Fit_Predictions.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Amazon Business Research Analyst Hiring Challenge: Round 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Gender Neutrality and Inclusion\n",
    "#### Problem\n",
    "Companies are expected to be equal opportunity employers. Their recruitment and selection process is\n",
    "expected to provide equal opportunity irrespective of gender identity or expression, religion, color,\n",
    "sex, age, physical or mental disability, sexual orientation, or any\n",
    "other basis covered by local law. Removal of bias and providing equal opportunity to all applicants\n",
    "promotes access to the widest pool of talent.\n",
    "XyX corporation recruits employees for fixed Job Codes every year. Basis the applications received last\n",
    "year for each job code profile, a ‘fitment %’ percentage was determined based on selections made. XyX followed a fair and equitable approach by personally\n",
    "looking at all parameters and determining the right fit. This year the number of applicants has\n",
    "multiplied and they are looking at an ML model to predict the ‘fitment %’ for the applications received\n",
    "\n",
    "\n",
    "#### Task\n",
    "Build a model that calculates the ‘fitment %’ & detects the factor that influences relevancy and making sure that factor does not introduce inequality and/or bias in the’ fitment\n",
    "%’ by appropriate feature reengineering. Submit a presentation explaining how your model’s predictions will be used by business leaders to analyse and enable an equal-opportunity and\n",
    "bias-free recruitment process.\n",
    "\n",
    "\n",
    "The Code is structured as follows:\n",
    "\n",
    "1. EDA\n",
    "2. Feature Engineering\n",
    "3. Feature Selection\n",
    "4. Update BiasInfluentialFactor field to handle bias\n",
    "5. Model Training\n",
    "6. Model Testing\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ibaruah001\\AppData\\Local\\Annaconda\\lib\\site-packages\\dask\\dataframe\\utils.py:13: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
      "  import pandas.util.testing as tm\n"
     ]
    }
   ],
   "source": [
    "# Import Necessary Libraries\n",
    "\n",
    "#!pip install pyod\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn import linear_model\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.linear_model import Lasso, LassoCV\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from matplotlib import pyplot as plt\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from sklearn.feature_selection import RFE\n",
    "from sklearn.feature_selection import RFECV\n",
    "import statsmodels.api as sm\n",
    "from scipy.stats.mstats import zscore\n",
    "from pyod.models.abod import ABOD\n",
    "from pyod.models.knn import KNN\n",
    "import seaborn as sns\n",
    "from pandas.tseries.holiday import USFederalHolidayCalendar as calendar\n",
    "from pandas import Series\n",
    "from matplotlib import pyplot\n",
    "\n",
    "from category_encoders import TargetEncoder\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import RepeatedKFold\n",
    "from catboost import CatBoostRegressor\n",
    "from sklearn.model_selection import KFold, cross_val_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from hyperopt import fmin, tpe, hp, anneal, Trials\n",
    "import lightgbm as lgb\n",
    "import itertools\n",
    "import math\n",
    "import random\n",
    "import xgboost as xgb\n",
    "import tensorflow\n",
    "import keras\n",
    "from keras.layers import Dropout\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import LSTM\n",
    "from keras.preprocessing.sequence import TimeseriesGenerator\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.svm import SVR\n",
    "from sklearn.preprocessing import LabelEncoder\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>CurrentCTC</th>\n",
       "      <th>ExpectedCTC</th>\n",
       "      <th>MartialStatus</th>\n",
       "      <th>EmpScore</th>\n",
       "      <th>CurrentDesignation</th>\n",
       "      <th>CurrentCompanyType</th>\n",
       "      <th>DepartmentInCompany</th>\n",
       "      <th>TotalLeavesTaken</th>\n",
       "      <th>BiasInfluentialFactor</th>\n",
       "      <th>FitmentPercent</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>11041</td>\n",
       "      <td>John</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Electrical</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>21</td>\n",
       "      <td>26</td>\n",
       "      <td>Married</td>\n",
       "      <td>5</td>\n",
       "      <td>SSE</td>\n",
       "      <td>Enterprise</td>\n",
       "      <td>Design</td>\n",
       "      <td>20</td>\n",
       "      <td>YearsOfExperince</td>\n",
       "      <td>95.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15079</td>\n",
       "      <td>William</td>\n",
       "      <td>English</td>\n",
       "      <td>26</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>15</td>\n",
       "      <td>19</td>\n",
       "      <td>Married</td>\n",
       "      <td>5</td>\n",
       "      <td>BA</td>\n",
       "      <td>MidSized</td>\n",
       "      <td>Engineering</td>\n",
       "      <td>6</td>\n",
       "      <td>NaN</td>\n",
       "      <td>67.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18638</td>\n",
       "      <td>James</td>\n",
       "      <td>English</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>15</td>\n",
       "      <td>24</td>\n",
       "      <td>Single</td>\n",
       "      <td>5</td>\n",
       "      <td>SDE</td>\n",
       "      <td>MidSized</td>\n",
       "      <td>Engineering</td>\n",
       "      <td>19</td>\n",
       "      <td>Gender</td>\n",
       "      <td>91.26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3941</td>\n",
       "      <td>Charles</td>\n",
       "      <td>English</td>\n",
       "      <td>29</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>BCA</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>16</td>\n",
       "      <td>24</td>\n",
       "      <td>Married</td>\n",
       "      <td>5</td>\n",
       "      <td>SDE</td>\n",
       "      <td>Startup</td>\n",
       "      <td>Product</td>\n",
       "      <td>16</td>\n",
       "      <td>Gender</td>\n",
       "      <td>72.29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5936</td>\n",
       "      <td>George</td>\n",
       "      <td>English</td>\n",
       "      <td>25</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR70175</td>\n",
       "      <td>Dual M.Tech</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>24</td>\n",
       "      <td>32</td>\n",
       "      <td>Married</td>\n",
       "      <td>5</td>\n",
       "      <td>SDE</td>\n",
       "      <td>Enterprise</td>\n",
       "      <td>Engineering</td>\n",
       "      <td>10</td>\n",
       "      <td>DegreeBranch</td>\n",
       "      <td>86.34</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0  11041     John                 English   35    Male   \n",
       "1  15079  William                 English   26    Male   \n",
       "2  18638    James                 English   36  Female   \n",
       "3   3941  Charles                 English   29  Female   \n",
       "4   5936   George                 English   25    Male   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289        B.Tech               Electrical   \n",
       "1                 JR87525        B.Tech  Artificial Intelligence   \n",
       "2                 JR87525           PhD         Computer Science   \n",
       "3                 JR87525           BCA   Information Technology   \n",
       "4                 JR70175   Dual M.Tech         Computer Science   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  CurrentCTC  ExpectedCTC  \\\n",
       "0              Tier 1                 7  ...          21           26   \n",
       "1              Tier 3                 7  ...          15           19   \n",
       "2              Tier 1                 6  ...          15           24   \n",
       "3              Tier 2                 5  ...          16           24   \n",
       "4              Tier 3                 8  ...          24           32   \n",
       "\n",
       "   MartialStatus  EmpScore CurrentDesignation  CurrentCompanyType  \\\n",
       "0        Married         5                SSE          Enterprise   \n",
       "1        Married         5                 BA            MidSized   \n",
       "2         Single         5                SDE            MidSized   \n",
       "3        Married         5                SDE             Startup   \n",
       "4        Married         5                SDE          Enterprise   \n",
       "\n",
       "  DepartmentInCompany TotalLeavesTaken BiasInfluentialFactor  FitmentPercent  \n",
       "0              Design               20      YearsOfExperince           95.40  \n",
       "1         Engineering                6                   NaN           67.09  \n",
       "2         Engineering               19                Gender           91.26  \n",
       "3             Product               16                Gender           72.29  \n",
       "4         Engineering               10          DegreeBranch           86.34  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Input the dataset\n",
    "df1 = pd.read_csv('Train.csv')\n",
    "\n",
    "df1.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 13645 entries, 0 to 13644\n",
      "Data columns (total 22 columns):\n",
      " #   Column                   Non-Null Count  Dtype  \n",
      "---  ------                   --------------  -----  \n",
      " 0   EmpID                    13645 non-null  int64  \n",
      " 1   EmpName                  13645 non-null  object \n",
      " 2   LanguageOfCommunication  13645 non-null  object \n",
      " 3   Age                      13645 non-null  int64  \n",
      " 4   Gender                   13645 non-null  object \n",
      " 5   JobProfileIDApplyingFor  13645 non-null  object \n",
      " 6   HighestDegree            13645 non-null  object \n",
      " 7   DegreeBranch             13645 non-null  object \n",
      " 8   GraduatingInstitute      13645 non-null  object \n",
      " 9   LatestDegreeCGPA         13645 non-null  int64  \n",
      " 10  YearsOfExperince         13645 non-null  int64  \n",
      " 11  GraduationYear           13645 non-null  int64  \n",
      " 12  CurrentCTC               13645 non-null  int64  \n",
      " 13  ExpectedCTC              13645 non-null  int64  \n",
      " 14  MartialStatus            13645 non-null  object \n",
      " 15  EmpScore                 13645 non-null  int64  \n",
      " 16  CurrentDesignation       13645 non-null  object \n",
      " 17  CurrentCompanyType       13645 non-null  object \n",
      " 18  DepartmentInCompany      13645 non-null  object \n",
      " 19  TotalLeavesTaken         13645 non-null  int64  \n",
      " 20  BiasInfluentialFactor    10309 non-null  object \n",
      " 21  FitmentPercent           13645 non-null  float64\n",
      "dtypes: float64(1), int64(9), object(12)\n",
      "memory usage: 2.3+ MB\n"
     ]
    }
   ],
   "source": [
    "#Check number of rows and columns, type of each columns\n",
    "df1.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 13645 entries, 0 to 13644\n",
      "Data columns (total 22 columns):\n",
      " #   Column                   Non-Null Count  Dtype  \n",
      "---  ------                   --------------  -----  \n",
      " 0   EmpID                    13645 non-null  object \n",
      " 1   EmpName                  13645 non-null  object \n",
      " 2   LanguageOfCommunication  13645 non-null  object \n",
      " 3   Age                      13645 non-null  int64  \n",
      " 4   Gender                   13645 non-null  object \n",
      " 5   JobProfileIDApplyingFor  13645 non-null  object \n",
      " 6   HighestDegree            13645 non-null  object \n",
      " 7   DegreeBranch             13645 non-null  object \n",
      " 8   GraduatingInstitute      13645 non-null  object \n",
      " 9   LatestDegreeCGPA         13645 non-null  int64  \n",
      " 10  YearsOfExperince         13645 non-null  int64  \n",
      " 11  GraduationYear           13645 non-null  int64  \n",
      " 12  CurrentCTC               13645 non-null  int64  \n",
      " 13  ExpectedCTC              13645 non-null  int64  \n",
      " 14  MartialStatus            13645 non-null  object \n",
      " 15  EmpScore                 13645 non-null  int64  \n",
      " 16  CurrentDesignation       13645 non-null  object \n",
      " 17  CurrentCompanyType       13645 non-null  object \n",
      " 18  DepartmentInCompany      13645 non-null  object \n",
      " 19  TotalLeavesTaken         13645 non-null  int64  \n",
      " 20  BiasInfluentialFactor    13645 non-null  object \n",
      " 21  FitmentPercent           13645 non-null  float64\n",
      "dtypes: float64(1), int64(8), object(13)\n",
      "memory usage: 2.3+ MB\n"
     ]
    }
   ],
   "source": [
    "# Ensure Correct Column Types\n",
    "\n",
    "String_cols = ['EmpID', 'EmpName', 'LanguageOfCommunication', 'Gender',\n",
    "               'JobProfileIDApplyingFor', 'HighestDegree', 'DegreeBranch',\n",
    "               'GraduatingInstitute',  'MartialStatus',\n",
    "               'CurrentDesignation', 'CurrentCompanyType',\n",
    "               'DepartmentInCompany', 'BiasInfluentialFactor']\n",
    "\n",
    "Numeric_cols = ['Age', 'YearsOfExperince',\n",
    "               'GraduationYear', 'CurrentCTC', 'ExpectedCTC',\n",
    "               'EmpScore', 'TotalLeavesTaken','FitmentPercent']\n",
    "\n",
    "\n",
    "\n",
    "df1[String_cols] = df1[String_cols].astype(str)\n",
    "df1[Numeric_cols] = df1[Numeric_cols].apply(pd.to_numeric, errors='coerce')\n",
    "df1.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Age</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>YearsOfExperince</th>\n",
       "      <th>GraduationYear</th>\n",
       "      <th>CurrentCTC</th>\n",
       "      <th>ExpectedCTC</th>\n",
       "      <th>EmpScore</th>\n",
       "      <th>TotalLeavesTaken</th>\n",
       "      <th>FitmentPercent</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "      <td>13645.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>30.691242</td>\n",
       "      <td>7.100257</td>\n",
       "      <td>6.547746</td>\n",
       "      <td>2014.452254</td>\n",
       "      <td>18.076951</td>\n",
       "      <td>25.061488</td>\n",
       "      <td>3.392085</td>\n",
       "      <td>11.967021</td>\n",
       "      <td>75.880093</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>3.881378</td>\n",
       "      <td>1.346539</td>\n",
       "      <td>3.426920</td>\n",
       "      <td>3.426920</td>\n",
       "      <td>7.208130</td>\n",
       "      <td>7.478119</td>\n",
       "      <td>1.044676</td>\n",
       "      <td>4.897836</td>\n",
       "      <td>14.894207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>24.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2009.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>33.530000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>28.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>2012.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>68.720000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>31.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>2014.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>78.110000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>34.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>2017.000000</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>31.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>87.990000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>42.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>2020.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>95.500000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Age  LatestDegreeCGPA  YearsOfExperince  GraduationYear  \\\n",
       "count  13645.000000      13645.000000      13645.000000    13645.000000   \n",
       "mean      30.691242          7.100257          6.547746     2014.452254   \n",
       "std        3.881378          1.346539          3.426920        3.426920   \n",
       "min       24.000000          4.000000          1.000000     2009.000000   \n",
       "25%       28.000000          6.000000          4.000000     2012.000000   \n",
       "50%       31.000000          7.000000          7.000000     2014.000000   \n",
       "75%       34.000000          8.000000          9.000000     2017.000000   \n",
       "max       42.000000         10.000000         12.000000     2020.000000   \n",
       "\n",
       "         CurrentCTC   ExpectedCTC      EmpScore  TotalLeavesTaken  \\\n",
       "count  13645.000000  13645.000000  13645.000000      13645.000000   \n",
       "mean      18.076951     25.061488      3.392085         11.967021   \n",
       "std        7.208130      7.478119      1.044676          4.897836   \n",
       "min        6.000000     10.000000      1.000000          4.000000   \n",
