOIP_LAT'].apply(lambda x: 0 if x=='null' else x)\n",
    "data['GEOIP_LAT']=data['GEOIP_LAT'].astype(float)\n",
    "data['GEOIP_LNG'] = data['GEOIP_LNG'].apply(lambda x: 0 if x=='null' else x)\n",
    "data['GEOIP_LNG']=data['GEOIP_LNG'].astype(float)\n",
    "data['TIMESTAMP']=data['TIMESTAMP'].astype(int)\n",
    "data['HOSTWINDOWHEIGHT'] = data['HOSTWINDOWHEIGHT'].apply(lambda x: 0 if x=='null' else x)\n",
    "data['HOSTWINDOWHEIGHT']=data['HOSTWINDOWHEIGHT'].astype(int)\n",
    "data['HOSTWINDOWWIDTH'] = data['HOSTWINDOWWIDTH'].apply(lambda x: 0 if x=='null' else x)\n",
    "data['HOSTWINDOWWIDTH']=data['HOSTWINDOWWIDTH'].astype(int)\n",
    "data['TOPMOSTREACHABLEWINDOWHEIGHT'] = data['TOPMOSTREACHABLEWINDOWHEIGHT'].apply(lambda x: 0 if x=='null' else x)\n",
    "data['TOPMOSTREACHABLEWINDOWHEIGHT']=data['TOPMOSTREACHABLEWINDOWHEIGHT'].astype(int)\n",
    "data['TOPMOSTREACHABLEWINDOWWIDTH'] = data['TOPMOSTREACHABLEWINDOWWIDTH'].apply(lambda x: 0 if x=='null' else x)\n",
    "data['TOPMOSTREACHABLEWINDOWWIDTH']=data['TOPMOSTREACHABLEWINDOWWIDTH'].astype(int)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#####remove EXTERNALCREATIVEID as it is always null"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data.remove_column('EXTERNALCREATIVEID');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#####remove NETWORKTYPE as it is mainly null"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data.remove_column('NETWORKTYPE');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "remove top 5% of ADLOADINGTIME outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18.7192\n",
      "3.69722294021\n"
     ]
    }
   ],
   "source": [
    "# 3.7835\n",
    "perct = 95\n",
    "print(np.percentile(data['ADLOADINGTIME'],perct))\n",
    "data = data[data['ADLOADINGTIME']<np.percentile(data['ADLOADINGTIME'],perct)]\n",
    "print(graphlab.evaluation.rmse(data['ADLOADINGTIME'], graphlab.SArray.from_const(data['ADLOADINGTIME'].mean(),len(data['ADLOADINGTIME']))))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "#####show data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Canvas is accessible via web browser at the URL: http://localhost:31720/index.html\n",
      "Opening Canvas in default web browser.\n"
     ]
    }
   ],
   "source": [
    "data.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "source": [
    "##Feature engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "combine TOPMOSTREACHABLEWINDOWHEIGHT and TOPMOSTREACHABLEWINDOWWIDTH by multiplication and computation of the ratio\n",
    "\n",
    "remove data with very large or negative data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data['TOPMOSTREACHABLEWINDOWAREA'] = data['TOPMOSTREACHABLEWINDOWWIDTH']*data['TOPMOSTREACHABLEWINDOWHEIGHT'];\n",
    "data = data[data['TOPMOSTREACHABLEWINDOWAREA']>=0]\n",
    "data = data[data['TOPMOSTREACHABLEWINDOWAREA']<1e7]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "tmp = data[data['TOPMOSTREACHABLEWINDOWAREA']>1e7]\n",
    "print(tmp.num_rows())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "remove all other screen features data except the area"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 250,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_column('TOPMOSTREACHABLEWINDOWWIDTH');\n",
    "data.remove_column('TOPMOSTREACHABLEWINDOWHEIGHT');\n",
    "data.remove_column('HOSTWINDOWWIDTH');\n",
    "data.remove_column('HOSTWINDOWHEIGHT');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "remove device orientation as it should not affect loading time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_column('DEVICEORIENTATION');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "UA_MOBILEDEVICE can be assesed from PLATFORM -> remove"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_column('UA_MOBILEDEVICE');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = data[data['TIMESTAMP']>1458500000]\n",
    "data = data[data['TIMESTAMP']<1461000000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "tmp = data[data['TIMESTAMP']<=1458500000]\n",
    "print(tmp.num_rows())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "GEOIP: to start combine COUNTRY with REGION for USA\n",
    "In my expectation the data is to scarce to have enough entries for exact geolocation and time assignment\n",
    "-> remove DMACODE, METROCODE, TIMEZONE, CITY, AREACODE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data.remove_column('GEOIP_TIMEZONE');\n",
    "data.remove_column('GEOIP_METROCODE');\n",
    "data.remove_column('GEOIP_AREACODE');\n",
    "data.remove_column('GEOIP_CITY');\n",
    "data.remove_column('GEOIP_DMACODE');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data['GEOIP_COUNTRY'] = data['GEOIP_COUNTRY'] + '_' + data['GEOIP_REGION'] \n",
    "data['GEOIP_COUNTRY'] = data['GEOIP_COUNTRY'].apply(lambda x: x if 'United States' in x.split('_')[0] else x.split('_')[0])\n",
    "data.remove_column('GEOIP_REGION');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID: 3.4685\n",
      "COMBINEDID and GEOIP_LAT: 3.4681\n",
      "COMBINEDID and GEOIP_LNG: 3.4684\n",
      "COMBINEDID and GEOIP_LNG and GEOIP_LAT: 3.4680\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID',rms))\n",
    "#LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'GEOIP_COUNTRY'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "#model = LR_model\n",
    "#rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "#print(\"{}: {:.4f}\".format('COMBINEDID and GEOIP_COUNTRY',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'GEOIP_LAT'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and GEOIP_LAT',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'GEOIP_LNG'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and GEOIP_LNG',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'GEOIP_LNG', 'GEOIP_LAT'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and GEOIP_LNG and GEOIP_LAT',rms))\n",
    "#LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'GEOIP_LNG', 'GEOIP_LAT','GEOIP_COUNTRY'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "#model = LR_model\n",
    "#rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "#print(\"{}: {:.4f}\".format('COMBINEDID and GEOIP_LNG and GEOIP_LAT and GEOIP_COUNTRY',rms))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "ename": "KeyError",
     "evalue": "'Cannot find column GEOIP_COUNTRY'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-102-5b43cae7a4a2>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[0mdata\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mremove_columns\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'GEOIP_LNG'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'GEOIP_LAT'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m;\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m \u001b[0mdata\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mremove_columns\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'GEOIP_COUNTRY'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m;\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32mC:\\Users\\Tim\\AppData\\Local\\Dato\\Dato Launcher\\lib\\site-packages\\graphlab\\data_structures\\sframe.pyc\u001b[0m in \u001b[0;36mremove_columns\u001b[1;34m(self, column_names)\u001b[0m\n\u001b[0;32m   3876\u001b[0m         \u001b[1;32mfor\u001b[0m \u001b[0mname\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mcolumn_names\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3877\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mexisting_columns\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 3878\u001b[1;33m                 \u001b[1;32mraise\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'Cannot find column %s'\u001b[0m \u001b[1;33m%\u001b[0m \u001b[0mname\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   3879\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3880\u001b[0m         \u001b[1;31m# Delete it going backwards so we don't invalidate indices\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyError\u001b[0m: 'Cannot find column GEOIP_COUNTRY'"
     ]
    }
   ],
   "source": [
    "data.remove_columns(['GEOIP_LNG','GEOIP_LAT']);\n",
    "data.remove_columns(['GEOIP_COUNTRY']);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " **CONCLUSIONS**\n",
    "\n",
    "GEOIP_DATA just barely results in a reduction of RMSE -> not needed"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data['ERRORSJSON'] = data['ERRORSJSON'].apply(lambda x: x.replace('\"',''))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "crate new feature by extracting sizes of images from FILESJSON"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def extractImageSize(sIn):\n",
    "    sIn = sIn.replace('[','')\n",
    "    sIn = sIn.replace('}','')\n",
    "    sIn = sIn.replace(']','')\n",
    "    sIn = sIn.replace('{\"type\":\"image\",\"size\":','')\n",
    "    numberList = [int(s) for s in sIn.split(',') if s.isdigit()]\n",
    "    if len(numberList) == 0:\n",
    "        numberList = 0\n",
    "    return np.sum(numberList)\n",
    "\n",
    "data['FILESJSON_SIZE'] = data['FILESJSON'].apply(extractImageSize)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "remove outliers from FILESJSON_SIZE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#tmp = data[data['FILESJSON_SIZE']>np.percentile(data['FILESJSON_SIZE'],99)]\n",
    "#print(np.percentile(data['FILESJSON_SIZE'],99))\n",
    "#print(tmp.num_rows())\n",
    "data = data[data['FILESJSON_SIZE']<np.percentile(data['FILESJSON_SIZE'],99)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The new feature is not very helpful"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_column('FILESJSON_SIZE');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "FILESJSON_SIZE is useful to remove some outliers but not as a features -> remove after removal of outliers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#####Create new fetures by combining IDS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 257,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data['COMBINEDID'] = data['SDK']+'_'+data['ACCOUNTID']+'_'+data['CAMPAIGNID']+'_'+data['CREATIVEID']+'_'+data['EXTERNALADSERVER']\n",
    "data['COMBINEDIDWITHPLACEMENTID'] = data['SDK']+'_'+data['ACCOUNTID']+'_'+data['CAMPAIGNID']+'_'+data['CREATIVEID']+'_'+data['EXTERNALADSERVER']+'_'+data['PLACEMENTID']\n",
    "data['COMBINEDEXTERNALID'] = data['EXTERNALPLACEMENTID']+'_'+data['EXTERNALSITEID']+'_'+data['EXTERNALSUPPLIERID']\n",
    "data['COMBINEDEXTERNALID'] = data['COMBINEDEXTERNALID'].apply(lambda x : 'null' if x == 'null_null' or x== 'null_null_null' else x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "test with simple linear regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID: 3.4685\n",
      "COMBINEDID and PLACEMENTID: 3.4629\n",
      "COMBINEDIDWITHPLACEMENTID: 3.4630\n",
      "CREATIVEID: 3.4818\n",
      "CREATIVEID and PLACEMENTID: 3.4705\n",
      "PLACEMENTID: 3.4748\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'PLACEMENTID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and PLACEMENTID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDIDWITHPLACEMENTID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDIDWITHPLACEMENTID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['CREATIVEID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('CREATIVEID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['CREATIVEID', 'PLACEMENTID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('CREATIVEID and PLACEMENTID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['PLACEMENTID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('PLACEMENTID',rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "use COMBINEDID and PLACEMENTID as separate features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID: 3.4685\n",
      "EXTERNALPLACEMENTID and EXTERNALSITEID and EXTERNALSUPPLIERID: 3.5237\n",
      "COMBINEDIDWITHPLACEMENTID: 3.4630\n",
      "COMBINEDEXTERNALID: 3.5291\n",
      "COMBINEDIDWITHPLACEMENTID and COMBINEDEXTERNALID2: 3.4659\n",
