Visa_predict_undersample_ensemble.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sqlalchemy import create_engine\n",
    "from sqlalchemy_utils import database_exists, create_database\n",
    "import psycopg2\n",
    "import pandas as pd\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import accuracy_score, recall_score, precision_score\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn import preprocessing\n",
    "from lime.lime_tabular import LimeTabularExplainer\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.pipeline import make_pipeline, Pipeline\n",
    "from sklearn.preprocessing import StandardScaler, OneHotEncoder, LabelEncoder\n",
    "from sklearn.pipeline import FeatureUnion\n",
    "from sklearn.base import BaseEstimator, TransformerMixin\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import f1_score\n",
    "\n",
    "from lime.lime_tabular import LimeTabularExplainer\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "\n",
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load data with headers\n",
    "\n",
    "df = pd.read_csv('/Users/tarabaris/Insight_Data_Science/project/perm_visa_all_years_C_D_limes.csv')\n",
    "\n",
    "# Order features (numeric first, categorical second) since it's convenient later\n",
    "feature_order = ['salary',\n",
    "                 'state',\n",
    "                 'job_title',\n",
    "                 'level',\n",
    "                 'education',\n",
    "                 'citizenship',\n",
    "                 'admission',\n",
    "                 'ownership',\n",
    "                 'language',\n",
    "                 'outcome']\n",
    "\n",
    "\n",
    "data = df[feature_order]\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Class to help select categorical vs. continuous data as part of the pipeline (see below)\n",
    "class DataSelector(BaseEstimator, TransformerMixin):\n",
    "    '''Select columns of numpy arrays based on attribute_indices.'''\n",
    "\n",
    "    def __init__(self, attribute_indices):\n",
    "        self.attribute_indices = attribute_indices\n",
    "\n",
    "    def fit(self, X, y=None):\n",
    "        return self\n",
    "\n",
    "    def transform(self, X):\n",
    "        return np.array(X)[:,self.attribute_indices]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>salary</th>\n",
       "      <th>state</th>\n",
       "      <th>job_title</th>\n",
       "      <th>level</th>\n",
       "      <th>education</th>\n",
       "      <th>citizenship</th>\n",
       "      <th>admission</th>\n",
       "      <th>ownership</th>\n",
       "      <th>language</th>\n",
       "      <th>outcome</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>83366.0</td>\n",
       "      <td>WI</td>\n",
       "      <td>Industrial_Engineers</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>49982.0</td>\n",
       "      <td>CO</td>\n",
       "      <td>Biological_Scientists_All_Other</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Doctorate</td>\n",
       "      <td>GERMANY</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>43514.0</td>\n",
       "      <td>WA</td>\n",
       "      <td>Market_Research_Analysts_and_Marketing_Special...</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Master's</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>E-2</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    salary state                                          job_title     level  \\\n",
       "0  83366.0    WI                               Industrial_Engineers  Level IV   \n",
       "1  16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   Level I   \n",
       "2  49982.0    CO                    Biological_Scientists_All_Other   Level I   \n",
       "3  43514.0    WA  Market_Research_Analysts_and_Marketing_Special...   Level I   \n",
       "4  16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   Level I   \n",
       "\n",
       "    education  citizenship admission ownership language  outcome  \n",
       "0  Bachelor's        INDIA      H-1B         N        N      0.0  \n",
       "1        None  SOUTH KOREA       NaN         N        N      1.0  \n",
       "2   Doctorate      GERMANY      H-1B         N        N      1.0  \n",
       "3    Master's  SOUTH KOREA       E-2         N        N      0.0  \n",
       "4        None  SOUTH KOREA       NaN         N        N      1.0  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(230233, 10)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "salary         float64\n",
       "state           object\n",
       "job_title       object\n",
       "level           object\n",
       "education       object\n",
       "citizenship     object\n",
       "admission       object\n",
       "ownership       object\n",
       "language        object\n",
       "outcome        float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>salary</th>\n",
       "      <th>state</th>\n",
       "      <th>job_title</th>\n",
       "      <th>level</th>\n",
       "      <th>education</th>\n",
       "      <th>citizenship</th>\n",
       "      <th>admission</th>\n",
       "      <th>ownership</th>\n",
       "      <th>language</th>\n",
       "      <th>outcome</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>83366.0</td>\n",
       "      <td>WI</td>\n",
       "      <td>Industrial_Engineers</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>49982.0</td>\n",
       "      <td>CO</td>\n",
       "      <td>Biological_Scientists_All_Other</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Doctorate</td>\n",
       "      <td>GERMANY</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>43514.0</td>\n",
       "      <td>WA</td>\n",
       "      <td>Market_Research_Analysts_and_Marketing_Special...</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Master's</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>E-2</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>24710.0</td>\n",
       "      <td>MI</td>\n",
       "      <td>Cooks_Restaurant</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>F-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>55682.0</td>\n",
       "      <td>NY</td>\n",
       "      <td>Public_Relations_Specialists</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>CANADA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>72467.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Computer_Systems_Analysts</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>46301.0</td>\n",
       "      <td>KS</td>\n",
       "      <td>Computer_Programmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>53560.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Market_Research_Analysts_and_Marketing_Special...</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>BRAZIL</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>16640.0</td>\n",
       "      <td>SC</td>\n",
       "      <td>Janitors_and_Cleaners_Except_Maids_and_Houseke...</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>CHINA</td>\n",
       "      <td>Not_in_USA</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>72467.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Computer_Systems_Analysts</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>98675.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>BELARUS</td>\n",
       "      <td>L-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>99424.0</td>\n",
       "      <td>WA</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level III</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>98675.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>CHINA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>16973.0</td>\n",
       "      <td>MS</td>\n",
       "      <td>Meat_Poultry_and_Fish_Cutters_and_Trimmers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>69600.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Computer_Systems_Analysts</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>60944.0</td>\n",
       "      <td>MO</td>\n",
       "      <td>Financial_Analysts</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>CHINA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>97032.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>Computer_Systems_Analysts</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>GERMANY</td>\n",
       "      <td>L-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>104790.0</td>\n",
       "      <td>NJ</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>PAKISTAN</td>\n",
       "      <td>F-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>122658.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Software_Developers_Systems_Software</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>visa_parolee</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>46093.0</td>\n",
       "      <td>GA</td>\n",
       "      <td>Farmers_Ranchers_and_Other_Agricultural_Managers</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>GUATEMALA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>72467.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Computer_Systems_Analysts</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>50024.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>Financial_Analysts</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Master's</td>\n",
       "      <td>CHINA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>26000.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Cooks_Restaurant</td>\n",
       "      <td>Level II</td>\n",
       "      <td>None</td>\n",
       "      <td>MEXICO</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>54059.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Other</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230203</th>\n",
       "      <td>113381.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Software_Developers_Systems_Software</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>F-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230204</th>\n",
       "      <td>102877.0</td>\n",
       "      <td>IL</td>\n",
       "      <td>Chemists</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Master's</td>\n",
       "      <td>CANADA</td>\n",
       "      <td>TN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230205</th>\n",
       "      <td>124093.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Computer_Occupations_All_Other</td>\n",
       "      <td>Level III</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230206</th>\n",
       "      <td>82805.0</td>\n",
       "      <td>NJ</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230207</th>\n",
       "      <td>32781.0</td>\n",
       "      <td>UT</td>\n",
       "      <td>Fine_Artists_Including_Painters_Sculptors_and_...</td>\n",
       "      <td>Level III</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>SOUTH KOREA</td>\n",
       "      <td>B-2</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230208</th>\n",
       "      <td>66997.0</td>\n",
       "      <td>MO</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230209</th>\n",
       "      <td>110448.0</td>\n",
       "      <td>NC</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230210</th>\n",
       "      <td>101566.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level III</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230211</th>\n",
       "      <td>85862.0</td>\n",
       "      <td>TX</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>CHINA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230212</th>\n",
       "      <td>113381.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Software_Developers_Systems_Software</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>IRAN</td>\n",
       "      <td>F-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230213</th>\n",
       "      <td>67288.0</td>\n",
       "      <td>MD</td>\n",
       "      <td>Network_and_Computer_Systems_Administrators*</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>TURKMENISTAN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230214</th>\n",
       "      <td>100006.0</td>\n",
       "      <td>GA</td>\n",
       "      <td>Software_Developers_Systems_Software</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230215</th>\n",
       "      <td>99195.0</td>\n",
       "      <td>MA</td>\n",
       "      <td>Software_Developers_Systems_Software</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>CHINA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230216</th>\n",
       "      <td>94203.0</td>\n",
       "      <td>MA</td>\n",
       "      <td>Software_Developers_Systems_Software</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>MEXICO</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230217</th>\n",
       "      <td>103168.0</td>\n",
       "      <td>GA</td>\n",
       "      <td>Electrical_Engineers</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>MEXICO</td>\n",
       "      <td>L-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230218</th>\n",
       "      <td>127379.0</td>\n",
       "      <td>VA</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230219</th>\n",
       "      <td>124675.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230220</th>\n",
       "      <td>46904.0</td>\n",
       "      <td>FL</td>\n",
       "      <td>Software_Developers_Applications_Non_R_&amp;_D_(AC...</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>ITALY</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230221</th>\n",
       "      <td>64272.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>Procurement_Clerks</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>GERMANY</td>\n",
       "      <td>L-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230222</th>\n",
       "      <td>17139.0</td>\n",
       "      <td>LA</td>\n",
       "      <td>Cleaners_of_Vehicles_and_Equipment</td>\n",
       "      <td>Level I</td>\n",
       "      <td>None</td>\n",
       "      <td>VIETNAM</td>\n",
       "      <td>Not_in_USA</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230223</th>\n",
       "      <td>93163.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>CHINA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230224</th>\n",
       "      <td>152131.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Electronics_Engineers_Except_Computer</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230225</th>\n",