       "25%       12.000000     19.000000      3.000000          8.000000   \n",
       "50%       18.000000     25.000000      3.000000         12.000000   \n",
       "75%       24.000000     31.000000      4.000000         16.000000   \n",
       "max       30.000000     40.000000      5.000000         20.000000   \n",
       "\n",
       "       FitmentPercent  \n",
       "count    13645.000000  \n",
       "mean        75.880093  \n",
       "std         14.894207  \n",
       "min         33.530000  \n",
       "25%         68.720000  \n",
       "50%         78.110000  \n",
       "75%         87.990000  \n",
       "max         95.500000  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Understanding the numeric fields\n",
    "df1.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3cAAANeCAYAAACxkPfoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3XucZFV57//PV8YLAsrVkZuOSfCCYoggaDwxE68IRvSIBiUIhoSYQNQ458TRXDDegp7jNRoNRgQNgogaiRAVkdHjL4KKoohoQBxlAEHkOqDo4PP7Y68ONU13T0/fqnr35/161au7Vu2969lVXU+vZ++1V6WqkCRJkiQtbvcYdgCSJEmSpNmzuJMkSZKkHrC4kyRJkqQesLiTJEmSpB6wuJMkSZKkHrC4kyRJkqQesLiTJEmSFoEkeyb5ZpJbk/z5sOMZk2SLJOuTPGjYsSx1FneaM0nWJLkxyb2HHYukhZNkbZKnbOY6a5L88Rw898ok68a1vSbJL1vn59Yk/5XkXUl2nu3zzZckD03y0STXJ7k5ybeSvCLJFu3xeyX5uyTfS3JbkquS/EeSpw1sY22Sn7UO1rVJPpBk63HPc1KSDUl2Weh9lEZVklOSnDiu7XeT/HQh80aS7ZO8t31+b2954EXjFnsl8Nmq2qaq/inJvyb5Rfvcj90uXKiYx1TVnVW1dVX9aKGfWxuzuNOcSLIC+B2ggGcNNRhJgo9U1TbA9sBzgAcCF85HRy3Jslmu/+vABcCVwF5VdX/gecC+wDZtsTOAg4EXAdsBDwHeARw0bnO/X1VbA48BHgv8zcDzbAU8F7gZOGw2MUs981LgwCRPBUhyH+B9wKqqumaunmTsYM0kj90bOBfYDdgP2BZYDbwlyUsHFn0wcMm41d/YCqux2z5zFfN0zDYHam5Z3GmuvAg4HzgJOGKsMckOSf49yS1Jvprk9Um+NPD4w5Ock+SGdkT6+QsfuqS5lmS7JJ9K8pN2Rv9TSXZrj72B7mDQu9pR5ne19knzQZIDk3ynnYm7Ksn/asXKfwC7DByx3uiMVFX9sqouAf4A+AmwamCbz0xyUZKbkvxnkkcPPPaYJN9oz/fRJB9J8vr22Mok65K8MsmPgQ9MY3u7JPlYez1+MK6z9vfAf1bVK8Y6klX1vap6YVXd1M6KPhU4uKouqKpftNunq+plE73+VXVVe20eNdD8XOAm4LUM5GlpqauqnwJ/AZzQ8spxwPer6qQk90jy6iTfb2fWT0uyHUB77IwkP26f+zVJHjG23XZW7d1JPp3kNuB3Wp64tOWWdUn+si1+JLAz8Pyq+mH7jJ8NvBx4fZKtk3yRLne+t+W7X5tqv5IcluTysTP4SX4/ydWtb7YsSSX5i5aTrk9yfJJ7DKz/x0m+23L4fyTZvbWPrfvnSS4HvjvQtmJg39/Z1rs1yZeTPGRg23sl+VzL9z9O8lcDr+mEr7emx+JOc+VFwCnt9vQky1v7u4Hb6I6aH8HGhd9WwDnAh4EHAC8A/inJIxcwbknz4x50Rc+DgQcBPwPeBVBVfw38P+DYdpT52Gnkg/cDf9rOxj0K+HxV3QY8A7h64Ij11RMFU1V3Ap+k6xiR5DHAicCfAjsA/wycmeTeSe4FfILuYNX2wKl0Z/8GPbA99mDg6E1s7x7AvwPfBHYFngy8PMnT27aeQndmbjJPAS6oqnVTLLOR1gk7EPjGQPMRbV9OAx7eYpYEVNVHgQvpPiNH032WAV5Bd4b8iXRn1W4D3jmw6qeAPehywreBD43b9AvpDuBsA3yZLi8e1XLZo4EvtOWeCpxVVbePW/8MYGtg/6p6YtvGS1q+u2IT+3RK26e3J9mJ7mzkH7VidszBdGf69wUOoevPkeQQ4H+3x3eiG13w4XFP8Sy6EQJ7TRLCC4G/pcuVPwJe17Z9f+BzdHlxZ+ChwJq2zqZeb22CxZ1mLcn/oOvgnF5VFwLfB16YbvjBc4Hjqur2qvoOcPLAqs8E1lbVB6pqQ1V9HfgYXXKRtIhV1U+r6mPts38r8Abgd6dYZVP54JfAnknuV1U3tsc319V0nQyAPwH+uZ0Ju7OqTgbuAB7XbsuAd7Yzfx8HvjJuW7+iy213VNXPNrG9xwI7VdVr29H4K+g6WYe2be0ATDX0a0fgx2N30l2Xc1O6a/N+Pm7Zf0tyE/Aluk7jG9s6DwJ+D/hwVV1LN/zLs3fSxo4BngS8duDasT8FXl1VV1XVz4HXAM9Pco+q+lVVnVRVtw48tk87WDXmE1X15bbsHdyVy7apqhsGctmOTJAH2jo3tMcns7rlhLHb+wceewlwAPB54GNV9elx6x7fcupauiLqBQP7/cY2imAD8HpgvyS7Dqz7xrbuzyaJ64yq+lpV/ZLu4P/erf1ZwJVV9Y6WQ2+pqrEcO+nrPcX+a4AvlObCEXQX917f7n+4te1E10G6cmDZwd8fDOw/mJDorgN54ALELGkeJblvkn9O8sMktwBfBLbN5NecbCofPJfuTNQPk3whyeNnENaudJ2ksedbNe75dgd2aberqqoG1r1y403xk9bxGIx/su09mG7o6OBjrwbGRjj8lO7o9WQ2erx1CLcF9gHGT2D17KratqoeXFV/PtDpOhy4tKouavdPoTsId88pnldaUtqBj+vZ+Jq2BwH/PvDZvZhufoEHpJsh8s1Jrmh57vK2zmAhNj53PIeuuPlRG8a5f2u/ngnyQLpr8bZvj0/m+Pa5H7sdNbBPN9IdKHsU8JYJ1h2M74d0OQu6vPXugf2+nu6g1m5T7Nt4Px74/Xa6M5DQ5cbL7744MMXrvYnnUmNxp1lJsiXwfOB325jpHwN/CfwmXcdlAxsngt0Hfr8S+MK4hLR1Vf3ZQsUvad6sAh5GN5TofnRDbADSfta45afMB1X11ao6mO4f/L8Bp0+ynQm1o76/TzccdOz53jDu+e5bVafSHT3fNUkGNrH7uE1OFP9k27sS+MG4x7apqgPbup+jK14ncy7w2LRrFmfoRcCvDeTpt9J1QJ8xi21KS8E64KnjPr/3qaof032uDqQ723d/4DfaOoO5Y6Nc0c7uP4sul32Kbpg0dHngoCT3Hff8h9ANTRw/emBakuxDd3DnI0w8vHEwtz2IboQDdHnrqHH7vWVVXTC4OzOJqW371yd5bKrXW9NgcafZejZwJ7An3en2vYFH0HWgXgR8HHhNO4r/8NY25lPAQ5McnuSe7fbYDFyMLGnRuGeS+4zd6GZ0/BlwU5Lt6SYoGHQtMDgZwKT5IN3XAByW5P5teM8tdHlnbDs7tGs47qZt5xF019E8kK6ogW5Y5EuS7J/OVkkOSjJ2XcydwLHpJgk4mG72uqlMtb2vALekm4Bly3a0/1FJHtvWPQ747ST/J8kDW9y/kW5Cgm2r6rPAeXRDLvdvr8c96YZ8blI7y/nrbR/G8vSjuGuUhaTJvRd4YxvaTJIHJBmbFXwbuuHXPwXuSzf8fFLt8//CNrz8l8Ct3JXLTqbLZx9J8uCWu54BvA34uza8fbO0A/D/Svf1CUfSHeA5etxif5Vk27Z/L6UrAsf2+6/H+mRtmbm6bOZM4EFJjm357H5JxnLsVK+3psHiTrN1BPCBqvpRVf147EY3ccJhwLF0R7N+THeR8al0iZCWqJ5Gd93J1W2ZN3H3YUaSRt/ZdMXc2G1bYEu6oTznA+Ov83gHcEi6WdjeOY18cDiwtg19egnwhwBV9V26vHJFG8YzNqToD5Ksp5sd8ky6ztc+1SZcqaqv0V0n9y7gRrohQke2x34B/E/gqLb+H9IVn3dMtvOb2N6ddGcN9wZ+0F6Tf6HLjVTV94HHAyuAS5LcTDeM6mt0nT9aPJ+i66jd1LZzGN21NJtyBPDJqrp4XJ5+B/DMVnxLmthb6fLXuUluBf6T7jpa6CZHubrdLmmPbcoRdMPLb6HLMYcDtCHUT6LLfV+lO4j1ZuCVVfW2TWzz1dn4e+7GznK9mW7Wz/e1YeR/CByf7utXxvw7cBHd5EtjE0mNTTDzVuCjLdZvAU9nDlTVzXQTyDwXuA74L+66Jnuq11vTkI0vKZDmV5I3AQ+sKo8WS1o0klwAvLeqPjDsWCRpttJ9N90vgYe0yVTUE56507xK971Vj27DlPajO0r1iWHHJUlTSfK7SR7YhmUeQTdl+fizj5IkjRS/UV7zbRu6IVO70J16fwvdd01J0ih7GN2kLVvTfb3LIdW+YFySpFHlsExJkiRJ6gGHZUqSJElSD4z0sMwdd9yxVqxYMS/bvu2229hqq63mZdujZKnsJ7ivw3DhhRdeX1U7DTuOUTPd3DUq7yOMViwwWvEYy8RGKRaYfjzmrYnNZ59rKqP2dzRmVOOC0Y3NuDbfvOStqhrZ2z777FPz5bzzzpu3bY+SpbKfVe7rMABfqxHIFaN2m27uGpX3sWq0YqkarXiMZWKjFEvV9OMxb80ub821Ufs7GjOqcVWNbmzGtfnmI285LFOSJEmSesDiTpIkSZJ6wOJOkiRJknrA4k6SJEmSesDiTpIkSZJ6wOJOkiRJknrA4k6SJEmSesDiTpIkSZJ6wOJOkiRJknpg2bAD0NKyYvVZc77NtccfNOfblKQxF191M0fOce4yb0nS4jIffdiTDthqzrfpmTtJkiRJ6gGLO0mSJEnqAYs7SZIkSeoBiztJkiRJ6gEnVFlgTigiSZIkaT545k6SJEmSesDiTpIkSZJ6wOJOkiRJknrA4k6SJEmSesAJVSRJkjRtczk53Kq9NnDk6rOcHE6aI565kyRJkqQe8MxdD0x1BG3siNhMeBRNkiRJWjw8cydJkiRJPWBxJ0mSJEk9YHEnSZIkST1gcSdJkiRJPWBxJ0mSJEk9YHEnSZIkST1gcSdJkiRJPWBxJ0mSJEk9YHEnSZIkST1gcSdJkiRJPWBxJ0mSJEk9YHEnSZI0ApLsnuS8JJcmuSTJy1r79knOSXJZ+7lda0+Sdya5PMm3kjxmYFtHtOUvS3LEsPZJ0sLaZHFnopEkSVoQG4BVVfUI4HHAMUn2BFYD51bVHsC57T7AM4A92u1o4D3Q9dGA44D9gf2A48b6aZL6bTpn7kw0kiRJ86yqrqmqr7ffbwUuBXYFDgZOboudDDy7/X4w8MHqnA9sm2Rn4OnAOVV1Q1XdCJwDHLCAuyJpSJZtaoGquga4pv1+a5LBRLOyLXYysAZ4JQOJBjg/yViiWUlLNABJxhLNqXO4P5IkSYtekhXAbwEXAMtbf4yquibJA9piuwJXDqy2rrVN1j7+OY6mOxDP8uXLWbNmzbRiW7XXhunvyCYs37Lb3nSfe6GsX79+5GIaM6qx9T2uufy7HzMfr9kmi7tBo5xoNtew/gDn4w9jKmNJcybm4/WZj/0fi3NUk8p8WEr7KklLTZKtgY8BL6+qW5JMuugEbTVF+8YNVScAJwDsu+++tXLlymnFd+Tqs6a13HSs2msDb7l4GWsPm95zL5Q1a9Yw3ddjoY1qbH2Pay7/7secdMBWc/6aTbu4G/VEs7mG9Qc4H38YUxlLmjMxH4l2PvZ/LM5RTSrzYSntqyQtJUnuSdffOqWqPt6ar02yczuYvjNwXWtfB+w+sPpuwNWtfeW49jXzGbek0TCt2TKnSjTt8ekmmonaJUmSlrx0R87fD1xaVW8deOhMYGwiuiOATw60v6hNZvc44OY2quozwNOSbNfmN3haa5PUc9OZLdNEI2lRSrJFkm8k+VS7/5AkF7QZez+S5F6t/d7t/uXt8RUD23hVa/9ekqcPZ08kLRFPAA4HnpTkonY7EDgeeGqSy4CntvsAZwNXAJcD7wP+HKDNb/A64Kvt9tqxOQ8k9dt0xuuNJZqLk1zU2l5Nl1hOT3IU8CPgee2xs4ED6RLN7cCLoUs0ScYSDZhoJM2/l9HNNne/dv9NwNuq6rQk7wWOopvR9yjgxqr6jSSHtuX+oM0MfCjwSGAX4HNJHlpVdy70jkjqv6r6EhNfxgLw5AmWL+CYSbZ1InDi3EUnaTGYzmyZJhpJi06S3YCDgDcAr2ijEJ4EvLAtcjLwGrri7uD2O8AZwLva8gcDp1XVHcAPklxO91UuX16g3ZAkSZq2mc20IUmj7+3AXwHbtPs7ADdV1diUrYMz9v73bL5VtSHJzW35XYHzB7Y54Sy/MLOZfkdp1tNRigVGK57ZzDo8mZnu2yi9LqMUC4xePJI0DBZ3knonyTOB66rqwiQrx5onWLQ28di0ZvmFmc30O0qzno5SLDBa8fzjKZ+c8azDk5npbMSj9LqMUiwwevFI0jBY3EnqoycAz2oTEdyH7pq7twPbJlnWzt4Nztg7NpvvuiTLgPsDN+Asv5IkaRGZ1lchSNJiUlWvqqrdqmoF3YQon6+qw4DzgEPaYuNn+R2b/feQtny19kPbbJoPAfYAvrJAuyFJkrRZPHMnaSl5JXBaktcD36D7mhfazw+1CVNuoCsIqapLkpwOfAfYABzjTJmSJGlUWdxJ6rWqWgOsab9fQTfb5fhlfs5dX+cy/rE30M24KUmSNNIclilJkiRJPWBxJ0mSJEk9YHEnSZIkST3gNXdTWLH6rGGHIEmSJEnT4pk7SZIkSeoBiztJkiRJ6gGLO0mSJEnqAYs7SZIkSeoBiztJkiRJ6gGLO0mSJEnqAYs7SZIkSeoBiztJkiRJ6gGLO0mSJEnqAYs7SZIkSeoBiztJkiRJ6gGLO0mSJEnqAYs7SZIkSeqBZcMOQJKkpWbF6rNmtN6qvTZw5CTrrj3+oNmEJEnqAc/cSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiSNgCQnJrkuybcH2l6T5KokF7XbgQOPvSrJ5Um+l+TpA+0HtLbLk6xe6P2QNDybLO5MNJIkSQviJOCACdrfVlV7t9vZAEn2BA4FHtnW+ackWyTZAng38AxgT+AFbVlJS8B0ztydhIlGkiRpXlXVF4Ebprn4wcBpVXVHVf0AuBzYr90ur6orquoXwGltWUlLwCaLOxONJEnSUB2b5FttNNV2rW1X4MqBZda1tsnaJS0By2ax7rFJXgR8DVhVVTfSJY/zB5YZTCjjE83+E200ydHA0QDLly9nzZo1swhxcuvXr9/ktlfttWFennshLd9y5vsxH6/9fLymY3FO5z3ti6W0r5K0xL0HeB1Q7edbgD8CMsGyxcQH7muiDc+0zzWX/8vH+imj9j9tlP/PjmpsfY9rPvqw8/GazbS4m7dEU1UnACcA7LvvvrVy5coZhji1NWvWsKltH7n6rHl57oW0aq8NvOXimb3Naw9bObfBMD+v6Vic03lP+2Ip7etMJLkP8EXg3nR57oyqOi7JQ+hGDmwPfB04vKp+keTewAeBfYCfAn9QVWvbtl4FHAXcCby0qj6z0PsjaemqqmvHfk/yPuBT7e46YPeBRXcDrm6/T9Y+ftsz6nPN5f/ysX7KfPQ5ZmOU/8+Oamx9j2s++rAnHbDVnL9mM+r1z2eikTbXivZhW7XXhjn74K09/qA52Y6G5g7gSVW1Psk9gS8l+Q/gFXTXC5+W5L10Rdt72s8bq+o3khwKvAn4g3HXEe8CfC7JQ6vqzmHslKSlJ8nOVXVNu/scYGyCuzOBDyd5K11+2gP4Ct2B9j3awayr6HLYCxc2aknDMqOvQkiy88Dd8Ynm0CT3bkllLNF8lZZoktyLLtGcOfOwJWly1Vnf7t6z3Qp4EnBGaz8ZeHb7/eB2n/b4k5OEya8jlqQ5l+RU4MvAw5KsS3IU8OYkFyf5FvB7wF8CVNUlwOnAd4BPA8dU1Z1VtQE4FvgMcClweltW0hKwyTN3LdGsBHZMsg44DliZZG+6ztJa4E+hSzRJxhLNBlqiadsZSzRbACfOdaJZsZlnbObyLI+k0dNm6b0Q+A262Xq/D9zUOj6w8TXB/z0BQVVtSHIzsANTX0c8/vk2+9qVUbo+YZRigdGKZzbXLs+1qWJZ6NdrlN4jGL14ZqKqXjBB8/unWP4NwBsmaD8bOHsOQ5O0SGyyuDPRSFqM2oGlvZNsC3wCeMREi7Wfk10vPFn7RM+32deujNL1CaMUC4xWPP94yidnfO3yXJvqOuqFvmZplN4jGL14JGkYZjQsU5IWi6q6CVgDPA7YNslYz3jw2t//vl64PX5/uq+Ameo6YkmSpJFicSepd5Ls1M7YkWRL4Cl0156cBxzSFjsC+GT7/cx2n/b456uqmPw6YkmSpJEzGuNMJGlu7Qyc3K67uwfdhAKfSvId4LQkrwe+wV1DzN8PfCjJ5XRn7A6Fqa8jliRJGjUWd5J6p6q+BfzWBO1XMMFsl1X1c+B5k2xrwuuIJUmSRo3DMiVJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpB5YNOwBJkiRJS8+K1WfNy3bXHn/QvGx3MfDMnSRJkiT1gGfuNKn5OpoiSZKkzub2t1bttYEjN7HOUj5ztdR55k6SJEmSesDiTpIkSZJ6wOJOkiRJknrA4k6SJEmSesAJVSRJktQ7s5kYbjqTlkijyOJOkiRJ6pH5mPHcGTgXB4dlSpIkSVIPeOZOkiRJQ+V3646+Tb1HozSUdTDWUYprIXjmTpIkSZJ6wOJOkiRJknrA4k6SJEmSesDiTpIkSZJ6wOJOkiRJknrA4k6SJEmSesDiTpIkaQQkOTHJdUm+PdC2fZJzklzWfm7X2pPknUkuT/KtJI8ZWOeItvxlSY4Yxr5IGo5NFncmGkmSpAVxEnDAuLbVwLlVtQdwbrsP8Axgj3Y7GngPdH004Dhgf2A/4Lixfpqk/pvOmbuTMNFIkiTNq6r6InDDuOaDgZPb7ycDzx5o/2B1zge2TbIz8HTgnKq6oapuBM7h7v04ST21yeLORCNJkjQ0y6vqGoD28wGtfVfgyoHl1rW2ydolLQHLZrjeRokmyZwlmiRH0531Y/ny5axZs2ZaAa3aa8NmhA/Lt9z8dRajpbKfMLf7Ot2/u2FZv379yMc4TEl2Bz4IPBD4FXBCVb2jjSL4CLACWAs8v6puTBLgHcCBwO3AkVX19batI4C/aZt+fVWdjCQNXyZoqyna776BBepzTWVU+ymjGheMbmzGtfnmoz830+JuMrNONFV1AnACwL777lsrV66c1hMfufqs6UXYrNprA2+5eK53f/Qslf2Eud3XtYetnJPtzJc1a9Yw3c/GErUBWFVVX0+yDXBhknOAI+mGlB+fZDXdkPJXsvGQ8v3phpTvPzCkfF+6nHVhkjPbCARJWgjXJtm5HUzfGbiuta8Ddh9Ybjfg6ta+clz7mok2vFB9rqmMaj9lVOOC0Y3NuDbfSQdsNef9uZnOlnltSzBsRqKZqF2S5lxVXTN25q2qbgUupRst4JBySYvNmcDYRHRHAJ8caH9Rm8zuccDNbVTVZ4CnJdmuzW/wtNYmaQmYaRk7lmiO5+6J5tgkp9Ed/b65HWn6DPDGgUlUnga8auZhS9L0JFkB/BZwASM2pHyUhteOUiwwWvGM0pCeqWJZ6NdrlN4jGL14ZiLJqXRn3XZMso5u5MDxwOlJjgJ+BDyvLX423VDyy+mGk78YoKpuSPI64KttuddW1fi5EyT11CaLOxONpMUqydbAx4CXV9Ut3aV1Ey86Qdu8DykfpeG1oxQLjFY8/3jKJ0dmSM9Uw4sWejj5KL1HMHrxzERVvWCSh548wbIFHDPJdk4ETpzD0CQtEpv8b2WikbQYJbknXWF3SlV9vDXP27UrkiRJwzbTa+4kaWS12S/fD1xaVW8deMhrVyRJUm+NxjgTSZpbTwAOBy5OclFrezUOKZckST1mcSepd6rqS0x8vRw4pFySJPWUwzIlSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQecUEWSNBQrVp816WOr9trAkVM8Ppm1xx80m5AkSVrUPHMnSZIkST1gcSdJkiRJPWBxJ0mSJEk9YHEnSZIkST1gcSdJkiRJPWBxJ0mSJEk9YHEnSZIkST1gcSdJkiRJPWBxJ0mSJEk9sGzYAUiSpNG0YvVZkz62aq8NHDnF45NZe/xBswlJkjQFiztpAlN1aGbKDo0kSZLmk8MyJUmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHls1m5SRrgVuBO4ENVbVvku2BjwArgLXA86vqxiQB3gEcCNwOHFlVX5/N80uSNrZi9VkzWm/VXhs4cop11x5/0ExDkjQH7HNJmo65OHP3e1W1d1Xt2+6vBs6tqj2Ac9t9gGcAe7Tb0cB75uC5JWlCSU5Mcl2Sbw+0bZ/knCSXtZ/btfYkeWeSy5N8K8ljBtY5oi1/WZIjhrEvktTY55I0pfkYlnkwcHL7/WTg2QPtH6zO+cC2SXaeh+eXJICTgAPGtW1WR6gdFT8O2B/YDzhurCCUpBFgn0vSRmY1LBMo4LNJCvjnqjoBWF5V1wBU1TVJHtCW3RW4cmDdda3tmsENJjmarnPF8uXLWbNmzbQCWbXXhs0KfPmWm7/OYrRU9hNGf1+n+7c8HevXr5/T7fVRVX0xyYpxzQcDK9vvJwNrgFcy0BECzk8y1hFaCZxTVTcAJDmHrmA8dZ7Dl6TxFm2fayqj+r97VOOC0Y3NuDbffPTnZlvcPaGqrm7J5Jwk351i2UzQVndr6JLVCQD77rtvrVy5clqBTHWtyERW7bWBt1w8290ffUtlP2H093XtYSvnbFtr1qxhup8NbWRzO0KTtd/NTDpJ85HUZ/oPbFP//ObjYMJUzzc35bGDAAAgAElEQVTTf8bzEecodQymimUpv0ewJA56Ldo+11RG9X/3qMYFoxubcW2+kw7Yas77c7Pa06q6uv28Lskn6IYtXZtk59Zx2hm4ri2+Dth9YPXdgKtn8/ySNEcm6whNq4MEM+skzUeRPtNO16b++c3lwYkxU8U603/G8xHnP57yyZHpGEz1uizl9wj6f9DLPpek6ZjxNXdJtkqyzdjvwNOAbwNnAmOTDhwBfLL9fibwojZxweOAm8eOoEvSArl27LqTaXaE7CBJGjr7XJKmazaHIpcDn+hm22UZ8OGq+nSSrwKnJzkK+BHwvLb82XRT8l5ONy3vi2fx3JI0E2MdoeO5e0fo2CSn0U2ecnM7Ev4Z4I0Dk6g8DXjVAscsSfa5JE3LjIu7qroC+M0J2n8KPHmC9gKOmenzSdLmSHIq3YQoOyZZRzfr5fFsRkeoqm5I8jrgq225145NriJJC8U+l6TpGo2LCCRpjlXVCyZ5aLM6QlV1InDiHIYmSZI0L+bje+4kSZIkSQvM4k6SJEmSesDiTpIkSZJ6wOJOkiRJknrA4k6SJEmSesDiTpIkSZJ6wOJOkiRJknrA77mTpCG5+KqbOXL1WcMOQ5Ik9YRn7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpByzuJEmSJKkHLO4kSZIkqQcs7iRJkiSpB5YNOwBJkrR0rFh91rxs96QDtpqX7UrSYuKZO0mSJEnqAYs7SZIkSeoBh2VKC2QuhyKt2msDR87T0Ka1xx80L9uVJEnS/PLMnSRJkiT1gMWdJEmSJPWAxZ0kSZIk9YDFnSRJkiT1gMWdJEmSJPWAxZ0kSZIk9cCCF3dJDkjyvSSXJ1m90M8vSZvLvCVpsTFvSUvTghZ3SbYA3g08A9gTeEGSPRcyBknaHOYtSYuNeUtauhb6zN1+wOVVdUVV/QI4DTh4gWOQpM1h3pK02Ji3pCUqVbVwT5YcAhxQVX/c7h8O7F9Vxw4sczRwdLv7MOB78xTOjsD187TtUbJU9hPc12F4cFXtNOwg5tN08lZrn0nuGpX3EUYrFhiteIxlYqMUC0w/HvPWXcstVJ9rKqP2dzRmVOOC0Y3NuDbfnOetZbOLZ7NlgraNqsuqOgE4Yd4DSb5WVfvO9/MM21LZT3BfNW82mbdgZrlrlN7HUYoFRiseY5nYKMUCoxfPkM1b3ppro/q+jWpcMLqxGdfmm4/YFnpY5jpg94H7uwFXL3AMkrQ5zFuSFhvzlrRELXRx91VgjyQPSXIv4FDgzAWOQZI2h3lL0mJj3pKWqAUdlllVG5IcC3wG2AI4saouWcgYBgx1GMICWir7Ce6r5sE8561Reh9HKRYYrXiMZWKjFAuMXjxDM2L9rU0Z1fdtVOOC0Y3NuDbfnMe2oBOqSJIkSZLmx4J/ibkkSZIkae5Z3EmSJElSD/S+uEuye5Lzklya5JIkL2vtr0lyVZKL2u3AYcc6W0nuk+QrSb7Z9vXvW/tDklyQ5LIkH2kXVy9qU+zrSUl+MPC+7j3sWOdCki2SfCPJp9r93r2nS9H493WIcaxNcnH7zHxtyLFsm+SMJN9tefvxQ4zlYQO55KIktyR5+RDj+cuW776d5NQk9xliLC9rcVyy0K9JkhOTXJfk2wNt2yc5p+XEc5Jst5AxafNM1jcbFaOSm8cbpfw43qjkp1HOD5PE9n/a+/mtJJ9Isu1sn6f3xR2wAVhVVY8AHgcck2TP9tjbqmrvdjt7eCHOmTuAJ1XVbwJ7AwckeRzwJrp93QO4EThqiDHOlcn2FeB/D7yvFw0vxDn1MuDSgft9fE+XovHv6zD9XvvMDPu7gN4BfLqqHg78JkN8farqe2O5BNgHuB34xDBiSbIr8FJg36p6FN0kGYcOKZZHAX8C7Ef3Hj0zyR4LGMJJwAHj2lYD57aceG67r9E1Vd9sFIxSbh40Mvlx0CjlJ0Y7P5zE3WM7B3hUVT0a+C/gVbN9kt4Xd1V1TVV9vf1+K90HYdfhRjU/qrO+3b1nuxXwJOCM1n4y8OwhhDenptjX3kmyG3AQ8C/tfujhe7rUjH9fBUnuBzwReD9AVf2iqm4ablT/7cnA96vqh0OMYRmwZZJlwH0Z3veWPQI4v6pur6oNwBeA5yzUk1fVF4EbxjUfTJcLwZw48ka5bzaquXnE8yOMSH4a5fwwUWxV9dmWRwHOp/tOylnpfXE3KMkK4LeAC1rTse006Il9GcLRhhJcBFxHdzTg+8BNA3846xiRBDpb4/e1qsbe1ze09/VtSe49xBDnytuBvwJ+1e7vQE/f0yVm/Ps6TAV8NsmFSY4eYhy/BvwE+EAbEvUvSbYaYjyDDgVOHdaTV9VVwP8FfgRcA9xcVZ8dUjjfBp6YZIck9wUOZOMvzB6G5VV1DXSFA/CAIcejaZqgbzZso5SbB41sfhyx/DSRxZIf/gj4j9luZMkUd0m2Bj4GvLyqbgHeA/w63ZC+a4C3DDG8OVNVd7YhRLvRDZl5xESLLWxU82P8vrahQq8CHg48FtgeeOUQQ5y1JM8ErquqCwebJ1i0F+/pUjHJ+zpMT6iqxwDPoBse9cQhxbEMeAzwnqr6LeA2RmB4Xbum9VnAR4cYw3Z0R58fAuwCbJXkD4cRS1VdSjc0/Bzg08A36YbZSZtlgr7ZsOMZtdw8aCTzI4xWflqskvw1XR49ZbbbWhLFXZJ70iWPU6rq4wBVdW0rDn4FvI+uEOqNdqp+Dd1Y9m3baXLoCqFhDeWZFwP7ekAb6lFVdQfwARb/+/oE4FlJ1gKn0Q3HfDs9f0+XgLu9r0n+dVjBVNXV7ed1dNeUDetzsw5YN3AW/gy6zsywPQP4elVdO8QYngL8oKp+UlW/BD4O/Pawgqmq91fVY6rqiXTDjC4bVizNtUl2Bmg/rxtyPNqEifpmI2CkcvM4o5ofYcTy0wRGOj8kOQJ4JnBYzcEXkPe+uGvXJ70fuLSq3jrQvvPAYs+hG2ayqCXZaWyWnSRb0n3YLgXOAw5pix0BfHI4Ec6dSfb1uwMf3tCNqV7U72tVvaqqdquqFXTDwj5fVYfRw/d0KZnkfR3KUc4kWyXZZux34GkM6XNTVT8GrkzysNb0ZOA7w4hlnBcwxCGZzY+AxyW5b8tvT2aIkykkeUD7+SDgfzL81+dMulwI5sSRN1nfbNhGKTePN8L5EUYsP01gZPNDkgPoRpk9q6pun4ttLtv0IoveE4DDgYvb9VkArwZekG6a/ALWAn86nPDm1M7AyUm2oCvcT6+qTyX5DnBaktcD36BdjLvITbavn0+yE93QxYuAlwwzyHn0Svr3nmo4lgOf6P4fswz4cFV9eojx/AVwShsKeQXw4iHGQrum7KkM+X9EVV2Q5Azg63RDd74BnDDEkD6WZAfgl8AxVXXjQj1xklOBlcCOSdYBxwHHA6cnOYquo/m8hYpHMzJh36wnM5fPp5HKj2NGKT+Ncn6YJLZXAfcGzmn/h8+vqln1XTMHZ/8kSZIkSUPW+2GZkiRJkrQUWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcSZIkSVIPWNxJkiRJUg9Y3EmSJElSD1jcaVJJ1iZ5yjxt+71J/nY+ti1JkqTpS3Jkki8NOw7NnsXdIpbk0CQXJLktyXXt9z9PkmHHNmiihFFVL6mq181yu49McnOSh45rPzfJP8xm25JGR5IXJvlakvVJrknyH0n+x7DjgokPgiW5V5LXJLms5ee1SU5MsiLJJW0/1ie5M8nPB+6/uq2/c5L3t329Ncl3k/x9kq2Gs5eSJtI+2z8b+AyvT/KuIcWyJskfz+H2kuSlSb7d8ti6JB9NslfLwWP7+8skvxi4/962/v2SvD3Jj1r75e3+jnMVoyZmcbdIJVkFvAP4P8ADgeXAS4AnAPeaYPktFjTABVBVlwD/F3j/WEGb5ChgV+Dv5/K5kiyby+1Jmp4krwDeDryRLs89CPgn4ODN3M7dPsPz+Lk+A3gW8ELg/sBvAhcCT66qR1bV1lW1NfD/gGPH7lfVG5NsD3wZ2BJ4fFVtAzwV2Bb49XmKV9LM/f7AZ3jrqjp22AHNkXcALwNeCmwPPBT4N+CgqnrGQB47BXjzwP6/JMm9gHOBRwIHAPcDfhv4KbDfEPZlaakqb4vsRtdZuA147hTLnAS8Bzi7LfsU4CDgG8AtwJXAa8atczjwQ7oP318Da4GnDGzv9QPLrgTWDdxfDXwfuBX4DvCc1v4I4OfAncB64KZJtvcnwOXADcCZwC4DjxVd4XoZcCPwbiDtsWVtn46h6/hdD/z2wLp7Ap9r2/3u4GtG1/m6qMX8I+BvBx77jfa8L26PfX7Y77s3b0vt1nLdeuB5kzy+qby0Fngl8C3gjpYvJmrbBfgY8BPgB8BLB7bxGuB04IMtV1wC7Nse+xDwK+BnLc6/arn2Z8Du09i/NcAfj2t7PXAxcI9hv/7evHmb+jbYTxrX/h7gjIH7b6IrdjKWp4BXtz7LWuCwgWXvTXfg+kfAtcB7gS0HHj+49V1uoet3HQC8ga6f9fOWi97Vln04cE7rA30PeP7Adnag62/dAnwFeB3wpfbYHm17+03jNdgoD7e2P26xbz3s92gp3jxztzg9nu7D/8lNLPdCug/8NsCX6Iq8F9EdAT4I+LMkzwZIsiddMjqcrqOzA7DbZsT0feB36Dpjfw/8a5Kdq+pSusLsy9Ud0dl2/IpJngT8A/B8YGe6AvO0cYs9E3gs3RHw5wNPB6iqDcAf0SWlfwX+tar+s213G7qk9kHgAcBhwAlJHta2uR74wxbz7wMvS/LMcc/7RLrkeNBmvBaS5sbjgfsAn5jFNl5A9/ndtuWLjdroirN/B75Jd9b/ycDLkzx9YBvPostJ29J1ht4FUFWH03XAxo7cv5muuPtKVV05w3ifAny8qn41w/UlDd8q4NHtspTfAY4CjqhW+dCNuNqRLuccwcZ9kzfRnSXbm+5A867A3wEk2Y+uT/O/6fLRE4G1VfXXbDwS4Ng2jPsc4MN0faAXAP+U5JHted5NVwzuTNeP+qOB+J9Md6DsKzPc/6cAn66q9TNcX7Ngcbc47QhcP9BRIcl/Jrmpjf1+Ymv+ZFX9f1X1q6r6eVWtqaqL2/1vAacCv9uWPQT4VFV9saruAP6WrtMzLVX10aq6um37I3Rn2aZ76v0w4MSq+np77lcBj0+yYmCZ46vqpqr6EXAeXdIbe+5vAO+nO0v46oF1ngX8V1V9sKo2VNWFdEMKDmnrfb6qvt1i/iZd5+132dhxVXV7Vf1smvsiae7swLhcNwPvrKorx32GB9seC+xUVa+tql9U1RXA+4BDB5b/UlWdXVV30p2t+81NxHzNLOKd7fqSFta/tf7X2O1Pqup2uoPHb6U78PwXVbVu3Hp/W1V3VNUXgLOA57dLTP4E+MuquqGqbqUbkj6Wj46i6y+d0/ouV1XVdyeJ65l0hd8HWh/o63QjFA5pl+o8F/i7qrqtqr4NnDywrnlsEfM6osXpp8COSZaNdXqq6rcBkqzjrqJ9oyPHSfYHjgceRXdd3r2Bj7aHdxlcvqpuS/LT6QaU5EXAK4AVrWlruiJ0OnYBvj7w3Ovbc+9KN1wB4McDy9/etj/oErokdvtA24OBJyS5aaBtGd0QApI8nu6M4SO56/U4ddx2Z3r0XdLs3S3XzcBEn+HBtgcDu4zLE1vQHQUfMz7/3GeKmH5Kd9R9pn5KdyRd0uLw7Kr63PjGqvpKkivozpqdPu7hG6vqtoH7P6TrC+0E3Be4cGBuvNDlJIDd6S63mY4HA/tP0Af6UHueZWycC3848Pts85B5bIg8c7c4fZnuWpFNTShQ4+5/mG5I0e5VdX+6cdxj2eMauqQBQJL70h15GXMbXcIZ88CBZR9Md