      "COMBINEDEXTERNALID2: 3.5291\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['EXTERNALPLACEMENTID', 'EXTERNALSITEID', 'EXTERNALSUPPLIERID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('EXTERNALPLACEMENTID and EXTERNALSITEID and EXTERNALSUPPLIERID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDIDWITHPLACEMENTID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDIDWITHPLACEMENTID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDEXTERNALID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDEXTERNALID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDIDWITHPLACEMENTID', 'COMBINEDEXTERNALID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDIDWITHPLACEMENTID and COMBINEDEXTERNALID',rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "do not use any of the following external features: EXTERNALPLACEMENTID, EXTERNALSITEID, EXTERNALSUPPLIERID"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "######Remove features based on above two conlclusions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data.remove_column('SDK');\n",
    "data.remove_column('ACCOUNTID');\n",
    "data.remove_column('CAMPAIGNID');\n",
    "data.remove_column('CREATIVEID');\n",
    "data.remove_column('EXTERNALADSERVER');\n",
    "data.remove_column('COMBINEDIDWITHPLACEMENTID');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data['COMBINEDID'] = data['SDK']+'_'+data['ACCOUNTID']+'_'+data['CAMPAIGNID']+'_'+data['CREATIVEID']+'_'+data['EXTERNALADSERVER']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_column('EXTERNALPLACEMENTID');\n",
    "data.remove_column('EXTERNALSITEID');\n",
    "data.remove_column('EXTERNALSUPPLIERID');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "######Test the need for COMBINEDEXTERNALID"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID: 3.4685\n",
      "COMBINEDEXTERNALID: 3.5291\n",
      "COMBINEDID and COMBINEDEXTERNALID: 3.4701\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDEXTERNALID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDEXTERNALID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'COMBINEDEXTERNALID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and COMBINEDEXTERNALID',rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "COMBINEDEXTERNALID doesn't help with the linear regression, however, it is essential for random forest -> do not remove it!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#data.remove_column('COMBINEDEXTERNALID');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "######Test the need for TOPMOSTREACHABLEWINDOWAREA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID: 3.4685\n",
      "TOPMOSTREACHABLEWINDOWAREA: 3.6797\n",
      "COMBINEDID and TOPMOSTREACHABLEWINDOWAREA: 3.4645\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['TOPMOSTREACHABLEWINDOWAREA'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('TOPMOSTREACHABLEWINDOWAREA',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'TOPMOSTREACHABLEWINDOWAREA'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and TOPMOSTREACHABLEWINDOWAREA',rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "TOPMOSTREACHABLEWINDOWAREA can help sometimes, especially for RF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#data.remove_column('TOPMOSTREACHABLEWINDOWAREA');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "######Combine PLATFORM and PLATFORMVERSION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data['PLATFORMCOMBINED'] = data['PLATFORM']+'_'+data['PLATFORMVERSION']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PLATFORM: 3.6262\n",
      "PLATFORMVERSION: 3.5990\n",
      "PLATFORMCOMBINED: 3.5979\n",
      "PLATFORM and PLATFORMVERSION: 3.5982\n",
      "COMBINEDID and PLATFORM: 3.3705\n",
      "COMBINEDID and PLATFORMVERSION: 3.3455\n",
      "COMBINEDID and PLATFORMCOMBINED: 3.3448\n",
      "COMBINEDID and PLATFORM and PLATFORMVERSION: 3.3451\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['PLATFORM'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('PLATFORM',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['PLATFORMVERSION'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('PLATFORMVERSION',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['PLATFORMCOMBINED'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('PLATFORMCOMBINED',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['PLATFORM', 'PLATFORMVERSION'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('PLATFORM and PLATFORMVERSION',rms))\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'PLATFORM'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and PLATFORM',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID','PLATFORMVERSION'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and PLATFORMVERSION',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID','PLATFORMCOMBINED'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and PLATFORMCOMBINED',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'PLATFORM', 'PLATFORMVERSION'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and PLATFORM and PLATFORMVERSION',rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "PLATFORMCOMBINED is better than separately use PLATFORM and PLATFORMVERSION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_columns(['PLATFORM','PLATFORMVERSION']);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "UA_PLATFORM and PLATFORM, and UA_PLATFORMVERSION and PLATFORMVERSION are practically the same"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 262,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_columns(['UA_PLATFORM','UA_PLATFORMVERSION']);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID: 3.4685\n",
      "COMBINEDID and PLATFORMCOMBINED: 3.3448\n",
      "COMBINEDID: 3.4685\n",
      "COMBINEDID and PLATFORMCOMBINED: 3.3449\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID','PLATFORMCOMBINED'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and PLATFORMCOMBINED',rms))\n",
    "\n",
    "train_data = train_data[train_data['PLATFORMCOMBINED'] != 'null_null']\n",
    "#test_data = test_data[test_data['PLATFORMCOMBINED'] != 'null_null']\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID',rms))\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID','PLATFORMCOMBINED'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('COMBINEDID and PLATFORMCOMBINED',rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "No need to remove 'null_null' from PLATFORMCOMBINED"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#####Device type from UA"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "COMBINEDID: 3.4685\n",
    "COMBINEDID and UA_VENDOR: 3.3694\n",
    "COMBINEDID and UA_MODEL: 3.3565\n",
    "COMBINEDID and UA_OS: 3.3693\n",
    "COMBINEDID and UA_OSVERSION: 3.3419\n",
    "COMBINEDID and UA_BROWSER: 3.3737\n",
    "COMBINEDID and UA_BROWSERVERSION: 3.3385"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Try combining this features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data['UA_VENDORMODEL'] = data['UA_VENDOR']+'_'+data['UA_MODEL']\n",
    "data['UA_OSCOMB'] = data['UA_OS']+'_'+data['UA_OSVERSION']\n",
    "data['UA_BROWSERCOMB'] = data['UA_BROWSER']+'_'+data['UA_BROWSERVERSION']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID and UA_VENDOR: 3.3694\n",
      "COMBINEDID and UA_MODEL: 3.3565\n",
      "COMBINEDID and UA_OS: 3.3693\n",
      "COMBINEDID and UA_OSVERSION: 3.3419\n",
      "COMBINEDID and UA_BROWSER: 3.3737\n",
      "COMBINEDID and UA_BROWSERVERSION: 3.3385\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "def eval_LR_features(feat, name):\n",
    "    LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=feat,validation_set=test_data, max_iterations=100,verbose=False)\n",
    "    model = LR_model\n",
    "    rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "    print(\"{}: {:.4f}\".format(name,rms))\n",
    "\n",
    "eval_LR_features(['COMBINEDID', 'UA_VENDOR'], 'COMBINEDID and UA_VENDOR')\n",
    "eval_LR_features(['COMBINEDID', 'UA_MODEL'], 'COMBINEDID and UA_MODEL')\n",
    "eval_LR_features(['COMBINEDID', 'UA_OS'], 'COMBINEDID and UA_OS')\n",
    "eval_LR_features(['COMBINEDID', 'UA_OSVERSION'], 'COMBINEDID and UA_OSVERSION')\n",
    "eval_LR_features(['COMBINEDID', 'UA_BROWSER'], 'COMBINEDID and UA_BROWSER')\n",
    "eval_LR_features(['COMBINEDID', 'UA_BROWSERVERSION'], 'COMBINEDID and UA_BROWSERVERSION')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID and UA_VENDOR and UA_MODEL: 3.3554\n",
      "COMBINEDID and UA_OS and UA_OSVERSION: 3.3415\n",
      "COMBINEDID and UA_BROWSERVERSION and UA_BROWSER: 3.3359\n"
     ]
    }
   ],
   "source": [
    "eval_LR_features(['COMBINEDID', 'UA_VENDOR', 'UA_MODEL'], 'COMBINEDID and UA_VENDOR and UA_MODEL')\n",
    "eval_LR_features(['COMBINEDID', 'UA_OS', 'UA_OSVERSION'], 'COMBINEDID and UA_OS and UA_OSVERSION')\n",
    "eval_LR_features(['COMBINEDID', 'UA_BROWSERVERSION', 'UA_BROWSER'], 'COMBINEDID and UA_BROWSERVERSION and UA_BROWSER')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID and UA_VENDORMODEL: 3.3567\n",
      "COMBINEDID and UA_OSCOMB: 3.3412\n",
      "COMBINEDID and UA_BROWSERCOMB: 3.3361\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "eval_LR_features(['COMBINEDID', 'UA_VENDORMODEL'], 'COMBINEDID and UA_VENDORMODEL')\n",
    "eval_LR_features(['COMBINEDID', 'UA_OSCOMB'], 'COMBINEDID and UA_OSCOMB')\n",
    "eval_LR_features(['COMBINEDID', 'UA_BROWSERCOMB'], 'COMBINEDID and UA_BROWSERCOMB')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "Do not combine UA_BROWSERVERSION and UA_BROWSER\n",
    "\n",
    "Do not combine UA_VENDOR and UA_MODEL\n",
    "\n",
    "Combine UA_OS and UA_OSVERSION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_columns(['UA_BROWSERCOMB', 'UA_VENDORMODEL']);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 265,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_columns(['UA_OS', 'UA_OSVERSION']);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### New feature: waht if INTENDEDDEVICETYPE and ACTUALDEVICETYPE differs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data['INTENDENTISACTUALDEVICETYPE'] = data['INTENDEDDEVICETYPE']==data['ACTUALDEVICETYPE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "COMBINEDID: 3.4749\n",
      "COMBINEDID and INTENDEDDEVICETYPE: 3.4749\n",
      "COMBINEDID and ACTUALDEVICETYPE: 3.4720\n",
      "COMBINEDID and INTENDEDDEVICETYPE and ACTUALDEVICETYPE: 3.4720\n",
      "COMBINEDID and INTENDENTISACTUALDEVICETYPE: 3.4733\n",
      "COMBINEDID and INTENDENTISACTUALDEVICETYPE and ACTUALDEVICETYPE: 3.4718\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "def eval_LR_features(feat, name):\n",
    "    LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=feat,validation_set=test_data, max_iterations=100,verbose=False)\n",
    "    model = LR_model\n",
    "    rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "    print(\"{}: {:.4f}\".format(name,rms))\n",
    "\n",
    "eval_LR_features(['COMBINEDID'], 'COMBINEDID')\n",
    "eval_LR_features(['COMBINEDID', 'INTENDEDDEVICETYPE'], 'COMBINEDID and INTENDEDDEVICETYPE')\n",
    "eval_LR_features(['COMBINEDID', 'ACTUALDEVICETYPE'], 'COMBINEDID and ACTUALDEVICETYPE')\n",
    "eval_LR_features(['COMBINEDID', 'INTENDEDDEVICETYPE', 'ACTUALDEVICETYPE'], 'COMBINEDID and INTENDEDDEVICETYPE and ACTUALDEVICETYPE')\n",
    "eval_LR_features(['COMBINEDID', 'INTENDENTISACTUALDEVICETYPE'], 'COMBINEDID and INTENDENTISACTUALDEVICETYPE')\n",
    "eval_LR_features(['COMBINEDID', 'INTENDENTISACTUALDEVICETYPE', 'ACTUALDEVICETYPE'], 'COMBINEDID and INTENDENTISACTUALDEVICETYPE and ACTUALDEVICETYPE')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "INTENDEDDEVICETYPE not really needed, retain only ACTUALDEVICETYPE\n",
    "\n",
    "INTENDENTISACTUALDEVICETYPE can help just a little bit with RF however not with LR -> retain it, why not"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remaining features:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'ACTUALDEVICETYPE': 'Phone',\n",
       " 'ADLOADINGTIME': 7.121,\n",
       " 'COMBINEDEXTERNALID': 'null',\n",
       " 'COMBINEDID': '6650fadcc0264109f8bd976558377652_a0776e9e61f79615d93907ec5710c9fa_ab0a0251add6aba71e800537c6a729c8_83361d7aa1757e5f8e9ed97b71a8c49b_b46e86833f3f102dfeca4b7a13d6db21',\n",
       " 'CREATIVETYPE': 'Interstitial',\n",
       " 'ERRORSJSON': '[]',\n",
       " 'FILESJSON': '[]',\n",
       " 'FILESJSON_SIZE': 0L,\n",
       " 'INTENDEDDEVICETYPE': 'Phone',\n",
       " 'INTENDENTISACTUALDEVICETYPE': 1L,\n",
       " 'PLACEMENTID': 'f6297752e6cd8cd455c55d95c085035d',\n",
       " 'PLATFORMCOMBINED': 'Android_4.3',\n",
       " 'TIMESTAMP': 1459512400L,\n",
       " 'TOPMOSTREACHABLEWINDOWAREA': 221400L,\n",
       " 'UA_BROWSER': 'Android Browser',\n",
       " 'UA_BROWSERRENDERINGENGINE': 'WebKit',\n",
       " 'UA_BROWSERVERSION': '4.0',\n",
       " 'UA_DEVICETYPE': 'Phone',\n",
       " 'UA_HARDWARETYPE': 'Mobile Phone',\n",
       " 'UA_MODEL': 'SM-G7102',\n",
       " 'UA_OSCOMB': 'Android_4.3',\n",