       "      <td>87422.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Electronics_Engineers_Except_Computer</td>\n",
       "      <td>Level I</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230226</th>\n",
       "      <td>109845.0</td>\n",
       "      <td>NY</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230227</th>\n",
       "      <td>123490.0</td>\n",
       "      <td>NJ</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230228</th>\n",
       "      <td>84635.0</td>\n",
       "      <td>KY</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230229</th>\n",
       "      <td>69098.0</td>\n",
       "      <td>OH</td>\n",
       "      <td>Software_Developers_Applications</td>\n",
       "      <td>Level II</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230230</th>\n",
       "      <td>114795.0</td>\n",
       "      <td>CA</td>\n",
       "      <td>Computer_Systems_Analysts</td>\n",
       "      <td>Level III</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>L-1</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230231</th>\n",
       "      <td>112354.0</td>\n",
       "      <td>NY</td>\n",
       "      <td>Information_Security_Analysts</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Master's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>230232</th>\n",
       "      <td>136219.0</td>\n",
       "      <td>NJ</td>\n",
       "      <td>Management_Analysts</td>\n",
       "      <td>Level IV</td>\n",
       "      <td>Bachelor's</td>\n",
       "      <td>INDIA</td>\n",
       "      <td>H-1B</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>230233 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          salary state                                          job_title  \\\n",
       "0        83366.0    WI                               Industrial_Engineers   \n",
       "1        16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   \n",
       "2        49982.0    CO                    Biological_Scientists_All_Other   \n",
       "3        43514.0    WA  Market_Research_Analysts_and_Marketing_Special...   \n",
       "4        16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   \n",
       "5        24710.0    MI                                   Cooks_Restaurant   \n",
       "6        55682.0    NY                       Public_Relations_Specialists   \n",
       "7        16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   \n",
       "8        16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   \n",
       "9        16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   \n",
       "10       72467.0    TX                          Computer_Systems_Analysts   \n",
       "11       46301.0    KS                               Computer_Programmers   \n",
       "12       53560.0    TX  Market_Research_Analysts_and_Marketing_Special...   \n",
       "13       16640.0    SC  Janitors_and_Cleaners_Except_Maids_and_Houseke...   \n",
       "14       72467.0    TX                          Computer_Systems_Analysts   \n",
       "15       16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   \n",
       "16       98675.0    PA                   Software_Developers_Applications   \n",
       "17       99424.0    WA                   Software_Developers_Applications   \n",
       "18       98675.0    CA                   Software_Developers_Applications   \n",
       "19       16973.0    MS         Meat_Poultry_and_Fish_Cutters_and_Trimmers   \n",
       "20       69600.0    TX                          Computer_Systems_Analysts   \n",
       "21       60944.0    MO                                 Financial_Analysts   \n",
       "22       97032.0    PA                          Computer_Systems_Analysts   \n",
       "23      104790.0    NJ                   Software_Developers_Applications   \n",
       "24      122658.0    CA               Software_Developers_Systems_Software   \n",
       "25       46093.0    GA   Farmers_Ranchers_and_Other_Agricultural_Managers   \n",
       "26       72467.0    TX                          Computer_Systems_Analysts   \n",
       "27       50024.0    PA                                 Financial_Analysts   \n",
       "28       26000.0    CA                                   Cooks_Restaurant   \n",
       "29       54059.0    TX                   Software_Developers_Applications   \n",
       "...          ...   ...                                                ...   \n",
       "230203  113381.0    CA               Software_Developers_Systems_Software   \n",
       "230204  102877.0    IL                                           Chemists   \n",
       "230205  124093.0    CA                     Computer_Occupations_All_Other   \n",
       "230206   82805.0    NJ                   Software_Developers_Applications   \n",
       "230207   32781.0    UT  Fine_Artists_Including_Painters_Sculptors_and_...   \n",
       "230208   66997.0    MO                   Software_Developers_Applications   \n",
       "230209  110448.0    NC                   Software_Developers_Applications   \n",
       "230210  101566.0    TX                   Software_Developers_Applications   \n",
       "230211   85862.0    TX                   Software_Developers_Applications   \n",
       "230212  113381.0    CA               Software_Developers_Systems_Software   \n",
       "230213   67288.0    MD       Network_and_Computer_Systems_Administrators*   \n",
       "230214  100006.0    GA               Software_Developers_Systems_Software   \n",
       "230215   99195.0    MA               Software_Developers_Systems_Software   \n",
       "230216   94203.0    MA               Software_Developers_Systems_Software   \n",
       "230217  103168.0    GA                               Electrical_Engineers   \n",
       "230218  127379.0    VA                   Software_Developers_Applications   \n",
       "230219  124675.0    CA                   Software_Developers_Applications   \n",
       "230220   46904.0    FL  Software_Developers_Applications_Non_R_&_D_(AC...   \n",
       "230221   64272.0    PA                                 Procurement_Clerks   \n",
       "230222   17139.0    LA                 Cleaners_of_Vehicles_and_Equipment   \n",
       "230223   93163.0    CA                   Software_Developers_Applications   \n",
       "230224  152131.0    CA              Electronics_Engineers_Except_Computer   \n",
       "230225   87422.0    CA              Electronics_Engineers_Except_Computer   \n",
       "230226  109845.0    NY                   Software_Developers_Applications   \n",
       "230227  123490.0    NJ                   Software_Developers_Applications   \n",
       "230228   84635.0    KY                   Software_Developers_Applications   \n",
       "230229   69098.0    OH                   Software_Developers_Applications   \n",
       "230230  114795.0    CA                          Computer_Systems_Analysts   \n",
       "230231  112354.0    NY                      Information_Security_Analysts   \n",
       "230232  136219.0    NJ                                Management_Analysts   \n",
       "\n",
       "            level   education   citizenship     admission ownership language  \\\n",
       "0        Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "1         Level I        None   SOUTH KOREA           NaN         N        N   \n",
       "2         Level I   Doctorate       GERMANY          H-1B         N        N   \n",
       "3         Level I    Master's   SOUTH KOREA           E-2         N        N   \n",
       "4         Level I        None   SOUTH KOREA           NaN         N        N   \n",
       "5        Level IV        None   SOUTH KOREA           F-1         N        N   \n",
       "6        Level II  Bachelor's        CANADA          H-1B         N        N   \n",
       "7         Level I        None   SOUTH KOREA           NaN         N        N   \n",
       "8         Level I        None   SOUTH KOREA           NaN         N        N   \n",
       "9         Level I        None   SOUTH KOREA           NaN         N        N   \n",
       "10       Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "11        Level I  Bachelor's         INDIA           NaN         N        N   \n",
       "12       Level II  Bachelor's        BRAZIL          H-1B         N        N   \n",
       "13        Level I        None         CHINA    Not_in_USA         N        N   \n",
       "14       Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "15        Level I        None   SOUTH KOREA           NaN         N        N   \n",
       "16       Level II    Master's       BELARUS           L-1         N        N   \n",
       "17      Level III  Bachelor's         INDIA          H-1B         N        N   \n",
       "18       Level II    Master's         CHINA          H-1B         N        N   \n",
       "19        Level I        None   SOUTH KOREA           NaN         N        N   \n",
       "20       Level II    Master's         INDIA          H-1B         N        N   \n",
       "21       Level II    Master's         CHINA          H-1B         N        N   \n",
       "22       Level IV  Bachelor's       GERMANY           L-1         N        N   \n",
       "23       Level IV  Bachelor's      PAKISTAN           F-1         N        N   \n",
       "24       Level IV    Master's         INDIA  visa_parolee         N        N   \n",
       "25        Level I  Bachelor's     GUATEMALA          H-1B         N        N   \n",
       "26       Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "27        Level I    Master's         CHINA          H-1B         N        N   \n",
       "28       Level II        None        MEXICO           NaN         N        N   \n",
       "29       Level II       Other         INDIA          H-1B         N        N   \n",
       "...           ...         ...           ...           ...       ...      ...   \n",
       "230203   Level II    Master's         INDIA           F-1         N        N   \n",
       "230204   Level IV    Master's        CANADA            TN         N        N   \n",
       "230205  Level III  Bachelor's         INDIA          H-1B         N        N   \n",
       "230206   Level II  Bachelor's         INDIA          H-1B         N        N   \n",
       "230207  Level III  Bachelor's   SOUTH KOREA           B-2         N        N   \n",
       "230208    Level I    Master's         INDIA          H-1B         N        N   \n",
       "230209   Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "230210  Level III    Master's         INDIA          H-1B         N        N   \n",
       "230211   Level II    Master's         CHINA          H-1B         N        N   \n",
       "230212   Level II    Master's          IRAN           F-1         N        N   \n",
       "230213    Level I  Bachelor's  TURKMENISTAN           NaN         N        N   \n",
       "230214   Level IV    Master's         INDIA          H-1B         N        N   \n",
       "230215   Level II    Master's         CHINA          H-1B         N        N   \n",
       "230216   Level II  Bachelor's        MEXICO          H-1B         N        N   \n",
       "230217   Level IV  Bachelor's        MEXICO           L-1         N        N   \n",
       "230218   Level IV    Master's         INDIA          H-1B         N        N   \n",
       "230219   Level II  Bachelor's         INDIA          H-1B         N        N   \n",
       "230220    Level I  Bachelor's         ITALY          H-1B         N        N   \n",
       "230221   Level IV  Bachelor's       GERMANY           L-1         N        N   \n",
       "230222    Level I        None       VIETNAM    Not_in_USA         N        N   \n",
       "230223   Level II    Master's         CHINA          H-1B         N        N   \n",
       "230224   Level IV    Master's         INDIA          H-1B         N        N   \n",
       "230225    Level I  Bachelor's         INDIA          H-1B         N        N   \n",
       "230226   Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "230227   Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "230228   Level IV    Master's         INDIA          H-1B         N        N   \n",
       "230229   Level II    Master's         INDIA          H-1B         N        N   \n",
       "230230  Level III    Master's         INDIA           L-1         N        N   \n",
       "230231        NaN    Master's         INDIA          H-1B         N        N   \n",
       "230232   Level IV  Bachelor's         INDIA          H-1B         N        N   \n",
       "\n",
       "        outcome  \n",
       "0           0.0  \n",
       "1           1.0  \n",
       "2           1.0  \n",
       "3           0.0  \n",
       "4           1.0  \n",
       "5           0.0  \n",
       "6           0.0  \n",
       "7           1.0  \n",
       "8           1.0  \n",
       "9           1.0  \n",
       "10          0.0  \n",
       "11          0.0  \n",
       "12          0.0  \n",
       "13          0.0  \n",
       "14          0.0  \n",
       "15          1.0  \n",
       "16          0.0  \n",
       "17          0.0  \n",
       "18          0.0  \n",
       "19          1.0  \n",
       "20          0.0  \n",
       "21          0.0  \n",
       "22          0.0  \n",
       "23          0.0  \n",
       "24          0.0  \n",
       "25          0.0  \n",
       "26          0.0  \n",
       "27          0.0  \n",
       "28          1.0  \n",
       "29          0.0  \n",
       "...         ...  \n",
       "230203      0.0  \n",
       "230204      0.0  \n",
       "230205      0.0  \n",
       "230206      0.0  \n",
       "230207      0.0  \n",
       "230208      0.0  \n",
       "230209      0.0  \n",
       "230210      0.0  \n",
       "230211      0.0  \n",
       "230212      0.0  \n",
       "230213      0.0  \n",
       "230214      0.0  \n",
       "230215      0.0  \n",
       "230216      0.0  \n",
       "230217      0.0  \n",
       "230218      0.0  \n",
       "230219      0.0  \n",
       "230220      0.0  \n",
       "230221      0.0  \n",
       "230222      0.0  \n",
       "230223      0.0  \n",
       "230224      0.0  \n",
       "230225      0.0  \n",
       "230226      0.0  \n",
       "230227      0.0  \n",
       "230228      0.0  \n",