6T7WGCHNvTy2wPbHh/HeFfTJaGx7W3VnvuqTay3KVcC51bVtgO3wYudT6M7ijX2evzLQMxd4HcNoZC08L5MN2zo2ZM8PmleGjDRZ3iw7UrgB+PyxDZVdeA0Yxy//c8B+yXZnGHt49d/ThL/P0uLWJJj6A4aX013Pe6g7cbNfvugttz1dNfsPnIgH92/uolLoMtXk02sND4XXQl8YYI+0J/RXV+8gYF+X4thzLnAbkn2nd7e3s3ngKc7w+9w+M9jEaqqm+iua/unJIck2TrJPZLsDUz1QdoGuKGqft7Gbb9w4LEzgGcm+R9tlqPXsvHfx0XAgUm2T/JA4OUDj21Fl1R+ApDkxXRnB8dcS5ck7jaLZ/Nh4MVJ9k5yb7ohCBdU1dqpXodpOBN4ZLpp1O/ZbvsNjGsffD0ex8bDsCQNWVXdTHetybuTPDvJfdvn+BlJ3szUeWm6vgLckuSVSbZMskWSRyV57DTXvxb4tYGYP0d3ncsnkuyTZFmSbZK8JMkfTbqVu7yVbma5k9uBM5LsmuStSR69ebsmaRjSfUXT6+mGZh4O/FXrow36+3Rfm/I7dEMoP9qutX0f8LYkD2jb2nXgGuD30/WXntz6fbsmeXh7bKNcBHwKeGiSwwf6QI9N8og2xPzjwGtaXt2T7to/AKrqMrpZiU9NsrLFeZ90X8G1ehovwYfoisuPJXl4i3WHJK9OMt0DZ5ohi7tFql24/wq6o0HX0X2o/5luFrj/nGS1Pwdem+RWug7Tfw8TqO5rBY6hK7SuoZuVcnB8+IfoJhxYC3wW+MjAut8B3kJ3lP1aYC/g/xtY9/N0wyZ/nOT6CfblXLpr/D7WnvvXmYNCq3UMn06XXK+hG1r1D3RH0gD+DPiH9nq8mrsPm5A0ZFX1Vrpc9zd0B5CupBsl8G9MkZc2Y/t30k2otDfdTJnX053Fv/80N/EPwN+0a23+V2s7hG7o1EeAm+lGMuxLdzR7U/HcQDdl+C+BC1p+Ordt5/JpxiRp4fx7Nv6eu0/QXWf3pqr6ZiuUXg18qB3Ahq4/ciPd2bpTgJcMXDv3SrrP+vlJbqHLGw+Dbqgn3Szeb6PLCV/grpFP76C7nu7GJO9s1+s9ja4/dXV7zjdxVx/oWLpLXH5Md7nKB8bt10vpJo96N3AT3cR5z6GbgGpKbf6Ep9DNUn4Od83IuSNwwabW1+yMTScvSZIkaR4lWUk3s/dMh25LU/LMnSRJkiT1gMWdJEmSJPWAwzIlSZIkqQc8cydJkiRJPTDSX2K+44471ooVKxb8eW+77Ta22mo0v5pjVGMb1bhgdGPrQ1wXXnjh9VW10zyHtOjsuOOOtdNOO43k+zuZUf17nMpii9l459d04zVvTWxz+lyj9LdhLBMzlsmNUjzzkreqamRv++yzTw3DeeedN5TnnY5RjW1U46oa3dj6EBfwtRqBXDFqt3322Wdk39/JLLZ4qxZfzMY7v6Ybr3lr8rw1XaP0t2EsEzOWyY1SPPORtxyWKUmSJEk9YHEnSZIkST1gcSdJkiRJPWBxJ0mSJEk9YHEnSZIkST1gcSdJkiRJPWBxJ0mSJEk9YHEnSZIkST1gcSdJkiRJPbBs2AHMlRWrz5qzba3aawNHrj6LtccfNGfblCRpPs3l/0Ho/heunNMtduY6zjEnHbDVvGxXWowuvupmjpzjz5r94sXBM3eSJEmS1AO9OXMnSZIkzaepzjyPjfzaXJ4R01zyzJ0kSZIk9YDFnSRJkiT1gMWdJEmSJPWA19xJkjZpPmY49DoTSVra5mv23KX8/8Uzd5IkSZLUA565kyRpCnN1ZHlwJr2lfFRZkjR/PHMnqZeSnJjkuiTfHtf+F0m+l+SSJG8eaH9VksvbY08faD+gtV2eZPVC7oOkpcW8JWm2PHMnLWLzMVb9pAO2mvNtDslJwLuAD441JPk94GDg0VV1R5IHtPY9gUOBRwK7AJ9L8tC22ruBpwLrgK8mObOqvrNgeyFpKTkJ85ZG1Ez7HDP9/j/NjMWdpF6qqi8mWTGu+c+A46vqjrbMda39YOC01v6DJJcD+7XHLq+qKwCSnNaWtZMkac6ZtyTNlsWdpKXkocDvJHkD8HPgf1XVV4FdgfMHllvX2gCuHNe+/0QbTnI0cDTA8uXLWb9+PWvWrJnb6OfRpuJdtdeGOX/O2b4+C/Uaz9W+L9/yrm3NR9xz/R4t33JxxDlmsX3mNsOC5a3pvn6j9FpPFcvFV90858+3aq/JHxv8jG+O+XgtZxrLfBhGLFO9povl73emLO4kLSXLgO2AxwGPBU5P8mtAJli2mPi65Jpow1V1AnACwL777ltbb701K1eunIuYF8SaNWumjHc+htSsPWzy55uOTcU8V+Zq31fttYG3XNz9253tvk9krt+jVXtt4Pnz8PrO1/Cskw7YalF95jbDguWt6b5+C/XZm46pYlnooYCDn/HNcvFt8xALM4tlHsz4dZmFqXLsYvn7nanReNclaWGsAz5eVQV8JcmvgB1b++4Dy+0GXN1+n6x91vzuOEnTMFJ5S9Joc7ZMSUvJvwFPAmgTD9wLuB44Ezg0yb2TPATYA/gK8FVgjyQPSXIvuskLzhxK5JKWKvOWpGnzzJ2kXkpyKrAS2DHJOuA44ETgxDbN+C+AI9rR8EuSnE434cAG4JiqurNt51jgM8AWwIlVdcmC74ykJcG8JWm2LO4k9VJVvWCSh/5wkuXfALxhgvazgbPnMDRJmpB5S9JsOSxTkiRJknrA4k6SJEmSesDiTpIkSZJ6wOJOkiRJknrA4k6SJEmSesDiTpIkSZJ6YJPFXZLdk5yX5NIklyR5WWvfPsk5SS5rP7dr7UnyziSXJ/lWkscMbOuItvxlSY6Yv92SJEmSpKVlOmfuNgCrquoRwOOAY5LsCawGzq2qPYBz232AZwB7tNvRwHugKwbpvoxzf2A/4LixglCSJEmSNDubLO6q6pqq+nr7/VbgUmBX4GDg5LbYycCz2+8HAx+szvnAtkl2Bp4OnFNVN1TVjcA5wAFzujeSJEmStEQt25yFk6wAfgu4AFheVddAVwAmeUBbbFfgyoHV1rW2ydrHP8fRdGf8+P/Zu+N42+q6zv+vdyCGXA2QOCIwXTKqwa4pc0co59ecYsILOmIzWThMglHUDFZO10eizmMojfnhFDlaZl3ldrUxkEzjjlJK2Pk5TYGoqYBo3JDkCoEKojdKvfr5/bG+RzaHfc7d55y999lnn9fz8diPs/d3fffan7XuWt+7Pmt913fNzMwwNzc3UGzbt+wffEEOYObQbn6D/vY47du3z7iWaVJjG0Zcw9zu503q+pIkSdLSBk7ukmwC/gh4cVV9McmiVfuU1RLlDy+o2gHsANi6dWvNzs4OFN95F717oHqD2L5lP5fddDB3nDPYb4/T3Nwcg66TcZrUuGByYxtGXMPc7uft2nbYRK4vSZIkLW2g0TKTPIousXtrVb2jFd/TulvS/t7byvcCx/d8/TjgriXKJUmSJEmrNMhomQEuB26tqt/ombQbmB/x8lzg6p7yF7RRM08FHmjdN98DnJ7kiDaQyumtTJIkSZK0SoN0y3wG8BPATUk+0speDlwKXJXkfODTwPPatGuAM4E9wIPACwGq6r4krwJubPVeWVX3DWUpJEmSJGmDO2ByV1V/Qf/75QBO61O/gAsXmddOYOdyApQkSZIkHdhA99xJkiRJkiabyZ2kqZRkZ5J7k9zcZ9pLklSSo9rnJHldkj1JPpbk5J665ya5rb3OXTgvSZKkSWFyJ2la7QK2LSxMcjzww3T3Cs87AzixvS4A3tDqHglcDJwCPB24uA0IJUlD50kpSatlcidpKlXV+4F+gza9BvglHv6czbOAt1TneuDw9oiXZwLXVtV9VXU/cC19EkZJGpJdeFJK0iqY3EnaMJI8B/hMVX10waRjgTt7Pu9tZYuVS9LQeVJK0moN8igESVr3kjwGeAXdMzYfMblPWS1R3m/+F9CdPWdmZoZ9+/YxNze3ZEzbt+xfcvpKHOg3F3OgeCcp1nmDrONhGNayzxz60LxGEfew/41mDl0fcc4b1/Ywbr0npbpHD3/Dqk9KLWy3Bl1/k7Sul4plVNvaYnr38bW20WNZavtcL9vvSpncSdoongScAMwfIB0HfDjJ0+kOfo7vqXsccFcrn11QPtdv5lW1A9gBsHXr1tq0aROzs7P9qn7DeRe9e/lLcQB3nLP0by5mbm5uyXgnKdZ5B4p5WIa17Nu37Oeym7r/dle77P0M+99o+5b9/NgI1u8otiWAXdsOG8v2ME6jPim1sN0adP2Na98bxFKxjGpbW0zvPr7WNnosS7Wx62X7XSm7ZUraEKrqpqo6uqo2V9VmusTt5Kr6e2A38II2QMGpwANVdTfwHuD0JEe0e1ZOb2WSNA69J6Xu4KGTUk9g6ZNS/colbQAmd5KmUpIrgL8CvivJ3iTnL1H9GuB2YA/wRuA/A1TVfcCrgBvb65WtTJJGzpNSkpZrMq7XStKQVdXzDzB9c8/7Ai5cpN5OYOdQg5OkPtpJqVngqCR7gYur6vJFql8DnEl3UupB4IXQnZRKMn9SCjwpJW0oJneSJEkTwJNS0nBsXuJ+y+1b9q/ofsw7Ln3WakIaG1kvZ/cAACAASURBVLtlSpIkSdIUMLmTJEmSpClgcidJkiRJU8DkTpIkSZKmgMmdJEmSJE0BkztJkiRJmgImd5IkSZI0BUzuJEmSJGkKmNxJkiRJ0hQwuZMkSZKkKWByJ0mSJElTwOROkiRJkqaAyZ0kSZIkTQGTO0mSJEmaAiZ3kiRJkjQFTO4kSZIkaQqY3EmaSkl2Jrk3yc09Zb+W5BNJPpbknUkO75n2siR7knwyyTN7yre1sj1JLhr3ckjaOGy3JK2WyZ2kabUL2Lag7Frge6rqKcDfAC8DSHIScDbw5Pad305yUJKDgNcDZwAnAc9vdSVpFHZhuyVpFUzuJE2lqno/cN+CsvdW1f728XrguPb+LODKqvpyVX0K2AM8vb32VNXtVfUV4MpWV5KGznZL0modvNYBSNIa+Ungbe39sXQHTfP2tjKAOxeUn9JvZkkuAC4AmJmZYd++fczNzS0ZwPYt+5ecvhIH+s3FHCjeSYp13iDreBiGtewzhz40r1HEPex/o5lD10ec88a1PayxkbZbg66/SVrXS8Uyqm1tMb37+FozlsWtNJ5RbPOj2JdM7iRtOEleAewH3jpf1Kda0b93Q/WbZ1XtAHYAbN26tTZt2sTs7OyScZx30bsHjHhwd5yz9G8uZm5ubsl4JynWeQeKeViGtezbt+znspu6/3ZXu+z9DPvfaPuW/fzYCNbvKLYlgF3bDhvL9rBWxtFuDbr+xrXvDWKpWEa1rS2mdx9fa8ayuJXGM4p2exT70uSsaUkagyTnAs8GTquq+QOevcDxPdWOA+5q7xcrl6SxsN2SNCjvuZO0YSTZBrwUeE5VPdgzaTdwdpJHJzkBOBH4AHAjcGKSE5IcQjd4we5xxy1p47LdkrQcXrmTNJWSXAHMAkcl2QtcTDfK3KOBa5MAXF9VP1tVtyS5Cvg4XbenC6vqa20+LwLeAxwE7KyqW8a+MJI2BNstSatlcidpKlXV8/sUX75E/UuAS/qUXwNcM8TQJKkv2y1Jq2VyJ0mSpDW1eYWDn2zfsn/sA6dIk8x77iRJkiRpChzwyl2SnXQjNN1bVd/Tyn4Z+Gngs63ay1sXAJK8DDgf+Brw81X1nla+DXgtXf/vN1XVpcNdFEnSRrfSs/+SJE2DQa7c7QK29Sl/TVU9tb3mE7uT6EZlenL7zm8nOSjJQcDrgTOAk4Dnt7qSJEmSpCE44JW7qnp/ks0Dzu8s4Mqq+jLwqSR7gKe3aXuq6naAJFe2uh9fdsSSJEmSpEdYzYAqL0ryAuCDwPaquh84Fri+p87eVgZw54LyU/rNNMkFwAUAMzMzzM3NDRTM9i37lxP7kmYO7eY36G+P0759+4xrmSY1tmHENcztft6kri9JkiQtbaXJ3RuAVwHV/l4G/CSQPnWL/t0/q9+Mq2oHsANg69atNTs7O1BAwxwpafuW/Vx208Hccc5gvz1Oc3NzDLpOxmlS44LJjW0YcY1ihLBd2w6byPUlSZKkpa0ouauqe+bfJ3kj8K72cS9wfE/V44C72vvFyiVJkiRJq7SiRyEkOabn448AN7f3u4Gzkzw6yQnAicAHgBuBE5OckOQQukFXdq88bEmSJElSr0EehXAFMAsclWQvcDEwm+SpdF0r7wB+BqCqbklyFd1AKfuBC6vqa20+LwLeQ/cohJ1VdcvQl0aSNjgfBCxJ0sY1yGiZz+9TfPkS9S8BLulTfg1wzbKikyRpCvk8PknSKKyoW6YkSZIkabKY3EmSJEnSFFjNc+4kSVqx1XZN9D5BSZIezuROkiT15b2BkrS+2C1TkiRJkqaAyZ2kqZRkZ5J7k9zcU3ZkkmuT3Nb+HtHKk+R1SfYk+ViSk3u+c26rf1uSc9diWSRtDLZbklbL5E7StNoFbFtQdhFwXVWdCFzXPgOcAZzYXhcAb4DuoIru2Z6nAE8HLp4/sJKkEdiF7ZakVTC5kzSVqur9wH0Lis8C3tzevxl4bk/5W6pzPXB4kmOAZwLXVtV9VXU/cC2PPPCSpKGw3ZK0Wg6oImkjmamquwGq6u4kR7fyY4E7e+rtbWWLlT9Ckgvozp4zMzPDvn37mJubWzKY7Vv2r2ARRmPm0MmKZxDrLWbjHa1B9rl1amTtlqTpY3InSZA+ZbVE+SMLq3YAOwC2bt1amzZtYnZ2dskfnaRh/Ldv2c9lN62v/xLWW8zGO1q7th12wH1uyqy63Vp4UmrQ5HgUifRKTyRM0kkIY+lvkmKBlcczipNHo9iX1k+rLUmrd0+SY9rZ72OAe1v5XuD4nnrHAXe18tkF5XNjiFOS5o2s3Vp4UmrQ5Hhubm7oifRKT3ZN0kkIY+lvkmKBlcdzxzmzQ49lFPuS99xJ2kh2A/Mjx50LXN1T/oI2+typwAOtG9R7gNOTHNEGJDi9lUnSuNhuSRrY5KTRkjRESa6gO3t9VJK9dKPHXQpcleR84NPA81r1a4AzgT3Ag8ALAarqviSvAm5s9V5ZVQsHO5CkobDdkrRaJneSplJVPX+RSaf1qVvAhYvMZyewc4ihSVJftluSVstumZIkSZI0BUzuJEmSJGkKmNxJkiRJ0hQwuZMkSZKkKWByJ0mSJElTwOROkiRJkqaAj0IYs80XvXtV39++ZT/nLZjHHZc+a1XzlCRJkrT+eeVOkiRJkqaAyZ0kSZIkTQG7ZWqsVtsttR+7pUqSJEleuZMkSZKkqWByJ0mSJElTwOROkiRJkqaA99xJfXhvoCRJktYbkzutewdKxPo9G1CSJEmaNiZ3WtRyrl6ZQEmSJElry+RuCaPomjcK6yVOSZIkaT0axfH2rm2HDX2eJnfSmCxsFLzauXaS/Bfgp4ACbgJeCBwDXAkcCXwY+Imq+kqSRwNvAf4F8Hngx6vqjrWIW9LGZbslaRCOlilpQ0lyLPDzwNaq+h7gIOBs4NXAa6rqROB+4Pz2lfOB+6vqO4DXtHqSNDa2W5IGZXInaSM6GDg0ycHAY4C7gR8C3t6mvxl4bnt/VvtMm35akowxVkkC2y1JA7BbpqQNpao+k+TXgU8D/wi8F/gQ8IWq2t+q7QWObe+PBe5s392f5AHg8cDneueb5ALgAoCZmRn27dvH3NzckrFs37J/yenjNHPoZMUziPUWs/GO1iD73Ho1rnZr0PU3inW90m1tkrZTY+lvkmKByYpnFPuSyZ2kDSXJEXRntU8AvgD8IXBGn6o1/5Ulpj1UULUD2AGwdevW2rRpE7Ozs0vGMkn3XG7fsp/Lblpf/yWst5iNd7R2bTvsgPvcejWudmvQ9Tc3Nzf0db3S9nCStlNj6W+SYoHJimcU7ZbdMiVtNP8G+FRVfbaqvgq8A/h+4PDW3QngOOCu9n4vcDxAm/4twH3jDVnSBme7JWkgJneSNppPA6cmeUy7B+U04OPAnwM/2uqcC1zd3u9un2nT31dVjzgDLkkjZLslaSAHTO6S7Exyb5Kbe8qOTHJtktva3yNaeZK8LsmeJB9LcnLPd85t9W9Lcm6/35KkUauqG+gGGPgw3XDi30TXLemlwC8m2UN3b8rl7SuXA49v5b8IXDT2oCVtaLZbkgY1SIfTXcBv0T0vZd5FwHVVdWmSi9rnl9L1/z6xvU4B3gCckuRI4GJgK12f7w8l2V1V9w9rQSRpUFV1MV2b1Ot24Ol96v4T8LxxxCVJi7HdkjSIA165q6r388h+2r1D7C4cevct1bmeri/4McAzgWur6r6W0F0LbBvGAkiSJEmSVj5a5kxV3Q1QVXcnObqVf2Po3WZ+WN7Fyh9hpcPyDnNI00kaInWhSY1tUuOCyY1tUuOa5uHEJUmSptmwxwFdbOjdgYbkhZUPyzvMIcUnaYjUhSY1tkmNCyY3tkmNa5qHE5ckSZpmKx0t857W3ZL2995W/o2hd5v5YXkXK5ckSZIkDcFKk7veIXYXDr37gjZq5qnAA6375nuA05Mc0UbWPL2VSZIkSZKG4IB9wpJcAcwCRyXZSzdS06XAVUnOp3v2yvyITNcAZwJ7gAeBFwJU1X1JXgXc2Oq9sqp8mKYkSZIkDckBk7uqev4ik07rU7eACxeZz05g57KikyRJkiQNZKXdMiVJkiRJE8TkTpIkSZKmgMmdJEmSJE0BkztJkiRJmgImd5IkSZI0BUzuJEmSJGkKmNxJkiRJ0hQwuZMkSZKkKWByJ0mSJElTwORO0oaT5PAkb0/yiSS3Jvm+JEcmuTbJbe3vEa1ukrwuyZ4kH0ty8lrHL2njsd2SNAiTO0kb0WuBP62q7wa+F7gVuAi4rqpOBK5rnwHOAE5srwuAN4w/XEmy3ZJ0YCZ3kjaUJI8DfgC4HKCqvlJVXwDOAt7cqr0ZeG57fxbwlupcDxye5Jgxhy1pA7PdkjSog9c6AEkas28HPgv8XpLvBT4E/AIwU1V3A1TV3UmObvWPBe7s+f7eVnZ370yTXEB3hpyZmRn27dvH3NzckoFs37J/1QszLDOHTlY8g1hvMRvvaA2yz61jY2m3Bl1/o1jXK93WJmk7NZb+JikWmKx4RrEvmdxJ2mgOBk4Gfq6qbkjyWh7qytRP+pTVIwqqdgA7ALZu3VqbNm1idnZ2yUDOu+jdg8Y8ctu37Oeym9bXfwnrLWbjHa1d2w474D63jo2l3Rp0/c3NzQ19Xa+0PZyk7dRY+pukWGCy4hlFu2W3TEkbzV5gb1Xd0D6/ne6g6Z75bkvt77099Y/v+f5xwF1jilWSwHZL0oBM7iRtKFX198CdSb6rFZ0GfBzYDZzbys4Frm7vdwMvaKPPnQo8MN8NSpLGwXZL0qAm45qkJI3XzwFvTXIIcDvwQrqTXVclOR/4NPC8Vvca4ExgD/BgqytJ42a7JemATO4kbThV9RFga59Jp/WpW8CFIw9KkpZguyVpEHbLlCRJkqQpYHInSZIkSVPA5E6SJEmSpoDJnSRJkiRNAZM7SZIkSZoCJneSJEmSNAVM7iRJkiRpCpjcSZIkSdIUMLmTJEmSpClgcidJkiRJU8DkTpIkSZKmgMmdJEmSJE0BkztJkiRJmgIHr3UAkiRJWj9u+swDnHfRu9c6DEl9eOVOkiRJkqaAyZ0kSZIkTQGTO0kbTpKDkvx1kne1zyckuSHJbUneluSQVv7o9nlPm755LeOWtLHZdkk6EJM7SRvRLwC39nx+NfCaqjoRuB84v5WfD9xfVd8BvKbVk6S1YtslaUkmd5I2lCTHAc8C3tQ+B/gh4O2typuB57b3Z7XPtOmntfqSNFa2XZIGYXInaaP5n8AvAV9vnx8PfKGq9rfPe4Fj2/tjgTsB2vQHWn1JGjfbLkkHtKpHISS5A/gS8DVgf1VtTXIk8DZgM3AH8GNVdX87Y/Ra4EzgQeC8qvrwan5fkpYjybOBe6vqQ0lm54v7VK0Bpi2c9wXABQAzMzPs27ePubm5JePZvmX/ktPHaebQyYpnEOstZuMdrUH2ufVqVG3XwnZr0PU3SduGsfRnLIubpHhG0W4N4zl3P1hVn+v5fBFwXVVdmuSi9vmlwBnAie11CvCG9leSxuUZwHOSnAl8M/A4urPhhyc5uJ3hPg64q9XfCxwP7E1yMPAtwH39ZlxVO4AdAFu3bq1NmzYxOzu7ZDCT9Jyo7Vv2c9lN6+vRp+stZuMdrV3bDjvgPreOjaTtWthuDbr+fvOtV0/MtjFJ26mx9DdJscBkxTOKdmsU3TJ7+3kv7P/9lupcT9cgHTOC35ekvqrqZVV1XFVtBs4G3ldV5wB/Dvxoq3YucHV7v7t9pk1/X1X1vXInSaNi2yVpUKtNWwt4b5ICfredAZqpqrsBquruJEe3ut/o/93M9w2/u3eGK+0iMMzLq5N0uXahSY1tUuOCyY1tUuOa5q5NS3gpcGWSXwX+Gri8lV8O/H6SPXRnvc9eo/gkqR/bLkkPs9rk7hlVdVdL4K5N8okl6g7U/3ulXQSG2b1pki7XLjSpsU1qXDC5sU1qXFPetekbqmoOmGvvbwee3qfOPwHPG2tgkrQE2y5JS1lVt8yquqv9vRd4J10Dc898d8v2995Wfb7/97zevuGSJEmSpFVYcXKX5LAkj51/D5wO3MzD+3kv7P/9gnROBR6Y774pSZIkSVqd1fQJmwHe2Z6JeTDwB1X1p0luBK5Kcj7waR7qFnAN3WMQ9tA9CuGFq/htSZIkSVKPFSd3rZ/39/Yp/zxwWp/yAi5c6e9JkiRJkhY3ikchSJIkSZLGzOROkiRJkqaAyZ0kSZIkTQGTO0mSJEmaAiZ3kiRJkjQFTO4kSZIkaQqY3EmSJEnSFDC5kyRJkqQpYHInSZIkSVPA5E6SJEmSpoDJnSRJkiRNAZM7SZIkSZoCJneSJEmSNAVM7iRtKEmOT/LnSW5NckuSX2jlRya5Nslt7e8RrTxJXpdkT5KPJTl5bZdA0kZjuyVpUCZ3kjaa/cD2qvrnwKnAhUlOAi4CrquqE4Hr2meAM4AT2+sC4A3jD1nSBme7JWkgJneSNpSquruqPtzefwm4FTgWOAt4c6v2ZuC57f1ZwFuqcz1weJJjxhy2pA3MdkvSoA5e6wAkaa0k2Qw8DbgBmKmqu6E7kEpydKt2LHBnz9f2trK7F8zrAroz5MzMzLBv3z7m5uaW/P3tW/avehmGZebQyYpnEOstZuMdrUH2uWkwynZr0PU3SduGsfRnLIubpHhG0W6Z3EnakJJsAv4IeHFVfTHJolX7lNUjCqp2ADsAtm7dWps2bWJ2dnbJGM676N3LCXmktm/Zz2U3ra//EtZbzMY7Wru2HXbAfW69G3W7Nej6+823Xj0x28YkbafG0t8kxQKTFc8o2i27ZUracJI8iu4A6a1V9Y5WfM98t6X2995Wvhc4vufrxwF3jStWSQLbLUmDMbmTtKGkO9V9OXBrVf1Gz6TdwLnt/bnA1T3lL2ijz50KPDDfDUqSxsF2S9KgJuOapCSNzzOAnwBuSvKRVvZy4FLgqiTnA58GntemXQOcCewBHgReON5wJcl2S9JgTO4kbShV9Rf0vx8F4LQ+9Qu4cKRBSdISbLckDcpumZIkSZI0BUzuJEmSJGkKmNxJkiRJ0hQwuZMkSZKkKWByJ0mSJElTwOROkiRJkqaAyZ0kSZIkTQGTO0mSJEmaAiZ3kiRJkjQFTO4kSZIkaQqY3EmSJEnSFDC5kyRJkqQpYHInSZIkSVPA5E6SJEmSpoDJnSRJkiRNAZM7SZIkSZoCY0/ukmxL8skke5JcNO7fl6Tlst2StN7Ybkkb01iTuyQHAa8HzgBOAp6f5KRxxiBJy2G7JWm9sd2SNq5xX7l7OrCnqm6vqq8AVwJnjTkGSVoO2y1J643tlrRBparG92PJjwLbquqn2uefAE6pqhf11LkAuKB9/C7gk2ML8CFHAZ9bg98dxKTGNqlxweTGNg1xfVtVfesog1lrg7RbrXxh2/V5JvPfdzGTuj0uZb3FbLyjNWi8tlsP1VvpMdckbRvG0p+xLG6S4hl6u3Xw6uJZtvQpe1h2WVU7gB3jCae/JB+sqq1rGcNiJjW2SY0LJjc241o3DthuwSPbrvW2HtdbvLD+Yjbe0Vpv8Y7YitqtgWc+QevaWPozlsVNUjyjiGXc3TL3Asf3fD4OuGvMMUjScthuSVpvbLekDWrcyd2NwIlJTkhyCHA2sHvMMUjScthuSVpvbLekDWqs3TKran+SFwHvAQ4CdlbVLeOMYUBr2i30ACY1tkmNCyY3NuNaB1bRbq239bje4oX1F7PxjtZ6i3dkxnC8NUnr2lj6M5bFTVI8Q49lrAOqSJIkSZJGY+wPMZckSZIkDZ/JnSRJkiRNgQ2R3CU5PsmfJ7k1yS1JfqGVH5nk2iS3tb9HtPLvTvJXSb6c5CUL5rUtySeT7Ely0aTEtth81jqunvkdlOSvk7xrNXENO7Ykhyd5e5JPtPl934TE9V/aPG5OckWSb15pXCuM7ZwkH2uvv0zyvT3zGuo+MI2S3JHkpiQfSfLBtY5noSQ7k9yb5Oaesr7bwiRYJN5fTvKZto4/kuTMtYyx13L3t7W2RLwTuY6TfHOSDyT5aIv3V1r5CUluaOv3bekGEtEqTFJbMUntwKTt45O0D0/S/rlELLuSfKpnvTx11LH0xPSw4+GRrJeqmvoXcAxwcnv/WOBvgJOA/wFc1MovAl7d3h8N/EvgEuAlPfM5CPhb4NuBQ4CPAidNSGx957PWcfXM7xeBPwDeNSn/nm3am4Gfau8PAQ5f67iAY4FPAYe2z1cB5415nX0/cER7fwZww6j2gWl8AXcAR611HEvE9wPAycDNPWV9t4VJeC0S7y/3a2sm4bXc/W2tX0vEO5HrmO4Zbpva+0cBNwCntrby7Fb+O8B/WutY1/trktqKSWoHJm0fn6R9eJL2zyVi2QX86Li3mxbHw46HR7FeNsSVu6q6u6o+3N5/CbiV7gD6LLqDe9rf57Y691bVjcBXF8zq6cCeqrq9qr4CXNnmseaxLTGfNY0LIMlxwLOAN600nlHEluRxdP9ZXN7qfaWqvrDWcTUHA4cmORh4DKt8PtEKYvvLqrq/lV9P94wkGME+oPGrqvcD9y0o7rstTIJF4p1Yy93f1tqw//8Ytersax8f1V4F/BDw9lY+Met3PZuktmKS2oFJ28cnaR+epP1ziVjWxMLj4SRhBOtlQyR3vZJsBp5Gl73PVNXd0O0YdFdSlnIscGfP570McedZZWyLzWcS4vqfwC8BXx9GPEOM7duBzwK/1y6RvynJYWsdV1V9Bvh14NPA3cADVfXeYcS1wtjOB/6kvR/pPjBFCnhvkg8luWCtgxnQitucNfSidF2Hd05KF8eFhtWuj0uf/z8mch23rk0fAe4FrqXrUfCFqtrfqtg2jc6kbcdruo1O2j4+CfvwJO2fC2Opqvn1cklbL69J8uhxxMIjj4cfzwjWy4ZK7pJsAv4IeHFVfXEls+hTNpQzAEOIbajzGdb8kjwbuLeqPrTaWIYdG93VsZOBN1TV04B/oOtGsaZxtcb3LOAE4InAYUn+42rjWklsSX6QLrl76XxRn2o+T+WRnlFVJ9N1ab0wyQ+sdUBT6A3Ak4Cn0p0EuWxtw3mkYbfHo9Yn3oldx1X1tap6Kl2vgqcD/7xftfFGpTWwptvopO3jk7IPT9L+uTCWJN8DvAz4brpbY47koWOckVnkeHgkx1QbJrlL8ii6Df6tVfWOVnxPkmPa9GPosvql7AWO7/l8HKvsLjfE2Babz1rH9QzgOUnuoOvC90NJ/teExLYX2NtzFuftdMneWsf1b4BPVdVnq+qrwDvo7oFbleXGluQpdF0Hzqqqz7fikewD06aq7mp/7wXeSfef26Rbdpuzlqrqnvaf9teBNzJh63hY7fq49It30tcxQOtKP0d3H83hrSs72DaN0sRsx2u5jU7aPj6J+/Ak7Z89sWxr3Virqr4M/B7jWS+POB6mu5I39PWyIZK71qf1cuDWqvqNnkm7gXPb+3OBqw8wqxuBE9vINocAZ7d5rHlsS8xnTeOqqpdV1XFVtZlufb2vqlZ1FWqIsf09cGeS72pFpwEfX+u46LpjnprkMW2ep9H1n1+x5caW5J/RJZU/UVV/01N/6PvAtElyWJLHzr8HTgduXvpbE2G52+mamj+Aan6ECVrHQ2wLxmKxeCd1HSf51iSHt/eH0p0QuxX4c+BHW7WJWb9TaGK247XaRidtH5+kfXiS9s9FYvlETwIeunvcRr5eFjkePodRrJdag5Fixv0C/hXdZc6PAR9przPp+rpeB9zW/h7Z6j+B7grFF4EvtPePa9POpBuF6G+BV0xKbIvNZ63jWjDPWYYzWuYw/z2fCnywzeuPaSNETkBcvwJ8gq7B+X3g0WNeZ28C7u+p+8GeeQ11H5i2F929nB9tr1smcR0BV9B10flq2+7OX2xbmITXIvH+PnBT26Z3A8esdZw98S5rf1vr1xLxTuQ6Bp4C/HWL62bgv7Xybwc+AOwB/nC17aavyWorJqkdmLR9fJL24UnaP5eI5X1tvdwM/C/aiJrjetFzPDyK9ZI2Y0mSJEnSOrYhumVKkiRJ0rQzuZMkSZKkKWByJ0mSJElTwOROkiRJkqaAyZ0kSZIkTQGTO0mSJEmaAiZ3kiRJkjQFTO4kSZIkaQqY3EmSJEnSFDC5kyRJkqQpYHInSZIkSVPA5E6SJEmSpoDJnSRJkiRNAZM7SZIkSZoCJneSJEmSNAVM7iRJkiRpCpjcSZIkSdIUMLmTJEmSpClgcidJkiRJU8DkTpIkSZKmgMmdJEmSJE0BkztJkiRJmgImd5IkSZI0BUzuJEmSJGkKmNxJkiRJ0hQwuZMkSZKkKWByJ0mSJElTwOROkiRJkqaAyZ0kSZIkTQGTO0mSJEmaAiZ3kiRJkjQFTO4kSZIkaQqY3EmSJEnSFDC5kyRJkqQpYHInSZIkSVPA5E6SJEmSpoDJnSRpXUnyzUkqyXFrHcu4JHlxknetdRySOkn+WZJ9SQ5a61ikXiZ3G1iSO5L8Y2uc5l+/NYLfOSvJR5J8McnnklyXZPOwf0fS2lnQjnx9QdtyzgG+uy3JnlX89pVJ/utKvz8OSc7vWR//2NbR/OfPrXV8khbX73gJ2F9Vm6rqa63OXJKfWqP4ZpPsXVD2y0m+2uL9QpK/TPJ9axHfYlqM/2ut45g2Jnf6t61xmn+9aJgzT/IdwFuA7cC3ACcAvw18fYi/kSRuy9Ia6m1HgE/z8LblrWsd31qrqst71s+PALf3rJ+j1jo+SQe08HjprrUOaABva23OtwJ/AbwjSZYzgyQHjyQyjYwHxHqEJOcl+b9JXtPO9tye5Ptb+Z1J7k1ybk/9XUl+J8m1Sb6U5P9L8m1t8lOBT1XVddX5UlX9UVV9un33oCQvT/K37bsfSnJ8m/b9SW5M8kD7+/09vzmX5JIk/xd4EPj2JN+S5PIkdyf5TJJftbuENBmSHJrk9W3/3Jvk15I8KsnjgXfS7cPzZ8Ufn+QZSW5oF+85RgAAIABJREFU+/9drT1a9kFGku9J8r4k9ye5Nclze6b9SJKPtl4Ff5fk5T3THnEWPsknk5w5wHzPSvKJ1qbdmeTnB4z1V9sVgi8l+ViS0xep901JdiR5T5LHtLKfS/I3Se5LsjvJE1r54em6sJ6f5FNt+quXsw4lPVKSzW3fOjjJJcD/A/xWenpBten/Ocltbb9+VZInJfmr1u5cleSQnnk+O11Pp/krbU/pmXZHkpe0tuGBJG9L10X9MOBPgCf2tKFP7I21qr4KvBl4AvD4Nr+fbG3X/a0t+bae36okFya5DbitlT25Hefdl+Se+faytUcXteO4z7dlOnLBOjo3yafT9d56RZu2DXg58OMt5o8O+99oozK502JOAT5G1wj8AXAl8C+B7wD+I10Dtqmn/jnAq4CjgI8A82fqPwx8dzsw+8EF3wH4ReD5wJnA44CfBB5sDcO7gde1GH4DeHc7EJz3E8AFwGOBv6NruPa3GJ8GnA6sSRcJSY/wK8BTgC3AvwBmgV+qqs/zyCtZnwe+CrwIOJLuoOnfssz9OcnjgGuBy+naphcAO9P1KAD4IvAfgMNbDC9pBxzQtXvP75nXv2ixvHeA+e4EXlBVj6U7wfV/Bgz5Frq293DgN4Grkhy+YJke1WJ7PN2VhAeTnAf8NHAG3cHbJ4HfWzDvH6Zb96cAFyY5dcCYJB1AVb2Cbj9/UZ9eUNvo2rxTgV8CdtAdMx0PfA+tnUlyMl3b8TN0+/fvAruTPLpnXj/W5ncCXXt6XlX9A92+f9diVxXbPM4D9lbV59rJqJcD/47uqt7/Aa5YsFjPpWsvTkryWODPgD8Fnkh3nHVdq/fzre6/btPuB16/YF7/Cvgu4DTgvyX551X1p8B/p11drKrvXWT1aplM7vTH7QzR/OunW/mnqur3Wl/yt9E1Qq+sqi9X1XuBr9Dt3PPeXVXvr6ovA68Avi/J8VV1O91B3LHAVcDn0l3pm0/yfgr4r1X1yXZl76PtwO5ZwG1V9ftVtb+qrgA+QXeAN29XVd1SVfvpDrrOAF5cVf9QVfcCrwHOHvYKk7Qi5wAXV9Xnquoe4FfpTtD0VVUfqKobq+prVfW3wJvoDh6W40eAm6vqrW0+NwL/G/j37Teua23I16vqw3Rt1PxvvB34/iTHtM//AfjD1t4sOV+6k0xPTvLYqvp8Vf31IMFW1RVVdU+b5xuBLwC9BzyHAn8M/BPwY1X1lVb+M8CvVNXftrKLgWcm+Zae7/5qVe2rqtuAv6RLOiUNrvd46Y+X8b1XV9UXq+oW4GbgvVV1e1U9QHfF7Wmt3k8Dv1tVN7Q24M3Al+mSwnmvq6q7quo+ujbnQPvxjyX5AnAnXYI538PgZ4D/t6pubW3afwee2nv1rk2/r6r+EXg28PdVdVlV/VPrhXVDz7xeUVV72zHgLwM/mof3tPiVqvrHqvoo8FEe3q5pyEzu9NyqOrzn9cZWfk9PnX8EaAdkvWW9V+HunH9TVfuA++jO4FBV11fVj1XVt9Kdgf8BugQQuqTxb/vE9US6q3G9/o4uSXzEbwLfBjwKuHu+8aU763V0/8WWNC5JQndFqXefXrg/L/zOSUn+pHX/+SLw3+iuki3HtwE/0HsCiy4BO6b9xjPSdSP/bJIH6M5sHwXQDp6upTs4+ibgx3moR8KS86U7gPr3wKfTdd3cOkiwSX42yc098zx+wTI/he5k2avaibfe5dzV87276A4Ke0cT/fue9w/y8PZb0oH1Hi8998DVv2HhsdNix1LfBmxf0K4cTzuWapa7H1/V4j26qn6oqj7U81