       " 'UA_VENDOR': 'Samsung'}"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###Effect of progressively adding more features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PLACEMENTID: 3.4629\n",
      "TIMESTAMP: 3.4684\n",
      "INTENDEDDEVICETYPE: 3.4685\n",
      "ACTUALDEVICETYPE: 3.4655\n",
      "CDNNAME: 3.4687\n",
      "CREATIVETYPE: 3.4685\n",
      "UA_HARDWARETYPE: 3.4648\n",
      "UA_DEVICETYPE: 3.4655\n",
      "UA_PLATFORM: 3.3706\n",
      "UA_PLATFORMVERSION: 3.3457\n",
      "UA_VENDOR: 3.3694\n",
      "UA_MODEL: 3.3565\n",
      "UA_OS: 3.3693\n",
      "UA_OSVERSION: 3.3419\n",
      "UA_BROWSER: 3.3737\n",
      "UA_BROWSERVERSION: 3.3385\n",
      "UA_BROWSERRENDERINGENGINE: 3.4672\n",
      "FILESJSON: 3.4686\n",
      "ERRORSJSON: 3.4676\n",
      "COMBINEDEXTERNALID: 3.4701\n",
      "TOPMOSTREACHABLEWINDOWAREA: 3.4645\n",
      "PLATFORMCOMBINED: 3.3448\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "for col in data.column_names():\n",
    "    if col != 'ADLOADINGTIME' and  col != 'COMBINEDID':# and col != 'UA_BROWSERVERSION':\n",
    "        #sqft_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'CREATIVEID','UA_BROWSERVERSION'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "        LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'COMBINEDID'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "        model = LR_model\n",
    "        rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "        print(\"{}: {:.4f}\".format(col,rms))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ": 3.2613\n",
      "PLACEMENTID: 3.2630\n",
      "INTENDEDDEVICETYPE: 3.2616\n",
      "CDNNAME: 3.2618\n",
      "UA_HARDWARETYPE: 3.2618\n",
      "UA_DEVICETYPE: 3.2616\n",
      "UA_MODEL: 3.2732\n",
      "UA_BROWSERRENDERINGENGINE: 3.2615\n",
      "TOPMOSTREACHABLEWINDOWAREA: 3.2675\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID', 'UA_BROWSERVERSION', 'UA_OSCOMB', 'UA_VENDOR', 'PLATFORMCOMBINED', 'COMBINEDEXTERNALID', 'FILESJSON','ACTUALDEVICETYPE','ERRORSJSON','UA_BROWSER','TIMESTAMP','CREATIVETYPE'],validation_set=test_data, max_iterations=50,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('',rms))\n",
    "\n",
    "for col in data.column_names():\n",
    "    #if col != 'ADLOADINGTIME' and  col != 'COMBINEDID' and col != 'UA_BROWSERVERSION' and col != 'UA_PLATFORMCOMBINED':\n",
    "    if col != 'ADLOADINGTIME' and  col != 'COMBINEDID' and col != 'UA_BROWSERVERSION' and col != 'UA_OSCOMB'and col != 'UA_VENDOR' and col != 'PLATFORMCOMBINED' and col != 'COMBINEDEXTERNALID' and col != 'FILESJSON' and col!='ACTUALDEVICETYPE' and col!= 'ERRORSJSON' and col != 'UA_BROWSER' and col != 'TIMESTAMP' and col != 'CREATIVETYPE':\n",
    "        #sqft_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'CREATIVEID','UA_BROWSERVERSION'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "        #LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'COMBINEDID', 'UA_BROWSERVERSION', 'UA_PLATFORMCOMBINED'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "        LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'COMBINEDID', 'UA_BROWSERVERSION', 'UA_OSCOMB', 'UA_VENDOR', 'PLATFORMCOMBINED', 'COMBINEDEXTERNALID', 'FILESJSON','ACTUALDEVICETYPE','ERRORSJSON','UA_BROWSER','TIMESTAMP','CREATIVETYPE'],validation_set=test_data, max_iterations=50,verbose=False)\n",
    "        model = LR_model\n",
    "        rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "        print(\"{}: {:.4f}\".format(col,rms))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ": 3.2615\n",
      "PLACEMENTID: 3.2630\n",
      "PLATFORM: 3.2613\n",
      "PLATFORMVERSION: 3.2616\n",
      "INTENDEDDEVICETYPE: 3.2612\n",
      "ACTUALDEVICETYPE: 3.2615\n",
      "UA_DEVICETYPE: 3.2617\n",
      "UA_PLATFORM: 3.2619\n",
      "FILESJSON: 3.2613\n",
      "FILESJSON_SIZE: 3.2618\n",
      "PLATFORMCOMBINED: 3.2614\n"
     ]
    }
   ],
   "source": [
    "#train_data,test_data = data.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=['COMBINEDID','UA_BROWSERVERSION','UA_OSCOMB','COMBINEDEXTERNALID','TIMESTAMP','ERRORSJSON','UA_BROWSER','UA_HARDWARETYPE','UA_VENDOR','UA_BROWSERRENDERINGENGINE','CREATIVETYPE','UA_PLATFORMVERSION'],validation_set=test_data, max_iterations=50,verbose=False)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('',rms))\n",
    "\n",
    "for col in data.column_names():\n",
    "    #if col != 'ADLOADINGTIME' and  col != 'COMBINEDID' and col != 'UA_BROWSERVERSION' and col != 'UA_PLATFORMCOMBINED':\n",
    "    if col != 'ADLOADINGTIME' and col!='COMBINEDID' and col!='UA_BROWSERVERSION' and col!='UA_OSCOMB' and col!='COMBINEDEXTERNALID' and col != 'TIMESTAMP' and col!='ERRORSJSON' and col != 'UA_BROWSER' and col != 'UA_HARDWARETYPE' and col != 'UA_VENDOR' and col != 'UA_BROWSERRENDERINGENGINE' and col != 'CREATIVETYPE' and col != 'UA_PLATFORMVERSION':\n",
    "        #sqft_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'CREATIVEID','UA_BROWSERVERSION'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "        #LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'COMBINEDID', 'UA_BROWSERVERSION', 'UA_PLATFORMCOMBINED'],validation_set=test_data, max_iterations=100,verbose=False)\n",
    "        LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',features=[col,'COMBINEDID','UA_BROWSERVERSION','UA_OSCOMB','COMBINEDEXTERNALID','TIMESTAMP','ERRORSJSON','UA_BROWSER','UA_HARDWARETYPE','UA_VENDOR','UA_BROWSERRENDERINGENGINE','CREATIVETYPE','UA_PLATFORMVERSION'],validation_set=test_data, max_iterations=50,verbose=False)\n",
    "        model = LR_model\n",
    "        rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "        print(\"{}: {:.4f}\".format(col,rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CONCLUSIONS**\n",
    "\n",
    "Minimum set of features for linear regression:\n",
    "\n",
    "'ADLOADINGTIME','COMBINEDID', 'UA_BROWSERVERSION', 'UA_OSCOMB', 'UA_VENDOR', 'PLATFORMCOMBINED', 'COMBINEDEXTERNALID', 'FILESJSON','ACTUALDEVICETYPE','ERRORSJSON','UA_BROWSER','TIMESTAMP','CREATIVETYPE'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##Evaluation using linear regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tmp = data['ADLOADINGTIME','COMBINEDID', 'UA_BROWSERVERSION', 'UA_OSCOMB', 'UA_VENDOR', 'PLATFORMCOMBINED', 'COMBINEDEXTERNALID', 'FILESJSON','ACTUALDEVICETYPE','ERRORSJSON','UA_BROWSER','TIMESTAMP','CREATIVETYPE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre>Linear regression:</pre>"
      ],
      "text/plain": [
       "Linear regression:"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of examples          : 116524</pre>"
      ],
      "text/plain": [
       "Number of examples          : 116524"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of features          : 13</pre>"
      ],
      "text/plain": [
       "Number of features          : 13"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of unpacked features : 13</pre>"
      ],
      "text/plain": [
       "Number of unpacked features : 13"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of coefficients    : 9501</pre>"
      ],
      "text/plain": [
       "Number of coefficients    : 9501"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Starting L-BFGS</pre>"
      ],
      "text/plain": [
       "Starting L-BFGS"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| Iteration | Passes   | Step size | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |</pre>"
      ],
      "text/plain": [
       "| Iteration | Passes   | Step size | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 1         | 6        | 0.000000  | 0.265015     | 16.378900          | 16.494300            | 3.497875      | 3.519177        |</pre>"
      ],
      "text/plain": [
       "| 1         | 6        | 0.000000  | 0.265015     | 16.378900          | 16.494300            | 3.497875      | 3.519177        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 2         | 9        | 5.000000  | 0.469027     | 17.085073          | 17.133558            | 3.153341      | 3.300617        |</pre>"
      ],
      "text/plain": [
       "| 2         | 9        | 5.000000  | 0.469027     | 17.085073          | 17.133558            | 3.153341      | 3.300617        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 3         | 10       | 5.000000  | 0.584034     | 24.620016          | 39.000159            | 4.844549      | 5.067211        |</pre>"
      ],
      "text/plain": [
       "| 3         | 10       | 5.000000  | 0.584034     | 24.620016          | 39.000159            | 4.844549      | 5.067211        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 4         | 12       | 1.000000  | 0.758044     | 17.614629          | 17.478601            | 3.044576      | 3.263763        |</pre>"
      ],
      "text/plain": [
       "| 4         | 12       | 1.000000  | 0.758044     | 17.614629          | 17.478601            | 3.044576      | 3.263763        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 5         | 13       | 1.000000  | 0.880051     | 17.626021          | 17.391837            | 3.032534      | 3.260502        |</pre>"
      ],
      "text/plain": [
       "| 5         | 13       | 1.000000  | 0.880051     | 17.626021          | 17.391837            | 3.032534      | 3.260502        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 6         | 14       | 1.000000  | 0.997057     | 17.842394          | 18.264920            | 3.018821      | 3.275606        |</pre>"
      ],
      "text/plain": [
       "| 6         | 14       | 1.000000  | 0.997057     | 17.842394          | 18.264920            | 3.018821      | 3.275606        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 10        | 18       | 1.000000  | 1.437083     | 17.392661          | 17.953992            | 3.003519      | 3.261255        |</pre>"
      ],
      "text/plain": [
       "| 10        | 18       | 1.000000  | 1.437083     | 17.392661          | 17.953992            | 3.003519      | 3.261255        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 11        | 19       | 1.000000  | 1.541088     | 17.356951          | 17.512168            | 3.002222      | 3.255243        |</pre>"
      ],
      "text/plain": [
       "| 11        | 19       | 1.000000  | 1.541088     | 17.356951          | 17.512168            | 3.002222      | 3.255243        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 20        | 28       | 1.000000  | 2.442140     | 17.247095          | 17.838724            | 2.995633      | 3.255223        |</pre>"
      ],
      "text/plain": [
       "| 20        | 28       | 1.000000  | 2.442140     | 17.247095          | 17.838724            | 2.995633      | 3.255223        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 30        | 38       | 1.000000  | 3.590206     | 17.301070          | 18.720726            | 2.994953      | 3.257004        |</pre>"
      ],
      "text/plain": [
       "| 30        | 38       | 1.000000  | 3.590206     | 17.301070          | 18.720726            | 2.994953      | 3.257004        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 40        | 48       | 1.000000  | 4.678268     | 17.278442          | 21.720858            | 2.994678      | 3.260014        |</pre>"
      ],
      "text/plain": [
       "| 40        | 48       | 1.000000  | 4.678268     | 17.278442          | 21.720858            | 2.994678      | 3.260014        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 50        | 58       | 1.000000  | 5.837334     | 17.288097          | 22.289371            | 2.994503      | 3.261961        |</pre>"
      ],
      "text/plain": [
       "| 50        | 58       | 1.000000  | 5.837334     | 17.288097          | 22.289371            | 2.994503      | 3.261961        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 51        | 59       | 1.000000  | 5.949341     | 17.285156          | 22.158846            | 2.994492      | 3.262187        |</pre>"
      ],
      "text/plain": [
       "| 51        | 59       | 1.000000  | 5.949341     | 17.285156          | 22.158846            | 2.994492      | 3.262187        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 60        | 68       | 1.000000  | 6.902395     | 17.284969          | 19.991480            | 2.994375      | 3.264163        |</pre>"
      ],
      "text/plain": [
       "| 60        | 68       | 1.000000  | 6.902395     | 17.284969          | 19.991480            | 2.994375      | 3.264163        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 70        | 78       | 1.000000  | 8.026459     | 17.285815          | 20.017761            | 2.994334      | 3.264978        |</pre>"