       "230229      0.0  \n",
       "230230      0.0  \n",
       "230231      0.0  \n",
       "230232      0.0  \n",
       "\n",
       "[230233 rows x 10 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#We want to replace any infinity values with \"nan\" so that we can drop them along with any other \"nan\" in order to run \n",
    "#throught our model\n",
    "data.replace([np.inf, -np.inf], np.nan)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Here we drop 'nan', which in this case also includes the converte infiinity values\n",
    "data = data.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([u'salary', u'state', u'job_title', u'level', u'education',\n",
       "       u'citizenship', u'admission', u'ownership', u'language', u'outcome'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/tarabaris/anaconda/envs/my_projects_env/lib/python2.7/site-packages/sklearn/model_selection/_split.py:2010: FutureWarning: From version 0.21, test_size will always complement train_size unless both are specified.\n",
      "  FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "# Define categorical names and indices\n",
    "categorical_features = list(data.columns[1:])\n",
    "categorical_idx = list(range(1, len(data.columns)-1))\n",
    "continuous_features = list(data.columns[0:1])\n",
    "continuous_idx = list(range(0,1))\n",
    "\n",
    "Y = data.outcome.values\n",
    "X = data.drop(\"outcome\", axis=1).values\n",
    "\n",
    "# Get feature names and their values for categorical data (needed for LIME)\n",
    "categorical_names = {}\n",
    "for idx, feature in zip(categorical_idx, categorical_features):\n",
    "    le = LabelEncoder()\n",
    "    X[:, idx] = le.fit_transform(X[:, idx])\n",
    "    categorical_names[idx] = le.classes_\n",
    "\n",
    "# To suppress a warning later (not strictly necessary)\n",
    "X = X.astype(float)\n",
    "\n",
    "# Train test split\n",
    "x_train, x_test, y_train, y_test = train_test_split(\n",
    "    X, Y, train_size=0.70, random_state=42\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#In order to undersample, I separate out the data by the class, 1 is \"True\"/\"Denied\", and 0 is \"False\"/\"Certified\"\n",
    "X_true = x_train[y_train == 1]\n",
    "X_false = x_train[y_train == 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6608, 9)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_true.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(133680, 9)"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_false.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#I specify that I want the True class size to match the False class size when I oversample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_false_sample = X_false[np.random.choice(len(X_false), size=6608, replace=False)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6608, 9)"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_false_sample.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6608, 9)"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_true.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Split the majority set into 20 groups\n",
    "kf = KFold(n_splits=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "split_data = [X_false[split[1]] for split in kf.split(X_false)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n",
      "(6684, 9)\n"
     ]
    }
   ],
   "source": [
    "for split in kf.split(X_false):\n",
    "    print(X_false[split[1]].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Create the accompanying y values for the training sets\n",
    "train_x_sets = [np.concatenate([split, X_true]) for split in split_data]\n",
    "train_y_sets = [[0] * len(split) + [1]*len(X_true) for split in split_data]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# models = []\n",
    "# predictions = []\n",
    "# for xx, yy in zip(train_x_sets, train_y_sets):\n",
    "#     model = RandomForestClassifier()\n",
    "#     model.fit(xx, yy)\n",
    "#     pred = model.predict(x_test)\n",
    "#     models.append(model)\n",
    "#     predictions.append(pred)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run grid search to tune hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 48 candidates, totalling 240 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  1.0min\n",
      "[Parallel(n_jobs=-1)]: Done 184 tasks      | elapsed:  5.4min\n",
      "[Parallel(n_jobs=-1)]: Done 240 out of 240 | elapsed:  7.2min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=1,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False),\n",
       "       fit_params=None, iid=True, n_jobs=-1,\n",
       "       param_grid={'n_estimators': [200, 500, 1000], 'min_samples_split': [8, 12, 16, 20], 'max_depth': [25, 50, 75, 150]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring='f1', verbose=1)"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_grid = {'n_estimators': [200, 500, 1000],\n",
    "               'min_samples_split': [8, 12, 16, 20],\n",
    "               'max_depth': [25, 50, 75, 150]}\n",
    "\n",
    "#scoring=f1 to tune for better precision and recall\n",
    "\n",
    "rf = RandomForestClassifier()\n",
    "grid_search = GridSearchCV(rf, param_grid, cv=5, scoring='f1', n_jobs=-1, verbose=1)\n",
    "grid_search.fit(xx, yy)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mean_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>mean_train_score</th>\n",
       "      <th>param_max_depth</th>\n",
       "      <th>param_min_samples_split</th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>params</th>\n",
       "      <th>rank_test_score</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split0_train_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split1_train_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split2_train_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split3_train_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>split4_train_score</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>std_train_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>4.415908</td>\n",
       "      <td>0.165963</td>\n",
       "      <td>0.694561</td>\n",
       "      <td>0.890593</td>\n",
       "      <td>50</td>\n",
       "      <td>12</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 20...</td>\n",
       "      <td>1</td>\n",
       "      <td>0.698084</td>\n",
       "      <td>0.891538</td>\n",
       "      <td>0.704354</td>\n",
       "      <td>0.891104</td>\n",
       "      <td>0.682718</td>\n",
       "      <td>0.891692</td>\n",
       "      <td>0.693785</td>\n",
       "      <td>0.889869</td>\n",
       "      <td>0.693862</td>\n",
       "      <td>0.888761</td>\n",
       "      <td>0.233003</td>\n",
       "      <td>0.009319</td>\n",
       "      <td>0.007067</td>\n",
       "      <td>0.001117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>21.089066</td>\n",
       "      <td>0.929221</td>\n",
       "      <td>0.693942</td>\n",
       "      <td>0.892352</td>\n",
       "      <td>50</td>\n",
       "      <td>12</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 10...</td>\n",
       "      <td>2</td>\n",
       "      <td>0.695089</td>\n",
       "      <td>0.894129</td>\n",
       "      <td>0.700769</td>\n",
       "      <td>0.892624</td>\n",
       "      <td>0.684313</td>\n",
       "      <td>0.892682</td>\n",
       "      <td>0.693032</td>\n",
       "      <td>0.891675</td>\n",
       "      <td>0.696510</td>\n",
       "      <td>0.890652</td>\n",
       "      <td>0.166508</td>\n",
       "      <td>0.086053</td>\n",
       "      <td>0.005442</td>\n",
       "      <td>0.001157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>11.556772</td>\n",
       "      <td>0.439900</td>\n",
       "      <td>0.693899</td>\n",
       "      <td>0.920415</td>\n",
       "      <td>150</td>\n",
       "      <td>8</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 500...</td>\n",
       "      <td>3</td>\n",
       "      <td>0.697905</td>\n",
       "      <td>0.921277</td>\n",
       "      <td>0.701447</td>\n",
       "      <td>0.923283</td>\n",
       "      <td>0.682344</td>\n",
       "      <td>0.920680</td>\n",
       "      <td>0.688180</td>\n",
       "      <td>0.918395</td>\n",
       "      <td>0.699623</td>\n",
       "      <td>0.918438</td>\n",
       "      <td>0.193501</td>\n",
       "      <td>0.048327</td>\n",
       "      <td>0.007376</td>\n",
       "      <td>0.001846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>4.188209</td>\n",
       "      <td>0.160574</td>\n",
       "      <td>0.693752</td>\n",
       "      <td>0.890960</td>\n",
       "      <td>150</td>\n",
       "      <td>12</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 20...</td>\n",
       "      <td>4</td>\n",
       "      <td>0.696018</td>\n",
       "      <td>0.892378</td>\n",
       "      <td>0.702951</td>\n",
       "      <td>0.890381</td>\n",
       "      <td>0.679642</td>\n",
       "      <td>0.891235</td>\n",
       "      <td>0.691667</td>\n",
       "      <td>0.890340</td>\n",
       "      <td>0.698485</td>\n",
       "      <td>0.890465</td>\n",
       "      <td>0.195631</td>\n",
       "      <td>0.020484</td>\n",
       "      <td>0.007946</td>\n",
       "      <td>0.000781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>23.045973</td>\n",
       "      <td>0.878270</td>\n",
       "      <td>0.693726</td>\n",
       "      <td>0.920941</td>\n",
       "      <td>50</td>\n",
       "      <td>8</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 100...</td>\n",
       "      <td>5</td>\n",
       "      <td>0.693009</td>\n",
       "      <td>0.921915</td>\n",
       "      <td>0.701141</td>\n",
       "      <td>0.923269</td>\n",
       "      <td>0.681361</td>\n",
       "      <td>0.921224</td>\n",
       "      <td>0.694674</td>\n",
       "      <td>0.918753</td>\n",
       "      <td>0.698449</td>\n",
       "      <td>0.919545</td>\n",
       "      <td>0.097924</td>\n",
       "      <td>0.032627</td>\n",
       "      <td>0.006806</td>\n",
       "      <td>0.001624</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>21.478017</td>\n",
       "      <td>0.868469</td>\n",
       "      <td>0.693599</td>\n",
       "      <td>0.892433</td>\n",
       "      <td>150</td>\n",
       "      <td>12</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 10...</td>\n",
       "      <td>6</td>\n",
       "      <td>0.696947</td>\n",
       "      <td>0.894399</td>\n",
       "      <td>0.698462</td>\n",
       "      <td>0.891791</td>\n",
       "      <td>0.685625</td>\n",
       "      <td>0.892819</td>\n",
       "      <td>0.692337</td>\n",
       "      <td>0.891125</td>\n",
       "      <td>0.694624</td>\n",
       "      <td>0.892031</td>\n",
       "      <td>0.125898</td>\n",
       "      <td>0.044304</td>\n",
       "      <td>0.004497</td>\n",
       "      <td>0.001122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>10.687245</td>\n",
       "      <td>0.452287</td>\n",
       "      <td>0.693382</td>\n",
       "      <td>0.891686</td>\n",
       "      <td>75</td>\n",
       "      <td>12</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 50...</td>\n",
       "      <td>7</td>\n",
       "      <td>0.694127</td>\n",
       "      <td>0.893226</td>\n",
       "      <td>0.700809</td>\n",
       "      <td>0.892166</td>\n",
       "      <td>0.685848</td>\n",
       "      <td>0.892452</td>\n",
       "      <td>0.691005</td>\n",
       "      <td>0.890017</td>\n",
       "      <td>0.695122</td>\n",
       "      <td>0.890568</td>\n",
       "      <td>0.107898</td>\n",
       "      <td>0.022386</td>\n",
       "      <td>0.004923</td>\n",
       "      <td>0.001202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>10.982650</td>\n",
       "      <td>0.440209</td>\n",
       "      <td>0.693324</td>\n",
       "      <td>0.892077</td>\n",
       "      <td>50</td>\n",
       "      <td>12</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 50...</td>\n",
       "      <td>8</td>\n",
       "      <td>0.697639</td>\n",
       "      <td>0.894244</td>\n",
       "      <td>0.702848</td>\n",
       "      <td>0.892751</td>\n",
       "      <td>0.679172</td>\n",
       "      <td>0.892933</td>\n",
       "      <td>0.691610</td>\n",
       "      <td>0.890634</td>\n",
       "      <td>0.695354</td>\n",
       "      <td>0.889824</td>\n",
       "      <td>0.179170</td>\n",
       "      <td>0.044461</td>\n",
       "      <td>0.007959</td>\n",
       "      <td>0.001615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>10.993991</td>\n",
       "      <td>0.457164</td>\n",
       "      <td>0.693212</td>\n",
       "      <td>0.920001</td>\n",
       "      <td>75</td>\n",
       "      <td>8</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 500...</td>\n",
       "      <td>9</td>\n",
       "      <td>0.698941</td>\n",
       "      <td>0.921557</td>\n",
       "      <td>0.702846</td>\n",
       "      <td>0.922399</td>\n",
       "      <td>0.680341</td>\n",
       "      <td>0.920172</td>\n",
       "      <td>0.687195</td>\n",
       "      <td>0.916995</td>\n",
       "      <td>0.696740</td>\n",
       "      <td>0.918883</td>\n",
       "      <td>0.328056</td>\n",
       "      <td>0.052335</td>\n",
       "      <td>0.008246</td>\n",
       "      <td>0.001923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>12.035331</td>\n",
       "      <td>0.477261</td>\n",
       "      <td>0.692980</td>\n",
       "      <td>0.920289</td>\n",
       "      <td>50</td>\n",
       "      <td>8</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 500...</td>\n",
       "      <td>10</td>\n",
       "      <td>0.690937</td>\n",
       "      <td>0.920750</td>\n",
       "      <td>0.705123</td>\n",
       "      <td>0.922134</td>\n",
       "      <td>0.681818</td>\n",
       "      <td>0.919973</td>\n",
       "      <td>0.688697</td>\n",
       "      <td>0.919531</td>\n",
       "      <td>0.698328</td>\n",
       "      <td>0.919059</td>\n",
       "      <td>0.182710</td>\n",
       "      <td>0.053293</td>\n",
       "      <td>0.008040</td>\n",
       "      <td>0.001077</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>4.331092</td>\n",
       "      <td>0.172203</td>\n",
       "      <td>0.692849</td>\n",
       "      <td>0.890217</td>\n",