uv7fmd+4Cw+DHWYsdq8/N6Z8+8bgW+BsysIm6tgsmdhuX4+TftqtyRdAcYD9POcL+DrisCdI3Hk/rM7y66BqPXPwM+0zu7nvd30h3MHNXT+D6uqp683AWRNFxVVXT/uffu0737cz3iS/BGum7dT6qqxwGvpDv4WI476c6S957A2lRVL27Tr6L1TKiqbwF2LfiNK+i6TP1ruqtxfznIfKvqr6rq2XQHN+/lkd2dHiHJFuDX6BLMI6vq8PY7vfHcQNeV/c8WnGG/Ezh7QTyHtisFksajXzu2HHcClyzYjx/Tei4N+7fvBH6mT5vxlz11Fh5j9TtWm592xoJ5fXNVfWaR+quJWwMwudOwnJnkX7Ubg18F3FBVd7ayn05yNECS7waeA1zfvvcm4FVJTkznKe2+umuA70zyH9LdrPzjwElA3+c8VdXddAdRlyV5XLvB90lJltuNS9JoXAFcnG6wlKPprt7PD4F9D3D0gntyHws8UFX7kjyZrsvSUg5ON7jA/OtRdF0Yn5bkx9MN3nJIklOTfGe7mrgJ+HxV/VO6AZuet2CeVwNPbrFe0ZJUDjDfw5Kc3e7L+yrwJbqz2Afy2Fbvs8A3JfkFHn7lDYCq+l26LufXJZk/y/47dOv2RIAkRyb5dwP8pqThuQf49lV8/43AzyY5pR0PHZbkWe1+t0F++/ELumIv5XeAl7W2lXQD0i1s/3q9C3hCuudtPjrJY5Oc0jOvS+ZPOCX51iRnDRjHPcDmOOL5ULky9b/z8OdTvXOF8/kDuvs87qPr1z3/XKsv0CVzN6V7Lsyf0o2M9z/a9N+gO3v+XrrBDS4HDm333T2b7hEKn6e7CfnZVbXU86BeABwCfJzuht6381A3KUlr67/R7Zu30A269H95qB34KLAb+LvWtedI4L8AP9XajdfTXWFbysV0XZzmX39SVfcDzwReCNxN1yPgV4FHtUTtZ4FfT/IlujbmD3tnWFUPtrhOo2vj5ssXnW+r8pN03U4foGuXzuXA/orusTEfpbuieRTd/TmPUFWvozugel+SJ1TV79ENxHB168L6Ybrum5LG57V095rdn+R1y/1yVX2Q7iTWb9Edw+yhu5I/yHc/QXcC7fbWhj7xAPXfCbwauLK1GTfTjVuwWP0v0Q3K9G/pemHcBvxgm/xaunbyva0tvZ5uIJZBzLe5n0/y4QG/owPIQycipZVJsotuBKaJfoiwJEmSNM28cidJkiRJU8DkTpIkSZKmgN0yJUmSJGkKeOVOkiRJkqbAwQeusnaOOuqo2rx580B1/+Ef/oHDDjtstAGNmcu0PkzbMi1neT70oQ99rj2cXj2W03Yt17Rtb6vhuui4HjqDrgfbrf5G2W4NalK2ZeMwjkmLYznt1kQnd5s3b+aDH/zgQHXn5uaYnZ0dbUBj5jKtD9O2TMtZniR/N9po1qfltF3LNW3b22q4Ljquh86g68F2q79RtluDmpRt2TiMY9LiWE67ZbdMSZIkSZoCJneSJEmSNAVM7iRJkiRpCpjcSZIkSdIUMLmTJEmSpClgcidJkiRJU8DkTpIkSZKmgMmdJEmSJE0BkztJkiRJmgIHr3UA2lhu+swDnHfRu4c6zzsufdZQ5ydJvTYPuc0C2y1JWm/m/y/YvmX/0I5lR/F/gVfuJEmSJGkKmNxJkiRJ0hQwuZMkSZKkKWByJ0mSJElTwORO0lRKsjPJvUluXlD+c0k+meSWJP+jp/xlSfa0ac/sKd/WyvYkuWicyyBJkrQcjpYpaVrtAn4LeMt8QZIfBM4CnlJVX05ydCs/CTgbeDLwRODPknxn+9rrgR8G9gI3JtldVR8fRoArGYXxQKN0OQqjJt0oRh8F2LXtsJHMV5LWE5M7SVOpqt6fZPOC4v8EXFpVX2517m3lZwFXtvJPJdkDPL1N21NVtwMkubLVHUpyNwqjOnA2aZQkafKZ3EnaSL4T+H+SXAL8E/CSqroROBa4vqfe3lYGcOeC8lP6zTjJBcAFADMzM8zNzR0wmO1b9i8zfJg5dGXfW61Blmfc9u3bN5a4RrG+hxn3uNbDsIxq+11v60GSRsHkTtJGcjBwBHAq8C+Bq5J8O5A+dYv+9yVXvxlX1Q5gB8DWrVtrdnb2gMGs5CGo27fs57Kbxt9033HO7Nh/80Dm5uYYZD2v1rAeVttrmOtzXOthWEaxPqHrlrme1oMkjYLJnaSNZC/wjqoq4ANJvg4c1cqP76l3HHBXe79YuVZptV1I+91/uF66jw6z++z8elgvyy5JGh1Hy5S0kfwx8EMAbcCUQ4DPAbuBs5M8OskJwInAB4AbgROTnJDkELpBV3avSeSSJEkH4JU7SVMpyRXALHBUkr3AxcBOYGd7PMJXgHPbVbxbklxFN1DKfuDCqvpam8+LgPcABwE7q+qWsS+MtEZGNUCPJGk0TO4kTaWqev4ik/7jIvUvAS7pU34NcM0QQ5MkSRoJu2VKkiRJ0hQYKLlLckeSm5J8JMkHW9mRSa5Nclv7e0QrT5LXJdmT5GNJTu6Zz7mt/m1Jzh3NIkmSJEnSxrOcK3c/WFVPraqt7fNFwHVVdSJwXfsMcAbdYAQn0j3z6Q3QJYN097ycQvdw4IvnE0JJkiRJ0uqs5p67s+gGKwB4MzAHvLSVv6UNUnB9ksOTHNPqXltV9wEkuRbYBlyxihgkSZIk6Rs28mBQgyZ3Bbw3SQG/2x7WO1NVdwNU1d1Jjm51jwXu7Pnu3la2WPnDJLmA7oofMzMzzM3NDRTgvn37Bq67XkzjMs0c2j2TaZjWeh1N27/TtC2PJK0HSY4H3gI8Afg6sKOqXpvkl4GfBj7bqr68DfREkpcB5wNfA36+qt7TyrcBr6Ub5fdNVXXpOJdF0toZNLl7RlXd1RK4a5N8Yom66VNWS5Q/vKBLHHcAbN26tWZnZwcKcG5ujkHrrhfTuEy/+darueym4Q7Sesc5s0Od33JN27/TtC2PJK0T+4HtVfXhJI8FPtR6OQG8pqp+vbdykpPonr35ZOCJwJ+153cCvB74YboT6Tcm2V1VHx/LUkhaUwPdc1dVd7W/9wLvpLtn7p7W3ZL2995WfS9wfM/XjwPuWqJckiRpQ6uqu6vqw+39l4Bb6dPDqcdZwJVV9eWq+hSwh+747OnAnqq6vaq+AlzZ6kraAA54CSXJYcA3VdWX2vvTgVcCu4FzgUvb36vbV3YDL0pyJd3gKQ+0bpvvAf57zyAqpwMvG+rSSJIkrXNJNgNPA24AnkF3XPUC4IN0V/fup0v8ru/5Wu/tLgtvgzllkd9Z0a0wozIptwUYx/qPY9i3APUa5i1Go1ivg/SPmwHemWS+/h9U1Z8muRG4Ksn5wKeB57X61wBn0p1BehB4IUBV3ZfkVcCNrd4r5wdXkSRJEiTZBPwR8OKq+mKSNwCvoruV5VXAZcBPsvjtLv16ZT3iNhhY+a0wozIptwUYx/qP47wRDqiyfcv+od1iNIpbiw4YWVXdDnxvn/LPA6f1KS/gwkXmtRPYufwwJUmSpluSR9Eldm+tqncAVNU9PdPfCLyrfVzqdhdvg5E2qOU8506SJEkjkK6L1OXArVX1Gz3lx/RU+xHg5vZ+N3B2kkcnOYHu+cIfoOshdWKSE5IcQjfoyu5xLIOktTfcYQslSZK0Es8AfgK4KclHWtnLgecneSpd18o7gJ8BqKpbklwFfJxupM0Lq+prAEleBLyH7lEIO6vqlnEuiKS1Y3InSZK0xqrqL+h/H901S3znEuCSPuXXLPU9SdPLbpmSJEmSNAVM7iRJkiRpCpjcSZIkSdIUMLmTJEmSpClgcidpKiXZmeTeJDf3mfaSJJXkqPY5SV6XZE+SjyU5uafuuUlua69zx7kMkiRJy2FyJ2la7QK2LSxMcjzww8Cne4rPoHtG1InABcAbWt0jgYuBU4CnAxcnOWKkUUuSJK2QyZ2kqVRV7wfu6zPpNcAv0T0zat5ZwFuqcz1weHtw8DOBa6vqvqq6H7iWPgmjJEnSJPA5d5I2jCTPAT5TVR9NHvY4qWOBO3s+721li5X3m/cFdFf9mJmZYW5u7oDxbN+yfxnRd2YOXdn3VmuQ5Vmu1S5Hv3UxiXGO2vx62IjL3mvfvn0jWQeSOpsvevfQ53nHpc8a+jw3OpM7SRtCkscArwBO7ze5T1ktUf7IwqodwA6ArVu31uzs7AFjOm8F/1Fu37Kfy24af9N9xzmzQ5/nSpa/V791MYlxjtr8etiIy95r17bDGGS/k6RpZrdMSRvFk4ATgI8muQM4DvhwkifQXZE7vqfuccBdS5RLkiRNHJM7SRtCVd1UVUdX1eaq2kyXuJ1cVX8P7AZe0EbNPBV4oKruBt4DnJ7kiDaQyumtTJIkaeKY3EmaSkmuAP4K+K4ke5Ocv0T1a4DbgT3AG4H/DFBV9wGvAm5sr1e2MkmSpInjPXeSplJVPf8A0zf3vC/gwkXq7QR2DjU4SZKkEfDKnSRJkiRNAZM7SZIkSZoCJneSJEmSNAVM7iRJkiRpCpjcSZIkSdIUMLmTJEmSpClgcidJkiRJU8DkTpIkSZKmgMmdJEmSJE0BkztJkiRJmgImd5IkSZI0BQ5e6wCkSbT5oncPXHf7lv2cN0D9Oy591mpCkiRJkpbklTtJkiRJmgImd5IkSZI0BUzuJEmSJGkKDJzcJTkoyV8neVf7fEKSG5LcluRtSQ5p5Y9un/e06Zt75vGyVv7JJM8c9sJIkiRJ0ka1nCt3vwDc2vP51cBrqupE4H7g/FZ+PnB/VX0H8JpWjyQnAWcDTwa2Ab+d5KDVhS9JkiRJggGTuyTHAc8C3tQ+B/gh4O2typuB57b3Z7XPtOmntfpnAVdW1Zer6lPAHuDpw1gISVooyc4k9ya5uafs15J8IsnHkrwzyeE90/r2LEiyrZXtSXLRuJdD0saR5Pgkf57k1iS3JPmFVn5kkmtbb6lrkxzRypPkda19+liSk3vmdW6rf1uSc9dqmSSN16CPQvifwC8Bj22fHw98oar2t897gWPb+2OBOwGqan+SB1r9Y4Hre+bZ+51vSHIBcAHAzMwMc3NzAwW4b9++geuuF9O4TDOHdo8OGKZRrKPlxDjoMq2Xf8sp2u52Ab8FvKWn7FrgZa1tejXwMuClC3oWPBH4syTf2b7zeuCH6dqsG5PsrqqPj2kZJG0s+4HtVfXhJI8FPpTkWuA84LqqurSdZLoIeClwBnBie50CvAE4JcmRwMXAVqDafHZX1f1jXyJJY3XA5C7Js4F7q+pDSWbni/tUrQNMW+o7DxVU7QB2AGzdurVmZ2cXVulrbm6OQeuuF9O4TL/51qu57KbhPl7xjnNmhzo/YKDn1s3bvmX/QMs0ijhHYVq2u6p6f+89v63svT0frwd+tL3/Rs8C4FNJensW7Kmq2wGSXNnqmtxJGrqquhu4u73/UpJb6U6EnwXMtmpvBubokruzgLdUVQHXJzk8yTGt7rVVdR9ASxC3AVeMbWEkrYlBjrKfATwnyZnANwOPo7uSd3iSg9vVu+OAu1r9vcDxwN4kBwPfAtzXUz6v9zuSNG4/CbytvV+qZ8GdC8pP6TezlfQ6WMlV7FFc/R7EWl8h76ffupjEOEdtfj1sxGXvNUW9DgBoJ6eeBtwAzLTEj6q6O8nRrdo3eks1823XYuULf2NFvaVGZVL+DY2jfxyjaA+Ws3zLWR+jbLuG+f/wKP59D5jcVdXL6Lou0a7cvaSqzknyh3Rnva8EzgWubl/Z3T7/VZv+vqqqJLuBP0jyG3Tdnk4EPjDcxZGkA0vyCrruT2+dL+pTreh/X/IjehzAynodLOcK8bxBrxQP21pfIe+n37qYxDhHbX49bMRl77Vr22FT0esAIMkm4I+AF1fVF7uhC/pX7VM28t5SozIpPUeMo38co2gPltNuLWd9jLLtGub/w6Not1fznLuXAr/Yui89Hri8lV8OPL6V/yJdv3Cq6hbgKrruTH8KXFhVX1vF70vSsrWBBZ4NnNO6MsHiPQvscSBprJI8ii6xe2tVvaMV39O6W9L+3tvKbbskPcyykruqmquqZ7f3t1fV06vqO6rqee1eFarqn9rn72jTb+/5/iVV9aSq+q6q+pPhLookLS3JNroTU8+pqgd7Ju0Gzm7P6TyBh3oW3Aic2J7reQjdoCu7xx23pI2hjS5+Ofz/7d19kGTVed/x78+sXgC/8GZN4QVnUbRRpHgjRLYAWynVGGy8YJXBKZGIYGlRcNapgI2sTVkr/YMtRSlUZYQlv1BZmzVLgkEYS4GSKMkbxER2VcAyEhYgpGKNCKxYs3J4kdbEksZ+8kefMb2zM0zP7Ez3zO3vp2qq7z339J3n3r59up9zz73NI1X14b5FM6Oi4PDRUu9od808G3i+Dd/8DHBekuPbnTXPa2WSOm74Y3skaQiS3ELvpgInJdlH785x7wVeAexpw5zurar/UFUPJ5kZWTBN38iCJFfS+1J0FLCrjUKQpJXwJuDtwINJHmhl7wOuAW5LcjnwBHBxW3YXcAG9n5d6AXgnQFU9k+QD9DqoAN4/c3MVSd1mciepk6rqkjmKb5ijbKb+B4EPzlF+F70vUJK0oqrqT5n7ejmAc+eoX8AV86xrF7Br+aKTtBYcyTV3kiRJkqRVwuROkiRJkjrA5E6SJEmSOsBr7iRJkqQO2bCMv/O2fdP0mvrNy3HnmTtJkiRJ6gCTO0mSJEnqAJM7SZIkSeoAr7mTJEmSNHSLuTbQa/8G45k7SZIkSeoAkztJkiRJ6gCTO0mSJEnqAJM7SZIkSeoAkztJkiRJ6gCTO0mSJEnqAJM7SZIkSeoAkztJkiRJ6gCTO0mSJEnqAJM7SZ2UZFeSA0ke6is7IcmeJI+2x+NbeZJ8NMneJF9Kckbfc7a2+o8m2TqKbZEkSRqEyZ2krroR2DKrbAdwd1VtBO5u8wDnAxvb3zbgeuglg8DVwFnAmcDVMwmhJEnSamNyJ6mTqupzwDOzii8Edrfp3cBFfeU3Vc+9wHFJTgZ+CthTVc9U1bPAHg5PGCVJklaFdaMOQJKGaKKq9gNU1f4kr2rl64En++rta2XzlR8myTZ6Z/2YmJhgampqwWC2b5peZPgwcfTSnnekBtmexTrS7ZhrX6zGOFfazH4Yx23vd/DgwRXZB5K0lpjcSRJkjrJ6ifLDC6t2AjsBNm/eXJOTkwv+08t2fGrwCJvtm6a59sHhN92PXzq57Otcyvb3m2tfrMY4V9rMfhjHbe9345ZjGeR9J0ld5rBMSePk6TbckvZ4oJXvA07tq3cK8NRLlEuSJK06JneSxsmdwMwdL7cCd/SVv6PdNfNs4Pk2fPMzwHlJjm83UjmvlUmSJK06DsuU1ElJbgEmgZOS7KN318trgNuSXA48AVzcqt8FXADsBV4A3glQVc8k+QDw+Vbv/VU1+yYtkiRJq4LJnaROqqpL5ll07hx1C7hinvXsAnYtY2iSJEkrwmGZkiRJktQBJneSJEmS1AELJndJXpnkz5L8RZKHk/xaKz8tyX1JHk3ysSQvb+WvaPN72/INfet6byv/apKfWqmNkiRJkqRxM8iZu28D51TVG4DTgS3tbnIfAq6rqo3As8Dlrf7lwLNV9RrgulaPJK8H3gb8M2AL8DtJjlrOjZEkSZKkcbVgclc9B9vsy9pfAecAt7fy3cBFbfrCNk9bfm6StPJbq+rbVfU1enelO3NZtkKSJGmNS7IryYEkD/WV/WqSryd5oP1d0LdszhFRSba0sr1Jdgx7OySNzkB3y2xn2O4HXgP8NvCXwHNVNd2q7APWt+n1wJMAVTWd5HngxFZ+b99q+5/T/7+2AdsAJiYmmJqaGmhDDh48OHDdtaKL2zRxNGzfNL1wxUVYiX20mBgH3aa18lp28biTpDXiRuC3gJtmlV9XVb/eXzBrRNQPAf8zyT9pi38b+El637U+n+TOqvrySgYuaXUYKLmrqr8DTk9yHPAJ4HVzVWuPmWfZfOWz/9dOYCfA5s2ba3JycpAQmZqaYtC6a0UXt+k3b76Dax9c3l/gePzSyWVdH8BlOz41cN3tm6YH2qaViHMldPG4k6S1oKo+13+vggX8w4go4GtJ+kdE7a2qxwCS3NrqmtxJY2BR37Kr6rkkU8DZwHFJ1rWzd6cAT7Vq+4BTgX1J1gE/ADzTVz6j/zmSJEma25VJ3gH8ObC9qp7lpUdEPTmr/Ky5VrrU0VIrZbWMHOlCHMs5SmolRl0ZR89KHGcLJndJfhD4bkvsjgZ+gt5NUu4B3grcCmwF7mhPubPN/++2/LNVVUnuBP4gyYfpDR/YCPzZMm+PJElSl1wPfIDeaKcPANcC/475R0TNdT+Fw0ZKwdJHS62U1TJypAtxLGYE0kIGHaG00roYx0qM6hokspOB3e26u+8BbquqTyb5MnBrkv8MfBG4odW/AfhvbXjAM/TGg1NVDye5jd6wgGngijbcU5IkSXOoqqdnppP8LvDJNvtSI6IcKSWNqQWTu6r6EvDGOcofY467XVbV3wIXz7OuDwIfXHyYkiRJ4yfJyVW1v83+LDBzJ835RkQF2JjkNODr9DrZ/+1wo5Y0KqM/tylJkiSS3AJMAicl2QdcDUwmOZ3e0MrHgV+Alx4RleRK4DPAUcCuqnp4yJsiaURM7iRJnbFhGa8zkYatqi6Zo/iGOcpm6s85Iqqq7gLuWsbQJK0RC/6IuSRJkiRp9TO5kyRJkqQOMLmTJEmSpA4wuZMkSZKkDjC5kzR2kvxykoeTPJTkliSvTHJakvuSPJrkY0le3uq+os3vbcs3jDZ6SZKkuZncSRorSdYDvwRsrqofoXer8LcBHwKuq6qNwLPA5e0plwPPVtVrgOtaPUmSpFXH5E7SOFoHHJ1kHXAMsB84B7i9Ld8NXNSmL2zztOXnJskQY5UkSRqIv3MnaaxU1deT/DrwBPD/gD8G7geeq6rpVm0fsL5NrweebM+dTvI8cCLw1/3rTbIN2AYwMTHB1NTUgrFs3zS9YJ3ZJo5e2vOO1CDbs1hHuh2j2herzcx+WI2v0TAdPHhwRfaBJK0lJneSxkqS4+mdjTsNeA74Q+D8OarWzFNeYtmLBVU7gZ0AmzdvrsnJyQVjuWwJP7i9fdM01z44gqb7wb9ZgZUe2XaMbF+sMjP74fFLJ5d93Us5Rkflxi2ZWSfmAAAUMElEQVTHMsj7TpK6zGGZksbNTwBfq6pvVNV3gY8DPwYc14ZpApwCPNWm9wGnArTlPwA8M9yQJUmSFmZyJ2ncPAGcneSYdu3cucCXgXuAt7Y6W4E72vSdbZ62/LNVddiZO0mSpFEzuZM0VqrqPno3RvkC8CC9dnAn8B7g3Un20rum7ob2lBuAE1v5u4EdQw9akiRpAF6sIGnsVNXVwNWzih8Dzpyj7t8CFw8jLulIbFhD18dJklaGZ+4kSZIkqQNM7iRJkiSpA0zuJEmSJKkDTO4kSZIkqQNM7iRJkiSpA0zuJEmSJKkDTO4kSZIkqQNM7iRJkiSpA0zuJEmSJKkDTO4kSZIkqQNM7iRJkiSpA0zuJEmSJKkDTO4kSZIkqQNM7iRJkiSpA0zuJEmSJKkDTO4kSZIkqQMWTO6SnJrkniSPJHk4yVWt/IQke5I82h6Pb+VJ8tEke5N8KckZfeva2uo/mmTrym2WJEmSJI2XQc7cTQPbq+p1wNnAFUleD+wA7q6qjcDdbR7gfGBj+9sGXA+9ZBC4GjgLOBO4eiYhlKRhSnJcktuTfKV1XP3oUjqsJGk5JdmV5ECSh/rK7EyXNLB1C1Woqv3A/jb9rSSPAOuBC4HJVm03MAW8p5XfVFUF3Nu+RJ3c6u6pqmcAkuwBtgC3LOP2SGNlw45PLfs6b9xy7LKvcxX6CPDpqnprkpcDxwDvo9dhdU2SHfQ6rN7DoR1WZ9HrsDprNGFL6rgbgd8Cbuorm+lMH6ht6utM3wwUcH+SO6vq2aFthaSRWTC565dkA/BG4D5goiV+VNX+JK9q1dYDT/Y9bV8rm6989v/YRu+MHxMTE0xNTQ0U28GDBweuu1Z0cZsmjobtm6aXdZ0rsY8WE+Og2zTqOAfVxeOuX5LvB94MXAZQVd8BvpNkUR1WM+2fJC2Xqvpc+67Vz850SQMbOLlL8r3AHwHvqqpvJpm36hxl9RLlhxZU7QR2AmzevLkmJycHim9qaopB664VXdym37z5Dq59cFF9Cgt6/NLJZV0fwGWLOCO2fdP0QNs06jgHdeOWYzt33M3yauAbwO8neQNwP3AVi++wOiS5W0rH1FKS85XoIFmr3Bc97oeeDndMrUhnOiy9Q32lrJbXsAtxLGebsFramC7GsRLH2UDfspO8jF5id3NVfbwVPz3Te916ig608n3AqX1PPwV4qpVPziqfWnrokrQk64AzgF+sqvuSfIQXrxmey4p1TC0lOR+0M2EcuC963A89Y9AxNdsRdabD0jvUV8pq6dTuQhzL2fm7WtqYLsaxEh3/g9wtM8ANwCNV9eG+RXcCMxfpbgXu6Ct/R7vQ92zg+dbj9BngvCTHt4uBz2tlkjRM+4B9VXVfm7+dXrL3dOuoYsAOK0kahsW2TbZZ0hgb5G6ZbwLeDpyT5IH2dwFwDfCTSR4FfrLNA9wFPAbsBX4X+I8Abez3B4DPt7/3z4wHl6Rhqaq/Ap5M8tpWdC7wZRbfYSVJw2BnuqSBDXK3zD9l7lP80PtSNLt+AVfMs65dwK7FBChJK+AXgZvbnTIfA95Jr7PrtiSXA08AF7e6dwEX0OuweqHVlaRll+QWepewnJRkH727Xl7DItqmqnomyUxnOtiZLo2V0Q9claQhq6oH6N0mfLZFdVhJ0nKqqkvmWWRnuqSBDDIsU5IkSZK0ypncSZIkSVIHmNxJkiRJUgeY3EmSJElSB5jcSZIkSVIHmNxJkiRJUgeY3EmSJElSB5jcSZIkSVIHmNxJkiRJUgeY3EmSJElSB5jcSZIkSVIHmNxJkiRJUgeY3EmSJElSB5jcSZIkSVIHmNxJkiRJUgeY3EmSJElSB5jcSZIkSVIHmNxJGjtJjkryxSSfbPOnJbkvyaNJPpbk5a38FW1+b1u+YZRxS5IkvRSTO0nj6Crgkb75DwHXVdVG4Fng8lZ+OfBsVb0GuK7VkyRJWpVM7iSNlSSnAD8N/F6bD3AOcHurshu4qE1f2OZpy89t9SVJkladdaMOQJKG7DeAXwG+r82fCDxXVdNtfh+wvk2vB54EqKrpJM+3+n89e6VJtgHbACYmJpiamlowkO2bphesM9vE0Ut7Xhe5L3rcDz0HDx4c6H0nSV1mcidpbCR5C3Cgqu5PMjlTPEfVGmDZoYVVO4GdAJs3b67Jycm5qh3ish2fWrDObNs3TXPtgzbd4L6Y4X7ouXHLsQzyvpOkLvPTQNI4eRPwM0kuAF4JfD+9M3nHJVnXzt6dAjzV6u8DTgX2JVkH/ADwzPDDliRJWpjX3EkaG1X13qo6pao2AG8DPltVlwL3AG9t1bYCd7TpO9s8bflnq2rOM3eSJEmjZnInSfAe4N1J9tK7pu6GVn4DcGIrfzewY0TxSZIkLchhmZLGUlVNAVNt+jHgzDnq/C1w8VADkyRJWiLP3EmSJElSB5jcSZIkSVIHmNxJkiRJUgcsmNwl2ZXkQJKH+spOSLInyaPt8fhWniQfTbI3yZeSnNH3nK2t/qNJts71vyRJkiRJSzPIDVVuBH4LuKmvbAdwd1Vdk2RHm38PcD6wsf2dBVwPnJXkBOBqYDO9HwC+P8mdVfXscm3IWrFhET9avH3T9EA/cvz4NT99JCFJkqRVLsnjwLeAvwOmq2pz+371MWAD8Djwr6vq2SQBPgJcALwAXFZVXxhF3JKGa8Ezd1X1OQ7/0d4Lgd1tejdwUV/5TdVzL70fBj4Z+ClgT1U90xK6PcCW5dgASZKkMfHjVXV6VW1u8zOd7RuBu3nx51r6O9u30etslzQGlvpTCBNVtR+gqvYneVUrXw882VdvXyubr/wwSbbRa4iYmJhgampqoIAOHjw4cN1R2r5peuC6E0cPVn8tbPeMQbdpMVZi+9fK67Tc+xLWzntJksSFwGSb3k3v513eQ19nO3BvkuOSnDzz3U1Sdy3379xljrJ6ifLDC6t2AjsBNm/eXJOTkwP946mpKQatO0qDDLOcsX3TNNc+uPBL9Pilk0cQ0XD95s13DLRNi7ES279WXqfFxDmoG7ccuybeS5I0Zgr44yQF/Nf2fWmxne2HJHdL7VBfKaulc7ELcSxn5+9KdMwbR89KHGdL/Zb99EwPUBt2eaCV7wNO7at3CvBUK5+cVT61xP8tSZI0bt5UVU+1BG5Pkq+8RN2BOtWX2qG+UlZLR30X4ljOzt9BO7FXWhfjWImO/6VGdiewFbimPd7RV35lklvp3VDl+ZYAfgb4LzN31QTOA9679LAlSZLGR1U91R4PJPkEcCaL72zXEVrMjfEG5Y3xtJwWTO6S3ELvrNtJSfbRu+vlNcBtSS4HngAubtXvondnpr307s70ToCqeibJB4DPt3rvr6rZN2k5Ig9+/fllH6Lmm02SJI1akmOB76mqb7Xp84D3s8jO9uWKxwRnec23Pwe9a7rUb8HkrqoumWfRuXPULeCKedazC9i1qOgkSZI0AXyi9wsHrAP+oKo+neTzLKKzfTXbsONTy57MjHPCqPE1+oGrkiRJmldVPQa8YY7y/8siO9sldZvJnSRJkjpnqcNHHQ6ptWzBHzGXJEmSJK1+JneSJEmS1AEmd5IkSZLUASZ3ksZKklOT3JPkkSQPJ7mqlZ+QZE+SR9vj8a08ST6aZG+SLyU5Y7RbIEmSNDeTO0njZhrYXlWvA84GrkjyemAHcHdVbQTubvMA5wMb29824PrhhyxJkrQwkztJY6Wq9lfVF9r0t4BHgPXAhcDuVm03cFGbvhC4qXruBY5LcvKQw5YkSVqQP4UgaWwl2QC8EbgPmKiq/dBLAJO8qlVbDzzZ97R9rWz/rHVto3dmj4mJCaamphb8/9s3TS865omjl/a8LnJf9Lgfeg4ePDjQ+06SuszkTtJYSvK9wB8B76qqbyaZt+ocZXVYQdVOYCfA5s2ba3JycsEYlvI7Sts3TXPtgzbd4L6Y4X7ouXHLsQzyvpOkLnNYpqSxk+Rl9BK7m6vq46346Znhlu3xQCvfB5za9/RTgKeGFaskSdKgTO4kjZX0TtHdADxSVR/uW3QnsLVNbwXu6Ct/R7tr5tnA8zPDNyVJklYTx3FIGjdvAt4OPJjkgVb2PuAa4LYklwNPABe3ZXcBFwB7gReAdw43XEmSpMGY3EkaK1X1p8x9HR3AuXPUL+CKFQ1KkiRpGTgsU5IkSZI6wOROkiRJkjrA5E6SJEmSOsDkTpIkSZI6wOROkiRJkjrA5E6SJEmSOsDkTpIkSZI6wOROkiRJkjrA5E6SJEmSOsDkTpIkSZI6wOROkiRJkjrA5E6SJEmSOsDkTpIkSZI6wOROkiRJkjrA5E6SJEmSOsDkTpIkSZI6wOROkiRJkjpg6Mldki1Jvppkb5Idw/7/krRYtluS1hrbLWk8DTW5S3IU8NvA+cDrgUuSvH6YMUjSYthuSVprbLek8TXsM3dnAnur6rGq+g5wK3DhkGOQpMWw3ZK01thuSWMqVTW8f5a8FdhSVT/f5t8OnFVVV/bV2QZsa7OvBb464OpPAv56GcNdDdymtaFr27SY7flHVfWDKxnMqA3SbrXypbZdi9W14+1IuC963A89g+4H260X6w2r3RrUajmWjeNQxnGoUcQxcLu1bqUjmSVzlB2SXVbVTmDnolec/HlVbV5qYKuR27Q2dG2burY9y2DBdguW3nYtOhhfn3/gvuhxP/S4Hw6xqtqtQa2W19A4jGMtxDGfYQ/L3Aec2jd/CvDUkGOQpMWw3ZK01thuSWNq2Mnd54GNSU5L8nLgbcCdQ45BkhbDdkvSWmO7JY2poQ7LrKrpJFcCnwGOAnZV1cPLtPpVM6xgGblNa0PXtqlr23NEVrjdWgpfnxe5L3rcDz3uh2YVtluDWi2voXEcyjgOtVrimNNQb6giSZIkSVoZQ/8Rc0mSJEnS8jO5kyRJkqQOWPPJXZJdSQ4keWjUsSyHJKcmuSfJI0keTnLVqGM6UklemeTPkvxF26ZfG3VMyyXJUUm+mOSTo45lOSR5PMmDSR5I8uejjkeHGtfXZ652PskJSfYkebQ9Hj/KGIdhnv3wq0m+3o6JB5JcMMoYh2G+z8lxPCbWutmfoe0GMPe11/Bj7WYwKx3DYe3qKI6lJMcluT3JV9qx/aPDjiPJa/vakgeSfDPJu0a0P365vb8fSnJL+y45iuPjqhbDw0ne1cpWdVuz5pM74EZgy6iDWEbTwPaqeh1wNnBFktePOKYj9W3gnKp6A3A6sCXJ2SOOablcBTwy6iCW2Y9X1emr+Tdcxtw4vj43cng7vwO4u6o2Ane3+a67kbk/765rx8TpVXXXkGMahfk+J8fxmFjrZn+Gfoje8bwReBa4fEhxzG5XR3EsfQT4dFX9U+AN9PbLUOOoqq/OtCXAvwBeAD4x7DiSrAd+CdhcVT9C76ZAb2PIx0eSHwH+PXAmvdfkLUk2ssrbmjWf3FXV54BnRh3Hcqmq/VX1hTb9LXpv7vWjjerIVM/BNvuy9rfm7+ST5BTgp4HfG3UsUpfN085fCOxu07uBi4Ya1Ah07fNuqV7ic3Lsjom1bPZnaJIA5wC3tyqjfA2Heiwl+X7gzcANAFX1nap6bthxzHIu8JdV9X9GFMc64Ogk64BjgP0M//h4HXBvVb1QVdPA/wJ+llXe1qz55K7LkmwA3gjcN9pIjlwbevEAcADYU1VrfpuA3wB+Bfj7UQeyjAr44yT3J9k26mB0GF+fF01U1X7ofdkHXjXieEbpyiRfasM2V9XwoJU263PSY2Jtmf0ZeiLwXPsSDb0fYh9G5/Zc7eqwj6VXA98Afr8NU/29JMeOII5+bwNuadNDjaOqvg78OvAEvaTueeB+hn98PAS8OcmJSY4BLgBOZZW3NSZ3q1SS7wX+CHhXVX1z1PEcqar6u3aa/xTgzHaqe81K8hbgQFXdP+pYltmbquoM4Hx6Q53ePOqAdAhfH812PfCP6Q153w9cO9pwhqdrn5PjZJ7P0MxRdRijfFZDu7oOOAO4vqreCPwNIxzq165l+xngD0f0/4+nd3bsNOCHgGPpvT6zrejxUVWP0BsKugf4NPAX9IaFr2omd6tQkpfR+8C6uao+Pup4llMbZjDF2r9O8k3AzyR5HLgVOCfJfx9tSEeuqp5qjwfojbM/c7QRqZ+vzyGeTnIyQHs8MOJ4RqKqnm6dZ38P/C5jckzM8znpMbF2HPYZSu9M3nFtGB70OoOfWulA5mlXh30s7QP29Y1qup1esjeqY/p84AtV9XSbH3YcPwF8raq+UVXfBT4O/BijOT5uqKozqurN9IbFP8oqb2tM7laZNub8BuCRqvrwqONZDkl+MMlxbfpoem/ar4w2qiNTVe+tqlOqagO9oQufraqfG3FYRyTJsUm+b2YaOI/ekAStAr4+h7kT2NqmtwJ3jDCWkZn5gtH8LGNwTLzE56THxBoxz2fopcA9wFtbtRV/DV+iXR3qsVRVfwU8meS1rehc4MvDjqPPJbw4JJMRxPEEcHaSY9r7fWZ/DPX4AEjyqvb4w8C/ordfVnVbk6q1fV+LJLcAk8BJwNPA1VV1w0iDOgJJ/iXwJ8CDvDgO/X1r+Q5oSf45vQtOj6LXoXBbVb1/tFEtnySTwH+qqreMOpYjkeTV9HotoTdE5A+q6oMjDEl9xvn1maudB/4HcBvww/S+CFxcVZ2+2cg8+2GS3pDMAh4HfmHmWpCumu9zkt51d2N1THRB/2doa+duBU4Avgj8XFV9ewX/95ztapITGfKxlOR0ejeXeTnwGPBO2nemIcdxDPAk8Oqqer6VjWJ//Brwb+gNg/wi8PP0rrEb2vHR4vgTeteDfhd4d1XdPYr9sRhrPrmTJEmSJDksU5IkSZI6weROkiRJkjrA5E6SJEmSOsDkTpIkSZI6wOROkiRJkjrA5E6SJEmSOsDkTpIkSZI64P8D2iKsX5oJjP0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x1080 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Histogram of all numeric fields\n",
    "df1.hist(figsize=(15,15));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ibaruah001\\AppData\\Local\\Annaconda\\lib\\site-packages\\ipykernel_launcher.py:3: FutureWarning: Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>FitmentPercent</th>\n",
       "      <th>CurrentCTC</th>\n",
       "      <th>ExpectedCTC</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>JR70175</th>\n",
       "      <td>75.935455</td>\n",
       "      <td>18.165289</td>\n",
       "      <td>25.235537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR79193</th>\n",
       "      <td>76.979749</td>\n",
       "      <td>18.250737</td>\n",
       "      <td>25.231563</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR81165</th>\n",
       "      <td>75.348855</td>\n",
       "      <td>18.127753</td>\n",
       "      <td>25.162996</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR85289</th>\n",
       "      <td>75.646108</td>\n",
       "      <td>18.081967</td>\n",
       "      <td>25.003832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR87525</th>\n",
       "      <td>75.877453</td>\n",
       "      <td>17.987705</td>\n",