      ],
      "text/plain": [
       "| 70        | 78       | 1.000000  | 8.026459     | 17.285815          | 20.017761            | 2.994334      | 3.264978        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 80        | 90       | 1.000000  | 9.243529     | 17.284683          | 19.968059            | 2.994315      | 3.265240        |</pre>"
      ],
      "text/plain": [
       "| 80        | 90       | 1.000000  | 9.243529     | 17.284683          | 19.968059            | 2.994315      | 3.265240        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 90        | 100      | 1.000000  | 10.388594    | 17.281619          | 20.559533            | 2.994297      | 3.264813        |</pre>"
      ],
      "text/plain": [
       "| 90        | 100      | 1.000000  | 10.388594    | 17.281619          | 20.559533            | 2.994297      | 3.264813        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 100       | 110      | 1.000000  | 11.578663    | 17.288191          | 20.002165            | 2.994273      | 3.264513        |</pre>"
      ],
      "text/plain": [
       "| 100       | 110      | 1.000000  | 11.578663    | 17.288191          | 20.002165            | 2.994273      | 3.264513        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>TERMINATED: Iteration limit reached.</pre>"
      ],
      "text/plain": [
       "TERMINATED: Iteration limit reached."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>This model may not be optimal. To improve it, consider increasing `max_iterations`.</pre>"
      ],
      "text/plain": [
       "This model may not be optimal. To improve it, consider increasing `max_iterations`."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ALL FEATURES: 3.2645\n"
     ]
    }
   ],
   "source": [
    "train_data,test_data = tmp.random_split(.5,seed=1)\n",
    "\n",
    "LR_model = graphlab.linear_regression.create(train_data, target='ADLOADINGTIME',validation_set=test_data, max_iterations=100,verbose=True)\n",
    "model = LR_model\n",
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('ALL FEATURES',rms))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([  9.77000000e+02,   2.62900000e+03,   7.00500000e+03,\n",
       "          1.14380000e+04,   1.39900000e+04,   1.34610000e+04,\n",
       "          1.40080000e+04,   1.18870000e+04,   8.53000000e+03,\n",
       "          6.89400000e+03,   5.32700000e+03,   4.26600000e+03,\n",
       "          3.76400000e+03,   3.18200000e+03,   2.18200000e+03,\n",
       "          1.51100000e+03,   1.09700000e+03,   8.95000000e+02,\n",
       "          4.89000000e+02,   4.44000000e+02,   2.84000000e+02,\n",
       "          3.69000000e+02,   3.24000000e+02,   4.49000000e+02,\n",
       "          2.21000000e+02,   1.17000000e+02,   4.50000000e+01,\n",
       "          5.40000000e+01,   2.40000000e+01,   2.60000000e+01,\n",
       "          1.40000000e+01,   1.20000000e+01,   1.20000000e+01,\n",
       "          5.00000000e+00,   1.00000000e+01]),\n",
       " array([  0. ,   0.5,   1. ,   1.5,   2. ,   2.5,   3. ,   3.5,   4. ,\n",
       "          4.5,   5. ,   5.5,   6. ,   6.5,   7. ,   7.5,   8. ,   8.5,\n",
       "          9. ,   9.5,  10. ,  10.5,  11. ,  11.5,  12. ,  12.5,  13. ,\n",
       "         13.5,  14. ,  14.5,  15. ,  15.5,  16. ,  16.5,  17. ,  17.5]),\n",
       " <a list of 35 Patch objects>)"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEACAYAAABcXmojAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJLUMBUlSy1CQJLX+P6BDSOq8LwpIAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x825915f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(test_data['ADLOADINGTIME'],  bins=np.arange(0, int(max(test_data['ADLOADINGTIME'])),0.5),alpha=0.5)\n",
    "test_data_predict = LR_model.predict(test_data)\n",
    "plt.hist(test_data_predict,  bins=np.arange(0, int(max(test_data['ADLOADINGTIME'])),0.5), alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.264962651196299"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],LR_model.predict(test_data))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##Evaluation using Random Forests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre>Random forest regression:</pre>"
      ],
      "text/plain": [
       "Random forest regression:"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of examples          : 116524</pre>"
      ],
      "text/plain": [
       "Number of examples          : 116524"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of features          : 18</pre>"
      ],
      "text/plain": [
       "Number of features          : 18"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of unpacked features : 18</pre>"
      ],
      "text/plain": [
       "Number of unpacked features : 18"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| Iteration | Elapsed Time | Training-max_error | Training-rmse | Validation-max_error | Validation-rmse |</pre>"
      ],
      "text/plain": [
       "| Iteration | Elapsed Time | Training-max_error | Training-rmse | Validation-max_error | Validation-rmse |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 1         | 1.422081     | 18.122999          | 3.018238      | 17.864000            | 3.643402        |</pre>"
      ],
      "text/plain": [
       "| 1         | 1.422081     | 18.122999          | 3.018238      | 17.864000            | 3.643402        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 2         | 2.736157     | 17.701963          | 2.752370      | 17.828583            | 3.324879        |</pre>"
      ],
      "text/plain": [
       "| 2         | 2.736157     | 17.701963          | 2.752370      | 17.828583            | 3.324879        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 3         | 4.411252     | 16.726635          | 2.628983      | 17.917221            | 3.254141        |</pre>"
      ],
      "text/plain": [
       "| 3         | 4.411252     | 16.726635          | 2.628983      | 17.917221            | 3.254141        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 4         | 5.982342     | 16.722294          | 2.567770      | 17.527250            | 3.218089        |</pre>"
      ],
      "text/plain": [
       "| 4         | 5.982342     | 16.722294          | 2.567770      | 17.527250            | 3.218089        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 5         | 7.440426     | 16.792973          | 2.546544      | 17.551109            | 3.182338        |</pre>"
      ],
      "text/plain": [
       "| 5         | 7.440426     | 16.792973          | 2.546544      | 17.551109            | 3.182338        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 6         | 8.958513     | 16.829338          | 2.522612      | 17.584175            | 3.166701        |</pre>"
      ],
      "text/plain": [
       "| 6         | 8.958513     | 16.829338          | 2.522612      | 17.584175            | 3.166701        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 7         | 10.615607    | 16.853971          | 2.500581      | 17.495115            | 3.155537        |</pre>"
      ],
      "text/plain": [
       "| 7         | 10.615607    | 16.853971          | 2.500581      | 17.495115            | 3.155537        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 8         | 12.285703    | 16.839104          | 2.481401      | 17.552288            | 3.145441        |</pre>"
      ],
      "text/plain": [
       "| 8         | 12.285703    | 16.839104          | 2.481401      | 17.552288            | 3.145441        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 9         | 13.752787    | 16.677233          | 2.471089      | 17.494390            | 3.137892        |</pre>"
      ],
      "text/plain": [
       "| 9         | 13.752787    | 16.677233          | 2.471089      | 17.494390            | 3.137892        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 10        | 15.358879    | 16.663021          | 2.460709      | 17.561802            | 3.132221        |</pre>"
      ],
      "text/plain": [
       "| 10        | 15.358879    | 16.663021          | 2.460709      | 17.561802            | 3.132221        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 11        | 16.985972    | 16.681129          | 2.453058      | 17.494476            | 3.130102        |</pre>"
      ],
      "text/plain": [
       "| 11        | 16.985972    | 16.681129          | 2.453058      | 17.494476            | 3.130102        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 12        | 18.596064    | 16.683794          | 2.447213      | 17.534185            | 3.125413        |</pre>"
      ],
      "text/plain": [
       "| 12        | 18.596064    | 16.683794          | 2.447213      | 17.534185            | 3.125413        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 13        | 20.254159    | 16.100849          | 2.450538      | 17.502274            | 3.119143        |</pre>"
      ],
      "text/plain": [
       "| 13        | 20.254159    | 16.100849          | 2.450538      | 17.502274            | 3.119143        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 14        | 21.694241    | 15.684564          | 2.451077      | 17.519480            | 3.113966        |</pre>"
      ],
      "text/plain": [
       "| 14        | 21.694241    | 15.684564          | 2.451077      | 17.519480            | 3.113966        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 15        | 22.977314    | 15.707900          | 2.458548      | 17.502481            | 3.107342        |</pre>"
      ],
      "text/plain": [
       "| 15        | 22.977314    | 15.707900          | 2.458548      | 17.502481            | 3.107342        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 16        | 24.621408    | 15.719462          | 2.452132      | 17.509501            | 3.106408        |</pre>"
      ],
      "text/plain": [
       "| 16        | 24.621408    | 15.719462          | 2.452132      | 17.509501            | 3.106408        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 17        | 26.165497    | 15.734441          | 2.448692      | 17.484875            | 3.105715        |</pre>"
      ],
      "text/plain": [
       "| 17        | 26.165497    | 15.734441          | 2.448692      | 17.484875            | 3.105715        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 18        | 27.835592    | 15.427544          | 2.445457      | 17.454847            | 3.105400        |</pre>"
      ],
      "text/plain": [
       "| 18        | 27.835592    | 15.427544          | 2.445457      | 17.454847            | 3.105400        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 19        | 29.459685    | 15.432894          | 2.442490      | 17.419001            | 3.104491        |</pre>"
      ],
      "text/plain": [
       "| 19        | 29.459685    | 15.432894          | 2.442490      | 17.419001            | 3.104491        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 20        | 31.005774    | 15.450750          | 2.440558      | 17.456476            | 3.103307        |</pre>"
      ],
      "text/plain": [
       "| 20        | 31.005774    | 15.450750          | 2.440558      | 17.456476            | 3.103307        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_data,test_data = data.random_split(.5,seed=1)\n",
    "RF_model = graphlab.random_forest_regression.create(train_data, target='ADLOADINGTIME',validation_set=test_data,\n",
    "                                           max_iterations=20,\n",
    "                                           max_depth =  50, row_subsample=0.5, random_seed=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ADLOADINGTIME',\n",
       " 'PLACEMENTID',\n",
       " 'TIMESTAMP',\n",
       " 'INTENDEDDEVICETYPE',\n",
       " 'ACTUALDEVICETYPE',\n",
       " 'CREATIVETYPE',\n",
       " 'UA_HARDWARETYPE',\n",
       " 'UA_VENDOR',\n",
       " 'UA_MODEL',\n",
       " 'UA_BROWSER',\n",
       " 'UA_BROWSERVERSION',\n",
       " 'UA_BROWSERRENDERINGENGINE',\n",
       " 'FILESJSON',\n",
       " 'ERRORSJSON',\n",
       " 'TOPMOSTREACHABLEWINDOWAREA',\n",
       " 'FILESJSON_SIZE',\n",
       " 'COMBINEDID',\n",
       " 'COMBINEDEXTERNALID',\n",
       " 'PLATFORMCOMBINED',\n",
       " 'UA_OSCOMB',\n",
       " 'INTENDENTISACTUALDEVICETYPE']"