       "      <td>75</td>\n",
       "      <td>12</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 20...</td>\n",
       "      <td>11</td>\n",
       "      <td>0.699237</td>\n",
       "      <td>0.890126</td>\n",
       "      <td>0.696993</td>\n",
       "      <td>0.889059</td>\n",
       "      <td>0.684292</td>\n",
       "      <td>0.891425</td>\n",
       "      <td>0.690593</td>\n",
       "      <td>0.890930</td>\n",
       "      <td>0.693130</td>\n",
       "      <td>0.889547</td>\n",
       "      <td>0.213018</td>\n",
       "      <td>0.003773</td>\n",
       "      <td>0.005223</td>\n",
       "      <td>0.000868</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10.862266</td>\n",
       "      <td>0.412673</td>\n",
       "      <td>0.692691</td>\n",
       "      <td>0.887230</td>\n",
       "      <td>25</td>\n",
       "      <td>12</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 50...</td>\n",
       "      <td>12</td>\n",
       "      <td>0.697621</td>\n",
       "      <td>0.889103</td>\n",
       "      <td>0.700350</td>\n",
       "      <td>0.886550</td>\n",
       "      <td>0.677647</td>\n",
       "      <td>0.887770</td>\n",
       "      <td>0.693414</td>\n",
       "      <td>0.887413</td>\n",
       "      <td>0.694423</td>\n",
       "      <td>0.885312</td>\n",
       "      <td>0.231226</td>\n",
       "      <td>0.040667</td>\n",
       "      <td>0.007910</td>\n",
       "      <td>0.001263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>4.211588</td>\n",
       "      <td>0.177381</td>\n",
       "      <td>0.692655</td>\n",
       "      <td>0.866148</td>\n",
       "      <td>75</td>\n",
       "      <td>16</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 20...</td>\n",
       "      <td>13</td>\n",
       "      <td>0.695253</td>\n",
       "      <td>0.868336</td>\n",
       "      <td>0.700617</td>\n",
       "      <td>0.862711</td>\n",
       "      <td>0.678783</td>\n",
       "      <td>0.865767</td>\n",
       "      <td>0.698941</td>\n",
       "      <td>0.867320</td>\n",
       "      <td>0.689682</td>\n",
       "      <td>0.866609</td>\n",
       "      <td>0.078186</td>\n",
       "      <td>0.018825</td>\n",
       "      <td>0.007886</td>\n",
       "      <td>0.001915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.210468</td>\n",
       "      <td>0.198299</td>\n",
       "      <td>0.692631</td>\n",
       "      <td>0.912612</td>\n",
       "      <td>25</td>\n",
       "      <td>8</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 200...</td>\n",
       "      <td>14</td>\n",
       "      <td>0.696217</td>\n",
       "      <td>0.913761</td>\n",
       "      <td>0.698730</td>\n",
       "      <td>0.914120</td>\n",
       "      <td>0.681098</td>\n",
       "      <td>0.911586</td>\n",
       "      <td>0.688822</td>\n",
       "      <td>0.912825</td>\n",
       "      <td>0.698292</td>\n",
       "      <td>0.910769</td>\n",
       "      <td>0.023997</td>\n",
       "      <td>0.015800</td>\n",
       "      <td>0.006777</td>\n",
       "      <td>0.001272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>22.954586</td>\n",
       "      <td>0.964277</td>\n",
       "      <td>0.692546</td>\n",
       "      <td>0.920786</td>\n",
       "      <td>150</td>\n",
       "      <td>8</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 100...</td>\n",
       "      <td>15</td>\n",
       "      <td>0.695321</td>\n",
       "      <td>0.922236</td>\n",
       "      <td>0.702744</td>\n",
       "      <td>0.923283</td>\n",
       "      <td>0.681412</td>\n",
       "      <td>0.920356</td>\n",
       "      <td>0.688156</td>\n",
       "      <td>0.918571</td>\n",
       "      <td>0.695094</td>\n",
       "      <td>0.919483</td>\n",
       "      <td>0.142095</td>\n",
       "      <td>0.164548</td>\n",
       "      <td>0.007231</td>\n",
       "      <td>0.001739</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21.122218</td>\n",
       "      <td>0.921210</td>\n",
       "      <td>0.692185</td>\n",
       "      <td>0.915052</td>\n",
       "      <td>25</td>\n",
       "      <td>8</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 100...</td>\n",
       "      <td>16</td>\n",
       "      <td>0.693221</td>\n",
       "      <td>0.915542</td>\n",
       "      <td>0.700652</td>\n",
       "      <td>0.916892</td>\n",
       "      <td>0.680124</td>\n",
       "      <td>0.915388</td>\n",
       "      <td>0.692075</td>\n",
       "      <td>0.914115</td>\n",
       "      <td>0.694857</td>\n",
       "      <td>0.913324</td>\n",
       "      <td>0.247245</td>\n",
       "      <td>0.085392</td>\n",
       "      <td>0.006714</td>\n",
       "      <td>0.001233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>20.665334</td>\n",
       "      <td>0.728915</td>\n",
       "      <td>0.692141</td>\n",
       "      <td>0.868615</td>\n",
       "      <td>75</td>\n",
       "      <td>16</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 10...</td>\n",
       "      <td>17</td>\n",
       "      <td>0.694648</td>\n",
       "      <td>0.868366</td>\n",
       "      <td>0.700425</td>\n",
       "      <td>0.867895</td>\n",
       "      <td>0.680251</td>\n",
       "      <td>0.868550</td>\n",
       "      <td>0.691724</td>\n",
       "      <td>0.869423</td>\n",
       "      <td>0.693659</td>\n",
       "      <td>0.868840</td>\n",
       "      <td>0.124296</td>\n",
       "      <td>0.011925</td>\n",
       "      <td>0.006615</td>\n",
       "      <td>0.000507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>4.739599</td>\n",
       "      <td>0.189025</td>\n",
       "      <td>0.692070</td>\n",
       "      <td>0.919332</td>\n",
       "      <td>75</td>\n",
       "      <td>8</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 200...</td>\n",
       "      <td>18</td>\n",
       "      <td>0.695224</td>\n",
       "      <td>0.919318</td>\n",
       "      <td>0.702435</td>\n",
       "      <td>0.921796</td>\n",
       "      <td>0.679260</td>\n",
       "      <td>0.919648</td>\n",
       "      <td>0.686792</td>\n",
       "      <td>0.918587</td>\n",
       "      <td>0.696638</td>\n",
       "      <td>0.917312</td>\n",
       "      <td>0.140107</td>\n",
       "      <td>0.014707</td>\n",
       "      <td>0.008127</td>\n",
       "      <td>0.001470</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>21.870639</td>\n",
       "      <td>0.840272</td>\n",
       "      <td>0.691879</td>\n",
       "      <td>0.887042</td>\n",
       "      <td>25</td>\n",
       "      <td>12</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 10...</td>\n",
       "      <td>19</td>\n",
       "      <td>0.698425</td>\n",
       "      <td>0.889039</td>\n",
       "      <td>0.697638</td>\n",
       "      <td>0.886594</td>\n",
       "      <td>0.677673</td>\n",
       "      <td>0.887447</td>\n",
       "      <td>0.690909</td>\n",
       "      <td>0.886044</td>\n",
       "      <td>0.694753</td>\n",
       "      <td>0.886088</td>\n",
       "      <td>0.256144</td>\n",
       "      <td>0.083710</td>\n",
       "      <td>0.007577</td>\n",
       "      <td>0.001119</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>12.510526</td>\n",
       "      <td>0.543884</td>\n",
       "      <td>0.691818</td>\n",
       "      <td>0.868090</td>\n",
       "      <td>150</td>\n",
       "      <td>16</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 50...</td>\n",
       "      <td>20</td>\n",
       "      <td>0.698425</td>\n",
       "      <td>0.868733</td>\n",
       "      <td>0.696759</td>\n",
       "      <td>0.866892</td>\n",
       "      <td>0.680784</td>\n",
       "      <td>0.868401</td>\n",
       "      <td>0.691695</td>\n",
       "      <td>0.867896</td>\n",
       "      <td>0.691424</td>\n",
       "      <td>0.868530</td>\n",
       "      <td>1.650814</td>\n",
       "      <td>0.091103</td>\n",
       "      <td>0.006165</td>\n",
       "      <td>0.000660</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>21.259679</td>\n",
       "      <td>0.896056</td>\n",
       "      <td>0.691804</td>\n",
       "      <td>0.892404</td>\n",
       "      <td>75</td>\n",
       "      <td>12</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 10...</td>\n",
       "      <td>21</td>\n",
       "      <td>0.697604</td>\n",
       "      <td>0.894016</td>\n",
       "      <td>0.697460</td>\n",
       "      <td>0.893087</td>\n",
       "      <td>0.678335</td>\n",
       "      <td>0.892459</td>\n",
       "      <td>0.691293</td>\n",
       "      <td>0.890526</td>\n",
       "      <td>0.694328</td>\n",
       "      <td>0.891931</td>\n",
       "      <td>0.094656</td>\n",
       "      <td>0.067973</td>\n",
       "      <td>0.007123</td>\n",
       "      <td>0.001168</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>26.575069</td>\n",
       "      <td>1.024600</td>\n",
       "      <td>0.691787</td>\n",
       "      <td>0.868219</td>\n",
       "      <td>150</td>\n",
       "      <td>16</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 10...</td>\n",
       "      <td>22</td>\n",
       "      <td>0.696422</td>\n",
       "      <td>0.869598</td>\n",
       "      <td>0.701118</td>\n",
       "      <td>0.866718</td>\n",
       "      <td>0.677165</td>\n",
       "      <td>0.867644</td>\n",
       "      <td>0.692628</td>\n",
       "      <td>0.868997</td>\n",
       "      <td>0.691603</td>\n",
       "      <td>0.868138</td>\n",
       "      <td>0.055535</td>\n",
       "      <td>0.093843</td>\n",
       "      <td>0.008039</td>\n",
       "      <td>0.001010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.908106</td>\n",
       "      <td>0.153714</td>\n",
       "      <td>0.691784</td>\n",
       "      <td>0.884888</td>\n",
       "      <td>25</td>\n",
       "      <td>12</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 20...</td>\n",
       "      <td>23</td>\n",
       "      <td>0.697318</td>\n",
       "      <td>0.884241</td>\n",
       "      <td>0.692637</td>\n",
       "      <td>0.884247</td>\n",
       "      <td>0.682472</td>\n",
       "      <td>0.885385</td>\n",
       "      <td>0.692104</td>\n",
       "      <td>0.886386</td>\n",
       "      <td>0.694391</td>\n",
       "      <td>0.884180</td>\n",
       "      <td>0.138193</td>\n",
       "      <td>0.007919</td>\n",
       "      <td>0.005000</td>\n",
       "      <td>0.000874</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>4.248663</td>\n",
       "      <td>0.170544</td>\n",
       "      <td>0.691731</td>\n",
       "      <td>0.919668</td>\n",
       "      <td>150</td>\n",
       "      <td>8</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 200...</td>\n",
       "      <td>24</td>\n",
       "      <td>0.696544</td>\n",
       "      <td>0.920957</td>\n",
       "      <td>0.696313</td>\n",
       "      <td>0.922325</td>\n",
       "      <td>0.683456</td>\n",
       "      <td>0.919965</td>\n",
       "      <td>0.687783</td>\n",
       "      <td>0.917948</td>\n",
       "      <td>0.694561</td>\n",
       "      <td>0.917144</td>\n",
       "      <td>0.082240</td>\n",
       "      <td>0.010191</td>\n",
       "      <td>0.005220</td>\n",
       "      <td>0.001905</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>10.428074</td>\n",
       "      <td>0.408405</td>\n",
       "      <td>0.691640</td>\n",
       "      <td>0.868522</td>\n",
       "      <td>75</td>\n",
       "      <td>16</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 50...</td>\n",
       "      <td>25</td>\n",
       "      <td>0.694444</td>\n",
       "      <td>0.868494</td>\n",
       "      <td>0.699420</td>\n",
       "      <td>0.866782</td>\n",
       "      <td>0.679186</td>\n",
       "      <td>0.869615</td>\n",
       "      <td>0.691724</td>\n",
       "      <td>0.868368</td>\n",
       "      <td>0.693425</td>\n",
       "      <td>0.869349</td>\n",
       "      <td>0.114419</td>\n",
       "      <td>0.051597</td>\n",
       "      <td>0.006734</td>\n",
       "      <td>0.000993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>20.511648</td>\n",
       "      <td>0.736294</td>\n",
       "      <td>0.691600</td>\n",
       "      <td>0.868439</td>\n",
       "      <td>50</td>\n",
       "      <td>16</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 10...</td>\n",
       "      <td>26</td>\n",
       "      <td>0.695385</td>\n",
       "      <td>0.869549</td>\n",
       "      <td>0.694712</td>\n",
       "      <td>0.867001</td>\n",
       "      <td>0.680834</td>\n",
       "      <td>0.869473</td>\n",
       "      <td>0.690854</td>\n",
       "      <td>0.868019</td>\n",
       "      <td>0.696217</td>\n",
       "      <td>0.868156</td>\n",
       "      <td>0.154656</td>\n",
       "      <td>0.022947</td>\n",
       "      <td>0.005689</td>\n",
       "      <td>0.000962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10.751458</td>\n",
       "      <td>0.397120</td>\n",
       "      <td>0.691542</td>\n",
       "      <td>0.914502</td>\n",
       "      <td>25</td>\n",
       "      <td>8</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 500...</td>\n",
       "      <td>27</td>\n",
       "      <td>0.696449</td>\n",
       "      <td>0.915019</td>\n",
       "      <td>0.699731</td>\n",
       "      <td>0.915735</td>\n",
       "      <td>0.677808</td>\n",
       "      <td>0.914836</td>\n",
       "      <td>0.685822</td>\n",
       "      <td>0.914368</td>\n",
       "      <td>0.697905</td>\n",
       "      <td>0.912554</td>\n",
       "      <td>0.293775</td>\n",
       "      <td>0.012685</td>\n",
       "      <td>0.008402</td>\n",
       "      <td>0.001069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>21.823842</td>\n",
       "      <td>0.980334</td>\n",
       "      <td>0.691438</td>\n",
       "      <td>0.920752</td>\n",
       "      <td>75</td>\n",
       "      <td>8</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 100...</td>\n",
       "      <td>28</td>\n",
       "      <td>0.694825</td>\n",
       "      <td>0.921986</td>\n",
       "      <td>0.698171</td>\n",
       "      <td>0.922959</td>\n",
       "      <td>0.680325</td>\n",
       "      <td>0.921739</td>\n",
       "      <td>0.684250</td>\n",
       "      <td>0.917378</td>\n",
       "      <td>0.699623</td>\n",