       "      <td>25.031616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR88654</th>\n",
       "      <td>75.257857</td>\n",
       "      <td>18.380952</td>\n",
       "      <td>25.366667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR88873</th>\n",
       "      <td>76.679581</td>\n",
       "      <td>18.343722</td>\n",
       "      <td>25.352627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR88879</th>\n",
       "      <td>76.217409</td>\n",
       "      <td>17.727864</td>\n",
       "      <td>24.695370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JR89890</th>\n",
       "      <td>75.585480</td>\n",
       "      <td>18.171997</td>\n",
       "      <td>25.184097</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         FitmentPercent  CurrentCTC  ExpectedCTC\n",
       "JobProfileIDApplyingFor                                         \n",
       "JR70175                       75.935455   18.165289    25.235537\n",
       "JR79193                       76.979749   18.250737    25.231563\n",
       "JR81165                       75.348855   18.127753    25.162996\n",
       "JR85289                       75.646108   18.081967    25.003832\n",
       "JR87525                       75.877453   17.987705    25.031616\n",
       "JR88654                       75.257857   18.380952    25.366667\n",
       "JR88873                       76.679581   18.343722    25.352627\n",
       "JR88879                       76.217409   17.727864    24.695370\n",
       "JR89890                       75.585480   18.171997    25.184097"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking Average values of some of the categorical variables\n",
    "\n",
    "# JobProfileIDApplyingFor\n",
    "\n",
    "df1.groupby(['JobProfileIDApplyingFor'])['FitmentPercent','CurrentCTC','ExpectedCTC',].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ibaruah001\\AppData\\Local\\Annaconda\\lib\\site-packages\\ipykernel_launcher.py:5: FutureWarning: Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.\n",
      "  \"\"\"\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>FitmentPercent</th>\n",
       "      <th>CurrentCTC</th>\n",
       "      <th>ExpectedCTC</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Artificial Intelligence</th>\n",
       "      <td>75.712249</td>\n",
       "      <td>18.039480</td>\n",
       "      <td>24.962519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Computer Science</th>\n",
       "      <td>76.104620</td>\n",
       "      <td>18.088019</td>\n",
       "      <td>25.106549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Electrical</th>\n",
       "      <td>75.713683</td>\n",
       "      <td>18.115363</td>\n",
       "      <td>25.016639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <td>76.462175</td>\n",
       "      <td>18.435000</td>\n",
       "      <td>25.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Electronics</th>\n",
       "      <td>75.522331</td>\n",
       "      <td>18.193055</td>\n",
       "      <td>25.216598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Information Technology</th>\n",
       "      <td>75.861590</td>\n",
       "      <td>17.947544</td>\n",
       "      <td>24.969930</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            FitmentPercent  CurrentCTC  ExpectedCTC\n",
       "DegreeBranch                                                       \n",
       "Artificial Intelligence          75.712249   18.039480    24.962519\n",
       "Computer Science                 76.104620   18.088019    25.106549\n",
       "Electrical                       75.713683   18.115363    25.016639\n",
       "Electrical and Electronics       76.462175   18.435000    25.250000\n",
       "Electronics                      75.522331   18.193055    25.216598\n",
       "Information Technology           75.861590   17.947544    24.969930"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking Average values of some of the categorical variables\n",
    "\n",
    "# DegreeBranch\n",
    "\n",
    "df1.groupby(['DegreeBranch'])['FitmentPercent','CurrentCTC','ExpectedCTC',].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Ordinal Encoding for string variables with logical order\n",
    "\n",
    "# HighestDegree\n",
    "\n",
    "# create a list of our conditions\n",
    "conditions = [\n",
    "    (df1['HighestDegree'] == \"B.Tech\") | (df1['HighestDegree'] == \"BCA\") ,\n",
    "    (df1['HighestDegree'] == \"M.Tech\") | (df1['HighestDegree'] == \"MCA\") | (df1['HighestDegree'] == \"MS\"),\n",
    "    (df1['HighestDegree'] == \"Dual M.Tech\") | (df1['HighestDegree'] == \"Dual MBA\") | (df1['HighestDegree'] == \"PhD\"),\n",
    "    ]\n",
    "\n",
    "# create a list of the values we want to assign for each condition\n",
    "values = [1,2,3]\n",
    "\n",
    "# create a new column and use np.select to assign values to it using our lists as arguments\n",
    "df1['HighestDegree_Ordinal'] = np.select(conditions, values)\n",
    "\n",
    "\n",
    "# GraduatingInstitute\n",
    "\n",
    "# create a list of our conditions\n",
    "conditions = [\n",
    "    (df1['GraduatingInstitute'] == \"Tier 3\"),\n",
    "    (df1['GraduatingInstitute'] == \"Tier 2\"),\n",
    "    (df1['GraduatingInstitute'] == \"Tier 1\")]\n",
    "\n",
    "# create a list of the values we want to assign for each condition\n",
    "values = [1,2,3]\n",
    "\n",
    "# create a new column and use np.select to assign values to it using our lists as arguments\n",
    "df1['GraduatingInstitute_Ordinal'] = np.select(conditions, values)\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13645\n",
      "13645\n",
      "13645\n",
      "13645\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>Design</th>\n",
       "      <th>Engineering</th>\n",
       "      <th>Finance</th>\n",
       "      <th>Product</th>\n",
       "      <th>Artificial Intelligence</th>\n",
       "      <th>Computer Science</th>\n",
       "      <th>Electrical</th>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <th>Electronics</th>\n",
       "      <th>Information Technology</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>11041</td>\n",
       "      <td>John</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Electrical</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15079</td>\n",
       "      <td>William</td>\n",
       "      <td>English</td>\n",
       "      <td>26</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18638</td>\n",
       "      <td>James</td>\n",
       "      <td>English</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3941</td>\n",
       "      <td>Charles</td>\n",
       "      <td>English</td>\n",
       "      <td>29</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>BCA</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5936</td>\n",
       "      <td>George</td>\n",
       "      <td>English</td>\n",
       "      <td>25</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR70175</td>\n",
       "      <td>Dual M.Tech</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 38 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0  11041     John                 English   35    Male   \n",
       "1  15079  William                 English   26    Male   \n",
       "2  18638    James                 English   36  Female   \n",
       "3   3941  Charles                 English   29  Female   \n",
       "4   5936   George                 English   25    Male   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289        B.Tech               Electrical   \n",
       "1                 JR87525        B.Tech  Artificial Intelligence   \n",
       "2                 JR87525           PhD         Computer Science   \n",
       "3                 JR87525           BCA   Information Technology   \n",
       "4                 JR70175   Dual M.Tech         Computer Science   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  Design  Engineering  Finance  \\\n",
       "0              Tier 1                 7  ...       1            0        0   \n",
       "1              Tier 3                 7  ...       0            1        0   \n",
       "2              Tier 1                 6  ...       0            1        0   \n",
       "3              Tier 2                 5  ...       0            0        0   \n",
       "4              Tier 3                 8  ...       0            1        0   \n",
       "\n",
       "   Product Artificial Intelligence  Computer Science Electrical  \\\n",
       "0        0                       0                 0          1   \n",
       "1        0                       1                 0          0   \n",
       "2        0                       0                 1          0   \n",
       "3        1                       0                 0          0   \n",
       "4        0                       0                 1          0   \n",
       "\n",
       "  Electrical and Electronics Electronics  Information Technology  \n",
       "0                          0           0                       0  \n",
       "1                          0           0                       0  \n",
       "2                          0           0                       0  \n",
       "3                          0           0                       1  \n",
       "4                          0           0                       0  \n",
       "\n",
       "[5 rows x 38 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# One Hot Encoding\n",
    "\n",
    "# CurrentCompanyType has only 3 types of values\n",
    "\n",
    "# Get one hot encoding \n",
    "one_hot = pd.get_dummies(df1['CurrentCompanyType'])\n",
    "\n",
    "# Join the encoded df\n",
    "df2 = df1.join(one_hot)\n",
    "print(len(df1))\n",
    "print(len(df2))\n",
    "\n",
    "# DepartmentInCompany has only 5 types of values\n",
    "\n",
    "# Get one hot encoding \n",
    "one_hot2 = pd.get_dummies(df1['DepartmentInCompany'])\n",
    "\n",
    "# Join the encoded df\n",
    "df3 = df2.join(one_hot2)\n",
    "print(len(df3))\n",
    "\n",
    "# DepartmentInCompany has only 5 types of values\n",
    "\n",
    "# Get one hot encoding \n",
    "one_hot3 = pd.get_dummies(df1['DegreeBranch'])\n",
    "\n",
    "# Join the encoded df\n",
    "df4 = df3.join(one_hot3)\n",
    "print(len(df4))\n",
    "\n",
    "df4.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>Engineering</th>\n",
       "      <th>Finance</th>\n",
       "      <th>Product</th>\n",
       "      <th>Artificial Intelligence</th>\n",
       "      <th>Computer Science</th>\n",
       "      <th>Electrical</th>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <th>Electronics</th>\n",
       "      <th>Information Technology</th>\n",
       "      <th>Current_less_Grad_Year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>11041</td>\n",
       "      <td>John</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Electrical</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15079</td>\n",
       "      <td>William</td>\n",
       "      <td>English</td>\n",
       "      <td>26</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18638</td>\n",
       "      <td>James</td>\n",
       "      <td>English</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3941</td>\n",
       "      <td>Charles</td>\n",
       "      <td>English</td>\n",
       "      <td>29</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>BCA</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5936</td>\n",
       "      <td>George</td>\n",
       "      <td>English</td>\n",
       "      <td>25</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR70175</td>\n",
       "      <td>Dual M.Tech</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 39 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0  11041     John                 English   35    Male   \n",
       "1  15079  William                 English   26    Male   \n",
       "2  18638    James                 English   36  Female   \n",
       "3   3941  Charles                 English   29  Female   \n",
       "4   5936   George                 English   25    Male   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289        B.Tech               Electrical   \n",
       "1                 JR87525        B.Tech  Artificial Intelligence   \n",
       "2                 JR87525           PhD         Computer Science   \n",
       "3                 JR87525           BCA   Information Technology   \n",
       "4                 JR70175   Dual M.Tech         Computer Science   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  Engineering  Finance  Product  \\\n",
       "0              Tier 1                 7  ...            0        0        0   \n",
       "1              Tier 3                 7  ...            1        0        0   \n",
       "2              Tier 1                 6  ...            1        0        0   \n",
       "3              Tier 2                 5  ...            0        0        1   \n",
       "4              Tier 3                 8  ...            1        0        0   \n",
       "\n",
       "   Artificial Intelligence Computer Science  Electrical  \\\n",
       "0                        0                0           1   \n",
       "1                        1                0           0   \n",
       "2                        0                1           0   \n",
       "3                        0                0           0   \n",
       "4                        0                1           0   \n",
       "\n",
       "  Electrical and Electronics Electronics Information Technology  \\\n",
       "0                          0           0                      0   \n",
       "1                          0           0                      0   \n",
       "2                          0           0                      0   \n",
       "3                          0           0                      1   \n",
       "4                          0           0                      0   \n",
       "\n",
       "   Current_less_Grad_Year  \n",
       "0                      12  \n",
       "1                       3  \n",
       "2                       6  \n",
       "3                       6  \n",
       "4                       2  \n",
       "\n",
       "[5 rows x 39 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Current Year - Graduation Year\n",
    "\n",
    "df4['Current_less_Grad_Year'] = 2021 - df4['GraduationYear']\n",
    "df4.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ibaruah001\\AppData\\Roaming\\Python\\Python37\\site-packages\\category_encoders\\utils.py:21: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead\n",
      "  elif pd.api.types.is_categorical(cols):\n",
      "C:\\Users\\ibaruah001\\AppData\\Roaming\\Python\\Python37\\site-packages\\category_encoders\\utils.py:21: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead\n",
      "  elif pd.api.types.is_categorical(cols):\n",
      "C:\\Users\\ibaruah001\\AppData\\Roaming\\Python\\Python37\\site-packages\\category_encoders\\utils.py:21: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead\n",
      "  elif pd.api.types.is_categorical(cols):\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>Computer Science</th>\n",
       "      <th>Electrical</th>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <th>Electronics</th>\n",
       "      <th>Information Technology</th>\n",
       "      <th>Current_less_Grad_Year</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>11041</td>\n",
       "      <td>John</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Electrical</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>12</td>\n",
       "      <td>JR85289SSEDesign</td>\n",
       "      <td>17.840000</td>\n",
       "      <td>24.613333</td>\n",
       "      <td>76.326833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>15079</td>\n",
       "      <td>William</td>\n",
       "      <td>English</td>\n",
       "      <td>26</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>JR87525BAEngineering</td>\n",
       "      <td>18.379679</td>\n",
       "      <td>25.465241</td>\n",
       "      <td>75.582888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>18638</td>\n",
       "      <td>James</td>\n",
       "      <td>English</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 1</td>\n",
       "      <td>6</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>JR87525SDEEngineering</td>\n",
       "      <td>17.802065</td>\n",
       "      <td>24.869191</td>\n",
       "      <td>75.790534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3941</td>\n",
       "      <td>Charles</td>\n",
       "      <td>English</td>\n",
       "      <td>29</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>BCA</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>JR87525SDEProduct</td>\n",
       "      <td>18.264957</td>\n",
       "      <td>25.307692</td>\n",
       "      <td>76.877094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5936</td>\n",
       "      <td>George</td>\n",
       "      <td>English</td>\n",
       "      <td>25</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR70175</td>\n",