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.column_names()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEACAYAAAC08h1NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n38ysxzjxm5n1GCd+M7Me8/8B3ZUcYEHVhgYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9aa545f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(test_data['ADLOADINGTIME'],  bins=np.arange(0, int(max(test_data['ADLOADINGTIME'])),0.1),alpha=0.5);\n",
    "test_data_predict = RF_model.predict(test_data)\n",
    "plt.hist(test_data_predict,  bins=np.arange(0, int(max(test_data['ADLOADINGTIME'])),0.1), alpha=0.5);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####Error distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([ 25274.,  24746.,  18692.,  11771.,   7688.,   5436.,   4232.,\n",
       "          3292.,   2349.,   1816.,   1552.]),\n",
       " array([ 0. ,  0.5,  1. ,  1.5,  2. ,  2.5,  3. ,  3.5,  4. ,  4.5,  5. ,\n",
       "         5.5]),\n",
       " <a list of 11 Patch objects>)"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEACAYAAAC+gnFaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\njft/8x/E+I2hDLcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x78027be0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_data_predict = RF_model.predict(test_data)\n",
    "\n",
    "#test_data['ADLOADINGTIME']\n",
    "\n",
    "pred_error = np.abs(test_data['ADLOADINGTIME']-test_data_predict)\n",
    "plt.hist(pred_error,  bins=np.arange(0, 6,0.5), alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "[histvals, histbins] = np.histogram(test_data['ADLOADINGTIME'], bins=np.arange(0, np.ceil(max(test_data['ADLOADINGTIME'])),0.5))\n",
    "minToBin = histbins-np.tile(np.array(test_data['ADLOADINGTIME']),(len(histbins),1)).transpose()\n",
    "minToBinIdx = np.argmin(np.abs(minToBin),axis=1)\n",
    "\n",
    "pred_error_loc = np.zeros_like(histbins)\n",
    "pred_error_loc_n = np.zeros_like(histbins)\n",
    "for i in range(len(minToBinIdx)):\n",
    "    pred_error_loc[minToBinIdx[i]] += pred_error[i]\n",
    "    pred_error_loc_n[minToBinIdx[i]] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  1.80727235,   1.3656214 ,   1.31241438,   1.22604653,\n",
       "         1.16086831,   1.18021638,   1.24068854,   1.38792183,\n",
       "         1.48763828,   1.62319369,   1.72230676,   1.89625418,\n",
       "         2.09762499,   2.33268414,   2.5250171 ,   2.84204889,\n",
       "         3.13240864,   3.4395743 ,   3.80150081,   4.25481851,\n",
       "         4.33970616,   4.81648892,   5.21751055,   5.50947954,\n",
       "         5.95859795,   6.21604543,   6.84784047,   7.39212076,\n",
       "         8.03045983,   8.57986475,   8.93725218,   9.43006671,\n",
       "         9.83246991,  10.54978905,  10.8886015 ,  11.39100219,\n",
       "        11.68420187,  12.2618772 ])"
      ]
     },
     "execution_count": 211,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pred_error_loc/pred_error_loc_n\n",
    "plt.plot(histbins, pred_error_loc/pred_error_loc_n)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Results** Larger times are badly detected!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##Data balancment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "dataE = data[0:1]\n",
    "\n",
    "[histvals, histbins] = np.histogram(data['ADLOADINGTIME'], bins=np.arange(0, np.ceil(max(data['ADLOADINGTIME'])),0.5))\n",
    "minToBin = histbins-np.tile(np.array(data['ADLOADINGTIME']),(len(histbins),1)).transpose()\n",
    "minToBinIdx = np.argmin(np.abs(minToBin),axis=1)\n",
    "\n",
    "totalEntriesPerBin = 10000\n",
    "\n",
    "for i in range(16):\n",
    "    sel = np.array(minToBinIdx==i)\n",
    "    subData = data[graphlab.SArray(sel)]\n",
    "    \n",
    "    dataE = dataE.append(subData)\n",
    "    \n",
    "for i in range(16, len(histvals)):\n",
    "#for i in range(len(histvals)):\n",
    "    diff = histvals[i]-totalEntriesPerBin\n",
    "    sel = np.array(minToBinIdx==i)\n",
    "    subData = data[graphlab.SArray(sel)]\n",
    "    \n",
    "    if diff > 0:\n",
    "        #dataE = dataE.append(subData[0:totalEntriesPerBin])\n",
    "        dataE = dataE.append(subData)\n",
    "    else:\n",
    "        if histvals[i]>0:\n",
    "            diff = -totalEntriesPerBin\n",
    "            while diff <0:\n",
    "                if histvals[i]+diff>0:\n",
    "                    dataE = dataE.append(subData[0:np.abs(diff)+1])\n",
    "                    diff = 0\n",
    "                else:\n",
    "                    dataE = dataE.append(subData)\n",
    "                    diff += histvals[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEACAYAAABcXmojAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJEnN/wewh1yWRNVkqQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x808e1208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(dataE['ADLOADINGTIME'],  bins=np.arange(0, int(max(dataE['ADLOADINGTIME'])),0.5),alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data,test_data = data.random_split(.5,seed=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "#train_data,test_data = data.random_split(.5,seed=1)\n",
    "RF_model = graphlab.random_forest_regression.create(dataE, target='ADLOADINGTIME',validation_set=test_data,\n",
    "                                           max_iterations=20,\n",
    "                                           max_depth =  50, row_subsample=0.5, random_seed=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "dataE[0:1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEACAYAAAC+gnFaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nV4FpAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1261a8fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_data_predict = RF_model.predict(test_data)\n",
    "\n",
    "#test_data['ADLOADINGTIME']\n",
    "\n",
    "pred_error = np.abs(test_data['ADLOADINGTIME']-test_data_predict)\n",
    "plt.hist(pred_error,  bins=np.arange(0, 6,0.5), alpha=0.5)\n",
    "\n",
    "[histvals, histbins] = np.histogram(test_data['ADLOADINGTIME'], bins=np.arange(0, np.ceil(max(test_data['ADLOADINGTIME'])),0.5))\n",
    "minToBin = histbins-np.tile(np.array(test_data['ADLOADINGTIME']),(len(histbins),1)).transpose()\n",
    "minToBinIdx = np.argmin(np.abs(minToBin),axis=1)\n",
    "\n",
    "pred_error_loc = np.zeros_like(histbins)\n",
    "pred_error_loc_n = np.zeros_like(histbins)\n",
    "for i in range(len(minToBinIdx)):\n",
    "    pred_error_loc[minToBinIdx[i]] += pred_error[i]\n",
    "    pred_error_loc_n[minToBinIdx[i]] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x125d4d630>]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEACAYAAABMEua6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x126268e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(histbins, pred_error_loc/pred_error_loc_n)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x36547b00>]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEACAYAAABMEua6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x369212e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(histbins, pred_error_loc/pred_error_loc_n)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ALL FEATURES: 3.3250\n"
     ]
    }
   ],
   "source": [
    "rms = graphlab.evaluation.rmse(test_data['ADLOADINGTIME'],RF_model.predict(test_data))\n",
    "print(\"{}: {:.4f}\".format('ALL FEATURES',rms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "##Split prediction to two models: lower and upper times"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEACAYAAAC57G0KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAETRJREFUeJzt3X+MZWV9x/H3R5FQtelKrMuvbZc/IOk2jVDTrQltvE0p\nwX8A/+FHYroYYkywKqZNBP+QWU3WH4lk/2j0H8Bs/YHdaKTbNCoL4aZuIk4xrCArCgmTuBQGjW4L\n0YSlfPvHnIXrzp07O/fO3Jl95v1KJvvc55xnzrMnJ5/73Oc8506qCklSe1633h2QJK0NA16SGmXA\nS1KjDHhJapQBL0mNMuAlqVEjAz7JWUl+kORwkiNJPt3Vn53kYJKfJbkvyZaBNrcleTLJE0muWOv/\ngCRpuCy3Dj7JG6vqN0nOAA4B/wRcBfyyqj6X5GPAW6rq1iQ7gK8BfwGcD9wPXFxVr6zp/0KStMiy\nUzRV9ZuueCbweuDXLAT8vq5+H3BNV74auKeqjlfVHPAUsHM1OyxJOjXLBnyS1yU5DMwDD1bV48DW\nqprvdpkHtnbl84CjA82PsjCSlyRN2RnL7dBNr1yS5A+A7yb5m5O2V5JR8zx+F4IkrYNlA/6Eqvqf\nJP8BvAOYT3JOVT2X5Fzg+W63Z4BtA80u6Op+xzJvCJKkJVRVTnXf5VbRvPXECpkkvwf8HfAIcADY\n1e22C7i3Kx8Ark9yZpILgYuA2SU66U8Vt99++7r3YaP8eC48F56L0T8rtdwI/lxgX5LXsfBm8OWq\neiDJI8D+JDcBc8C1XWgfSbIfOAK8DNxc4/RKkjSxkQFfVY8Bfz6k/lfA5Uu02QPsWZXeSZLG5pOs\n66zX6613FzYMz8VrPBev8VyMb9kHndbkoIkzN5K0Qkmo1brJKkk6fRnwktQoA16SGmXAS1KjDHhJ\napQBL0mNMuAlqVEGvCQ1yoCXpEYZ8JLUKANekhplwEtSowx4SWqUAS9JjTLgJalRBrwkNWq5v8l6\n2rvllhmOHVtcv2UL7N07M/X+SNK0NB/wx47B9u0zi+rn5hbXSVJLnKKRpEYZ8JLUKANekhplwEtS\nowx4SWqUAS9JjTLgJalRBrwkNWpkwCfZluTBJI8n+XGSD3f1M0mOJnmk+3n3QJvbkjyZ5IkkV6z1\nf0CSNNxyT7IeBz5aVYeTvBn4YZKDQAF3VNUdgzsn2QFcB+wAzgfuT3JxVb2yBn2XJI0wcgRfVc9V\n1eGu/CLwExaCGyBDmlwN3FNVx6tqDngK2Ll63ZUknapT/i6aJNuBS4GHgMuADyX5e+Bh4B+r6hhw\nXrf9hKO89oawZpb6QjGA2dnDbN++1j2QpI3nlAK+m575BvCRqnoxyReBT3abPwV8HrhpieY1cS+X\nsdQXigEcOnTNWh9ekjakZQM+yRuAbwJfqap7Aarq+YHtdwL/3r18Btg20PyCrm6RmZmZV8u9Xo9e\nr7eynktS4/r9Pv1+f+z2IwM+SYC7gCNVtXeg/tyqerZ7+R7gsa58APhakjtYmJq5CJgd9rsHA16S\ntNjJg9/du3evqP1yI/jLgPcCjyZ5pKv7OHBDkktYmH55GvgAQFUdSbIfOAK8DNxcVWs+RSNJWmxk\nwFfVIYavtPn2iDZ7gD0T9kuSNCGfZJWkRhnwktQoA16SGmXAS1KjDHhJapQBL0mNMuAlqVEGvCQ1\nyoCXpEYZ8JLUKANekhplwEtSowx4SWqUAS9JjTLgJalRBrwkNcqAl6RGGfCS1CgDXpIaZcBLUqMM\neElqlAEvSY0y4CWpUQa8JDXKgJekRhnwktQoA16SGmXAS1KjDHhJatTIgE+yLcmDSR5P8uMkH+7q\nz05yMMnPktyXZMtAm9uSPJnkiSRXrPV/QJI03HIj+OPAR6vqT4F3Ah9M8ifArcDBqroYeKB7TZId\nwHXADuBK4AtJ/JQgSetgZPhW1XNVdbgrvwj8BDgfuArY1+22D7imK18N3FNVx6tqDngK2LkG/ZYk\nLeOUR9dJtgOXAj8AtlbVfLdpHtjalc8Djg40O8rCG4IkacrOOJWdkrwZ+Cbwkap6Icmr26qqktSI\n5kO3zczMvFru9Xr0er1T6YokbRr9fp9+vz92+2UDPskbWAj3L1fVvV31fJJzquq5JOcCz3f1zwDb\nBppf0NUtMhjwkqTFTh787t69e0Xtl1tFE+Au4EhV7R3YdADY1ZV3AfcO1F+f5MwkFwIXAbMr6pEk\naVUsN4K/DHgv8GiSR7q624DPAPuT3ATMAdcCVNWRJPuBI8DLwM1VNWr6RpK0RkYGfFUdYulR/uVL\ntNkD7JmwX5KkCblGXZIaZcBLUqMMeElqlAEvSY0y4CWpUQa8JDXKgJekRhnwktQoA16SGmXAS1Kj\nDHhJapQBL0mNMuAlqVEGvCQ1yoCXpEYZ8JLUKANekhplwEtSowx4SWqUAS9JjTLgJalRBrwkNcqA\nl6RGGfCS1CgDXpIaZcBLUqMMeElqlAEvSY1aNuCT3J1kPsljA3UzSY4meaT7effAttuSPJnkiSRX\nrFXHJUmjncoI/kvAlSfVFXBHVV3a/XwbIMkO4DpgR9fmC0n8lCBJ62DZ8K2q7wG/HrIpQ+quBu6p\nquNVNQc8BeycqIeSpLFMMrr+UJIfJbkryZau7jzg6MA+R4HzJziGJGlMZ4zZ7ovAJ7vyp4DPAzct\nsW8Nq5yZmXm13Ov16PV6Y3ZFktrU7/fp9/tjtx8r4Kvq+RPlJHcC/969fAbYNrDrBV3dIoMBL0la\n7OTB7+7du1fUfqwpmiTnDrx8D3Bihc0B4PokZya5ELgImB3nGJKkySw7gk9yD/Au4K1Jfg7cDvSS\nXMLC9MvTwAcAqupIkv3AEeBl4OaqGjpFI0laW8sGfFXdMKT67hH77wH2TNIpSdLkXKMuSY0y4CWp\nUQa8JDVq3HXwp73Z2Ye48caZodu2bIG9e4dvk6TTxaYN+JdeOovt22eGbpubG14vSacTp2gkqVEG\nvCQ1yoCXpEYZ8JLUKANekhplwEtSowx4SWqUAS9JjTLgJalRBrwkNcqAl6RGGfCS1CgDXpIaZcBL\nUqMMeElqlAEvSY0y4CWpUQa8JDXKgJekRhnwktQoA16SGmXAS1KjDHhJatSyAZ/k7iTzSR4bqDs7\nycEkP0tyX5ItA9tuS/JkkieSXLFWHZckjXYqI/gvAVeeVHcrcLCqLgYe6F6TZAdwHbCja/OFJH5K\nkKR1sGz4VtX3gF+fVH0VsK8r7wOu6cpXA/dU1fGqmgOeAnauTlclSSsx7uh6a1XNd+V5YGtXPg84\nOrDfUeD8MY8hSZrAGZP+gqqqJDVql2GVMzMzr5Z7vR69Xm/SrkhSU/r9Pv1+f+z24wb8fJJzquq5\nJOcCz3f1zwDbBva7oKtbZDDgJUmLnTz43b1794rajztFcwDY1ZV3AfcO1F+f5MwkFwIXAbNjHkOS\nNIFlR/BJ7gHeBbw1yc+BTwCfAfYnuQmYA64FqKojSfYDR4CXgZuratT0jSRpjSwb8FV1wxKbLl9i\n/z3Ankk6JUmanGvUJalRBrwkNWriZZItmp19iBtvnBm6bcsW2Lt3+DZJ2kgM+CFeeukstm+fGbpt\nbm54vSRtNE7RSFKjDHhJapQBL0mNMuAlqVEGvCQ1yoCXpEYZ8JLUKANekhplwEtSowx4SWqUAS9J\njTLgJalRBrwkNcqAl6RGGfCS1CgDXpIaZcBLUqMMeElqlAEvSY06bf4m6y23zHDs2PBts7OH2b59\nqt2RpA3vtAn4Y8dY8g9hHzp0zXQ7I0mnAadoJKlRBrwkNeq0maLZKGZnH+LGG2cW1W/ZAnv3Lq6X\npPUyUcAnmQP+F/g/4HhV7UxyNvCvwB8Dc8C1VbXE7dHTz0svnTX0XsDc3OI6SVpPk07RFNCrqkur\namdXdytwsKouBh7oXkuSpmw15uBz0uurgH1deR/gEhdJWgerMYK/P8nDSd7f1W2tqvmuPA9snfAY\nkqQxTHqT9bKqejbJHwIHkzwxuLGqKklNeAxJ0hgmCviqerb79xdJvgXsBOaTnFNVzyU5F3h+WNuZ\nmZlXy71ej16vN0lXJKk5/X6ffr8/dvuxAz7JG4HXV9ULSd4EXAHsBg4Au4DPdv/eO6z9YMBLkhY7\nefC7e/fuFbWfZAS/FfhWkhO/56tVdV+Sh4H9SW6iWyY5wTEkSWMaO+Cr6mngkiH1vwIun6RTp6Ol\nHoACH4KStD58knWVLPUAFPgQlKT14XfRSFKjDHhJapQBL0mNcg5+CrwBK2k9GPBT4A1YSevBKRpJ\napQBL0mNMuAlqVEGvCQ1yoCXpEYZ8JLUKJdJrrOl1si7Pl7SpAz4dbbUGnnXx0ualFM0ktQoA16S\nGuUUzQbl99dImtSGC/hbbpnh2LHF9bOzh9m+ferdWTd+f42kSW24gD92jKHBdujQNdPvjCSdxpyD\nl6RGGfCS1CgDXpIateHm4DWZpW5Sw+m7+qbF/5M0DQb8aWjUEsrZ2cNce+29Q7cttfpmowfoUjfe\nwRVF0igG/Glo1BLKcVYbjRugS70xbIQ3BUkGvCaw1BuDo2ppY/AmqyQ1yhG8NgyfYpZW15oEfJIr\ngb3A64E7q+qza3EcrcxSN2dHBehyN3RXM3h9illaXase8EleD/wzcDnwDPBfSQ5U1U9W+1gt+O1v\nfzm1Yy11c3ZUgK72Dd1RK3YefPB+3ve+4ccax0ZfHTRKv9+n1+utdzc2BM/F+NZiBL8TeKqq5gCS\nfB24GjDgh5hmwG8Eo1bsvPDC16d2rI1+I9hQe43nYnxrEfDnAz8feH0U+MuTd/r+978/tPErr7yy\nBl3SNE1zWmdc4/ypxFGfCH70o4d4+9vfuaLfN02r/WnmdP50tJmsRcDXqez02c9+d1Hd29626n3R\nOljtaZ1Rxn0zGedPJY76RHDo0DVDt+3ff+VY3+v/ne/0h/Zl3Deg1X4AbpzfN67VPher/QY0zrGm\n1b9UnVIen/ovTN4JzFTVld3r24BXBm+0Jlndg0rSJlFVOdV91yLgzwB+Cvwt8N/ALHCDN1klabpW\nfYqmql5O8g/Ad1lYJnmX4S5J07fqI3hJ0sYw1a8qSHJlkieSPJnkY9M89kaTZC7Jo0keSTK73v2Z\npiR3J5lP8thA3dlJDib5WZL7kmxZzz5OyxLnYibJ0e7aeKR7cLB5SbYleTDJ40l+nOTDXf2muzZG\nnIsVXRtTG8F3D0D9lIEHoNjEc/NJngbeUVW/Wu++TFuSvwZeBP6lqv6sq/sc8Muq+lz35v+Wqrp1\nPfs5DUuci9uBF6rqjnXt3JQlOQc4p6oOJ3kz8EPgGuB9bLJrY8S5uJYVXBvTHMG/+gBUVR0HTjwA\ntZmd8t3wllTV94Bfn1R9FbCvK+9j4WJu3hLnAjbhtVFVz1XV4a78IgsPR57PJrw2RpwLWMG1Mc2A\nH/YA1PlL7LsZFHB/koeTvH+9O7MBbK2q+a48D2xdz85sAB9K8qMkd22GKYmTJdkOXAr8gE1+bQyc\ni4e6qlO+NqYZ8N7N/V2XVdWlwLuBD3Yf1QXUwrzhZr5evghcCFwCPAt8fn27M13dlMQ3gY9U1QuD\n2zbbtdGdi2+wcC5eZIXXxjQD/hlg28DrbSyM4jelqnq2+/cXwLdYmMLazOa7eUeSnAs8v879WTdV\n9Xx1gDvZRNdGkjewEO5frqoTj8puymtj4Fx85cS5WOm1Mc2Afxi4KMn2JGcC1wEHpnj8DSPJG5P8\nfld+E3AF8NjoVs07AOzqyruA4c/BbwJdiJ3wHjbJtZEkwF3AkaraO7Bp010bS52LlV4bU10Hn+Td\nvPY98XdV1aendvANJMmFLIzaYeFhs69upnOR5B7gXcBbWZhT/QTwb8B+4I+AOeDaqlri2zraMeRc\n3A70WPgIXsDTwAcG5qCbleSvgP8EHuW1aZjbWHgaflNdG0uci48DN7CCa8MHnSSpUf5NVklqlAEv\nSY0y4CWpUQa8JDXKgJekRhnwktQoA16SGmXAS1Kj/h+GolKStmCH3AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x825b55c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tmp = data.groupby(key_columns='PLACEMENTID',\n",