       "      <td>0.919698</td>\n",
       "      <td>0.136729</td>\n",
       "      <td>0.124109</td>\n",
       "      <td>0.007732</td>\n",
       "      <td>0.001993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>4.189693</td>\n",
       "      <td>0.157662</td>\n",
       "      <td>0.691417</td>\n",
       "      <td>0.866052</td>\n",
       "      <td>150</td>\n",
       "      <td>16</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 20...</td>\n",
       "      <td>29</td>\n",
       "      <td>0.691176</td>\n",
       "      <td>0.866131</td>\n",
       "      <td>0.697602</td>\n",
       "      <td>0.864245</td>\n",
       "      <td>0.680583</td>\n",
       "      <td>0.866892</td>\n",
       "      <td>0.692540</td>\n",
       "      <td>0.865795</td>\n",
       "      <td>0.695187</td>\n",
       "      <td>0.867197</td>\n",
       "      <td>0.217777</td>\n",
       "      <td>0.002387</td>\n",
       "      <td>0.005851</td>\n",
       "      <td>0.001035</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>4.009128</td>\n",
       "      <td>0.152456</td>\n",
       "      <td>0.691237</td>\n",
       "      <td>0.868333</td>\n",
       "      <td>50</td>\n",
       "      <td>16</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 20...</td>\n",
       "      <td>30</td>\n",
       "      <td>0.697389</td>\n",
       "      <td>0.869074</td>\n",
       "      <td>0.700193</td>\n",
       "      <td>0.867495</td>\n",
       "      <td>0.676827</td>\n",
       "      <td>0.868778</td>\n",
       "      <td>0.690854</td>\n",
       "      <td>0.868590</td>\n",
       "      <td>0.690923</td>\n",
       "      <td>0.867729</td>\n",
       "      <td>0.257482</td>\n",
       "      <td>0.007396</td>\n",
       "      <td>0.008075</td>\n",
       "      <td>0.000613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>21.236080</td>\n",
       "      <td>0.877175</td>\n",
       "      <td>0.690426</td>\n",
       "      <td>0.863032</td>\n",
       "      <td>25</td>\n",
       "      <td>16</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 10...</td>\n",
       "      <td>31</td>\n",
       "      <td>0.695216</td>\n",
       "      <td>0.864122</td>\n",
       "      <td>0.694466</td>\n",
       "      <td>0.860772</td>\n",
       "      <td>0.678193</td>\n",
       "      <td>0.864279</td>\n",
       "      <td>0.690647</td>\n",
       "      <td>0.863188</td>\n",
       "      <td>0.693610</td>\n",
       "      <td>0.862798</td>\n",
       "      <td>0.181050</td>\n",
       "      <td>0.059777</td>\n",
       "      <td>0.006311</td>\n",
       "      <td>0.001259</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>10.203875</td>\n",
       "      <td>0.427069</td>\n",
       "      <td>0.690232</td>\n",
       "      <td>0.868298</td>\n",
       "      <td>50</td>\n",
       "      <td>16</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 50...</td>\n",
       "      <td>32</td>\n",
       "      <td>0.693784</td>\n",
       "      <td>0.868917</td>\n",
       "      <td>0.696525</td>\n",
       "      <td>0.865818</td>\n",
       "      <td>0.674774</td>\n",
       "      <td>0.870094</td>\n",
       "      <td>0.692657</td>\n",
       "      <td>0.867044</td>\n",
       "      <td>0.693425</td>\n",
       "      <td>0.869615</td>\n",
       "      <td>0.172048</td>\n",
       "      <td>0.042656</td>\n",
       "      <td>0.007840</td>\n",
       "      <td>0.001616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>10.841403</td>\n",
       "      <td>0.411020</td>\n",
       "      <td>0.690017</td>\n",
       "      <td>0.892178</td>\n",
       "      <td>150</td>\n",
       "      <td>12</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 12, u'n_estimators': 50...</td>\n",
       "      <td>33</td>\n",
       "      <td>0.694561</td>\n",
       "      <td>0.894007</td>\n",
       "      <td>0.694615</td>\n",
       "      <td>0.891490</td>\n",
       "      <td>0.677596</td>\n",
       "      <td>0.892850</td>\n",
       "      <td>0.690045</td>\n",
       "      <td>0.890911</td>\n",
       "      <td>0.693273</td>\n",
       "      <td>0.891631</td>\n",
       "      <td>0.664654</td>\n",
       "      <td>0.046813</td>\n",
       "      <td>0.006430</td>\n",
       "      <td>0.001112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>4.875343</td>\n",
       "      <td>0.186968</td>\n",
       "      <td>0.689988</td>\n",
       "      <td>0.919074</td>\n",
       "      <td>50</td>\n",
       "      <td>8</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 8, u'n_estimators': 200...</td>\n",
       "      <td>34</td>\n",
       "      <td>0.692776</td>\n",
       "      <td>0.919877</td>\n",
       "      <td>0.699620</td>\n",
       "      <td>0.920869</td>\n",
       "      <td>0.680109</td>\n",
       "      <td>0.920781</td>\n",
       "      <td>0.688302</td>\n",
       "      <td>0.916843</td>\n",
       "      <td>0.689133</td>\n",
       "      <td>0.917001</td>\n",
       "      <td>0.153976</td>\n",
       "      <td>0.017715</td>\n",
       "      <td>0.006352</td>\n",
       "      <td>0.001792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>19.524125</td>\n",
       "      <td>0.849536</td>\n",
       "      <td>0.689747</td>\n",
       "      <td>0.848176</td>\n",
       "      <td>75</td>\n",
       "      <td>20</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 10...</td>\n",
       "      <td>35</td>\n",
       "      <td>0.690881</td>\n",
       "      <td>0.848090</td>\n",
       "      <td>0.700583</td>\n",
       "      <td>0.845065</td>\n",
       "      <td>0.678206</td>\n",
       "      <td>0.849836</td>\n",
       "      <td>0.690152</td>\n",
       "      <td>0.847935</td>\n",
       "      <td>0.688914</td>\n",
       "      <td>0.849952</td>\n",
       "      <td>0.316267</td>\n",
       "      <td>0.083136</td>\n",
       "      <td>0.007111</td>\n",
       "      <td>0.001769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>18.423834</td>\n",
       "      <td>0.575878</td>\n",
       "      <td>0.689714</td>\n",
       "      <td>0.848354</td>\n",
       "      <td>150</td>\n",
       "      <td>20</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 10...</td>\n",
       "      <td>36</td>\n",
       "      <td>0.692130</td>\n",
       "      <td>0.848665</td>\n",
       "      <td>0.698450</td>\n",
       "      <td>0.846428</td>\n",
       "      <td>0.676887</td>\n",
       "      <td>0.849352</td>\n",
       "      <td>0.689917</td>\n",
       "      <td>0.848233</td>\n",
       "      <td>0.691188</td>\n",
       "      <td>0.849091</td>\n",
       "      <td>0.466081</td>\n",
       "      <td>0.012321</td>\n",
       "      <td>0.007056</td>\n",
       "      <td>0.001035</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>5.218190</td>\n",
       "      <td>0.178517</td>\n",
       "      <td>0.689266</td>\n",
       "      <td>0.847537</td>\n",
       "      <td>150</td>\n",
       "      <td>20</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 20...</td>\n",
       "      <td>37</td>\n",
       "      <td>0.691444</td>\n",
       "      <td>0.849522</td>\n",
       "      <td>0.696934</td>\n",
       "      <td>0.843122</td>\n",
       "      <td>0.678993</td>\n",
       "      <td>0.848192</td>\n",
       "      <td>0.690232</td>\n",
       "      <td>0.848421</td>\n",
       "      <td>0.688726</td>\n",
       "      <td>0.848427</td>\n",
       "      <td>0.956120</td>\n",
       "      <td>0.020670</td>\n",
       "      <td>0.005837</td>\n",
       "      <td>0.002255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>9.696128</td>\n",
       "      <td>0.391955</td>\n",
       "      <td>0.689209</td>\n",
       "      <td>0.848210</td>\n",
       "      <td>50</td>\n",
       "      <td>20</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 50...</td>\n",
       "      <td>38</td>\n",
       "      <td>0.697836</td>\n",
       "      <td>0.849165</td>\n",
       "      <td>0.697241</td>\n",
       "      <td>0.846109</td>\n",
       "      <td>0.675601</td>\n",
       "      <td>0.848982</td>\n",
       "      <td>0.687191</td>\n",
       "      <td>0.847654</td>\n",
       "      <td>0.688172</td>\n",
       "      <td>0.849138</td>\n",
       "      <td>0.101203</td>\n",
       "      <td>0.023155</td>\n",
       "      <td>0.008116</td>\n",
       "      <td>0.001191</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>9.454104</td>\n",
       "      <td>0.365208</td>\n",
       "      <td>0.689160</td>\n",
       "      <td>0.848448</td>\n",
       "      <td>75</td>\n",
       "      <td>20</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 50...</td>\n",
       "      <td>39</td>\n",
       "      <td>0.691921</td>\n",
       "      <td>0.847607</td>\n",
       "      <td>0.696124</td>\n",
       "      <td>0.847772</td>\n",
       "      <td>0.676090</td>\n",
       "      <td>0.849864</td>\n",
       "      <td>0.691958</td>\n",
       "      <td>0.848356</td>\n",
       "      <td>0.689708</td>\n",
       "      <td>0.848642</td>\n",
       "      <td>0.073717</td>\n",
       "      <td>0.021722</td>\n",
       "      <td>0.006858</td>\n",
       "      <td>0.000802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>4.173165</td>\n",
       "      <td>0.159855</td>\n",
       "      <td>0.688921</td>\n",
       "      <td>0.861995</td>\n",
       "      <td>25</td>\n",
       "      <td>16</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 20...</td>\n",
       "      <td>40</td>\n",
       "      <td>0.692544</td>\n",
       "      <td>0.862960</td>\n",
       "      <td>0.690541</td>\n",
       "      <td>0.857226</td>\n",
       "      <td>0.677900</td>\n",
       "      <td>0.864374</td>\n",
       "      <td>0.691048</td>\n",
       "      <td>0.862385</td>\n",
       "      <td>0.692574</td>\n",
       "      <td>0.863032</td>\n",
       "      <td>0.097278</td>\n",
       "      <td>0.005079</td>\n",
       "      <td>0.005570</td>\n",
       "      <td>0.002472</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>4.041806</td>\n",
       "      <td>0.159113</td>\n",
       "      <td>0.688881</td>\n",
       "      <td>0.847108</td>\n",
       "      <td>50</td>\n",
       "      <td>20</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 20...</td>\n",
       "      <td>41</td>\n",
       "      <td>0.691387</td>\n",
       "      <td>0.846676</td>\n",
       "      <td>0.695754</td>\n",
       "      <td>0.844698</td>\n",
       "      <td>0.683157</td>\n",
       "      <td>0.848403</td>\n",
       "      <td>0.687666</td>\n",
       "      <td>0.848239</td>\n",
       "      <td>0.686441</td>\n",
       "      <td>0.847523</td>\n",
       "      <td>0.118824</td>\n",
       "      <td>0.015877</td>\n",
       "      <td>0.004330</td>\n",
       "      <td>0.001351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>10.993028</td>\n",
       "      <td>0.423171</td>\n",
       "      <td>0.688761</td>\n",
       "      <td>0.847684</td>\n",
       "      <td>150</td>\n",
       "      <td>20</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 50...</td>\n",
       "      <td>42</td>\n",
       "      <td>0.689655</td>\n",
       "      <td>0.847409</td>\n",
       "      <td>0.695416</td>\n",
       "      <td>0.845693</td>\n",
       "      <td>0.676633</td>\n",
       "      <td>0.848221</td>\n",
       "      <td>0.689917</td>\n",
       "      <td>0.847878</td>\n",
       "      <td>0.692190</td>\n",
       "      <td>0.849220</td>\n",
       "      <td>0.611760</td>\n",
       "      <td>0.031134</td>\n",
       "      <td>0.006408</td>\n",
       "      <td>0.001160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>10.479195</td>\n",
       "      <td>0.398551</td>\n",
       "      <td>0.688741</td>\n",
       "      <td>0.862119</td>\n",
       "      <td>25</td>\n",
       "      <td>16</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 16, u'n_estimators': 50...</td>\n",
       "      <td>43</td>\n",
       "      <td>0.692159</td>\n",
       "      <td>0.862262</td>\n",
       "      <td>0.694401</td>\n",
       "      <td>0.859884</td>\n",
       "      <td>0.675069</td>\n",
       "      <td>0.864586</td>\n",
       "      <td>0.691283</td>\n",
       "      <td>0.862775</td>\n",
       "      <td>0.690797</td>\n",
       "      <td>0.861087</td>\n",
       "      <td>0.341546</td>\n",
       "      <td>0.028570</td>\n",
       "      <td>0.006948</td>\n",
       "      <td>0.001587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>20.837171</td>\n",
       "      <td>0.772504</td>\n",
       "      <td>0.688339</td>\n",
       "      <td>0.848671</td>\n",
       "      <td>50</td>\n",
       "      <td>20</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 10...</td>\n",
       "      <td>44</td>\n",
       "      <td>0.690909</td>\n",
       "      <td>0.850029</td>\n",
       "      <td>0.695820</td>\n",
       "      <td>0.845665</td>\n",
       "      <td>0.676644</td>\n",
       "      <td>0.850087</td>\n",
       "      <td>0.691638</td>\n",
       "      <td>0.847772</td>\n",
       "      <td>0.686682</td>\n",
       "      <td>0.849803</td>\n",
       "      <td>0.097821</td>\n",
       "      <td>0.018975</td>\n",
       "      <td>0.006527</td>\n",
       "      <td>0.001731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>4.151654</td>\n",
       "      <td>0.161405</td>\n",
       "      <td>0.687328</td>\n",
       "      <td>0.848498</td>\n",
       "      <td>75</td>\n",
       "      <td>20</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 20...</td>\n",
       "      <td>45</td>\n",
       "      <td>0.687379</td>\n",
       "      <td>0.850206</td>\n",
       "      <td>0.694606</td>\n",
       "      <td>0.846361</td>\n",
       "      <td>0.680616</td>\n",
       "      <td>0.849826</td>\n",
       "      <td>0.685628</td>\n",
       "      <td>0.848637</td>\n",
       "      <td>0.688411</td>\n",
       "      <td>0.847458</td>\n",
       "      <td>0.181558</td>\n",
       "      <td>0.022557</td>\n",
       "      <td>0.004519</td>\n",
       "      <td>0.001439</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10.589169</td>\n",
       "      <td>0.429202</td>\n",
       "      <td>0.686657</td>\n",
       "      <td>0.844182</td>\n",
       "      <td>25</td>\n",