       "      <td>Dual M.Tech</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>JR70175SDEEngineering</td>\n",
       "      <td>18.130435</td>\n",
       "      <td>25.152174</td>\n",
       "      <td>76.341739</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 43 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0  11041     John                 English   35    Male   \n",
       "1  15079  William                 English   26    Male   \n",
       "2  18638    James                 English   36  Female   \n",
       "3   3941  Charles                 English   29  Female   \n",
       "4   5936   George                 English   25    Male   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289        B.Tech               Electrical   \n",
       "1                 JR87525        B.Tech  Artificial Intelligence   \n",
       "2                 JR87525           PhD         Computer Science   \n",
       "3                 JR87525           BCA   Information Technology   \n",
       "4                 JR70175   Dual M.Tech         Computer Science   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  Computer Science  Electrical  \\\n",
       "0              Tier 1                 7  ...                 0           1   \n",
       "1              Tier 3                 7  ...                 0           0   \n",
       "2              Tier 1                 6  ...                 1           0   \n",
       "3              Tier 2                 5  ...                 0           0   \n",
       "4              Tier 3                 8  ...                 1           0   \n",
       "\n",
       "   Electrical and Electronics  Electronics Information Technology  \\\n",
       "0                           0            0                      0   \n",
       "1                           0            0                      0   \n",
       "2                           0            0                      0   \n",
       "3                           0            0                      1   \n",
       "4                           0            0                      0   \n",
       "\n",
       "   Current_less_Grad_Year Job_Title_DegreeBrach_CurrentDesignation  \\\n",
       "0                      12                         JR85289SSEDesign   \n",
       "1                       3                     JR87525BAEngineering   \n",
       "2                       6                    JR87525SDEEngineering   \n",
       "3                       6                        JR87525SDEProduct   \n",
       "4                       2                    JR70175SDEEngineering   \n",
       "\n",
       "  Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded  \\\n",
       "0                                          17.840000             \n",
       "1                                          18.379679             \n",
       "2                                          17.802065             \n",
       "3                                          18.264957             \n",
       "4                                          18.130435             \n",
       "\n",
       "  Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded  \\\n",
       "0                                          24.613333              \n",
       "1                                          25.465241              \n",
       "2                                          24.869191              \n",
       "3                                          25.307692              \n",
       "4                                          25.152174              \n",
       "\n",
       "   Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded  \n",
       "0                                          76.326833   \n",
       "1                                          75.582888   \n",
       "2                                          75.790534   \n",
       "3                                          76.877094   \n",
       "4                                          76.341739   \n",
       "\n",
       "[5 rows x 43 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Target Encoding\n",
    "\n",
    "df4['Job_Title_DegreeBrach_CurrentDesignation'] = df4['JobProfileIDApplyingFor']+df4['CurrentDesignation'] + df4['DepartmentInCompany'] \n",
    "encoder = TargetEncoder()\n",
    "df4['Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded'] = encoder.fit_transform(df4['Job_Title_DegreeBrach_CurrentDesignation'], df4['CurrentCTC'])\n",
    "df4['Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded'] = encoder.fit_transform(df4['Job_Title_DegreeBrach_CurrentDesignation'], df4['ExpectedCTC'])\n",
    "df4['Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded'] = encoder.fit_transform(df4['Job_Title_DegreeBrach_CurrentDesignation'], df4['FitmentPercent'])\n",
    "\n",
    "df4.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['EmpID', 'EmpName', 'LanguageOfCommunication', 'Age', 'Gender',\n",
       "       'JobProfileIDApplyingFor', 'HighestDegree', 'DegreeBranch',\n",
       "       'GraduatingInstitute', 'LatestDegreeCGPA', 'YearsOfExperince',\n",
       "       'GraduationYear', 'CurrentCTC', 'ExpectedCTC', 'MartialStatus',\n",
       "       'EmpScore', 'CurrentDesignation', 'CurrentCompanyType',\n",
       "       'DepartmentInCompany', 'TotalLeavesTaken', 'BiasInfluentialFactor',\n",
       "       'FitmentPercent', 'HighestDegree_Ordinal',\n",
       "       'GraduatingInstitute_Ordinal', 'Enterprise', 'MidSized', 'Startup',\n",
       "       'Customer Success', 'Design', 'Engineering', 'Finance', 'Product',\n",
       "       'Artificial Intelligence', 'Computer Science', 'Electrical',\n",
       "       'Electrical and Electronics', 'Electronics', 'Information Technology',\n",
       "       'Current_less_Grad_Year', 'Job_Title_DegreeBrach_CurrentDesignation',\n",
       "       'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
       "       'Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded',\n",
       "       'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df4.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors = df4[['LatestDegreeCGPA', 'YearsOfExperince','CurrentCTC', 'ExpectedCTC', \n",
    "                   'EmpScore','TotalLeavesTaken', 'HighestDegree_Ordinal',\n",
    "                   'GraduatingInstitute_Ordinal', 'Enterprise', 'MidSized', 'Startup',\n",
    "                   'Customer Success', 'Design', 'Engineering', 'Finance', 'Product',\n",
    "                   'Artificial Intelligence', 'Computer Science', 'Electrical',\n",
    "                   'Electrical and Electronics', 'Electronics', 'Information Technology',\n",
    "                   'Current_less_Grad_Year',\n",
    "                   'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
    "                   'Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded',\n",
    "                   'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded']]\n",
    "Target = df4['FitmentPercent'] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x21e802c2e10>"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 1.Select the top n features based on feature importance from random forest\n",
    "\n",
    "np.random.seed(10)\n",
    "plt.rcParams[\"figure.figsize\"] = (10,20)\n",
    "# define the model\n",
    "model = RandomForestRegressor(random_state = random.seed(10))\n",
    "# fit the model\n",
    "model.fit(predictors, Target)\n",
    "\n",
    "# get importance\n",
    "features = predictors\n",
    "importances = model.feature_importances_\n",
    "indices = np.argsort(importances)\n",
    "\n",
    "feat_importances = pd.Series(model.feature_importances_, index=predictors.columns)\n",
    "feat_importances.nlargest(30).plot(kind='barh')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Enterprise',\n",
       " 'CurrentCTC',\n",
       " 'EmpScore',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded',\n",
       " 'TotalLeavesTaken',\n",
       " 'ExpectedCTC',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
       " 'LatestDegreeCGPA',\n",
       " 'Current_less_Grad_Year',\n",
       " 'YearsOfExperince',\n",
       " 'GraduatingInstitute_Ordinal',\n",
       " 'HighestDegree_Ordinal']"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Final Features from Random Forest (Select Features with highest feature importance)\n",
    "rf_top_features = pd.DataFrame(feat_importances.nlargest(13)).axes[0].tolist()\n",
    "rf_top_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x21e802c29e8>"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAARiCAYAAADFk4hxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Xm0nVV9//H3xwQZZFJBGqmYKgHKZISAMlWwFKvRItWKSn+CVVMcStWqpVUUrdZYreKIja3gWBFRq2IFVBBEGZKQAXDgJ8SfUlukamSIIPH7++PsK4dL7r0ZNiS5eb/Wuivn7Gc/372f5y7v8sPezzmpKiRJkiRJ6uUB63sCkiRJkqTJxaApSZIkSerKoClJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqaur6noAkbep22GGHmj59+vqehiRJ0oQWLFhwc1XtOFE/g6YkrWfTp09n/vz563sakiRJE0ryw9Xp59ZZSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1pA5Pkd5J8KskPklyb5MtJdrsfx98+yUsm6DM9ydX3w1xemeS7SZYmWZzknUk2W4d6Y847ycOS3JDkd4baPpDk5LUdT5IkaVNl0JQ2IEkCfA64qKoeXVV7An8P7LSa508Z9X7qWkxje2DcoHl/SHIicBTw+KraBzgAuAnYchV9p4xuW1NVdRPwNuAdreZ+wKHAP69L3Qz4t1aSJG1S/D8/0oblCODXVfXBkYaqWgRMSfKlkbYk70tyQnu9LMnrk3wT+LMkFyX5xyTfAP46yY5JzklyZfs5pJ13apIPt/7XJzmplZ8LPDrJoiRvn2jCSaYkeXurvSTJX7b2aUkubnWuTnJY63tme780ySvGKf1a4MVV9Yt2H+6sqrlV9ctW/9Ykb0pyOXBQuwdXttrzWmgnyf5tNfTbwEsnuJx57dqPAN4HvKyqfp1kaltNvaJd4wtb7W2TfD3Jwtb+1Na+a5vHB4GFwLSJ7qMkSdJksjarHZLuO3sDC9bivF9V1aHw25XA7avqCe39J4F3VdU3k+wCnAf8fjtvDwbhdhvge0lOB04G9q6qmas59guA5VV1QJLNgUuTnA/8KXBeVb2lrThuBcwEdq6qvdvctl9VwSTbAFtX1Q3jjPsg4Oqqen0759qqelN7/THgqcAXgTOAv6qqb0wUnKvqN0leDHwd+EJVXdwOzQFuqqoD2zVe1q7xJ8DRVXVLkocBlwIj/0FgT+D5VXXiGNc4p9Vll112GW9akiRJGx2DpjQ5nDXO+yOBPdsCH8C2LcgBnFtVdwB3JLmJ1dyiO8pRwL5JntnebwfMAK4EPtyeqfx8VS1Kcj3wqCTvBc4Fzh+jZoD67ZvkSQy2tW4PPLeqvgWsBM4ZOueIJK9hEGgfAlyT5GIGofsbrc/HgCePdzFtnlcDHxh1jb+f5NmjrvG/gbclORT4DfCIJDu0Pj+oqivHGWcegxVUZs2aVWP1kyRJ2hgZNKUNyzXAM1fRfhf33Oq+xajjt43z/gHAQVW1YrhDC553DDWtZO3+JoTBiuF59zqQ/AEwG/hYkrdX1UeTPAZ4EoNtrM8C/mL0eVX1yyS3Jfm9qrqh1T6vbR9+YOv2q6pa2cbZgkEwnFVVP0pyKoN7dI/AugZ+036Gr/ElVfW1Udf3Qgahc7+quivJj7n7dzP6dyJJkrTJ8BlNacPydWDzJC8aaUhyADCFwark5km2A/5wDWqeD7xsqN5EW2JvYbCVdnWdB7y4rVySZLckD0rySAbbTT8E/BuwX1vte0BVnQOcAuw3Tt23AqePbK9tz1yODtgjRtpvTrI1Lay35zuXtxVHgOPW4LpGX+NL0j5cKcnuSbZkEDJvaiHzj4Cd17K+JEnSpOKKprQBqapKcgxwWgZfq/ErYBnwcuDTwBLgOuCqNSh7EvD+JEsY/G/+YmCVzw22Ofxvkkvb9tH/rKpXT1D/X4HpwMIWBn8KPB04HHh1kl8DtwLPYxDEzsjdn8L6d+PUPZ3BNtjLk9zRalzKKq69qn6R5EPAUgb3a3jL6vMZbOG9nUFgXBv/AuwCLGorwTcBRzPYivvFJPMZfOjPdWtZX5IkaVJJlY8GSdL6NGvWrJo/f/76noYkSdKEkiyoqlkT9XPrrCRJkiSpK7fOShpTkn0YbA8ddkdVPa7jGO8HDhnV/O6qOqPXGKsYc+RTbIfdUFXH3FdjSpIkbUoMmpLGVFVLGXz35X05xkvvy/pjjHkea/+8piRJkibg1llJkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUldT1/cEJGlTt/TG5Uw/+dx1qrFs7uxOs5EkSVp3rmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSiID30zy5KG2ZyX5yn045p8mWZrkO0mWJHna0LE9kyxOclWSXZOsTLJo6OfV99W82vjH3NdjSJIkTWZ+vYkkqqqSnAicneRCYArwFuCP16VukqlVddcq2vcD3gYcWVU/TPJo4IIk11fVNcCfAp+pqn9IMhW4papmrstc1nDOn7s/xpIkSZqsXNGUBEBVXQ18Efhb4A3AR6vqB0mOT3JFW0n8QJIHACSZl2R+kmuSvH6kTpIfJzklyaXAMUlekeTatkL58dbt1cA/VNUP29g/YBA8X5XkT4CXAScm+epY802yfZLvJ9m1vT87yfOTTE3yiyTvSrIwyQVJHtr6zEhyXpIFSS5Osltr/3iSf24h+x+TvDDJaUPH3p3kW0muT3LM0Bz+vq3KLk7ylvHGkCRJ2pS4oilp2BuBhcCdwKwkewPHAAdX1V1J5gHPBj4JnFxVP2srjhcm+UxVXdvq3FZVhwAk+QnwyKq6M8n27fhewJtHjT0feEFVPT/JgcDNVXVaq79NkkVDfd9cVZ9JchJwZpLTga2q6ozWfzvgsqp6RZI3AacALwfmAS9sAfoQ4H3AUa3mo4E/rKrfJHnhqLk9DDgE2Af4NPC5ttX3ycCBVbUiyUNa3/HGkCRJ2iQYNCX9VlXdluQs4NaquiPJkcABwPwkAFsCP2rdn5PkBQz+jjwc2BMYCZpnDZW9Bvh4kv8APt/aAtSo4VfVNmKVW2er6itJngW8G9h36NBdwNnt9ceBT7aQ+3jgnHYtcM+/gWdX1W/GGP/zVVXAkiQ7t7YjgQ9X1Yo2l5+txhh3X2wyB5gDMGXbHccYVpIkaeNk0JQ02m/aDwzC34er6pThDklmAH/NYDXvF21L7BZDXW4bev0k4AnA0cDr2irpNcAs7g6mAPuNej+hJFOAPYAVwIOB/2qHRgfWatdy8zjPet42RjvAHcPDDv27qrA83hh3T6hqHoPVTzafNmOsgC1JkrRR8hlNSeP5KvCsJDsAJHlokl2AbYFbgF8mmcYgTN5LC4K/W1VfZ/Bc5o7AVsA7GITOXVq/RzF4NvSf13B+rwIWAc8DRrbNAmzG4AOFAJ4LfLOqfg78ZOQZyyQPSPKYNRxv2PnAC5Js2eo95D4YQ5IkaaPkiqakMVXV0iRvBL7aPgTo18CJDJ6nvBa4GrgeuHSMElMZbFvdhsF/2HpbVd3CYCvua4Evt3D4a+Bv2gcSrcroZzTPZbAl9ngGq6q3JrkM+DvgrcByYL8kfw/8DDi2nfds4PQkpwIPbDUWr9FNaarqSy1Ezk/yawYfpHRKzzEkSZI2Vhk8diRJk0MLrjdX1fYTdt5AbD5tRk07/rR1qrFs7uxOs5EkSRpbkgVVNWuifm6dlSRJkiR15dZZSZNKVd0FbDSrmZIkSZORK5qSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK78MCBJWs/22Xk75vv1JJIkaRJxRVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVOSJEmS1JVBU5IkSZLU1dT1PQFJ2tQtvXE5008+d53rLJs7u8NsJEmS1p0rmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJrSJibJrWvQ9/AkB6/lONOTPHdUreVJrkryvSQXJ3nq2tTuKcmrknw3ydVJFid5XmufmuQfk1yXZFH7ee3QeStb29VJzk6y1dCxY5JUkj3WxzVJkiStbwZNSeM5HFiroAlMB547qu2SqnpsVe0OnAS8L8kfrv30BjKwxn/PkpwI/BFwYFXtDfwBkHb4zcDDgX2qaiZwGLDZ0OkrqmpmO+9O4MShY88Bvgk8e40vRpIkaRIwaEoiydOSXN5WG7+aZKck0xmEp1e0lbvDkuyY5JwkV7afQ9r5Txha9bsqyTbAXOCw1vaK0WNW1SLgTcDLWo2xau+Y5IIkC5P8S5IfJtmhrZh+J8kHgIXAI5IcleTbre/ZSbZuNfZP8o0kC5Kcl2Ram8bfAy+pql+2OS2vqo+01ckXAX9VVb9qx26pqlPHuIWXALu2sbYGDgFegEFTkiRtogyakmCw+vb4qnos8CngNVW1DPgg8K62cncJ8O72/gDgGcC/tvNfBbx0aOVvBXAygxXMmVX1rjHGXQiMbC8dq/YbgK9X1X7A54Bdhs7fHfhom/dtwOuAI1vf+cArk2wGvBd4ZlXtD3wYeEsLw9tU1Q9WMa9dgf9XVbdMdOOSTAWeDCxtTU8HvlJV3wd+lmS/iWpIkiRNNlPX9wQkbRB+FzirrfQ9ELhhjH5HAnsmI7tL2bYFtkuBdyb5BPDZqvrxUJ/xDHcaq/ahwDEAVfWVJD8fOueHVXVZe/14YE/g0lbjgcC3GYTRvYELWvsU4Cdt7FqtSSbPB/4aeChwcFX9CNgyyaLW5RLg39rr5wCntdefau8XrqLmHGAOwJRtd1ydaUiSJG00DJqSYLDi986q+kKSw4FTx+j3AOCgqloxqn1uknOBpwCXJTlyNcd9LPCd8Wpn/MR623BX4IKqes6o8/cBrqmqg0afnOS2JI+qqutHHfq/wC5JtmlbZs8AzkhyNYOgCu0ZzVH1Hgo8Edg7SbW+leQ1VXWPUFtV84B5AJtPm7FagVeSJGlj4dZZSQDbATe218cPtd8CbDP0/nzaM5UASWa2fx9dVUur6m0MtqzusYpz7yHJvsApwPvHq81gW++zWttRwIPHKHkZcEiSkWclt0qyG/A9YMckB7X2zZLs1c55K/D+JNu2Y9smmVNVtzNYoXxfki3asSkMVknH80wGW3kfWVXTq+oRDFaHD53gPEmSpEnFoClterZK8uOhn1cyWME8O8klwM1Dfb8IHDPyYUAMPil2VpIlSa7l7k9affnI14MweD7zP4ElwF0ZfGXIyIcBHdY+LOh7DALmSVX1tXZsrNpvBI5KspDBs5A/YRBi76GqfgqcAPx7kiUMguceVXUngwD4tja/Rdz9SbqnAxcCV7bVym8At7djr21jXZ3kKgbbYz8C/Nc49/Y5DJ4jHXYO9/70XUmSpEkto3ZzSdIGJcnmwMqququtSp4+esvqxm7zaTNq2vGnTdxxAsvmzu4wG0mSpLElWVBVsybq5zOakjZ0uwCfbt+TeSeDrx2RJEnSBsygKWmDVlXXMfjQIEmSJG0kfEZTkiRJktSVQVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVHwYkSevZPjtvx3y/mkSSJE0irmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkrqaur4nIEmbuqU3Lmf6yed2q7ds7uxutSRJktaGK5qSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSu/HoTSZNKkpXA0qGmpwM7AM+rqpPWz6wkSZI2LQZNSZPNiqqaOaptGTB/PcxFkiRpk+TWWUmTXpLDk3ypvT41yYeTXJTk+iQnDfX7fJIFSa5JMmeo/dYkb0myOMllSXZq7Tsl+VxrX5zk4Nb+50muSLIoyb8kmXJ/X7MkSdL6ZNCUNNls2QLeoiSfG6PPHsCTgAOBNyTZrLX/RVXtD8wCTkry0Nb+IOCyqnoMcDHwotb+HuAbrX0/4Jokvw8cCxzSVlZXAsd1vkZJkqQNmltnJU02q9o6O9q5VXUHcEeSm4CdgB8zCJfHtD6PAGYA/wvcCXyptS8A/qi9fiLwPICqWgksT/J/gP2BK5MAbAncNHoCbcV0DsCUbXdci8uUJEnacBk0JW2K7hh6vRKYmuRw4EjgoKq6PclFwBatz6+rqob7j1M7wEeq6u/Gm0BVzQPmAWw+bUaN11eSJGlj49ZZSRrYDvh5C5l7AI9fjXO+BrwYIMmUJNu2tmcmeVhrf0iSR95Xk5YkSdoQGTQlaeArDFY2lwD/AFy2Guf8NXBEkqUMttTuVVXXAq8Dzm+1LgCm3UdzliRJ2iDl7t1gkqT1YfNpM2ra8ad1q7ds7uxutSRJkoYlWVBVsybq54qmJEmSJKkrg6YkSZIkqSuDpiRJkiSpK4OmJEmSJKkrg6YkSZIkqavxvnRcknQ/2Gfn7ZjvJ8VKkqRJxBVNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUldT1/cEJGlTt/TG5Uw/+dxu9ZbNnd2tliRJ0tpwRVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVOSJEmS1JVBU5IkSZLUlUFTkiRJktSVQVPahCV5aJJF7ee/k9w49P6Bq+j/kCQnrkbdqUl+0V7vmmTRfTH/NZFk5tC1/SzJDe31eeOcs0eS+ffnPCVJkiYDv0dT2oRV1f8CMwGSnArcWlXvGOeUhwAnAh+872fXV1Ut4u5r/Tjwmar6/PqdlSRJ0uTkiqakVUrymiRXt5+/as1zgd3bSuDcJNsm+XqShUmWJHnqGtSfkeS8JAuSXJxkt9Z+dJLLk1yV5PwkD0syJckPk2zb+iTJ9Ul2SLJTks8mmZ/kiiSPb32emGRxm+vCJA8aZy4PTnLh0HU8eRV9Rq57ZpLNkpzWxluc5Pmtzx8nuSDJ55J8L8kZq3/HJUmSJg9XNCXdS5IDgeOAA4EpwBVJvgGcDOxaVSMrg5sBR1fVLUkeBlwKfGk1h5kHvLCqfpDkEOB9wFHAxcAXqqraNt2/qaq/TfIl4GjgY8DBwPer6uYkZwH/VFWXJZnext8beDUwp6ouT7I18Ktx5nIb8LSqujXJTsAlwH8O3Y+9gI8Dx1XVNUlOAn5cVQcm2QK4PMn5rft+wF7AT4Erk8yqKrffSpKkTYpBU9KqHAacU1W3AyT5PHAocP6ofgHeluRQ4DfAI5LsAPxivOJJtgceD5yTZKR55O/RLsCnk/wOsDnw/dZ+FvAaBkHz2e09wJEMVllH6jw4yZYMQu9pST7ZruXW8aYEvKMF3pXAI9scAR4OnAP8SVWNzOUoYNckf97ebwvs2l5/q6r+u13nYmA6cK+gmWQOMAdgyrY7jjM1SZKkjY9BU9KqZOIuADwP2A7Yr6ruSvJjYIvVrH/zyMroKO8H/rGqvpzkSAarqDBYZTwzyUOBPwFOGap1YFXdOarOm5N8AZjNYGXx8Kq6boz5PB/YCnhsu47/HrqOnwE3Awdxd+gN8JdV9Y17XFTyx8AdQ00rGePvbFXNY7Cqy+bTZtQY85IkSdoo+YympFW5GDgmyZZt2+nRDILeLcA2Q/22A25q4eyPgJ1Xp3hV/Rz4SZJjAJI8IMljhmremMES5fFD5xTwH8BpwOKqGlk1/Srw0pF+SUa29T66qpZU1VuBq4Ddx5nSdsD/tOv4Y2CnoWO/YhBsX5zkGa3tPOAlSaa2sX6/baGVJEkSrmhKWoWquiLJvwNXtqbTq2opQPvQnaXAucA7gS+2rwBZCIy1YrhnW+0c8VcMtr+e3j7t9oEMnoFcDJwKfA74MXAFMG3ovLOAbwN/PtT20lbn+Qz+pl3Y2l6V5DAGW3qXcO9tv8M+AnwhyZUMtrneMOp+/DLJbOCrSW4DPgA8AriqBeL/YRBGJUmSBGSwSCBJWl82nzajph1/Wrd6y+bO7lZLkiRpWJIFVTVron5unZUkSZIkdWXQlCRJkiR1ZdCUJEmSJHVl0JQkSZIkdWXQlCRJkiR15debSNJ6ts/O2zHfT4qVJEmTiCuakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK6mru8JSNKmbumNy5l+8rlday6bO7trPUmSpDXhiqYkSZIkqSuDpiRJkiSpK4OmJEmSJKkrg6YkSZIkqSuDpiRJkiSpK4OmJEmSJKkrg6a0HiRZmWTR0M/Jrf2iJLPWot7MJE8Z5/isJO9Zy7mu1ZzWtm5r/97QvflMaz81yavWYpztk7ykx5yHar4pyZE9a0qSJE0mfo+mtH6sqKqZHevNBGYBXx59IMnUqpoPzO843n3tuDbnHrYHXgJ8YPSBJFOqauWaFqyq1/eYmCRJ0mTliqa0gUpyVJJvJ1mY5OwkW7f2A5J8K8niJFck2Q54E3BsWwE8tq3+zUtyPvDRJIcn+VI7f+skZyRZmmRJkme09tOTzE9yTZI3rsb8Xp/kyiRXt7HS2i9K8rY2t+8nOay1b5nkU23Ms4At1+HePDrJV5IsSHJJkj1a+05JPtfuzeIkBwNzgUe3e/P2di8uTPJJYGk775XtOq5O8vLWNj3Jd5J8qN2T85Ns2Y6dmeSZY/w+tkmyV3u9qF3vjLW9VkmSpI2RK5rS+rFlkkVD799aVWeNvEmyA/A64Miqui3J3wKvTDIXOAs4tqquTLItcDvwemBWVb2snX8