    "    operations={'ADLOADINGTIME_MEDIAN': graphlab.aggregate.QUANTILE('ADLOADINGTIME',0.5),'ADLOADINGTIME_STD': graphlab.aggregate.STD('ADLOADINGTIME')})\n",
    "\n",
    "plt.hist(np.array(tmp['ADLOADINGTIME_MEDIAN']).flatten(),  bins=np.arange(0, 25,0.5),alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEACAYAAAC57G0KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkkrKgJekkjLgJamk/j92num2ZR93oAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x82696748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for f in ['COMBINEDID','PLACEMENTID','FILESJSON',]:\n",
    "    tmp = data.groupby(key_columns=f,\n",
    "        operations={'ADLOADINGTIME_MEAN': graphlab.aggregate.MEAN('ADLOADINGTIME')})\n",
    "\n",
    "    plt.hist(np.array(tmp['ADLOADINGTIME_MEAN']).flatten(),  bins=np.arange(0, 25,0.5),alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAEACAYAAABF+UbAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/w+Lt5Uez5qsIQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x82c913c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for f in ['PLATFORMCOMBINED','UA_BROWSER']:\n",
    "    tmp = data.groupby(key_columns=f,\n",
    "        operations={'ADLOADINGTIME_MEAN': graphlab.aggregate.MEAN('ADLOADINGTIME')})\n",
    "\n",
    "    plt.hist(np.array(tmp['ADLOADINGTIME_MEAN']).flatten(),  bins=np.arange(0, 25,0.5),alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "'UA_VENDOR','UA_MODEL','UA_BROWSERVERSION','PLATFORMCOMBINED','UA_BROWSER','COMBINEDID','PLACEMENTID','FILESJSON'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAEACAYAAABYq7oeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nVAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x849604a8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data['GROUPDIVIDER'] = data['COMBINEDID'] + data['PLACEMENTID'] + data['PLATFORMCOMBINED'] + data['FILESJSON']+data['UA_VENDOR'] + data['UA_MODEL'] + data['UA_BROWSER'] + data['UA_BROWSERVERSION']\n",
    "groupdivider = data.groupby(key_columns='GROUPDIVIDER',\n",
    "    operations={'ADLOADINGTIME_MEDIAN': graphlab.aggregate.QUANTILE('ADLOADINGTIME',0.5),'ADLOADINGTIME_STD': graphlab.aggregate.STD('ADLOADINGTIME')})\n",
    "groupdivider['ADLOADINGTIME_MEDIAN'] = np.array(groupdivider['ADLOADINGTIME_MEDIAN']).flatten()\n",
    "\n",
    "plt.hist(np.array(groupdivider['ADLOADINGTIME_MEDIAN']).flatten(),  bins=np.arange(0, np.ceil(max(groupdivider['ADLOADINGTIME_MEDIAN'])),0.5),alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEACAYAAAC57G0KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x15f34eef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data['GROUPDIVIDER'] = data['COMBINEDID'] + data['PLACEMENTID']  + data['FILESJSON']\n",
    "groupdivider = data.groupby(key_columns='GROUPDIVIDER',\n",
    "    operations={'ADLOADINGTIME_MEDIAN': graphlab.aggregate.QUANTILE('ADLOADINGTIME',0.5),'ADLOADINGTIME_STD': graphlab.aggregate.STD('ADLOADINGTIME')})\n",
    "groupdivider['ADLOADINGTIME_MEDIAN'] = np.array(tmp['ADLOADINGTIME_MEDIAN']).flatten()\n",
    "\n",
    "plt.hist(np.array(tmp['ADLOADINGTIME_MEDIAN']).flatten(),  bins=np.arange(0, np.ceil(max(tmp['ADLOADINGTIME_MEDIAN'])),0.5),alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "groupdivider=groupdivider[groupdivider['ADLOADINGTIME_MEDIAN']>8]\n",
    "data['GROUPDIVIDER']\n",
    "groupdivider = graphlab.SArray.to_numpy(groupdivider['GROUPDIVIDER'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "data['GROUPDIVIDER_LOGICAL'] = data['GROUPDIVIDER'].apply(lambda x : x in groupdivider)\n",
    "data1 = data[data['GROUPDIVIDER_LOGICAL'] == False]\n",
    "data2 = data[data['GROUPDIVIDER_LOGICAL'] == True]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data.remove_columns(['GROUPDIVIDER', 'GROUPDIVIDER_LOGICAL'])\n",
    "dataCut1train, dataCut1test = data1.random_split(.5,seed=1)\n",
    "dataCut2train, dataCut2test = data2.random_split(.5,seed=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dataCut1train = dataCut1train[dataCut1train['ADLOADINGTIME']<18.7]\n",
    "dataCut2train = dataCut2train[dataCut2train['ADLOADINGTIME']<18.7]\n",
    "dataCut1test = dataCut1test[dataCut1test['ADLOADINGTIME']<18.7]\n",
    "dataCut2test = dataCut2test[dataCut2test['ADLOADINGTIME']<18.7]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre>WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset.</pre>"
      ],
      "text/plain": [
       "WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Linear regression:</pre>"
      ],
      "text/plain": [
       "Linear regression:"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of examples          : 102055</pre>"
      ],
      "text/plain": [
       "Number of examples          : 102055"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of features          : 22</pre>"
      ],
      "text/plain": [
       "Number of features          : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of unpacked features : 22</pre>"
      ],
      "text/plain": [
       "Number of unpacked features : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of coefficients    : 48076</pre>"
      ],
      "text/plain": [
       "Number of coefficients    : 48076"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Starting Accelerated Gradient (FISTA)</pre>"
      ],
      "text/plain": [
       "Starting Accelerated Gradient (FISTA)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| Iteration | Passes   | Step size | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |</pre>"
      ],
      "text/plain": [
       "| Iteration | Passes   | Step size | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Tuning step size. First iteration could take longer than subsequent iterations.</pre>"
      ],
      "text/plain": [
       "Tuning step size. First iteration could take longer than subsequent iterations."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 1         | 2        | 0.000000  | 2.426138     | 16.907385          | 17.149706            | 2.626332      | 2.676119        |</pre>"
      ],
      "text/plain": [
       "| 1         | 2        | 0.000000  | 2.426138     | 16.907385          | 17.149706            | 2.626332      | 2.676119        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 2         | 3        | 0.000000  | 2.633150     | 16.608286          | 17.038367            | 2.523948      | 2.613446        |</pre>"
      ],
      "text/plain": [
       "| 2         | 3        | 0.000000  | 2.633150     | 16.608286          | 17.038367            | 2.523948      | 2.613446        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 3         | 4        | 0.000000  | 2.838162     | 16.580231          | 17.034180            | 2.445783      | 2.584500        |</pre>"
      ],
      "text/plain": [
       "| 3         | 4        | 0.000000  | 2.838162     | 16.580231          | 17.034180            | 2.445783      | 2.584500        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 4         | 5        | 0.000000  | 3.048174     | 16.585083          | 17.052601            | 2.378428      | 2.570567        |</pre>"
      ],
      "text/plain": [
       "| 4         | 5        | 0.000000  | 3.048174     | 16.585083          | 17.052601            | 2.378428      | 2.570567        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 5         | 6        | 0.000000  | 3.263186     | 16.635147          | 17.061331            | 2.320409      | 2.566928        |</pre>"
      ],
      "text/plain": [
       "| 5         | 6        | 0.000000  | 3.263186     | 16.635147          | 17.061331            | 2.320409      | 2.566928        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 6         | 7        | 0.000000  | 3.499200     | 16.728395          | 17.205034            | 2.271277      | 2.571038        |</pre>"
      ],
      "text/plain": [
       "| 6         | 7        | 0.000000  | 3.499200     | 16.728395          | 17.205034            | 2.271277      | 2.571038        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 10        | 11       | 0.000000  | 4.339248     | 16.967817          | 17.694719            | 2.155702      | 2.632996        |</pre>"
      ],
      "text/plain": [
       "| 10        | 11       | 0.000000  | 4.339248     | 16.967817          | 17.694719            | 2.155702      | 2.632996        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 11        | 12       | 0.000000  | 4.576261     | 16.996050          | 17.782061            | 2.143220      | 2.653200        |</pre>"
      ],
      "text/plain": [
       "| 11        | 12       | 0.000000  | 4.576261     | 16.996050          | 17.782061            | 2.143220      | 2.653200        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>SUCCESS: Optimal solution found.</pre>"
      ],
      "text/plain": [
       "SUCCESS: Optimal solution found."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre></pre>"
      ],
      "text/plain": []
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>WARNING: The number of feature dimensions in this problem is very large in comparison with the number of examples. Unless an appropriate regularization value is set, this model may not provide accurate predictions for a validation/test set.</pre>"
      ],
      "text/plain": [
       "WARNING: The number of feature dimensions in this problem is very large in comparison with the number of examples. Unless an appropriate regularization value is set, this model may not provide accurate predictions for a validation/test set."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset.</pre>"
      ],
      "text/plain": [
       "WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Linear regression:</pre>"
      ],
      "text/plain": [
       "Linear regression:"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of examples          : 14596</pre>"
      ],
      "text/plain": [
       "Number of examples          : 14596"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of features          : 22</pre>"
      ],
      "text/plain": [
       "Number of features          : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of unpacked features : 22</pre>"
      ],
      "text/plain": [
       "Number of unpacked features : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of coefficients    : 15829</pre>"
      ],
      "text/plain": [
       "Number of coefficients    : 15829"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Starting Accelerated Gradient (FISTA)</pre>"
      ],
      "text/plain": [
       "Starting Accelerated Gradient (FISTA)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| Iteration | Passes   | Step size | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |</pre>"
      ],
      "text/plain": [
       "| Iteration | Passes   | Step size | Elapsed Time | Training-max_error | Validation-max_error | Training-rmse | Validation-rmse |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Tuning step size. First iteration could take longer than subsequent iterations.</pre>"
      ],
      "text/plain": [
       "Tuning step size. First iteration could take longer than subsequent iterations."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 1         | 2        | 0.000002  | 0.485028     | 13.436371          | 14.157954            | 4.848498      | 5.290267        |</pre>"
      ],
      "text/plain": [