       "      <td>20</td>\n",
       "      <td>500</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 50...</td>\n",
       "      <td>46</td>\n",
       "      <td>0.689200</td>\n",
       "      <td>0.844735</td>\n",
       "      <td>0.692278</td>\n",
       "      <td>0.840113</td>\n",
       "      <td>0.676899</td>\n",
       "      <td>0.844635</td>\n",
       "      <td>0.685150</td>\n",
       "      <td>0.845677</td>\n",
       "      <td>0.689761</td>\n",
       "      <td>0.845752</td>\n",
       "      <td>0.199180</td>\n",
       "      <td>0.041142</td>\n",
       "      <td>0.005389</td>\n",
       "      <td>0.002087</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>21.873220</td>\n",
       "      <td>0.885542</td>\n",
       "      <td>0.686338</td>\n",
       "      <td>0.843029</td>\n",
       "      <td>25</td>\n",
       "      <td>20</td>\n",
       "      <td>1000</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 10...</td>\n",
       "      <td>47</td>\n",
       "      <td>0.689069</td>\n",
       "      <td>0.842865</td>\n",
       "      <td>0.695212</td>\n",
       "      <td>0.839154</td>\n",
       "      <td>0.673759</td>\n",
       "      <td>0.843271</td>\n",
       "      <td>0.687476</td>\n",
       "      <td>0.844975</td>\n",
       "      <td>0.686176</td>\n",
       "      <td>0.844883</td>\n",
       "      <td>0.143269</td>\n",
       "      <td>0.061466</td>\n",
       "      <td>0.007012</td>\n",
       "      <td>0.002113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>4.163548</td>\n",
       "      <td>0.150920</td>\n",
       "      <td>0.686287</td>\n",
       "      <td>0.842566</td>\n",
       "      <td>25</td>\n",
       "      <td>20</td>\n",
       "      <td>200</td>\n",
       "      <td>{u'min_samples_split': 20, u'n_estimators': 20...</td>\n",
       "      <td>48</td>\n",
       "      <td>0.685315</td>\n",
       "      <td>0.843657</td>\n",
       "      <td>0.695041</td>\n",
       "      <td>0.838328</td>\n",
       "      <td>0.678233</td>\n",
       "      <td>0.843014</td>\n",
       "      <td>0.685780</td>\n",
       "      <td>0.843768</td>\n",
       "      <td>0.687064</td>\n",
       "      <td>0.844062</td>\n",
       "      <td>0.098572</td>\n",
       "      <td>0.004723</td>\n",
       "      <td>0.005354</td>\n",
       "      <td>0.002146</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mean_fit_time  mean_score_time  mean_test_score  mean_train_score  \\\n",
       "15       4.415908         0.165963         0.694561          0.890593   \n",
       "17      21.089066         0.929221         0.693942          0.892352   \n",
       "37      11.556772         0.439900         0.693899          0.920415   \n",
       "39       4.188209         0.160574         0.693752          0.890960   \n",
       "14      23.045973         0.878270         0.693726          0.920941   \n",
       "41      21.478017         0.868469         0.693599          0.892433   \n",
       "28      10.687245         0.452287         0.693382          0.891686   \n",
       "16      10.982650         0.440209         0.693324          0.892077   \n",
       "25      10.993991         0.457164         0.693212          0.920001   \n",
       "13      12.035331         0.477261         0.692980          0.920289   \n",
       "27       4.331092         0.172203         0.692849          0.890217   \n",
       "4       10.862266         0.412673         0.692691          0.887230   \n",
       "30       4.211588         0.177381         0.692655          0.866148   \n",
       "0        4.210468         0.198299         0.692631          0.912612   \n",
       "38      22.954586         0.964277         0.692546          0.920786   \n",
       "2       21.122218         0.921210         0.692185          0.915052   \n",
       "32      20.665334         0.728915         0.692141          0.868615   \n",
       "24       4.739599         0.189025         0.692070          0.919332   \n",
       "5       21.870639         0.840272         0.691879          0.887042   \n",
       "43      12.510526         0.543884         0.691818          0.868090   \n",
       "29      21.259679         0.896056         0.691804          0.892404   \n",
       "44      26.575069         1.024600         0.691787          0.868219   \n",
       "3        3.908106         0.153714         0.691784          0.884888   \n",
       "36       4.248663         0.170544         0.691731          0.919668   \n",
       "31      10.428074         0.408405         0.691640          0.868522   \n",
       "20      20.511648         0.736294         0.691600          0.868439   \n",
       "1       10.751458         0.397120         0.691542          0.914502   \n",
       "26      21.823842         0.980334         0.691438          0.920752   \n",
       "42       4.189693         0.157662         0.691417          0.866052   \n",
       "18       4.009128         0.152456         0.691237          0.868333   \n",
       "8       21.236080         0.877175         0.690426          0.863032   \n",
       "19      10.203875         0.427069         0.690232          0.868298   \n",
       "40      10.841403         0.411020         0.690017          0.892178   \n",
       "12       4.875343         0.186968         0.689988          0.919074   \n",
       "35      19.524125         0.849536         0.689747          0.848176   \n",
       "47      18.423834         0.575878         0.689714          0.848354   \n",
       "45       5.218190         0.178517         0.689266          0.847537   \n",
       "22       9.696128         0.391955         0.689209          0.848210   \n",
       "34       9.454104         0.365208         0.689160          0.848448   \n",
       "6        4.173165         0.159855         0.688921          0.861995   \n",
       "21       4.041806         0.159113         0.688881          0.847108   \n",
       "46      10.993028         0.423171         0.688761          0.847684   \n",
       "7       10.479195         0.398551         0.688741          0.862119   \n",
       "23      20.837171         0.772504         0.688339          0.848671   \n",
       "33       4.151654         0.161405         0.687328          0.848498   \n",
       "10      10.589169         0.429202         0.686657          0.844182   \n",
       "11      21.873220         0.885542         0.686338          0.843029   \n",
       "9        4.163548         0.150920         0.686287          0.842566   \n",
       "\n",
       "   param_max_depth param_min_samples_split param_n_estimators  \\\n",
       "15              50                      12                200   \n",
       "17              50                      12               1000   \n",
       "37             150                       8                500   \n",
       "39             150                      12                200   \n",
       "14              50                       8               1000   \n",
       "41             150                      12               1000   \n",
       "28              75                      12                500   \n",
       "16              50                      12                500   \n",
       "25              75                       8                500   \n",
       "13              50                       8                500   \n",
       "27              75                      12                200   \n",
       "4               25                      12                500   \n",
       "30              75                      16                200   \n",
       "0               25                       8                200   \n",
       "38             150                       8               1000   \n",
       "2               25                       8               1000   \n",
       "32              75                      16               1000   \n",
       "24              75                       8                200   \n",
       "5               25                      12               1000   \n",
       "43             150                      16                500   \n",
       "29              75                      12               1000   \n",
       "44             150                      16               1000   \n",
       "3               25                      12                200   \n",
       "36             150                       8                200   \n",
       "31              75                      16                500   \n",
       "20              50                      16               1000   \n",
       "1               25                       8                500   \n",
       "26              75                       8               1000   \n",
       "42             150                      16                200   \n",
       "18              50                      16                200   \n",
       "8               25                      16               1000   \n",
       "19              50                      16                500   \n",
       "40             150                      12                500   \n",
       "12              50                       8                200   \n",
       "35              75                      20               1000   \n",
       "47             150                      20               1000   \n",
       "45             150                      20                200   \n",
       "22              50                      20                500   \n",
       "34              75                      20                500   \n",
       "6               25                      16                200   \n",
       "21              50                      20                200   \n",
       "46             150                      20                500   \n",
       "7               25                      16                500   \n",
       "23              50                      20               1000   \n",
       "33              75                      20                200   \n",
       "10              25                      20                500   \n",
       "11              25                      20               1000   \n",
       "9               25                      20                200   \n",
       "\n",
       "                                               params  rank_test_score  \\\n",
       "15  {u'min_samples_split': 12, u'n_estimators': 20...                1   \n",
       "17  {u'min_samples_split': 12, u'n_estimators': 10...                2   \n",
       "37  {u'min_samples_split': 8, u'n_estimators': 500...                3   \n",
       "39  {u'min_samples_split': 12, u'n_estimators': 20...                4   \n",
       "14  {u'min_samples_split': 8, u'n_estimators': 100...                5   \n",
       "41  {u'min_samples_split': 12, u'n_estimators': 10...                6   \n",
       "28  {u'min_samples_split': 12, u'n_estimators': 50...                7   \n",
       "16  {u'min_samples_split': 12, u'n_estimators': 50...                8   \n",
       "25  {u'min_samples_split': 8, u'n_estimators': 500...                9   \n",
       "13  {u'min_samples_split': 8, u'n_estimators': 500...               10   \n",
       "27  {u'min_samples_split': 12, u'n_estimators': 20...               11   \n",
       "4   {u'min_samples_split': 12, u'n_estimators': 50...               12   \n",
       "30  {u'min_samples_split': 16, u'n_estimators': 20...               13   \n",
       "0   {u'min_samples_split': 8, u'n_estimators': 200...               14   \n",
       "38  {u'min_samples_split': 8, u'n_estimators': 100...               15   \n",
       "2   {u'min_samples_split': 8, u'n_estimators': 100...               16   \n",
       "32  {u'min_samples_split': 16, u'n_estimators': 10...               17   \n",
       "24  {u'min_samples_split': 8, u'n_estimators': 200...               18   \n",
       "5   {u'min_samples_split': 12, u'n_estimators': 10...               19   \n",
       "43  {u'min_samples_split': 16, u'n_estimators': 50...               20   \n",
       "29  {u'min_samples_split': 12, u'n_estimators': 10...               21   \n",
       "44  {u'min_samples_split': 16, u'n_estimators': 10...               22   \n",
       "3   {u'min_samples_split': 12, u'n_estimators': 20...               23   \n",
       "36  {u'min_samples_split': 8, u'n_estimators': 200...               24   \n",
       "31  {u'min_samples_split': 16, u'n_estimators': 50...               25   \n",
       "20  {u'min_samples_split': 16, u'n_estimators': 10...               26   \n",
       "1   {u'min_samples_split': 8, u'n_estimators': 500...               27   \n",
       "26  {u'min_samples_split': 8, u'n_estimators': 100...               28   \n",
       "42  {u'min_samples_split': 16, u'n_estimators': 20...               29   \n",
       "18  {u'min_samples_split': 16, u'n_estimators': 20...               30   \n",
       "8   {u'min_samples_split': 16, u'n_estimators': 10...               31   \n",
       "19  {u'min_samples_split': 16, u'n_estimators': 50...               32   \n",
       "40  {u'min_samples_split': 12, u'n_estimators': 50...               33   \n",
       "12  {u'min_samples_split': 8, u'n_estimators': 200...               34   \n",
       "35  {u'min_samples_split': 20, u'n_estimators': 10...               35   \n",
       "47  {u'min_samples_split': 20, u'n_estimators': 10...               36   \n",
       "45  {u'min_samples_split': 20, u'n_estimators': 20...               37   \n",
       "22  {u'min_samples_split': 20, u'n_estimators': 50...               38   \n",
       "34  {u'min_samples_split': 20, u'n_estimators': 50...               39   \n",
       "6   {u'min_samples_split': 16, u'n_estimators': 20...               40   \n",
       "21  {u'min_samples_split': 20, u'n_estimators': 20...               41   \n",
       "46  {u'min_samples_split': 20, u'n_estimators': 50...               42   \n",
       "7   {u'min_samples_split': 16, u'n_estimators': 50...               43   \n",
       "23  {u'min_samples_split': 20, u'n_estimators': 10...               44   \n",
       "33  {u'min_samples_split': 20, u'n_estimators': 20...               45   \n",
       "10  {u'min_samples_split': 20, u'n_estimators': 50...               46   \n",
       "11  {u'min_samples_split': 20, u'n_estimators': 10...               47   \n",