       "| 1         | 2        | 0.000002  | 0.485028     | 13.436371          | 14.157954            | 4.848498      | 5.290267        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 2         | 3        | 0.000002  | 0.528031     | 11.520316          | 13.077381            | 3.998137      | 4.572410        |</pre>"
      ],
      "text/plain": [
       "| 2         | 3        | 0.000002  | 0.528031     | 11.520316          | 13.077381            | 3.998137      | 4.572410        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 3         | 4        | 0.000002  | 0.573033     | 11.128049          | 12.718574            | 3.681140      | 4.407818        |</pre>"
      ],
      "text/plain": [
       "| 3         | 4        | 0.000002  | 0.573033     | 11.128049          | 12.718574            | 3.681140      | 4.407818        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 4         | 5        | 0.000002  | 0.620036     | 11.485599          | 13.555567            | 3.455524      | 4.373945        |</pre>"
      ],
      "text/plain": [
       "| 4         | 5        | 0.000002  | 0.620036     | 11.485599          | 13.555567            | 3.455524      | 4.373945        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 5         | 6        | 0.000002  | 0.669039     | 11.899064          | 14.287541            | 3.255046      | 4.380460        |</pre>"
      ],
      "text/plain": [
       "| 5         | 6        | 0.000002  | 0.669039     | 11.899064          | 14.287541            | 3.255046      | 4.380460        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 6         | 7        | 0.000002  | 0.715041     | 12.255851          | 14.976129            | 3.072819      | 4.403754        |</pre>"
      ],
      "text/plain": [
       "| 6         | 7        | 0.000002  | 0.715041     | 12.255851          | 14.976129            | 3.072819      | 4.403754        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 11        | 12       | 0.000002  | 0.921053     | 13.311475          | 17.418002            | 2.477832      | 4.602750        |</pre>"
      ],
      "text/plain": [
       "| 11        | 12       | 0.000002  | 0.921053     | 13.311475          | 17.418002            | 2.477832      | 4.602750        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 25        | 26       | 0.000002  | 1.441083     | 13.773678          | 20.889831            | 2.261938      | 4.879402        |</pre>"
      ],
      "text/plain": [
       "| 25        | 26       | 0.000002  | 1.441083     | 13.773678          | 20.889831            | 2.261938      | 4.879402        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+----------+-----------+--------------+--------------------+----------------------+---------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>SUCCESS: Optimal solution found.</pre>"
      ],
      "text/plain": [
       "SUCCESS: Optimal solution found."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre></pre>"
      ],
      "text/plain": []
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "LR_model1 = graphlab.linear_regression.create(dataCut1train, target='ADLOADINGTIME',validation_set=dataCut1test, max_iterations=100, verbose=True, solver='fista')\n",
    "LR_model2 = graphlab.linear_regression.create(dataCut2train, target='ADLOADINGTIME',validation_set=dataCut2test, max_iterations=100, verbose=True, solver='fista')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.69124942086\n",
      "4.90523158358\n",
      "3.05334648425\n"
     ]
    }
   ],
   "source": [
    "pred1 = LR_model1.predict(dataCut1test)\n",
    "pred2 = LR_model2.predict(dataCut2test)\n",
    "print(graphlab.evaluation.rmse(dataCut1test['ADLOADINGTIME'],pred1))\n",
    "print(graphlab.evaluation.rmse(dataCut2test['ADLOADINGTIME'],pred2))\n",
    "      \n",
    "testFinal = pred1.append(pred2)\n",
    "print(graphlab.evaluation.rmse(dataCut1test['ADLOADINGTIME'].append(dataCut2test['ADLOADINGTIME']),testFinal))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEACAYAAAC08h1NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4jczyxknfjOznHHiNzPLmf8PPMupyK+oZCUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x83096940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(dataCut1test['ADLOADINGTIME'],  bins=np.arange(0, int(max(dataCut1test['ADLOADINGTIME'])),0.2),alpha=0.5);\n",
    "test_data_predict = LR_model1.predict(dataCut1test)\n",
    "plt.hist(test_data_predict,  bins=np.arange(0, int(max(dataCut1test['ADLOADINGTIME'])),0.2), alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEACAYAAAC57G0KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nb2ZWUA54M7OCcsCbmRXU/wew+Oaa0pPgUwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xe21fdcc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(dataCut2test['ADLOADINGTIME'],  bins=np.arange(0, int(max(dataCut2test['ADLOADINGTIME'])),0.2),alpha=0.5);\n",
    "test_data_predict = LR_model2.predict(dataCut2test)\n",
    "plt.hist(test_data_predict,  bins=np.arange(0, int(max(dataCut2test['ADLOADINGTIME'])),0.2), alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre>WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset.</pre>"
      ],
      "text/plain": [
       "WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Random forest regression:</pre>"
      ],
      "text/plain": [
       "Random forest regression:"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of examples          : 104675</pre>"
      ],
      "text/plain": [
       "Number of examples          : 104675"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of features          : 22</pre>"
      ],
      "text/plain": [
       "Number of features          : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of unpacked features : 22</pre>"
      ],
      "text/plain": [
       "Number of unpacked features : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| Iteration | Elapsed Time | Training-max_error | Training-rmse | Validation-max_error | Validation-rmse |</pre>"
      ],
      "text/plain": [
       "| Iteration | Elapsed Time | Training-max_error | Training-rmse | Validation-max_error | Validation-rmse |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 1         | 2.708155     | 17.760000          | 2.512983      | 17.605499            | 2.902737        |</pre>"
      ],
      "text/plain": [
       "| 1         | 2.708155     | 17.760000          | 2.512983      | 17.605499            | 2.902737        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 2         | 5.787331     | 17.386250          | 2.293522      | 17.366167            | 2.748652        |</pre>"
      ],
      "text/plain": [
       "| 2         | 5.787331     | 17.386250          | 2.293522      | 17.366167            | 2.748652        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 3         | 7.945454     | 17.475166          | 2.244307      | 16.999012            | 2.663157        |</pre>"
      ],
      "text/plain": [
       "| 3         | 7.945454     | 17.475166          | 2.244307      | 16.999012            | 2.663157        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 4         | 11.186640    | 16.853661          | 2.197841      | 17.170279            | 2.638825        |</pre>"
      ],
      "text/plain": [
       "| 4         | 11.186640    | 16.853661          | 2.197841      | 17.170279            | 2.638825        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 5         | 14.085805    | 16.881750          | 2.173033      | 17.158583            | 2.622050        |</pre>"
      ],
      "text/plain": [
       "| 5         | 14.085805    | 16.881750          | 2.173033      | 17.158583            | 2.622050        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 6         | 17.143980    | 16.781830          | 2.152983      | 17.053867            | 2.610922        |</pre>"
      ],
      "text/plain": [
       "| 6         | 17.143980    | 16.781830          | 2.152983      | 17.053867            | 2.610922        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 7         | 20.043146    | 16.711281          | 2.142576      | 17.030731            | 2.604029        |</pre>"
      ],
      "text/plain": [
       "| 7         | 20.043146    | 16.711281          | 2.142576      | 17.030731            | 2.604029        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 8         | 23.219328    | 16.302317          | 2.130976      | 17.131910            | 2.596956        |</pre>"
      ],
      "text/plain": [
       "| 8         | 23.219328    | 16.302317          | 2.130976      | 17.131910            | 2.596956        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 9         | 25.693469    | 16.291615          | 2.132618      | 17.132753            | 2.590405        |</pre>"
      ],
      "text/plain": [
       "| 9         | 25.693469    | 16.291615          | 2.132618      | 17.132753            | 2.590405        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 10        | 28.270617    | 16.269194          | 2.131503      | 17.089596            | 2.585802        |</pre>"
      ],
      "text/plain": [
       "| 10        | 28.270617    | 16.269194          | 2.131503      | 17.089596            | 2.585802        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 11        | 30.712756    | 16.260521          | 2.134023      | 17.044395            | 2.578988        |</pre>"
      ],
      "text/plain": [
       "| 11        | 30.712756    | 16.260521          | 2.134023      | 17.044395            | 2.578988        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 12        | 33.305905    | 16.250105          | 2.136009      | 17.045340            | 2.577374        |</pre>"
      ],
      "text/plain": [
       "| 12        | 33.305905    | 16.250105          | 2.136009      | 17.045340            | 2.577374        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 13        | 36.281075    | 16.275871          | 2.129787      | 17.001801            | 2.576644        |</pre>"
      ],
      "text/plain": [
       "| 13        | 36.281075    | 16.275871          | 2.129787      | 17.001801            | 2.576644        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 14        | 39.452256    | 15.880025          | 2.121662      | 17.048691            | 2.575539        |</pre>"
      ],
      "text/plain": [
       "| 14        | 39.452256    | 15.880025          | 2.121662      | 17.048691            | 2.575539        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 15        | 42.464429    | 16.009933          | 2.115747      | 17.008194            | 2.574479        |</pre>"
      ],
      "text/plain": [
       "| 15        | 42.464429    | 16.009933          | 2.115747      | 17.008194            | 2.574479        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 16        | 45.674612    | 15.571314          | 2.110360      | 17.022228            | 2.573631        |</pre>"
      ],
      "text/plain": [
       "| 16        | 45.674612    | 15.571314          | 2.110360      | 17.022228            | 2.573631        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 17        | 48.303763    | 15.603245          | 2.109923      | 16.980274            | 2.570976        |</pre>"
      ],
      "text/plain": [
       "| 17        | 48.303763    | 15.603245          | 2.109923      | 16.980274            | 2.570976        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 18        | 50.631896    | 15.692157          | 2.112714      | 16.992147            | 2.568521        |</pre>"
      ],
      "text/plain": [
       "| 18        | 50.631896    | 15.692157          | 2.112714      | 16.992147            | 2.568521        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 19        | 53.302049    | 15.757780          | 2.114469      | 16.921322            | 2.566855        |</pre>"
      ],
      "text/plain": [
       "| 19        | 53.302049    | 15.757780          | 2.114469      | 16.921322            | 2.566855        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 20        | 56.262218    | 15.794964          | 2.110921      | 16.914885            | 2.565497        |</pre>"
      ],
      "text/plain": [
       "| 20        | 56.262218    | 15.794964          | 2.110921      | 16.914885            | 2.565497        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "RF_model1 = graphlab.random_forest_regression.create(dataCut1train, target='ADLOADINGTIME',validation_set=dataCut1test,\n",
    "                                           max_iterations=20,\n",
    "                                           max_depth =  50, row_subsample=0.5, random_seed=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre>WARNING: The number of feature dimensions in this problem is very large in comparison with the number of examples. Unless an appropriate regularization value is set, this model may not provide accurate predictions for a validation/test set.</pre>"
      ],
      "text/plain": [
       "WARNING: The number of feature dimensions in this problem is very large in comparison with the number of examples. Unless an appropriate regularization value is set, this model may not provide accurate predictions for a validation/test set."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset.</pre>"
      ],
      "text/plain": [
       "WARNING: Detected extremely low variance for feature(s) 'GROUPDIVIDER_LOGICAL' because all entries are nearly the same.\n",
       "Proceeding with model training using all features. If the model does not provide results of adequate quality, exclude the above mentioned feature(s) from the input dataset."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Random forest regression:</pre>"
      ],
      "text/plain": [
       "Random forest regression:"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>--------------------------------------------------------</pre>"
      ],
      "text/plain": [
       "--------------------------------------------------------"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of examples          : 11917</pre>"
      ],
      "text/plain": [
       "Number of examples          : 11917"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of features          : 22</pre>"
      ],
      "text/plain": [
       "Number of features          : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>Number of unpacked features : 22</pre>"
      ],
      "text/plain": [
       "Number of unpacked features : 22"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| Iteration | Elapsed Time | Training-max_error | Training-rmse | Validation-max_error | Validation-rmse |</pre>"
      ],
      "text/plain": [
       "| Iteration | Elapsed Time | Training-max_error | Training-rmse | Validation-max_error | Validation-rmse |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 1         | 0.267016     | 17.151333          | 3.980833      | 17.617001            | 4.692649        |</pre>"
      ],
      "text/plain": [
       "| 1         | 0.267016     | 17.151333          | 3.980833      | 17.617001            | 4.692649        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 2         | 0.506029     | 14.262667          | 3.639329      | 17.318333            | 4.462216        |</pre>"
      ],
      "text/plain": [
       "| 2         | 0.506029     | 14.262667          | 3.639329      | 17.318333            | 4.462216        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 3         | 0.741043     | 14.275021          | 3.530604      | 17.314888            | 4.369804        |</pre>"
      ],
      "text/plain": [
       "| 3         | 0.741043     | 14.275021          | 3.530604      | 17.314888            | 4.369804        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 4         | 0.986057     | 13.607147          | 3.451920      | 17.318333            | 4.325181        |</pre>"
      ],
      "text/plain": [
       "| 4         | 0.986057     | 13.607147          | 3.451920      | 17.318333            | 4.325181        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 5         | 1.230071     | 13.816278          | 3.433239      | 17.316267            | 4.287280        |</pre>"
      ],
      "text/plain": [
       "| 5         | 1.230071     | 13.816278          | 3.433239      | 17.316267            | 4.287280        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 6         | 1.493086     | 14.128733          | 3.390232      | 17.303972            | 4.269917        |</pre>"
      ],
      "text/plain": [
       "| 6         | 1.493086     | 14.128733          | 3.390232      | 17.303972            | 4.269917        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 10        | 2.500143     | 13.361681          | 3.323706      | 16.379873            | 4.228616        |</pre>"
      ],
      "text/plain": [
       "| 10        | 2.500143     | 13.361681          | 3.323706      | 16.379873            | 4.228616        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 11        | 2.752158     | 12.797817          | 3.310189      | 16.466127            | 4.224926        |</pre>"
      ],
      "text/plain": [
       "| 11        | 2.752158     | 12.797817          | 3.310189      | 16.466127            | 4.224926        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 15        | 3.764216     | 13.420606          | 3.289249      | 15.978630            | 4.207051        |</pre>"
      ],
      "text/plain": [
       "| 15        | 3.764216     | 13.420606          | 3.289249      | 15.978630            | 4.207051        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>| 20        | 5.035288     | 12.427284          | 3.286034      | 14.949943            | 4.204785        |</pre>"
      ],
      "text/plain": [
       "| 20        | 5.035288     | 12.427284          | 3.286034      | 14.949943            | 4.204785        |"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre>+-----------+--------------+--------------------+---------------+----------------------+-----------------+</pre>"
      ],
      "text/plain": [
       "+-----------+--------------+--------------------+---------------+----------------------+-----------------+"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "RF_model2 = graphlab.random_forest_regression.create(dataCut2train, target='ADLOADINGTIME',validation_set=dataCut2test,\n",
    "                                           max_iterations=20,\n",
    "                                           max_depth =  100, row_subsample=0.5, random_seed=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.56549679472\n",
      "4.20478495188\n",
      "2.77736364955\n"
     ]
    }
   ],
   "source": [
    "pred1 = RF_model1.predict(dataCut1test)\n",
    "pred2 = RF_model2.predict(dataCut2test)\n",
    "print(graphlab.evaluation.rmse(dataCut1test['ADLOADINGTIME'],pred1))\n",
    "print(graphlab.evaluation.rmse(dataCut2test['ADLOADINGTIME'],pred2))\n",
    "      \n",
    "testFinal = pred1.append(pred2)\n",
    "print(graphlab.evaluation.rmse(dataCut1test['ADLOADINGTIME'].append(dataCut2test['ADLOADINGTIME']),testFinal))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEACAYAAACznAEdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1cc0cbe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(dataCut1test['ADLOADINGTIME'],  bins=np.arange(0, int(max(dataCut1test['ADLOADINGTIME'])),0.2),alpha=0.5);\n",
    "test_data_predict = RF_model1.predict(dataCut1test)\n",
    "plt.hist(test_data_predict,  bins=np.arange(0, int(max(dataCut1test['ADLOADINGTIME'])),0.2), alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEACAYAAAC08h1NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x82521a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(dataCut2test['ADLOADINGTIME'],  bins=np.arange(0, int(max(dataCut2test['ADLOADINGTIME'])),0.2),alpha=0.5);\n",
    "test_data_predict = RF_model2.predict(dataCut2test)\n",
    "plt.hist(test_data_predict,  bins=np.arange(0, int(max(dataCut2test['ADLOADINGTIME'])),0.2), alpha=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x87e189e8>]"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAEACAYAAABF+UbAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhHAwSeJCCOFg/wf1CmoTn1CngQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1cc0c128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_data_predict = RF_model2.predict(dataCut2test)\n",
    "\n",
    "#test_data['ADLOADINGTIME']\n",
    "\n",
    "pred_error = np.abs(dataCut2test['ADLOADINGTIME']-test_data_predict)\n",
    "\n",
    "[histvals, histbins] = np.histogram(dataCut2test['ADLOADINGTIME'], bins=np.arange(0, np.ceil(max(dataCut2test['ADLOADINGTIME'])),0.5))\n",
    "minToBin = histbins-np.tile(np.array(dataCut2test['ADLOADINGTIME']),(len(histbins),1)).transpose()\n",
    "minToBinIdx = np.argmin(np.abs(minToBin),axis=1)\n",
    "\n",
    "pred_error_loc = np.zeros_like(histbins)\n",
    "pred_error_loc_n = np.zeros_like(histbins)\n",
    "for i in range(len(minToBinIdx)):\n",
    "    pred_error_loc[minToBinIdx[i]] += pred_error[i]\n",
    "    pred_error_loc_n[minToBinIdx[i]] += 1\n",
    "    \n",
    "plt.plot(histbins, pred_error_loc/pred_error_loc_n)\n",
    "pred_error_loc1 = pred_error_loc\n",
    "pred_error_loc_n1 = pred_error_loc_n\n",
    "test_data_predict = RF_model1.predict(dataCut1test)\n",
    "\n",
    "#test_data['ADLOADINGTIME']\n",
    "\n",
    "pred_error = np.abs(dataCut1test['ADLOADINGTIME']-test_data_predict)\n",
    "\n",
    "[histvals, histbins] = np.histogram(dataCut1test['ADLOADINGTIME'], bins=np.arange(0, np.ceil(max(dataCut1test['ADLOADINGTIME'])),0.5))\n",
    "minToBin = histbins-np.tile(np.array(dataCut1test['ADLOADINGTIME']),(len(histbins),1)).transpose()\n",
    "minToBinIdx = np.argmin(np.abs(minToBin),axis=1)\n",
    "\n",
    "pred_error_loc = np.zeros_like(histbins)\n",
    "pred_error_loc_n = np.zeros_like(histbins)\n",
    "for i in range(len(minToBinIdx)):\n",
    "    pred_error_loc[minToBinIdx[i]] += pred_error[i]\n",
    "    pred_error_loc_n[minToBinIdx[i]] += 1\n",
    "    \n",
    "plt.plot(histbins, pred_error_loc/pred_error_loc_n)\n",
    "plt.plot(histbins, (pred_error_loc+pred_error_loc1)/(pred_error_loc_n+pred_error_loc_n1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8.928173233270796"
      ]
     },
     "execution_count": 217,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataCut2test['ADLOADINGTIME'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##Predicting and saving"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([  0.00000000e+00,   4.06000000e+02,   3.56300000e+03,\n",
       "          1.53920000e+04,   2.01900000e+04,   1.80360000e+04,\n",
       "          1.26090000e+04,   8.98800000e+03,   6.67200000e+03,\n",
       "          5.46400000e+03,   4.54000000e+03,   3.95500000e+03,\n",
       "          3.23800000e+03,   2.89000000e+03,   2.59000000e+03,\n",
       "          2.14200000e+03,   1.42100000e+03,   1.01800000e+03,\n",
       "          6.89000000e+02,   5.88000000e+02,   6.07000000e+02,\n",
       "          5.46000000e+02,   5.48000000e+02,   4.14000000e+02,\n",
       "          7.30000000e+01,   5.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   0.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   0.00000000e+00,   0.00000000e+00,\n",
       "          0.00000000e+00,   0.00000000e+00]),\n",
       " array([  0. ,   0.5,   1. ,   1.5,   2. ,   2.5,   3. ,   3.5,   4. ,\n",
       "          4.5,   5. ,   5.5,   6. ,   6.5,   7. ,   7.5,   8. ,   8.5,\n",
       "          9. ,   9.5,  10. ,  10.5,  11. ,  11.5,  12. ,  12.5,  13. ,\n",
       "         13.5,  14. ,  14.5,  15. ,  15.5,  16. ,  16.5,  17. ,  17.5]),\n",
       " <a list of 35 Patch objects>)"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEACAYAAABcXmojAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkCS1DAVJUuv/Axy2QXUxnjszAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x825b7a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(test_data['ADLOADINGTIME'],  bins=np.arange(0, int(max(test_data['ADLOADINGTIME'])),0.5),alpha=0.5)\n",
    "test_data_predict = RF_model.predict(test_data)\n",
    "plt.hist(test_data_predict,  bins=np.arange(0, int(max(test_data['ADLOADINGTIME'])),0.5), alpha=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sqft_model.save('E:\\celtraMiningFrik\\lin_reg_creativeid_perct_97')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sqft_model = graphlab.load_model('E:\\celtraMiningFrik\\lin_reg_creativeid_perct_96')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.72755946569\n"
     ]
    }
   ],
   "source": [
    "print(graphlab.evaluation.rmse(test_data['ADLOADINGTIME'], graphlab.SArray.from_const(3.90365759806,len(test_data['ADLOADINGTIME']))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Canvas is accessible via web browser at the URL: http://localhost:59632/index.html\n",
      "Opening Canvas in default web browser.\n"
     ]
    }
   ],
   "source": [
    "train_data.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predTimes = sqft_model.predict(sales)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "predTimes.save('E:\\celtraMiningFrik\\solution4.tsv', format='csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}