       "9   {u'min_samples_split': 20, u'n_estimators': 20...               48   \n",
       "\n",
       "    split0_test_score  split0_train_score  split1_test_score  \\\n",
       "15           0.698084            0.891538           0.704354   \n",
       "17           0.695089            0.894129           0.700769   \n",
       "37           0.697905            0.921277           0.701447   \n",
       "39           0.696018            0.892378           0.702951   \n",
       "14           0.693009            0.921915           0.701141   \n",
       "41           0.696947            0.894399           0.698462   \n",
       "28           0.694127            0.893226           0.700809   \n",
       "16           0.697639            0.894244           0.702848   \n",
       "25           0.698941            0.921557           0.702846   \n",
       "13           0.690937            0.920750           0.705123   \n",
       "27           0.699237            0.890126           0.696993   \n",
       "4            0.697621            0.889103           0.700350   \n",
       "30           0.695253            0.868336           0.700617   \n",
       "0            0.696217            0.913761           0.698730   \n",
       "38           0.695321            0.922236           0.702744   \n",
       "2            0.693221            0.915542           0.700652   \n",
       "32           0.694648            0.868366           0.700425   \n",
       "24           0.695224            0.919318           0.702435   \n",
       "5            0.698425            0.889039           0.697638   \n",
       "43           0.698425            0.868733           0.696759   \n",
       "29           0.697604            0.894016           0.697460   \n",
       "44           0.696422            0.869598           0.701118   \n",
       "3            0.697318            0.884241           0.692637   \n",
       "36           0.696544            0.920957           0.696313   \n",
       "31           0.694444            0.868494           0.699420   \n",
       "20           0.695385            0.869549           0.694712   \n",
       "1            0.696449            0.915019           0.699731   \n",
       "26           0.694825            0.921986           0.698171   \n",
       "42           0.691176            0.866131           0.697602   \n",
       "18           0.697389            0.869074           0.700193   \n",
       "8            0.695216            0.864122           0.694466   \n",
       "19           0.693784            0.868917           0.696525   \n",
       "40           0.694561            0.894007           0.694615   \n",
       "12           0.692776            0.919877           0.699620   \n",
       "35           0.690881            0.848090           0.700583   \n",
       "47           0.692130            0.848665           0.698450   \n",
       "45           0.691444            0.849522           0.696934   \n",
       "22           0.697836            0.849165           0.697241   \n",
       "34           0.691921            0.847607           0.696124   \n",
       "6            0.692544            0.862960           0.690541   \n",
       "21           0.691387            0.846676           0.695754   \n",
       "46           0.689655            0.847409           0.695416   \n",
       "7            0.692159            0.862262           0.694401   \n",
       "23           0.690909            0.850029           0.695820   \n",
       "33           0.687379            0.850206           0.694606   \n",
       "10           0.689200            0.844735           0.692278   \n",
       "11           0.689069            0.842865           0.695212   \n",
       "9            0.685315            0.843657           0.695041   \n",
       "\n",
       "    split1_train_score  split2_test_score  split2_train_score  \\\n",
       "15            0.891104           0.682718            0.891692   \n",
       "17            0.892624           0.684313            0.892682   \n",
       "37            0.923283           0.682344            0.920680   \n",
       "39            0.890381           0.679642            0.891235   \n",
       "14            0.923269           0.681361            0.921224   \n",
       "41            0.891791           0.685625            0.892819   \n",
       "28            0.892166           0.685848            0.892452   \n",
       "16            0.892751           0.679172            0.892933   \n",
       "25            0.922399           0.680341            0.920172   \n",
       "13            0.922134           0.681818            0.919973   \n",
       "27            0.889059           0.684292            0.891425   \n",
       "4             0.886550           0.677647            0.887770   \n",
       "30            0.862711           0.678783            0.865767   \n",
       "0             0.914120           0.681098            0.911586   \n",
       "38            0.923283           0.681412            0.920356   \n",
       "2             0.916892           0.680124            0.915388   \n",
       "32            0.867895           0.680251            0.868550   \n",
       "24            0.921796           0.679260            0.919648   \n",
       "5             0.886594           0.677673            0.887447   \n",
       "43            0.866892           0.680784            0.868401   \n",
       "29            0.893087           0.678335            0.892459   \n",
       "44            0.866718           0.677165            0.867644   \n",
       "3             0.884247           0.682472            0.885385   \n",
       "36            0.922325           0.683456            0.919965   \n",
       "31            0.866782           0.679186            0.869615   \n",
       "20            0.867001           0.680834            0.869473   \n",
       "1             0.915735           0.677808            0.914836   \n",
       "26            0.922959           0.680325            0.921739   \n",
       "42            0.864245           0.680583            0.866892   \n",
       "18            0.867495           0.676827            0.868778   \n",
       "8             0.860772           0.678193            0.864279   \n",
       "19            0.865818           0.674774            0.870094   \n",
       "40            0.891490           0.677596            0.892850   \n",
       "12            0.920869           0.680109            0.920781   \n",
       "35            0.845065           0.678206            0.849836   \n",
       "47            0.846428           0.676887            0.849352   \n",
       "45            0.843122           0.678993            0.848192   \n",
       "22            0.846109           0.675601            0.848982   \n",
       "34            0.847772           0.676090            0.849864   \n",
       "6             0.857226           0.677900            0.864374   \n",
       "21            0.844698           0.683157            0.848403   \n",
       "46            0.845693           0.676633            0.848221   \n",
       "7             0.859884           0.675069            0.864586   \n",
       "23            0.845665           0.676644            0.850087   \n",
       "33            0.846361           0.680616            0.849826   \n",
       "10            0.840113           0.676899            0.844635   \n",
       "11            0.839154           0.673759            0.843271   \n",
       "9             0.838328           0.678233            0.843014   \n",
       "\n",
       "    split3_test_score  split3_train_score  split4_test_score  \\\n",
       "15           0.693785            0.889869           0.693862   \n",
       "17           0.693032            0.891675           0.696510   \n",
       "37           0.688180            0.918395           0.699623   \n",
       "39           0.691667            0.890340           0.698485   \n",
       "14           0.694674            0.918753           0.698449   \n",
       "41           0.692337            0.891125           0.694624   \n",
       "28           0.691005            0.890017           0.695122   \n",
       "16           0.691610            0.890634           0.695354   \n",
       "25           0.687195            0.916995           0.696740   \n",
       "13           0.688697            0.919531           0.698328   \n",
       "27           0.690593            0.890930           0.693130   \n",
       "4            0.693414            0.887413           0.694423   \n",
       "30           0.698941            0.867320           0.689682   \n",
       "0            0.688822            0.912825           0.698292   \n",
       "38           0.688156            0.918571           0.695094   \n",
       "2            0.692075            0.914115           0.694857   \n",
       "32           0.691724            0.869423           0.693659   \n",
       "24           0.686792            0.918587           0.696638   \n",
       "5            0.690909            0.886044           0.694753   \n",
       "43           0.691695            0.867896           0.691424   \n",
       "29           0.691293            0.890526           0.694328   \n",
       "44           0.692628            0.868997           0.691603   \n",
       "3            0.692104            0.886386           0.694391   \n",
       "36           0.687783            0.917948           0.694561   \n",
       "31           0.691724            0.868368           0.693425   \n",
       "20           0.690854            0.868019           0.696217   \n",
       "1            0.685822            0.914368           0.697905   \n",
       "26           0.684250            0.917378           0.699623   \n",
       "42           0.692540            0.865795           0.695187   \n",
       "18           0.690854            0.868590           0.690923   \n",
       "8            0.690647            0.863188           0.693610   \n",
       "19           0.692657            0.867044           0.693425   \n",
       "40           0.690045            0.890911           0.693273   \n",
       "12           0.688302            0.916843           0.689133   \n",
       "35           0.690152            0.847935           0.688914   \n",
       "47           0.689917            0.848233           0.691188   \n",
       "45           0.690232            0.848421           0.688726   \n",
       "22           0.687191            0.847654           0.688172   \n",
       "34           0.691958            0.848356           0.689708   \n",
       "6            0.691048            0.862385           0.692574   \n",
       "21           0.687666            0.848239           0.686441   \n",
       "46           0.689917            0.847878           0.692190   \n",
       "7            0.691283            0.862775           0.690797   \n",
       "23           0.691638            0.847772           0.686682   \n",
       "33           0.685628            0.848637           0.688411   \n",
       "10           0.685150            0.845677           0.689761   \n",
       "11           0.687476            0.844975           0.686176   \n",
       "9            0.685780            0.843768           0.687064   \n",
       "\n",
       "    split4_train_score  std_fit_time  std_score_time  std_test_score  \\\n",
       "15            0.888761      0.233003        0.009319        0.007067   \n",
       "17            0.890652      0.166508        0.086053        0.005442   \n",
       "37            0.918438      0.193501        0.048327        0.007376   \n",
       "39            0.890465      0.195631        0.020484        0.007946   \n",
       "14            0.919545      0.097924        0.032627        0.006806   \n",
       "41            0.892031      0.125898        0.044304        0.004497   \n",
       "28            0.890568      0.107898        0.022386        0.004923   \n",
       "16            0.889824      0.179170        0.044461        0.007959   \n",
       "25            0.918883      0.328056        0.052335        0.008246   \n",
       "13            0.919059      0.182710        0.053293        0.008040   \n",
       "27            0.889547      0.213018        0.003773        0.005223   \n",
       "4             0.885312      0.231226        0.040667        0.007910   \n",
       "30            0.866609      0.078186        0.018825        0.007886   \n",
       "0             0.910769      0.023997        0.015800        0.006777   \n",
       "38            0.919483      0.142095        0.164548        0.007231   \n",
       "2             0.913324      0.247245        0.085392        0.006714   \n",
       "32            0.868840      0.124296        0.011925        0.006615   \n",
       "24            0.917312      0.140107        0.014707        0.008127   \n",
       "5             0.886088      0.256144        0.083710        0.007577   \n",
       "43            0.868530      1.650814        0.091103        0.006165   \n",
       "29            0.891931      0.094656        0.067973        0.007123   \n",
       "44            0.868138      0.055535        0.093843        0.008039   \n",
       "3             0.884180      0.138193        0.007919        0.005000   \n",
       "36            0.917144      0.082240        0.010191        0.005220   \n",
       "31            0.869349      0.114419        0.051597        0.006734   \n",
       "20            0.868156      0.154656        0.022947        0.005689   \n",
       "1             0.912554      0.293775        0.012685        0.008402   \n",
       "26            0.919698      0.136729        0.124109        0.007732   \n",
       "42            0.867197      0.217777        0.002387        0.005851   \n",
       "18            0.867729      0.257482        0.007396        0.008075   \n",
       "8             0.862798      0.181050        0.059777        0.006311   \n",
       "19            0.869615      0.172048        0.042656        0.007840   \n",
       "40            0.891631      0.664654        0.046813        0.006430   \n",
       "12            0.917001      0.153976        0.017715        0.006352   \n",
       "35            0.849952      0.316267        0.083136        0.007111   \n",
       "47            0.849091      0.466081        0.012321        0.007056   \n",
       "45            0.848427      0.956120        0.020670        0.005837   \n",
       "22            0.849138      0.101203        0.023155        0.008116   \n",
       "34            0.848642      0.073717        0.021722        0.006858   \n",
       "6             0.863032      0.097278        0.005079        0.005570   \n",
       "21            0.847523      0.118824        0.015877        0.004330   \n",
       "46            0.849220      0.611760        0.031134        0.006408   \n",
       "7             0.861087      0.341546        0.028570        0.006948   \n",
       "23            0.849803      0.097821        0.018975        0.006527   \n",
       "33            0.847458      0.181558        0.022557        0.004519   \n",
       "10            0.845752      0.199180        0.041142        0.005389   \n",
       "11            0.844883      0.143269        0.061466        0.007012   \n",
       "9             0.844062      0.098572        0.004723        0.005354   \n",
       "\n",
       "    std_train_score  \n",
       "15         0.001117  \n",
       "17         0.001157  \n",
       "37         0.001846  \n",
       "39         0.000781  \n",
       "14         0.001624  \n",
       "41         0.001122  \n",
       "28         0.001202  \n",
       "16         0.001615  \n",
       "25         0.001923  \n",
       "13         0.001077  \n",
       "27         0.000868  \n",
       "4          0.001263  \n",
       "30         0.001915  \n",
       "0          0.001272  \n",
       "38         0.001739  \n",
       "2          0.001233  \n",
       "32         0.000507  \n",
       "24         0.001470  \n",
       "5          0.001119  \n",
       "43         0.000660  \n",
       "29         0.001168  \n",
       "44         0.001010  \n",
       "3          0.000874  \n",
       "36         0.001905  \n",
       "31         0.000993  \n",
       "20         0.000962  \n",
       "1          0.001069  \n",
       "26         0.001993  \n",
       "42         0.001035  \n",
       "18         0.000613  \n",
       "8          0.001259  \n",
       "19         0.001616  \n",
       "40         0.001112  \n",
       "12         0.001792  \n",
       "35         0.001769  \n",
       "47         0.001035  \n",
       "45         0.002255  \n",
       "22         0.001191  \n",
       "34         0.000802  \n",
       "6          0.002472  \n",
       "21         0.001351  \n",
       "46         0.001160  \n",
       "7          0.001587  \n",
       "23         0.001731  \n",
       "33         0.001439  \n",
       "10         0.002087  \n",
       "11         0.002113  \n",
       "9          0.002146  "
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(grid_search.cv_results_).sort_values(by='rank_test_score')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Create and ensemble of random forest models for each group frome the majority class having all the minority class in the set.\n",
    "models = []\n",
    "predictions = []\n",
    "probabilities = []\n",
    "for xx, yy in zip(train_x_sets, train_y_sets):\n",
    "    model = RandomForestClassifier(n_estimators=200, max_depth = 50, min_samples_split = 12)\n",
    "    model.fit(xx, yy)\n",
    "    pred = model.predict(x_test)\n",
    "    proba = model.predict_proba(x_test)\n",
    "    models.append(model)\n",
    "    predictions.append(pred)\n",
    "    probabilities.append(proba)\n",
    "\n",
    "predictions = np.array(predictions)\n",
    "probabilities = np.array(probabilities)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Get and plot model importances!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature ranking:\n",
      " 1. feature salary (0.380373)\n",
      " 2. feature state (0.164257)\n",
      " 3. feature job_title (0.141785)\n",
      " 4. feature level (0.109003)\n",
      " 5. feature education (0.072943)\n",
      " 6. feature citizenship (0.069066)\n",
      " 7. feature admission (0.048932)\n",
      " 8. feature ownership (0.009355)\n",
      " 9. feature language (0.004285)\n"
     ]
    }
   ],
   "source": [
    "importances = model.feature_importances_\n",
    "#std = np.std([tree.feature_importances_ for tree in best_estimator.estimators_],\n",
    "#             axis=0)\n",
    "indices = np.argsort(importances)[::-1]\n",
    "\n",
    "print(\"Feature ranking:\")\n",
    "feature_names = data.columns[:-1]\n",
    "for f in range(data.shape[1]-1):\n",
    "    print(\"%2d. feature %s (%f)\" %\n",
    "          (f + 1, feature_names[f], importances[indices[f]]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature_name</th>\n",
       "      <th>importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>salary</td>\n",
       "      <td>0.380373</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>state</td>\n",
       "      <td>0.164257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>job_title</td>\n",
       "      <td>0.141785</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>level</td>\n",
       "      <td>0.109003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>education</td>\n",
       "      <td>0.072943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>citizenship</td>\n",
       "      <td>0.069066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>admission</td>\n",
       "      <td>0.048932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ownership</td>\n",
       "      <td>0.009355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>language</td>\n",
       "      <td>0.004285</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  feature_name  importance\n",
       "0       salary    0.380373\n",
       "1        state    0.164257\n",
       "2    job_title    0.141785\n",
       "3        level    0.109003\n",
       "4    education    0.072943\n",
       "5  citizenship    0.069066\n",
       "6    admission    0.048932\n",
       "7    ownership    0.009355\n",
       "8     language    0.004285"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ranking_df = pd.DataFrame([(feature_names[f], importances[indices[f]]) for f in range(data.shape[1]-1)])\n",
    "ranking_df.columns = [\"feature_name\", \"importance\"]\n",
    "ranking_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "# ranking_df['feature_name'].replace('salary', 'Salary')\n",
    "# ranking_df['feature_name'].replace('state', 'State')\n",
    "# ranking_df['feature_name'].replace('job_title', 'Job title')\n",
    "# ranking_df['feature_name'].replace('level', 'Pay level')\n",
    "# ranking_df['feature_name'].replace('education', 'Education')\n",
    "# ranking_df['feature_name'].replace('citizenship', 'Citizenship')\n",
    "# ranking_df['feature_name'].replace('admission', 'Admission')\n",
    "# ranking_df['feature_name'].replace('refile', 'Refile')\n",
    "# ranking_df['feature_name'].replace('ownership', 'Ownership interest')\n",
    "# ranking_df['feature_name'].replace('language', 'Language Requirement')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x17b00af10>"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAApIAAAFWCAYAAADaGX1ZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112c8b910>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ranking_df.plot(kind=\"barh\", color ='m', figsize=(10,5), fontsize=(14), width = .7, stacked=True, grid=False, legend=False)\n",
    "plt.gca().set_yticklabels(ranking_df.feature_name)\n",
    "plt.gca().invert_yaxis()\n",
    "plt.title('Feature Importance', fontsize=(14))\n",
    "plt.xlabel('Relative Importance', fontsize=(14))\n",
    "plt.ylabel('', fontsize=(14))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define the model in order to use LIME and get precision, recall, accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "class RFEnsemble(object):\n",
    "    \n",
    "    def __init__(self, models):\n",
    "        self.models = models\n",
    "    \n",
    "    def predict_proba(self, X):\n",
    "        proba_all = np.array([md.predict_proba(X) for md in self.models])\n",
    "        proba = proba_all.mean(axis=0)\n",
    "        return proba\n",
    "\n",
    "    def predict(self, X):\n",
    "        proba = self.predict_proba(X)\n",
    "        return proba[:, 1] >= 0.5\n",
    "    \n",
    "    \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "my_model = RFEnsemble(models)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "my_predictions = my_model.predict(x_test)\n",
    "my_pred_proba = my_model.predict_proba(x_test)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.69354838709677424"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test, my_predictions)\n",
    "precision_score(y_test, my_predictions)\n",
    "recall_score(y_test, my_predictions)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "        0.0       0.98      0.73      0.84     57210\n",
      "        1.0       0.12      0.69      0.20      2914\n",
      "\n",
      "avg / total       0.94      0.73      0.81     60124\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, my_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False, False, False, ...,  True, False, False], dtype=bool)"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my_model.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAEOCAYAAABhOhcDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1119d38d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(y_test, my_predictions)\n",
    "sns.heatmap(cm, annot=True, cmap=\"Reds\")\n",
    "plt.xlabel(\"Predicted\", size=14)\n",
    "plt.ylabel(\"Actual\", size=14)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11c0b4ed0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred_rf = model.predict_proba(x_test)[:, 1]\n",
    "fpr_rf, tpr_rf, _ = roc_curve(y_test, my_predictions)\n",
    "\n",
    "roc_auc=auc(fpr_rf, tpr_rf)\n",
    "\n",
    "plt.plot(fpr_rf, tpr_rf, label='Random Forest, AUC=%0.2f'% roc_auc)\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc='best')\n",
    "plt.plot([0,1],[0,1],'r--')\n",
    "#roc_auc=auc(fpr_rf, tpr_rf)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Now we can use LIME to explain why an indiviual was either certified or denied"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use LIME to explain individual predictions, initialize explainer object\n",
    "explainer = LimeTabularExplainer(\n",
    "    X_true[:, :-1],\n",
    "    class_names=['Denied', 'Certified'],\n",
    "    feature_names=list(data.columns[:-1]),\n",
    "    categorical_features=categorical_idx,\n",
    "    categorical_names=categorical_names,\n",
    "    discretize_continuous=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Probability of receiving a visa:', 0.65000000000000002)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAEICAYAAACNqfTZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1185c6c90>"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Explain a prediction (\"local interpretability\"): \n",
    "# Now we see that the pipeline that takes raw data and returns the prediction \n",
    "# of the trained model now comes in conveniently.\n",
    "example = 3\n",
    "exp = explainer.explain_instance(test[:, :-1][example], my_model.predict_proba, num_features=10)\n",
    "\n",
    "\n",
    "# exp = explainer.explain_instance(test[example], model.predict_proba, num_features=5)\n",
    "print('Probability of receiving a visa:', exp.predict_proba[1])\n",
    "exp.as_pyplot_figure()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Probability of receiving a visa:', 0.78988095238095235)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmIAAAEICAYAAAD80ZhHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wcJ9eLkhdtbtQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11ec8bf50>"
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Explain a prediction (\"local interpretability\"): \n",
    "# Now we see that the pipeline that takes raw data and returns the prediction \n",
    "# of the trained model now comes in conveniently.\n",
    "example = 10\n",
    "exp = explainer.explain_instance(test[:, :-1][example], my_model.predict_proba, num_features=10)\n",
    "\n",
    "\n",
    "# exp = explainer.explain_instance(test[example], model.predict_proba, num_features=5)\n",
    "print('Probability of receiving a visa:', exp.predict_proba[1])\n",
    "exp.as_pyplot_figure()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Probability of receiving a visa:', 0.5)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAApkAAAEICAYAAAAUfe8pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJEnZfQCxDO6BQlJD/QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11f307d50>"
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Explain a prediction (\"local interpretability\"): \n",
    "# Now we see that the pipeline that takes raw data and returns the prediction \n",
    "# of the trained model now comes in conveniently.\n",
    "example = 608\n",
    "exp = explainer.explain_instance(test[:, :-1][example], my_model.predict_proba, num_features=10)\n",
    "\n",
    "\n",
    "# exp = explainer.explain_instance(test[example], model.predict_proba, num_features=5)\n",
    "print('Probability of receiving a visa:', exp.predict_proba[1])\n",
    "exp.as_pyplot_figure()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10,)"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(current).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Couples probability of staying together:', 0.37614967235285984)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdUAAAEICAYAAAAJGW4GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMzPLxEHVzMwsEwdVMzOzTP4XPvdr7AA9WacAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1206b6210>"
      ]
     },
     "execution_count": 299,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdUAAAEICAYAAAAJGW4GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMzPLxEHVzMwsEwdVMzOzTP4XPvdr7AA9WacAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1206b6210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Using LIME for for relationship management (not advised): Current chance of relationship success.\n",
    "current = [34, 36, 2, 0, 1, 0, 1, 0, 0, 0]\n",
    "exp = explainer.explain_instance(np.array(current), my_model.predict_proba, num_features=5)\n",
    "print('Couples probability of staying together:', exp.predict_proba[1])\n",
    "exp.as_pyplot_figure()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}