qsD9waFWtSHL40FinAMurap/W98Gt/bVV9bMkU4CvJdm3qpaMcw3vq6o3tRofA54KfLEdm1pVB7btvG8AjgReDNxeVfsm2RdYOE7tTyRZ0V5fUFWvHnV8HnBiVV2X5HEMViufCLwH+EZVHdOuY2vgZGDvkRXkdi8ObG03JNkfeD7wOCDA5Um+AfwcmAE8p6pelOTTwDOAj49MIskDuffvYwVwIvDuqvpE6zNlnGuVJEmadAya0vox0dbZxwN7Ape2hcIHAt8Gdgd+UlVXAlTVLwFan9G+UFUrVtF+JPDskTdV9fP28llJ5jD4uzCtjT9e0DwiyWuArYCHANdwd9D8bPt3ATC9vf4DBkGQqlqSZLzaY26dbSu7BwNnD1335u3fJwLPa2OsBJYPBelhV1TVDe31ocDnquq2Vv+zwGHAF4AbqmrkPwgMX8uIsX4f3wZem+R3gc9W1XWruI45wByAKdvuOMZtkCRJ2jgZNKUNUxis5D3nHo2DlcBazRq3jVP7HjWS/B7wKuCAqvp5kjOBLcacXLIFg1XEWVX1o7aCOtz/jvbvSu75d2Z15z6eBwC/WMdnXIfvzSpTenPH0OuV3Hu7773uJUBVfTLJ5cBs4LwkL6yqr4/qM4/ByiybT5vR475IkiRtMHxGU9owXQYckmRXgCRbJdkN+C7w8CQHtPZtkkwFbgG2Wc3a5wMvG3nTVvy2ZRC+lifZCXjyBDVGQuXNbYXxmasx7sXAcW3MvYF9V3O+99BWDW9I8metVpI8ph3+GoMtuiSZ0rayTnRvLgae3u7xg4BjgEtWczqr/H0keRRwfVW9h8HK6FpdqyRJ0sbKoCmtH1vmnl9vMnf4YFX9FDgB+Pe2xfQyYI+quhM4FnhvksXABQxC34XAnq3WsROM/Wbgwe2DbxYDR1TVYuAqBttfPwxcOl6BqvoF8CEGH6bzeeDK1bjm04Gt2/W8BrhinL6fGLo3X13F8eOAF7T5XwMc3dr/msGW3qUMtrruVVX/y2AL8tVJ3r6Ka1kInNnmcznwr1V11WpcD+P8Po4Frm7P4e4BfHR16kmSJE0WqXLHliStT5tPm1HTjj+ta81lc2d3rSdJkgSQZEFVTfgd665oSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkrqaOnEXSdJ9aZ+dt2O+nxIrSZImEVc0JUmSJEldGTQlSZIkSV0ZNCVJkiRJXRk0JUmSJEldGTQlSZIkSV0ZNCVJkiRJXRk0JUmSJEldGTQlSZIkSV0ZNCVJkiRJXRk0JUmSJEldGTQlSZIkSV0ZNCVJkiRJXRk0JUmSJEldGTQlSZIkSV0ZNCVJkiS1vyVkAAAgAElEQVRJXRk0JUmSJEldGTQlSZIkSV0ZNCVJkiRJXRk0JUmSJEldGTQlSZIkSV1NXd8TkKRN3dIblzP95HO71lw2d3bXepIkSWvCFU1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JkiRJUlcGTUmSJElSVwZNSZIkSVJXBk1JG5QkK5MsGvo5ubVflGTWWtSbmeQpnef45STb96wpSZI0mfg9mpI2NCuqambHejOBWcCXRx9IMrWq7lrTglXVNbhKkiRNNq5oStroJDkqybeTLExydpKtW/sBSb6VZHGSK5JsB7wJOLatjh6b5NQk85KcD3w0yRZJzkiyNMlVSY5otU5I8tkkX0lyXZJ/Ghp/WZId2uvnJVnSxvxYa/uzJFe3tovv9xskSZK0nrmiKWlDs2WSRUPv31pVZ428aQHvdcCRVXVbkr8FXplkLnAWcGxVXZlkW+B24PXArKp6WTv/VGB/4NCqWpHkbwCqap8kewDnJ9mtDTcTeCxwB/C9JO+tqh8NzWUv4LXAIVV1c5KHtEOvB55UVTeOtcU2yRxgDsCUbXdc23slSZK0QTJoStrQTLR19vHAnsClSQAeCHwb2B34SVVdCVBVvwRofUb7QlWtaK8PBd7bzvlukh8CI0Hza1W1vNW5Fngk8KOhOk8EPlNVN7fzf9baLwXOTPJp4LOrmkBVzQPmAWw+bUaNc72SJEkbHYOmpI1NgAuq6jn3aEz2BVY3sN02qt5Y7hh6vZJ7/83MqsasqhOTPA6YDSxKMrOq/nc15yZJkrTR8xlNSRuby4BDkuwKkGSrttX1u8DDkxzQ2rdJMhW4BdhmnHoXA8e1c3YDdgG+t5pz+RrwrCQPbec/pP376Kq6vKpeD9wMPGINr1GSJGmjZtCUtKHZctTXm8wdPlhVPwVOAP49yRIGwXOPqroTOBZ4b5LFwAXAFsCFwJ4jHwa0ivE+AExJspTBM54nVNUdq+h3L1V1DfAW4BttzHe2Q29vHy50NYMgu3iN7oAkSdJGLlU+GiRJ69Pm02bUtONP61pz2dzZXetJkiQBJFlQVRN+t7krmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuRn/5uCTpfrbPztsx30+JlSRJk4grmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSupq7vCUjSpm7pjcuZfvK5XWsumzu7az1JkqQ14YqmJEmSJKkrg6YkSZIkqSuDpiRJkiSpK4OmJEmSJKkrg6YkSZIkqSuDpqRJKclrk1yTZEmSRUkel+TlSbZai1onJHn4fTFPSZKkycigKWnSSXIQ8FRgv6raFzgS+BHwcmCNgmaSKcAJgEFTkiRpNRk0JU1G04Cbq+oOgKq6GXgmg7B4YZILAZKcnmR+W/l848jJSZYleX2SbwLPAWYBn2gro1u24zu0vrOSXNRen5rkY0m+nuS6JC+6Py9akiRpQ2HQlDQZnQ88Isn3k3wgyROq6j3AfwFHVNURrd9rq2oWsC/whCT7DtX4VVUdWlUfB+YDx1XVzKpaMcHY+wKzgYOA17vlVpIkbYoMmpImnaq6FdgfmAP8FDgryQmr6PqsJAuBq4C9gD2Hjp21lsP/R1WtaKuoFwIHrqpTkjltNXX+ytuXr+VQkiRJG6ap63sCknRfqKqVwEXARUmWAscPH0/ye8CrgAOq6udJzgS2GOpy2zjl7+Lu/1C3xahjNcH7kfnNA+YBbD5txir7SJIkbaxc0ZQ06STZPcmMoaaZwA+BW4BtWtu2DMLk8iQ7AU8ep+TweQDLGKyYAjxjVN+jk2yR5KHA4cCVa3MNkiRJGzNXNCVNRlsD702yPYPVx//LYBvtc4D/TPKTqjoiyVXANcD1wKXj1DsT+GCSFQyevXwj8G9J/h64fFTfK4BzgV2Af6iq/+p3WZIkSRsHg6akSaeqFgAHr+LQe9vPSL8Txjh/+qj35wDnDDVdAuw2xvDfr6o5azBdSZKkScets5IkSZKkrlzRlKROqurU9T0HSZKkDYErmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrvwwIElaz/bZeTvmz529vqchSZLUjSuakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK4MmpIkSZKkrgyakiRJkqSuDJqSJEmSpK6mru8JSNKmbumNy5l+8rnd6y6bO7t7TUmSpNXhiqYkSZIkqSuDpiRJkiSpK4OmJEmSJKkrg6YkSZIkqSuDpiRJkiSpK4OmJEmSJKkrg6a0jpLslOSTSa5PsiDJt5Mcsw71Tk3yqrU8d3qS5w69n5XkPeswlzOTPLPnPJIcnuTgNa2xFnPYK8nXk3w/yXVJTkmS1Tz3oiSz2usvJ9l+Leew1r9LSZKkjZlBU1oHLbh8Hri4qh5VVfsDzwZ+d1S/++s7a6cDvw1nVTW/qk66n8Ze3XkcDkwYNEfXWBNJtgS+AMytqt2Ax7QxX7KKvuP+bqrqKVX1i7WZhyRJ0qbKoCmtmycCd1bVB0caquqHVfXeJCckOTvJF4Hzk2yd5GtJFiZZmuTokXOSvDbJ95J8Fdh9qH14ZW2HJMva6+lJLmm1Fg6tEM4FDkuyKMkr2urhl9o5pyb5cKt5fZKThsY5Jcl3k1yQ5N9XtQqXZFmSNw7Nf4/W/oQ23qIkVyXZZqx5JJkOnAi8oh07bPSqaZJbx7iWKUnenuTKJEuS/OU4v5fnApdW1fntd3I78DLg5KF7MS/J+cBHk2yZ5FOt7lnAlqOue4d2z7+T5ENJrklyfgu0JHlRm9fiJOck2WqcuUmSJE16Bk1p3ewFLBzn+EHA8VX1ROBXwDFVtR9wBPDPGRhZBX0s8KfAAasx7k3AH7VaxwIj22NPBi6pqplV9a5VnLcH8CTgQOANSTZrQfYZQ+PPGmfcm9uYpwMjYfRVwEuraiZwGLBirHlU1TLgg8C72rFLxhlrdI0XAMur6gAG9+hFSX5vjHP3AhYMN1TVD4Ctk2zbmvYHjq6q5wIvBm6vqn2Bt7RjqzIDeH9V7QX8gsF9A/hsVR1QVY8BvtPmOq4kc5LMTzJ/5e3LJ+ouSZK0UTFoSh0leX9b1bqyNV1QVT8bOQz8Y5IlwFeBnYGdGISzz1XV7VX1SwZbPieyGfChJEuBs4E9V3OK51bVHVV1M4OwuhNwKPAfVbWiqm4BvjjO+Z9t/y5gsLUV4FLgnW2FdPuqums157KmjgKel2QRcDnwUAbBb1UC1BjHRtq/UFUr2us/AD4OUFVLgCVjnHtDVS1qr4fvwd5thXkpcByDoDuuqppXVbOqataUrbabqLskSdJG5f56bkyarK7h7lUtquqlSXYA5rem24b6HgfsCOxfVb9u22C3GDl1jPp3cfd/ENpiqP0VwP8wePbwAQxWS1fHHUOvVzL4G7BaH5Az6vyRc6mquUnOBZ4CXJbkyDWoB0PX2J55feAY/QL8VVWdtxo1r2EQHu8+OXkUcGtV3dI+E+i2UeeM9TsYNvr+jWyxPRN4elUtTnICg+dQJUmSNlmuaErr5uvAFklePNQ21vN52wE3tZB5BPDI1n4xcEx7TnAb4GlD5yzj7m2cw5/+uh3wk6r6DfB/gCmt/RZgmzW8hm8CT0uyRZKtgdlrcnKSR1fV0qp6G4OAvccE8xh9bBl3X+PRDFZrV9XvPODFSTZr4+6W5EFjjPEJ4NCR0NuepXwP8E9j9L+YwX8IIMnewL5j9BvLNsBP2tyOW8NzJUmSJh2DprQOqqqApwNPSHJDkiuAjwB/u4runwBmJZnPIIx8t9VYCJwFLALOAYafW3wHg3D1LWCHofYPAMcnuQzYjbtX55YAd7Xtu69YzWu4ksF23cUMtsbOB9bkocGXJ7k6yWIGz2f+5wTz+CKDYL0oyWHAhxjcvyuAx41zLf8KXAssTHI18C+MsSujbYk9Gnhdku8BS4ErgfeNcQ2nM3h+cwnwGuCKNbh+gFMYbOe9gPZ7lSRJ2pRl8P+TJW3KkmxdVbe2T0u9GJjTArDuB5tPm1HTjj+te91lc9docVqSJGlCSRZU1XgfHgn4jKakgXlJ9mTwHOhHDJmSJElaFwZNSbSv+NjoJNkH+Nio5juq6nHrYz6SJEkaMGhK2mhV1VJg5vqehyRJku7JDwOSJEmSJHXliqYkrWf77Lwd8/3gHkmSNIm4oilJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqyqApSZIkSerKoClJkiRJ6sqgKUmSJEnqaur6noAkbeqW3ric6Sef273usrmzu9eUJElaHa5oSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKkiRJkroyaEqSJEmSujJoSpIkSZK6MmhKG6gkt456f0KS97XXJyZ53gTn/7b/Os7j6Un2HHp/ZpIbkixO8v0kH02y87qO01uS303yH0muS/KDJO9O8sDVPPfMJM9sr/91+PrXcA5dfgeSJEkbG4OmtBGqqg9W1Ufvp+GeDowOWq+uqscAuwNXAReubogbT5Kp61qj1QnwWeDzVTUD2A3YGnjLmo5ZVS+sqmt7zEuSJGlTYdCUNkJJTk3yqvb6gCRLknw7yduTXD3U9eFJvtJW9f5p6PyjWv+FSc5OsnVrn5vk2lbvHUkOBv4EeHuSRUkePTyPGngX8N/Akyeo/ZQk303yzSTvSfKloWuZl+R84KNJprTruLLN4y+H5v3qofY3jnOLngj8qqrOaPNcCbwC+IskW7WVxrOTfBE4PwPva9d+LvCwoTEvSjKrvb41yVvaau5lSXZq7U9LcnmSq5J8daRdkiRpU2XQlDZcW7ZwtyjJIuBNY/Q7Azixqg4CVo46NhM4FtgHODbJI5LsALwOOLKq9gPmA69M8hDgGGCvqtoXeHNVfQv4AoMVzJlV9YMx5rAQ2GOc2lsA/wI8uaoOBXYcdf7+wNFV9VzgBcDyqjoAOAB4UZLfS3IUMAM4sF3X/kn+YIz57AUsGG6oql8C/w/YtTUdBBxfVU9s1717u08vAg4eo+6DgMvaau7FrS/AN4HHV9VjgU8Brxnj/N9KMifJ/CTzV96+fKLukiRJG5Uu29Qk3SdWVNXMkTdJTgBmDXdIsj2wTQuEAJ8EnjrU5Wv1/9m78zDLqvrs+9+bBhsQaBwQS55oK6IEGRoZouAAiiSKE4qCokJiJM6gAV8cEnEgdoJxwAEZnAdEUHwIEAWRWRBo6AmCotLmFcdEbAUZtPk9f+xVcrqssXs3PdT3c1111Tlrr7X22vucqqvuWmvvU7W01b0BeASwOd1S2Mu7FabcD7gC+C1wJ3BKm9U7ewpjTfv+hDH63hb4UVXd3OqdChw20P6sqrqjPd4X2HH4GklgFl3A3Ld9XdfKN2nll4wxnpqg/Pyq+nV7/BTg1Dbz+dMk3x7jOO/m3vMyD3hGe/x/gNOSDLVjvnmUtsupqpOAkwBmDm0z2lglSZLWWgZNae2WCbbfNfB4Gd3PfOhC1kv+rLNkd+DpwEHA6+mWoE7GzsAFY/WdZOcJ2t8+WB14Q1V9c0Qffw28r6pOnMR4rgdeOKL9ZsBfAD+km0G9fUSbyYS9P1TVcL3h8wnwEeADVXVWkr2AYybRlyRJ0jrLpbPSWqyqbgV+l+QJreigSTS7EtgzyaMB2jWLj2nXUs6qqnOBI+iWpwL8Dth0tI7atY1vBIaAb4zVN3Aj8Kgks1vTA8cZ3zeB1yTZoPXxmCT3b+V/N3DN51ZJHjJGHxcAG6fdmTfJDODfgc9U1e9HqX8JcFC7PnQI2Huc8Y1mFnBLe3zIFNtKkiStcwya0trvlcBJSa6gmw0c94K/qvoVcChwapKFdOFwW7oweXYru5ju5jnQXXN4VLvRzfDNgI5LsgD4Pt11lHtX1d1j9d2Wxb4W+EaSy4BfjDPOU4AbgGvbjY1OBNavqvPolgZfkWQRcAZjBOA267g/8KIkN7Vx3gm8bYx9ngncBCwCTmjHPxXHAKcnuRT4nym2lSRJWufk3lVgktZGSTapqtva46OBoao6fDUP688MjzPdxZsfA25qd6yd9mYObVNDh3yo936XzN2v9z4lSdL0lmReVe06UT1nNKW1337tzrSLgScD713dAxrDq9rdc6+nW2o6mWstJUmStBbyZkDSWq6qTgNOW93jmEibvex1BjPJg+iuxxzp6VX1v33uS5IkSZNn0JS01mphcs6EFSVJknSfcumsJEmSJKlXzmhK0mq2w1azuMYb90iSpHWIM5qSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1av3VPQBJmu4W3bKU2Uef03u/S+bu13ufkiRJk+GMpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClplUuyLMn8ga+j74N9bp7ktSvQ7pgkRw48PzLJjUkWJ1mQ5BVJzmzH8YMkSweOa48kGySZm+Sm1uaqJM/s9+gkSZLWbH6OpqT7wh1VNec+3ufmwGuBj69oB0leDTwD2L2qfptkFvD8qtq/bd8LOLKqnj3QZi4wBGxfVXcl2RJ46oofhiRJ0trHGU1Jq0WSWUm+l+Sx7fmpSV7VHt+W5N+TXJvkgiRbtPKtk3wjybwklybZtpVv2WYZF7SvPYC5wNZtpvG4Vu+oJFcnWZjkXQNjeXsby7eAxw4M823Aa6vqtwBVtbSqPjvOMW0MvAp4Q1Xd1dr8oqq+0tuJkyRJWgsYNCXdFzYasXT2wKpaCrwe+EySg4AHVNXJrf79gWur6vHAxcA7W/lJdCFuF+BI7p2tPB64uKp2Ah4PXA8cDfywquZU1VFJ9gW2AXYH5gC7JHlKkl2Ag4CdgRcAuwEk2RTYtKp+OIXjfDTw38PBVJIkabpy6ayk+8KoS2er6vwkLwI+Buw0sOke4LT2+AvA15JsAuwBnJ5kuN7M9v1pwCtan8uApUkeMGJ3+7av69rzTeiC56bAmVX1e4AkZ7XtAWrqhzo5SQ4DDgOYsdkWq2o3kiRJq4VBU9Jqk2Q94C+BO4AHAj8Zo2rRrcD4zUpc6xngfVV14ogxHMEogbJdk3l7kkdV1Y8muY8fAA9PsmlV/W68ilV1Et0MLTOHtlllgVaSJGl1cOmspNXpTcB/AS8BPpVkg1a+HnBAe/xS4LK2HPXmNgNKOsOzoBcAr2nlM5JsBvyObrZy2DeBv2szoyTZKslDgEuA/ZNs1JbLPmegzfuAj7X+SLJZm4kcVZsV/SRwfJL7tTZDSV425TMjSZK0FjNoSrovjLxGc26SxwB/D/xjVV1KF/je0erfDjwuyTy6ZbHvbuUHA69MsoDuOszntfLDgb2TLALmAY+rqv8FLm8fMXJcVZ0HfAm4otU7g+4azGvplunOB74KXDow7hOAC4Grkyymu1709xMc6zuAXwE3tDZfb88lSZKmjVS5YkvSmiXJbVW1yeoex31l5tA2NXTIh3rvd8nc/XrvU5IkTW9J5lXVrhPVc0ZTkiRJktQrg6akNc50ms2UJElaFxk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9Wr91T0ASZrudthqFtd4h1hJkrQOcUZTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXq2/ugcgSdPdoluWMvvoc1ZJ30vm7rdK+pUkSRqPM5qSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmtIZJctsk6jw5yfVJ5ifZ6D4a15wkzxp4/twkR69knw9qxzA/yc+T3DLw/H5T7Ou9SY5YmfEM9PWFJM/voy9JkqTpyM/RlNZOBwPvr6pPT6ZykhlVtWwl9zkH2BU4F6CqzgLOWpkOq+p/W78kOQa4rarev3LDlCRJ0urmjKa0hkqyV5KLkpyR5MYkX0zn74EXA/88UHZcksVJFiU5cKD9hUm+BCxKMrv1c0qr+8Uk+yS5PMlNSXZv7XZP8p0k17Xvj22zi+8GDmyzjQcmOTTJR1ubRyS5IMnC9v3hrfwzSY5v/fwoyQFTPAeHJLmq7fPjSdZr5fsluTbJgiTnDTTZIcnFbV+va3Uf3Y73k20W+D+TbNi2PT7Jd9u4v5pk1ihjeEbb/6IkJw/PtLYZ3e8luTTJR5J8PcmMJD9I8sBWZ0YbywOnctySJElrO4OmtGbbGTgC2A54FLBnVZ1CN5N4VFUdDLyAblZwJ2Af4LgkQ6397sDbq2q79vzRwIeBHYFtgZcCTwKOBN7W6twIPKWqdgb+GfiXqrq7PT6tquZU1WkjxvlR4HNVtSPwReD4gW1DbR/PBuZO9sCTbA/sD+xRVXPoVmAclOShwAnA/lW1E3DQQLPHAM8AngC8O8mMVv5Y4ENV9TjgDmB4WewXgH9s4/4e8E8jxrAx8CnghVW1A7AxcFgr/ziwL/AU4KEAbdb4VLrzCvDXwNVV9etRju+wJNckuWbZ75dO9rRIkiStFQya0prtqqr6SVXdA8wHZo9S50nAqVW1rKp+AVwM7DbQ/uaBujdX1aLW3/XABVVVwKKBvmcBpydZDHwQeNwkxvlE4Evt8efbmIZ9varuqaobgC0n0dewfdpxXJNkPvBUYOu2rwur6scAI0Lc2VV1d1X9Evg1sEUr/0FVLWqP5wGzkzwI2LCqLmvln6ULjYP+Eripqn7Ynn+u1dkO+F5V/bidv1MH2nwSOKQ9/jtg1OXNVXVSVe1aVbvO2PjPJlIlSZLWal6jKa3Z7hp4vIzRf2YzTvvbx+nvnoHn9wz0/R66ILd/ktnARZMc66AaY5/jjXWkAJ+qqpGzjC8Y0f+gsc7XaOWTGctYdcZsW1VLktyaZG+6GenzxqorSZK0rnJGU1r7XUJ37eSMJFvQzbhdtRL9zQJuaY8PHSj/HbDpGG2+w71LWA8GLhuj3lR8C3hxkgfDn+5Q+3DgcuBpSR7Rylfo+seq+h/gjiR7tKKX080GD7oB2CbJo9rzl7U61wOPTfIXSQIcOKLdJ+mWEH+5zR5LkiRNKwZNae13JrAQWAB8G3hLVf18Jfr7N+B9SS4HZgyUXwhsN3wzoBFt3gj8bZKFdIHt8JXYPwBtqeu7gG+1fs8DtmzLg18D/N8kC+gC3Yp6OfDB1v92wHtHjOH3wCuBryVZRDczenIrfz1dGL4U+CkweKHlmXSB/TMrMTZJkqS1VrrLiyRJU5Fkk6q6rc1onggsqqqPtG1PAN5XVXtPpq+ZQ9vU0CEfWiXjXDJ3v1XSryRJmp6SzKuqXSeq54ymJK2Y17SbFN0AbAScDJDk7cBp3HsXX0mSpGnHmwFJ0gqoquOA40YpPxY49r4fkSRJ0prDGU1JkiRJUq8MmpIkSZKkXhk0JUmSJEm98hpNSVrNdthqFtd4d1hJkrQOcUZTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXq2/ugcgSdPdoluWMvvoc1ZJ30vm7rdK+pUkSRqPM5qSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU9JaKUkl+fzA8/WT/CrJ2e35c5McPUbb29r39ZIcn2RxkkVJrk7yyLbt3CSbr+QY9xoejyRJ0nTix5tIWlvdDmyfZKOqugN4BnDL8MaqOgs4a4I+DgQeBuxYVfck+T+tX6rqWatm2JIkSes+ZzQlrc3+Exj+oMiXAKcOb0hyaJKPtsePTHJFm7F8z0D7IeBnVXUPQFX9pKpubW2WJHlwklcnmd++bk5yYdu+b+vz2iSnJ9mklf9NkhuTXAa8YFWfAEmSpDWRQVPS2uzLwEFJNgR2BL47Rr0PAydU1W7AzwfKvwI8p4XIf0+y88iGVfWJqpoD7Ab8BPhAkgcD7wD2qarHA9cAb27jOBl4DvBk4KG9HKUkSdJaxqApaa1VVQuB2XSzmeeOU3VP7p3t/NN1nVX1E+CxwFuBe4ALkgy164oAACAASURBVDx9jD4+DHy7qv4DeAKwHXB5kvnAIcAjgG2Bm6vqpqoq4AtjDSjJYUmuSXLNst8vnfBYJUmS1iZeoylpbXcW8H5gL+BB49SrUQur7qJbgvufSX4BPB+4YLBOkkPpguTrh4uA86vqJSPqzRlrP6Ps9yTgJICZQ9tMqo0kSdLawhlNSWu7TwHvrqpF49S5HDioPT54uDDJ45M8rD1ej2757Y8HGybZBTgSeNnwtZzAlcCeSR7d6myc5DHAjcAjk2zd6i0XRCVJkqYLg6aktVq7gc+HJ6h2OPC6JFcDswbKHwL8R5LFwELgj8BHR7R9PfBA4MJ2LecpVfUr4FDg1CQL6YLntlV1J3AYcE67GdCPkSRJmobSXUYkSVpdZg5tU0OHfGiV9L1k7n4TV5IkSZqkJPOqateJ6jmjKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb1af3UPQJKmux22msU1fgyJJElahzijKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq/WX90DkKTpbtEtS5l99DmrpO8lc/dbJf1KkiSNxxlNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6lVvQTPJbeNs2yvJ2ZPs57tJ5if57yS/ao/nJ5md5Nwkm7ev1w60mZ1k8QqMeXaSO5Jcl+S/klyV5JCp9tOnJMckuaUd841JTkjSy+s03ms0Rv1XJFmc5PokNyQ5so9xTHEMz0+y3cDzzyS5OcmCJN9P8rkkW61g369O8or+Rvunfo9IsvHA83OTbN7zPo5s74/F7Vz0fhwT7H+512Wcen82ziRntvf3D5IsHfgZ3yPJBknmJrmptbkqyTPH6X9JkkUDfRzf75Eut69J/x4baHNRkl1X1ZgkSZLWVGvc52hW1V8BJDkU2LWqXj+w+Vlt22zgtcDHe9jlD6tq59bvo4CvJVmvqj69sh0nmVFVy1ag6Qer6v0tYF4CPBW4cETf61fVH1d2jGNpf9wfAexbVT9NsiHw8im0X258K3Eung+cDdwwUHZUVZ2RJG2MFybZvqrunkrHVfWJFRjPZBwBfAH4fdvPs/rsPMmrgWcAu1fVb5PMojtPk20/8rVZkffSaK/LpMZZVfu37XsBR1bVswfazAWGgO2r6q4kW9K9/8ezd1X9zxTHL0mSpFWo16Wz6RzXZiIWJTlwYPNmbSbjhiSfWJFZujZ78WBgLrB1m8E4bkSdGW0MVydZmOQfJtt/Vf0IeDPwxtbX/ZN8qvV1XZLntfKNk3yl9X9aulnYXdu225K8O8l3gScm2SXJxUnmJflmkqFWb+sk32jllybZdpQh3Q/YELi1tbkoyb8kuRg4PMlz2r6vS/Kt9kc5STZJ8un2GixM8sKB83Nsm1m6crj+GN5KFwJ+2s7NnVV18sA4ho/3wUmWtMeHJjk9yX8A57UZoAuTfAlY1Oq8rM1SzU9yYpIZA+dtubEl2QN4LnBcq7/1iNerquqDwM+BZ7Z+9k1yRZJr21g2aeVz23tvYZL3t7Jj0mZpk+zWtl0x/B4eOKavtdfqpiT/NnAuT0hyTboZ33e1sjcCD6MLvxe2suH3LUne3H4+Fic5opXNTjejfnLr67wkG43z2rwNeG1V/badh6VV9dlR9rVrkosGjvWkJOcBnxv5WrU6Rw383Lxr4Dhf0coWJPn8RK/LZMY5mnSzwK8C3lBVd7U2v6iqr4xzLsbq66Ik/9rea99P8uRWPiPJ+wd+Nt7Qyp/efo4WpfuZn9nK/ybdjOxlwAsG+h/rd8NGSb7c+j4NGO91lCRJWmf1fY3mC4A5wE7APnR/iA61bbsD/wjsAGzNwB9tK+BoupnIOVV11IhtrwSWVtVuwG7Aq5I8cgp9XwsMh763A99ufe1Ndzz3p5tNvbWqdgTeA+wy0P7+wOI2M/td4CPAAVW1C/Ap4NhW7yS6P6h3AY5k+dnZNyWZD/wM+H5VzR/YtnlVPbWq/h24DHhCm5H9MvCWVuef2jnYoY3x2wNju7KqdqKbKX3VOOdhe2DeuGdqdE8EDqmqp7XnuwNvr6rtkvwlcCCwZ1XNAZYBB481tqr6DnAW3QzmnKr64Rj7vBbYtgWsdwD7VNXjgWuANyd5ILA/8Lh2Pt47Sh+fBl5dVU9s4xo0p417B+DAJH/Ryt9eVbsCOwJPTbJjVR0P/JRulm3vwU6S7AL8LfBXwBPo3ps7t83bAB+rqscBvwFeyCiSbApsOs65GM8uwPOq6qXt+Z9eqyT7tjHs3o53lyRPSfI4up+Dp7XX5vDJvC4rOM5HA/89HEyn4MLcu3T2TQPl61fV7nQzzO9sZYcBjwR2bu+FL6abrf8McGBV7UC30uM1rfxk4DnAk4GHDvQ91u+G1wC/b30fy/K/G5aT5LD2j4prlv1+6RQPWZIkac3W99LZJwGntiWSv0g387Yb8FvgqjZjSJJTW90zet4/wL7AjkkOaM9n0f0BffMk22dEX8/Nvdcmbgg8nG7sHwaoqsVJFg60WQZ8tT1+LF1gOz8JwAzgZ22WbQ/g9FYOMHOgj+GlsxsAZyQ5qKq+3LadNlDv/wCntTB/v4Fj3Ac4aLhSVd3aHt5Nt9wRuhD5jAnOxYo4v6p+PfD8qqoaHtfT6f7wvrod90bAL3sY2/BJfAKwHXB56/9+wBV07787gVOSnDOwn65xd/3kpi1AAXwJePZAlQuqammrewPwCOD/B16c5DC6n6Ohtu/B98JITwLOrKrbW19fowswZwE3D/xDYR4we5xjrXH2MZ6zquqOgeeDr9W+7eu69nwTup+bnYAzhpemjnhtx7My45yqsZbOfq19Hzyf+wCfGF4qXFW/TrIT3fn/fqvzWeB1wEWt/CaAJF+gC6ow9u+GpwDHt74XjvjdsJyqOonuH07MHNrmvjpXkiRJ94m+g2bG2TbyD6lV9YdV6GYKv7mC7XcG/mugrxdW1feW28FAOhzFnQPXIga4vs2SDbbfDPhNm9UbU1X9Ick36P54HQ6atw9U+Qjwgao6K931bscM7He08/uHqhouX8b4r//1dKHw26Ns+yP3zoZvOGLb7eM8D/DZqnrrSo5tpJ2BC1r/51fVS0ZWSLI7XdA9CHg98LTBzRP0f9fA42XA+m2W/Ehgt6q6Ncln+PNz8WfDmMI+Rl1y2a51vD3Jo4b/cTPCyrw276uqE5cbcLcUeMo/q5MY52h+ADw8yaZV9bup7nMUw+d08P002s/GVH5vDbYZ7XfDeG0kSZKmjb6Xzl5Ct7RwRpIt6ALSVW3b7kkeme7azAPpln2uqN8Bm46x7Zt0y942AEjymLakbULpbjL0froAN9zXG4aD5cAyx8uAF7ey7eiWVI7me8AWSZ7Y6m6Q5HFtaeDNSV7UytNmVUaOJ3Qzn2MtP5wF3NIeD94t9zy6MDXczwPGaD+e9wH/luShrY+ZLXQALOHeJYEHjNJ2LBcAByR5SOvzgUkeMUGbMV/rdt7eSDeb+A3gSmDPJI9u2zdur/8mwKyqOpduGeVyAb/N+P4uyRNa0UFMbDO6oLY03bWug3dGHWvMlwDPb+O6P91y3ksnsa+R3gd8rP3DgiSbtZlVWP61GXX57Ri+Cfxd7r2mdav2Ol1AN3P7oFb+wFZ/vJ/ByYzzz1TV74FPAscnuV9rM5TkZVM4jomcB7w6yfqt/wcCNwKzh983dDe9uriVPzL3XoM6+A+MsX43XEJbDp5ke7pl1ZIkSdNOXx+bsT7d7MGZdEsHF9DNhL2lqn7eql1BdxOfxXRLPM9c0f1V1f/SLY9cnBE3AwJOobsT5rXpbuhyIuPPjm2d9vEmwFeAj9S9d5x9D7ABsLD19Z5W/nG6ALkQ+P/aMf/ZRVbV3QX1AOBfkywA5tMFR+j+GH1lK78eeN5A0+FrNBe3sY91d91j6JbfXgoMLh18L/CAdn4W0F1DNiUtlH0M+FaS6+mWHw6fx/fThfnvAA+eQp830F1DeV47d+fThcTxfBk4qr1Gw3/wH9eO6/t0S7P3rqq7q+pXwKHAqa3/K+mut90UOLuVXQy8aeRO6K7tPSnJFXSzVeNeNFdVC+iWmV5Pd+3t5QObTwL+M+1mQANtrqW7FvAquut3T6mq65i6E+juQnx1e19eTLvDLfAu4MPtPTHpu/xW1Xl0S4avSLKIbln7plV1Pd21hhe3c/6B1mS012Uq4xzLO4BfATe0Nl9vz8czeI3m5yaoewrw33Q/0wuAl1bVnXTXzp7ejv0euuW1d9ItlT0n3c2AfjzQz1i/G04ANmnvtbdw7z/aJEmSppXcu1pxJTrpZuNObjfeWOelu1PqBlV1Z/sj+wLgMTXFj9fQmiPJJlV1W3t8NDBUVYev5mFpmpg5tE0NHfKhVdL3krn7rZJ+JUnS9JRkXrsh5rhW+hrNdJ+V90a6JYnTxcZ0sygb0M1+vcaQudbbL8lb6X4mfkw3MypJkiRpBax00KzuQ+9X6IPv033W5MwRxS+vqkUrO64R+9kB+PyI4rvaR5BMWbtRyYQpfm2Q5O3Ai0YUn15Vx45Wf11VVaex/B19V7skHwP2HFH84YGl3WuE+2Kc99XvCkmSJPWjl6WzkqQV59JZSZK0tpjs0tm+7zorSZIkSZrmDJqSJEmSpF6t9DWakqSVs8NWs7jGJa6SJGkd4oymJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF6tv7oHIEnT3aJbljL76HNWSd9L5u63SvqVJEkajzOakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpqR1WpJlSeYnuT7JgiRvTrJCv/uSnJtk877HKEmStK7xczQlrevuqKo5AEkeAnwJmAW8c6odVdWzeh6bJEnSOskZTUnTRlX9EjgMeH06M5Icl+TqJAuT/ANAkqEkl7SZ0MVJntzKlyR5cHv8T0luTHJ+klOTHNnKL0ryr0muSvL94baSJEnTiTOakqaVqvpRWzr7EOB5wNKq2i3JTODyJOcBLwC+WVXHJpkBbDzYR5JdgRcCO9P9Hr0WmDdQZf2q2j3Js+hmTvdZ5QcmSZK0BjFoSpqO0r7vC+yY5ID2fBawDXA18KkkGwBfr6r5I9o/Cfi/VXUHQJL/GLH9a+37PGD2qANIDqObXWXGZlus+JFIkiStgVw6K2laSfIoYBnwS7rA+YaqmtO+HllV51XVJcBTgFuAzyd5xchuJtjNXe37Msb4h15VnVRVu1bVrjM2nrXCxyNJkrQmMmhKmjaSbAF8AvhoVRXwTeA1beaSJI9Jcv8kjwB+WVUnA58EHj+iq8uA5yTZMMkmwH733VFIkiSt+Vw6K2ldt1GS+cAGwB+BzwMfaNtOoVvaem2SAL8Cng/sBRyV5A/AbcByM5pVdXWSs4AFwI+Ba4Clq/xIJEmS1hLp/qkvSZqKJJtU1W1JNgYuAQ6rqmtXpK+ZQ9vU0CEf6neAzZK5TrZKkqT+JJlXVbtOVM8ZTUlaMScl2Q7YEPjsioZMSZKkdZFBU5JWQFW9dHWPQZIkaU3lzYAkSZIkSb0yaEqSJEmSemXQlCRJkiT1yms0JWk122GrWVzj3WElSdI6xBlNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSerX+6h6AJE13i25Zyuyjz1ll/S+Zu98q61uSJGk0zmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwya0jooyf5JKsm249TZPMlrB54/LMkZA89PTbIwyZuSvDvJPuP0tWuS4ycY015Jzp5s+XhjnaDube377CSLJzs+SZIk9cfP0ZTWTS8BLgMOAo4ZuTHJDGBz4LXAxwGq6qfAAW37Q4E9quoRk9lZVV0DXNPHwMew3Fin6j4YnyRJkgY4oymtY5JsAuwJvJIuaA6X75XkwiRfAhYBc4Gtk8xPctzgDCBwHvCQtu3JST6TZDiE7pbkO0kWJLkqyaaDs5JJdm/br2vfHzuFsR+T5FNJLkryoyRvbJuWG2ure1SSq9us67sm6HdwfFskOT/JtUlOTPLjJA9u217Wjml+2zajld+W5Nh2zFcm2bKVb5nkzFa+IMke4/UjSZI0XRg0pXXP84FvVNX3gV8nefzAtt2Bt1fVdsDRwA+rak5VHTWij+cObLt0uDDJ/YDTgMOraidgH+COEW1vBJ5SVTsD/wz8yxTHvy3w122s70yywcixJtkX2KbVmQPskuQpk+z/ncC3q+rxwJnAw9ux/SVwILBnVc0BlgEHtzb3B65sx3wJ8KpWfjxwcSt/PHD9BP1IkiRNCy6dldY9LwE+1B5/uT2/tj2/qqpuXom+Hwv8rKquBqiq3wIkGawzC/hskm2AAjaY4j7Oqaq7gLuS/BLYcpQ6+7av69rzTeiC5yWT6P9JwP5t/N9IcmsrfzqwC3B1O56NgF+2bXcDw9eRzgOe0R4/DXhF62sZsDTJy8fp50+SHAYcBjBjsy0mMWxJkqS1h0FTWockeRBd+Nk+SQEzgEryllbl9pXdBV14HM97gAurav8ks4GLpriPuwYeL2P031MB3ldVJ06x7+G2Y5V/tqreOsq2P1TV8HGPNabJ9PMnVXUScBLAzKFtJjqnkiRJaxWXzkrrlgOAz1XVI6pqdlX9BXAz3SzeSL8DNp1i/zcCD0uyG0C7PnNk6JoF3NIeHzrF/scycqzfBP6uXY9Kkq2SPGSSfV0GvLi12xd4QCu/ADhguJ8kD0wy0c2QLgBe0+rPSLLZCvYjSZK0TjFoSuuWl9Bddzjoq8BLR1asqv8FLk+yePgGOxOpqrvprj/8SJIFwPnAhiOq/RvwviSX082orrSRY62q84AvAVckWQScweRD87uAfZNcCzwT+Bnwu6q6AXgHcF6ShXTHNjRBX4cDe7cxzAMet4L9SJIkrVNy72owSVr3JZkJLKuqPyZ5InBCu2nPajNzaJsaOuRDE1dcQUvm7rfK+pYkSdNLknlVtetE9bxGU9J083DgK0nWo7vJz6smqC9JkqQpMmhKmlaq6iZg59U9DkmSpHWZ12hKkiRJknpl0JQkSZIk9cqgKUmSJEnqlddoStJqtsNWs7jGO8NKkqR1iDOakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9Wr91T0ASZruFt2ylNlHn7PK+l8yd79V1rckSdJonNGUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0pWkuyUOTfDnJD5PckOTcJI9ZTWN5Ww99PCHJd5PMT/JfSY6ZoP65STZf2f1KkiTpXgZNaRpLEuBM4KKq2rqqtgPeBmy5moY05aCZZMaIos8Ch1XVHGB74Cvjta+qZ1XVb6a6X0mSJI3NoClNb3sDf6iqTwwXVNX8qro0neOSLE6yKMmBAEn2SnJxkq8k+X6SuUkOTnJVq7d1q/eZJJ9Icmmr9+xWfmiSjw7vL8nZrc+5wEZtJvKLbdvLWr/zk5w4HCqT3Jbk3Um+CzxxxDE9BPhZO5ZlVXVDa7NJkk+3MS5M8sJWviTJgyexv2OTLEhyZZItW/mWSc5s5QuS7DFeP5IkSdOFQVOa3rYH5o2x7QXAHGAnYB/guCRDbdtOwOHADsDLgcdU1e7AKcAbBvqYDTwV2A/4RJINxxpIVR0N3FFVc6rq4CR/CRwI7NlmJ5cBB7fq9wcWV9VfVdVlI7r6IPC9FgD/YWCf/wQsraodqmpH4NuDjSaxvyuraifgEuBVrfx44OJW/njg+gn6kSRJmhbWX90DkLTGehJwalUtA36R5GJgN+C3wNVV9TOAJD8EzmttFtHNkg77SlXdA9yU5EfAtlPY/9OBXYCruxW+bAT8sm1bBnx1tEZV9e42I7ov8FLgJcBedGH5oIF6t05hf3cDZ7fH84BntMdPA17R+lsGLE3y8nH6+ZMkhwGHAczYbItxToMkSdLax6ApTW/XAweMsS3jtLtr4PE9A8/vYfnfKzWiXQF/ZPnVFGPNcgb4bFW9dZRtd7ZgN6qq+iFwQpKTgV8leVDrb+R4Jru/P1TVcNtljP+7c7x+Bsd4EnASwMyhbcYblyRJ0lrHpbPS9PZtYGaS4aWgJNktyVPplogemGRGki2ApwBXTbH/FyVZr123+Sjge8ASYE4r/wtg94H6f0iyQXt8AXBAkoe0cT0wySMm2mGS/dpNjgC2oQuGv6GbdX39QL0HjGi6Ivu7AHhNqz8jyWYrOm5JkqR1iUFTmsbaLN3+wDPax5tcDxwD/JTubrQLgQV0gfQtVfXzKe7ie8DFwH8Cr66qO4HLgZvpltm+H7h2oP5JwMIkX2w38XkHcF6ShcD5wBATezndNZrzgc8DB7fZz/cCD2g3N1rA8kt8WcH9HQ7snWQR3ZLax63EuCVJktYZuXc1mCT1J8lngLOr6ozVPZY13cyhbWrokA+tsv6XzN1vlfUtSZKmlyTzqmrXieo5oylJkiRJ6pU3A5K0SlTVoat7DJIkSVo9nNGUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK6/RlKTVbIetZnGNd4aVJEnrEGc0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6tX6q3sAkjTdLbplKbOPPmeV9b9k7n6rrG9JkqTROKMpSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCWtFkmWJZk/8HX0SvT1nT7HNsY+nrsyY5QkSZpO/HgTSavLHVU1p4+OqmqPPvoZS5L1q+os4KxVuR9JkqR1hTOaktYoSZYkeVeSa5MsSrJtK98iyfmt/MQkP07y4LbttvZ9ryQXJTkjyY1JvpgkbdsuSS5OMi/JN5MMtfKtk3yjlV86sL/PJPlAkguBf01yaJKPDmw7Psl3kvwoyQGtfL0kH09yfZKzk5w7vE2SJGk6MWhKWl02GrF09sCBbf9TVY8HTgCObGXvBL7dys8EHj5GvzsDRwDbAY8C9kyyAfAR4ICq2gX4FHBsq38S8IZWfiTw8YG+HgPsU1X/OMp+hoAnAc8G5rayFwCzgR2AvweeOPFpkCRJWve4dFbS6jLe0tmvte/z6MIbdKFuf4Cq+kaSW8doe1VV/QQgyXy64PcbYHvg/DbBOQP4WZJNgD2A01s5wMyBvk6vqmVj7OfrVXUPcEOSLQfGeHor/3mbDR1VksOAwwBmbLbFWNUkSZLWSgZNSWuiu9r3Zdz7eypj1B2r7WD7ANdX1XIzjEk2A34zTuC9fZL7yYjvE6qqk+hmU5k5tE1Ntp0kSdLawKWzktYWlwEvBkiyL/CAKbT9HrBFkie29hskeVxV/Ra4OcmLWnmS7LSSY3xhu1ZzS2CvlehLkiRprWXQlLS6jLxGc+4E9d8F7JvkWuCZwM+A301mR1V1N3AA3U19FgDz6ZbMAhwMvLKVXw88bwWOZdhXgZ8Ai4ETge8CS1eiP0mSpLVSqlyxJWnNl2QmsKyq/thmJk/o6+NR+pRkk6q6LcmDgKuAPavq5+O1mTm0TQ0d8qFVNqYlc/dbZX1LkqTpJcm8qtp1onpeoylpbfFw4CtJ1gPuBl61msczlrOTbA7cD3jPRCFTkiRpXWTQlLRWqKqb6D66ZI1WVXut7jFIkiStbl6jKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSr7wZkCStZjtsNYtr/AgSSZK0DnFGU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF6tv7oHIEnT3aJbljL76HNW6T6WzN1vlfYvSZI0yBlNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkhK/HtAAAIABJREFUSZJ6ZdCUJEmSJPXKoClJkiRJ6tWkgmaS28bZtleSsyfZz3eTzE/y30l+1R7PTzI7yblJNm9frx1oMzvJ4sn0P2Jfs5PckeS6JP+V5Kokh0y1nz4lOSbJLe2Yb0xyQpJewv54r9EY9V+RZHGS65PckOTIPsYxxTE8P8l2A88/k+TmJAuSfD/J55JstYJ9vzrJK/ob7Z/6PSLJxgPPz02yeY/9D5+D4Z+N7/TV9wT7PTTJw6bYZsKfzSS7J7kkyffae/6UJK8bOL67kyxqj+e2Ns9Mck37ub0xyfvH6X/wZ2r4q7fXY5T9LUny4CnUPzTJR1fVeCRJktZU9+nnaFbVX0H3xxewa1W9fmDzs9q22cBrgY/3sMsfVtXOrd9HAV9Lsl5VfXplO04yo6qWrUDTD1bV+1vAvAR4KnDhiL7Xr6o/ruwYx5LkmcARwL5V9dMkGwIvn0L75ca3Eufi+cDZwA0DZUdV1RlJ0sZ4YZLtq+ruqXRcVZ9YgfFMxhHAF4Dft/08axXs46iqOmMV9DueQ4HFwE/76jDJlsDpwEFVdUV7TV8IXFpVH2t1lgB7V9X/tOfbAx8F9quqG5OsDxw2wa4+WFVjhlFJkiTd9yY9m5bOcW0WbFGSAwc2b5bkzDYz9okVmaUbmCmYC2zdZiaOG1FnRhvD1UkWJvmHyfZfVT8C3gy8sfV1/ySfan1dl+R5rXzjJF9p/Z+WbhZ217bttiTvTvJd4IlJdklycZJ5Sb6ZZKjV2zrJN1r5pUm2HWVI9wM2BG5tbS5K8i9JLgYOT/Kctu/rknyr/dFOkk2SfLq9BguTvHDg/BzbZgOvHK4/hrcCR1bVT9u5ubOqTh4Yx/DxPrgFgeGZmdOT/AdwXrqZ7AuTfAlY1Oq8LN3M8fwkJyaZMXDelhtbkj2A5wLHtfpbj3i9qqo+CPwceGbrZ98kVyS5to1lk1Y+t733Fg7PfrWZriPb493atiuG38MDx/S19lrdlOTfBs7lCelm1a5P8q5W9kbgYXTh98JW9qcZriRvbj8fi5Mc0cpmp5uZO7n1dV6SjcZ5bUaV5Pgk/9we/3W6WcL10s2AfqK9z76f5Nmtzpg/K0ne0t4/C9q5OwDYFfhiey02Gue9vUtrdwXwugmG/Trgs1V1xcBrekZV/WKcNm8Bjq2qG1ubP1bVlP/pNMFr+zftPbQgyQWt7IFJvt7O1ZVJdmzlD2qv2XVJTgQy0M9Y7/e/ba/FxcCeUx27JEnSumAqgfAFwBxgJ2AfuoAw1LbtDvwjsAOwdau7oo6mm4mcU1VHjdj2SmBpVe0G7Aa8Kskjp9D3tcBw6Hs78O3W1950x3N/utnUW6tqR+A9wC4D7e8PLG4zs98FPgIcUFW7AJ8Cjm31TgLe0MqPZPnZ2TclmQ/8DPh+Vc0f2LZ5VT21qv4duAx4QpuR/TLdH+AA/9TOwQ5tjN8eGNuVVbUT3Uzpq8Y5D9sD8/5fe/cebldV3/v//THBCAJBK2pEayyCioIRAoJQBaUebVrFioJVC9Zf+eGleClaLLbFWxsLR6yKIlpBPSp440iBchEFRLkFCAnBolViW0u9VAx3lPR7/phjl8VmX8Pce2fvvF/Psx/WGnPMMcecc62wPmuMOdeYR2pkewGHVNVz2/M9gKOraqckTwEOAvauqiXAeuCVo/Wtqr4DnEE3erekqn4wyjavBp7cwtw7gf2raldgBfDWJA8HXgI8tR2P947QxsnA4VW1V+vXoCWt3zsDByV5XCs/uqqWArsAz0myS1V9iG7Eb7+q2m+wkSS7Aa8BngnsSffafEZbvANwQlU9Ffgl3ajeWIbC98okn2tlR7X+7Qd8CHhNVf13W7aYbmR8GXBiuhHqEd8r6UazDwCe2c7H37XR0xXAK9u5u4fRX9snA0e0YzmeDXmdbcg6bxk4XoOzA+53bpNsC3wCeGnb/5e1uu8Crmmvob8APtPK/xq4pL0PzwB+E2C013v7N/FddAHzd4D/mRo+XJLD2pcZK9bfsW6SuyxJkrRxm8zU2X2AL7Qpkj9p39bvDtwCXNFGDEnyhVZ3Kqb+PR/YpY3AACyk+xB/4wTXz8Dj5wMvyr3XJj6E7kPkPsDfA1TVdUlWDayzHvhKe/wkug/F5ycBmAfc1EbZngV8qZUDLBhoY2jq7GbAl5McXFWntmWnDdR7LHBa++D64IF93B84eKhSVd3cHv6KbhoqdB/Uf2ecY7Ehzq+qXww8v6Kqhvr1PLpQfmXb782Bn/bQt6GDuCfdh/Zvt/YfDFxK9/q7C/hkkrMGttOt3F2vt1ULtgCfB35voMoFVbWu1b0eeDzwb8DLkxxG9x5Z1LY9+FoYbh/g9Kq6vbX1VeC36cLJjQNfKFxFFwzHcr+ps1V1R5I/oQvqbxkWzL/YQuf3k/yQ7suU0d4r+wMnV9XQ1N/B8zlktNf2QrovQy5q9T5LG22eYaNNnR3p3D4MuHjodTuw//vQvgCoqm+0kcyFwLNpX5xV1VlJht5vo73enwlcWFU/a9s9DdhxpE5X1Ul0X0qxYNEOteG7L0mStPGZTNDMGMuGf0iaqg9NoRspPHcD138G8N2Btl5aVTfcZwMD6XAEdw1cixhgzfCRnSRbA79soxyjqqpfJzmH7oPsUNC8faDKh4EPVNUZSfYFjhnY7kjH99dVNVS+nrHP7Rq6D8nfGGHZPdw70v2QYctuH+N56KZJvuMB9m24ZwAXtPbPr6pXDK+QZA+6D/4HA28Enju4eJz27x54vB6Y30bJjwR2r6qbk5zC/Y/F/boxiW1MeupsszPwX3TTdweN9P4b8b2S5AUj1B9utNf2NhNYd9DQ6+xrG7DOtZNYZzT3O7eM/v4Z6fzVsP8Or3+/13uSA0apL0mStEmZzNTZi+mmn81r08+eDVzRlu3RpuU9iG462SUPoE+3AluNsuxc4HVtNJAkO7bpruNKd5Oh4+gC3FBbfzoULAemOV4CvLyV7UT34X4kNwDbJtmr1d0syVOr6hbgxiQva+VJ8vQR+hO6kc/RpowuBH7cHg/eLfc8ujA11M7DRll/LH8L/F2SR7c2FqS7/hBgLfdOFz5whHVHcwFwYJJHtjYfnuTx46wz6rlux+0IutHEc4DLgL2TPLEt36Kd/y2BhVV1Nt2Neu4T8NuI761J9mxFBzO+relC9Lp017oOjtqN1ueLgQNavx5KN533WxPY1oS0Y/lndMH7hUmeObD4Zemu19we+C261+Zo75XzgD9Ou3Num3o8fL9Ge23/ku6Y7NPqDU2NHs1HgEMG+5ruusZHj7HOscBfJNmx1X9QkreOs53JuJRuKvQTWvtD+38xbX/aFzs/b+/lwfIX0o2Iwuiv98uBfduI6GbcOzVXkiRpkzJu0Ex318e7gdPppg5eSzcS9vaq+s9W7VK6m/hcRzfF8/QN7VBV/Rfd9MjrMuxmQMAn6e5QenW6G7p8nLFHx7ZP+3kT4IvAh+veO86+B9gMWNXaek8r/yjdh+xVwJ+3fb7fBVTtLqgHAu9Pci2wki44QvfB9LWtfA3w4oFVh67RvK71fbQbnRxDN/32W8DPB8rfCzysHZ9r6a4vnZQWyk4Avp5kDd10zqHjeBxdQPkOMOGfcaiq6+muoTyvHbvz6ULiWE4F3tbO0dDNgI5t+/U9uqnZ+1XVr9pUxEOBL7T2L6ObIroVcGYruwh4ywjbeS1wUrob2IQRzuewfbkWuIbu3H0K+PbA4pOAf8p9rwWkqq4GTqH78uVy4JNVdc04+z+awWs0VyZZAPwD997A6bV0U4WHRllvoNv3f6K7FvUuRnmvVNU5dNN5V7TX4dDU8VPoru9cSTdVdrTX9muAE9qxvHOsnWg3/TkYOC7dz5t8l2468S1jrLOK7guDL7T61zH+6+gtw47X4jHa/xndXWy/2vZtaLr6McDS9jpazr1f7rwLeHaSq+mmI/9ra2fE13tV3dTauhT4Ot01xpIkSZuc3DujcZQK3WjcJ6pqj+np0sxKd+fIzarqrhZ+LgB2rEn+vIY2Hkm2rKrb2uOj6ALBm2a4W71o03rPHH5Np2aXBYt2qEWHfHBKt7F2+bIpbV+SJG0aklzVbpo5pjGvlUtyON3Pgby5r47NAlvQ/XzFZnSjX68zZM56y5K8g+71/iO6kVFJkiRJU2TMoFndj95v0A/fp/utyQXDil9dVas3pL0xtrMz3d0vB93dfoJk0qrqVrrfFJz1khzN/a8R+1JVvW+k+nNVVZ3Gfe/oO+OSnMD9f2Px7wemdk9IVR3aW6c2UJL/Bbx/WPGNVfWSHrfha1mSJGkWGXfqrCRpajl1VpIkzRYTnTo7mbvOSpIkSZI0LoOmJEmSJKlXY16jKUmaejtvt5AVTm2VJElziCOakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9Wr+THdAkjZ1q3+8jsVHnTXl21m7fNmUb0OSJAkc0ZQkSZIk9cygKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakuaUJOuTrExyXZIvJdniAbR1aJKPPIB1H7Oh25YkSZrNDJqS5po7q2pJVT0N+BVw+ODCdKbj375DAYOmJEnaJBk0Jc1l3wKemGRxku8m+ShwNfC4JK9IsrqNfL5/aIUkr0nyvSQXAXsPlJ+S5MCB57cNPH57a+vaJMtbvaXA59ro6ubTsbOSJEkbi/kz3QFJmgpJ5gMvBM5pRU8CXlNVr29TWt8P7AbcDJyX5ADgcuBdrXwd8E3gmnG280LgAOCZVXVHkodX1S+SvBE4sqpWTMHuSZIkbdQMmpLmms2TrGyPvwX8A90U1h9V1WWtfHfgwqr6GUCSzwHPbssGy08Ddhxne/sDJ1fVHQBV9YuJdDLJYcBhAPO23nYiq0iSJM0aBk1Jc82dVbVksCAJwO2DRWOsX6OU30O73CBdgw8eaGu0dUbfSNVJwEkACxbtMOn1JUmSNmZeoylpU3Q58Jwkj0gyD3gFcFEr3zfJbyTZDHjZwDpr6abUArwY2Kw9Pg/446G72yZ5eCu/FdhqSvdCkiRpI+WIpqRNTlXdlOQddNdgBji7qr4GkOQY4FLgJrobB81rq30C+FqSK4ALaCOkVXVOkiXAiiS/As4G/gI4BTgxyZ3AXlV15zTtniRJ0oxLlTO2JGkmLVi0Qy065INTvp21y5dN+TYkSdLcluSqqlo6Xj2nzkqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpV/68iSTNsJ23W8gK7wgrSZLmEEc0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6tX8me6AJG3qVv94HYuPOmtatrV2+bJp2Y4kSdq0OaIpSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpuwJI9OcmqSHyS5PsnZSXbcgHYOSLLTVPRxEn3YM8nlSVYm+W6SY2ayP5IkSZsyf0dT2kQlCXA68OmqOriVLQEeBXxvks0dAJwJXN9rJ8eQZF5VrR8o+jTw8qq6Nsk84EnT1RdJkiTdlyOa0qZrP+DXVXXiUEFVrayqbyXZN8mZQ+VJPpLk0PZ4eRv9XJXkuCTPAl4EHNtGE7dPsiTJZa3O6Uke1ta9MMnxSS5uo467J/lqku8nee/A9l6V5IrW3sdbcCTJbUneneRyYK9h+/NI4Ka2H+ur6vq2zjFJjhxo+7oki9vjP2p9vDbJZ1vZo1qfr21/zxqtT+3vlNbm6iRvaXWPGDhGp/ZwriRJkmYVRzSlTdfTgKsms0KShwMvAZ5cVZVkm6r6ZZIzgDOr6sut3irgT6vqoiTvBv4aeHNr5ldV9ewkbwK+BuwG/AL4QZLj6QLjQcDeVfXrJB8FXgl8BngocF1V/dUI3TseuCHJhcA5dCO1d42xL08Fjm7b+XnbN4APARdV1UtawN0yyVNG6dMaYLuqelprc5vWxlHAE6rq7oEySZKkTYYjmpIm4xbgLuCTSf4AuGN4hSQLgW2q6qJW9Gng2QNVzmj/XQ2sqaqbqupu4IfA44Dn0YXPK5OsbM9/q62zHvjKSB2rqncDS4HzgD+kC5tjeS7w5ar6eVv/FwPlH2tl66tq3Rh9+iHwW0k+nOQF7fgArAI+l+RVwD0jbTzJYUlWJFmx/o5143RVkiRpdjFoSpuuNXThaST3cN9/Hx4CUFX3AHvQhb0DGD/MjeTu9t//Hng89Hw+ELrRyCXt70lVdUyrc9ew6zLvo6p+UFUfowuCT0/yG6PtS9tOTbDPI/apqm4Gng5cCLwB+GSrvww4ge74XpXkfrNHquqkqlpaVUvnbbFwgt2QJEmaHQya0qbrG8CCJH8yVNCumXwO8CNgpyQL2gjl89ryLYGFVXU23VTYJW3VW4GtANoI4M1JfrstezUwNLo5ERcAByZ5ZNvmw5M8fryVkixrNzgC2IFu9POXwFpg11ZnV+AJA9t5eQujQ9OCh8pf18rmJdl6tD4leQTwoKr6CvCXwK5JHgQ8rqq+Cbwd2AbYchL7L0mSNOt5jaa0iWrXWL4E+GCSo+imxK4F3lxV/5bki3RTQL8PXNNW2wr4WpKH0I3yvaWVnwp8IskRwIHAIcCJSbagm176mkn06/ok7wTOa6Ht13SjhT8aZ9VXA8cnuYNuFPOVVbU+yVeAP2pTXq+k3VG3qtYkeR9wUZL1bR8PBd4EnJTktXRh9XVVdekofboTOLmVAbwDmAf8nxbQAxxfVb+c6P5LkiTNBama6MwxSdJUWLBoh1p0yAenZVtrly+blu1IkqS5KclVVbV0vHpOnZUkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSr/x5E0maYTtvt5AV3g1WkiTNIY5oSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1Kv5M90BSdrUrf7xOhYfdda0bGvt8mXTsh1JkrRpc0RTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqA5QUluG2PZvknOnGA7lydZmeRfk/ysPV6ZZHGSs5Ns0/5eP7DO4iTXbUCfFye5M8k1Sb6b5Iokh0y2nT4lOSbJj9s+/3OSjyXp5XU41jkapf4fJbkuyZok1yc5so9+TLIPByTZaeD5KUluTHJtku8l+UyS7Taw7cOT/FF/vf2fdt+cZIuB52cn2aantg9LctrA862T/CDJE0apP3S8ht5H3+mjH2P0b7KvsWNm4nUlSZI00wya06yqnllVS4C/Ak6rqiXtb21V/W5V/RLYBnj92C1N2A+q6hlV9RTgYOAtSV7TR8NJ5m3gqse3Y7ATsDPwnBHanv9A+jaeJC8E3gw8v6qeCuwKrJvE+vOHPd/QY3EA3XEY9LaqejrwJOAa4JtJHjzZhqvqxKr6zAb2ayxvBv4naA68bvvwCeCxSfZvz98NfKqqbhxjnbcNvI+e1VM/JEmS9AAYNCchnWPbKNjqJAcNLN46yeltZOzEDRmlS7I2ySOA5cD2bYTm2GF15rU+XJlkVZL/f6LtV9UPgbcCR7S2HprkU62ta5K8uJVvkeSLrf3T2ijs0rbstiTvTnI5sFeS3ZJclOSqJOcmWdTqbZ/knFb+rSRPHqFLDwYeAtzc1rkwyd8kuQh4U5Lfb9u+JsnXkzyq1dsyycntHKxK8tKB4/O+Nhp42VD9UbwDOLKq/qMdm7uq6hMD/Rja30ckWdseH5rkS0n+ETgv3Uj2N5N8Hljd6ryqjRyvTPLxoQDajtt9+pbkWcCLgGNb/e2Hna+qquOB/wRe2Np5fpJLk1zd+rJlK1/eXnurkhzXyv5nNC3J7m3ZpUOv4YF9+mo7V99P8ncDx/JjSVakG/F9Vys7AngMXfj9Zisbet2S5K3t/XFdkje3ssXpRtQ/0do6L8nmI52UqirgdcAH2zl4HnDsSHXH0vb9U+1c/rD1e2jZH7VjcW2Sz7ayxye5oJVfkOQ3W/kT2jG7Msl7hm3jbQPvw3cNlB+d5IYkX6f7skCSJGmTY9CcnD8AlgBPB/anCwiL2rI9gD+jG6HbvtXdUEfRjUQuqaq3DVv2WmBdVe0O7A78SUaZVjiKq4Gh0Hc08I3W1n50+/NQutHUm6tqF+A9wG4D6z8UuK6qnglcDnwYOLCqdgM+Bbyv1TsJ+NNWfiTw0YE23pJkJXAT8L2qWjmwbJuqek5V/W/gEmDPqnoGcCrw9lbnL9sx2Ln18RsDfbusjQZeDPzJGMfhacBVYx6pke0FHFJVz23P9wCOrqqdkjwFOAjYu43YrgdeOVrfquo7wBncOyL3g1G2eTXw5Bbm3gnsX1W7AiuAtyZ5OPAS4KnteLx3hDZOBg6vqr1avwYtaf3eGTgoyeNa+dFVtRTYBXhOkl2q6kPAfwD7VdV+g40k2Q14DfBMYE+61+Yz2uIdgBPa6PEvgZcyiqpaBZwLXAAcUVW/Gq1uMxTUVyb53ED5k4H/RXeO/jrJZkmeSve6f247F29qdT8CfKYdv88BH2rlfw98rL1H/nNgX5/f9mmPdvx2S/LsdgwOBp5B92/A7qN1Ot004RVJVqy/Y8KD6ZIkSbPClE5PnIP2Ab5QVeuBn7SRt92BW4Ar2oghSb7Q6n55CvrwfGCXJAe25wvpPvCONbVwUIa19aLcew3ZQ4DfpOv73wNU1XVJVg2ssx74Snv8JLrAdn4SgHnATW2U7VnAl1o5wIKBNo6vquOSbAZ8OcnBVXVqW3baQL3HAqe1MP/ggX3cn+7DPK2PN7eHvwKGrpW9CvidcY7Fhji/qn4x8PyKgWmdz6ML5Ve2/d4c+GkPfRs6iHvSTbP9dmv/wcCldK+/u4BPJjlrYDvdyt31k1u1YAvweeD3BqpcUFXrWt3rgccD/wa8PMlhdP9OLGrbHnwtDLcPcHpV3d7a+irw23Rh+saBLxSuAhaPs88nAC+sqm+OUw+6oD7Se+2sqrobuDvJT4FHAc8FvlxVPwcYOJd7ce+XQ58FhkZ29+beUPxZ4P3t8fPb3zXt+ZZ078Ot6I7BHQBJzhit01V1Et0XMixYtENNYD8lSZJmDYPm5GSMZcM/KE7VB8fQjRSeu4HrPwP47kBbL62qG+6zgYF0OIK7WtAeWn9NGyUbXH9r4JdtVG9UVfXrJOcAz6YbsQS4faDKh4EPVNUZSfYFjhnY7kjH99dt6iV0gXis1/caulD4jRGW3cO9o/0PGbbs9jGeB/h0Vb3jAfZtuGfQje6FLui+YniFJHvQBd2DgTfSBarBfo3l7oHH64H5bZT8SGD3qro5ySnc/1jcrxuT2MaIU2cH/Hf7eyDut1+M/toZrkZ5PCTA31bVx+9T2E0XNjRKkqRNnlNnJ+diuqmF85JsSxeQrmjL9mjXcz2IbhriJQ9gO7fSjYyM5FzgdW00kCQ7tumu40qyGDiOLsANtfWnQ8FyYJrjJcDLW9nQDXtGcgOwbZK9Wt3Nkjy1qm4BbkzyslaeJE8foT+hG/kcbcroQuDH7fHg3XLPowtTQ+08bJT1x/K3wN8leXRrY8HAdXxruXe68IEjrDuaC4ADkzyytfnwJI8fZ51Rz3U7bkfQjSaeA1wG7J3kiW35Fu38bwksrKqz6W7Uc5+A30Z8b02yZys6mPFtTRei16W71vWFE+jzxcABrV8PpZvO+60JbGs6XUA3Uvsb0J2jVv4d7j0ur+Te9++3h5UPORf449x7jex27bxfDLwkyeZJtgJ+f8r2RJIkaSNm0JyAdHcYvRs4nW7q4LV0I2Fvr6qh67YupbuJz3V0UzxP39DtVdV/0U2PvC7DbgYEfBK4Hrg63Q1dPs7Yo2Pbp/28CfBF4MNVdXJb9h5gM2BVa2voZicfpQuQq4A/b/t8v4vI2rVzBwLvT3ItsJIuOEL3ofy1rXwN8OKBVYeu0byu9X3w+s1Bx9BNv/0W8POB8vcCD2vH51q660snpYWyE4CvJ1lDN51z6DgeRxfmvwM8YhJtXk93DeV57didTxcSx3Iq8LZ2joZuBnRs26/v0U3N3q+qflVVPwMOBb7Q2r+M7jrErYAzW9lFwFtG2M5rgZOSXEo3GjfmRYFVdS3dtNA1dNfefntg8UnAP6XdDGhgnauBU+i+fLkc+GRVXcPUG7xGc2XGuENvVa2hu474onaMP9AWHQG8ph3DV3PvtZtvAt6Q5Eq6Lz6G2jmPbgrypUlW002T36odg9Po3gtfYeML2pIkSdMi987m02jaaNwnqmqPme7LdEh3p9TNququFn4uAHacwE1ZtJFKsmVV3dYeHwUsqqo3jbOapsmCRTvUokM+OC3bWrt82bRsR5IkzU1Jrmo3jByT12iOI8nhdKMdb57pvkyjLeh+vmIzutGv1xkyZ72QXsfvAAATRUlEQVRlSd5B957/Ed3IqCRJkjQlDJrjqKoTgRM3ZN10vzW5YFjxq6tq9QPu2H23szPdHTEH3d1+gmTSqupWYNxvKWaDJEcDLxtW/KWqet9I9eeqqjqN+97Rd8YlOYHurq6D/n5gavcG1ZUkSdLMM2hOoQ0NehuwndUMuwGMOi1QblKhcraoqjdMRV1JkiTNPG8GJEmSJEnqlSOakjTDdt5uISu8SY8kSZpDHNGUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpV/NnugOStKlb/eN1LD7qrGnZ1trly6ZlO5IkadPmiKYkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApadZJsj7JyoG/o1r5hUmWbkB7S5L87hjLlyb50Ab2dYP6JEmSNJv5O5qSZqM7q2pJj+0tAZYCZw9fkGR+Va0AVvS4PUmSpDnNEU1Jc1KS5ye5NMnVSb6UZMtWvnuS7yS5NskVSRYC7wYOaqOjByU5JslJSc4DPpNk3yRntvW3THJyktVJViV5aSv/WJIVSdYkedeM7bgkSdJGwKApaTbafNjU2YMGFyZ5BPBOYP+q2pVuNPKtSR4MnAa8qaqeDuwP3A78FXBaVS2pqtNaM7sBL66qPxy27b8E1lXVzlW1C/CNVn50VS0FdgGek2SXsXYgyWEtmK5Yf8e6DT0OkiRJGyWnzkqajcabOrsnsBPw7SQADwYuBZ4E3FRVVwJU1S0Arc5wZ1TVnSOU7w8cPPSkqm5uD1+e5DC6f1cXte2vGq2DVXUScBLAgkU71Bj7IkmSNOsYNCXNRQHOr6pX3KewG2WcaKi7fYy279NGkicARwK7V9XNSU4BHjKpHkuSJM0hTp2VNBddBuyd5IkASbZIsiPwz8BjkuzeyrdKMh+4Fdhqgm2fB7xx6EmShwFb0wXTdUkeBbywtz2RJEmahQyakmaj4ddoLh9cWFU/Aw4FvpBkFV3wfHJV/Qo4CPhwkmuB8+lGHr8J7DTS9Z4jeC/wsCTXtTb2q6prgWuANcCngG/3t6uSJEmzT6q8NEiSZtKCRTvUokM+OC3bWrt82bRsR5IkzU1Jrmo3QByTI5qSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6tX8me6AJG3qdt5uISu8G6wkSZpDHNGUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpV/NnugOStKlb/eN1LD7qrJnuxqy1dvmyme6CJEkaxhFNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClpo5dkfZKVA39H9dTuHydZnWRVkuuSvLiPdiVJkjZ1/ryJpNngzqpa0meDSR4LHA3sWlXrkmwJbPsA25xXVet76aAkSdIs5oimpFkrydokf5Pk0iQrkuya5NwkP0hyeKuzb5KLk5ye5PokJyZ5EPBI4FbgNoCquq2qbmzrPDHJ15Ncm+TqJNunc2wb+Vyd5KCB9r+Z5PPA6lb2qiRXtNHXjyeZNxPHR5IkaaYYNCXNBpsPmzp70MCyf6uqvYBvAacABwJ7Au8eqLMH8GfAzsD2wB8A1wI/AW5McnKS3x+o/znghKp6OvAs4Ka2zhLg6cD+wLFJFg20f3RV7ZTkKcBBwN5tFHY98Mq+DoQkSdJs4NRZSbPBWFNnz2j/XQ1sWVW3ArcmuSvJNm3ZFVX1Q4AkXwD2qaovJ3kBsDvwPOD4JLsB/xvYrqpOB6iqu9p6+wBfaFNjf5LkorbuLa39G9u2ngfsBlyZBGBz4KfDO53kMOAwgHlbP6AZu5IkSRsdg6ak2e7u9t//Hng89Hzo37gatk4BVFUBVwBXJDkfOBn4wCjbyRh9uH1YvU9X1TvG6nRVnQScBLBg0Q7D+ydJkjSrOXVW0qZgjyRPaNdmHgRckuQxSXYdqLME+FFV3QL8e5IDAJIsSLIFcDFwUJJ5SbYFnk0XUoe7ADgwySPb+g9P8vgp3DdJkqSNjiOakmaDzZOsHHh+TlVN5idOLgWW012jeTFwOvA44LgkjwHuAn4GHN7qvxr4eJJ3A78GXtbW2Yvu2s4C3l5V/5nkyYMbqqrrk7wTOK8F218DbwB+NJkdliRJms3SzRyTpLkpyb7AkVX1ezPdl9EsWLRDLTrkgzPdjVlr7fJlM90FSZI2GUmuqqql49Vz6qwkSZIkqVdOnZU0p1XVhcCFM9wNSZKkTYojmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqbAUnSDNt5u4Ws8Cc6JEnSHOKIpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF4ZNCVJkiRJvTJoSpIkSZJ6ZdCUJEmSJPXKoClJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb2aP9MdkKRN3eofr2PxUWfNdDfmlLXLl810FyRJ2qQ5oilJkiRJ6pVBU5IkSZLUK4OmJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JW0Ukjw6yalJfpDk+iRnJ9lxGre/TZLXDyvbsfXjX5J8N8kXkxyUZGX7uy3JDe3xZ9o6eyS5uJX/c5JPJtliuvZDkiRpY+DPm0iacUkCnA58uqoObmVLgEcB35vA+vOqav3A8/lVdc8ku7EN8Hrgo62NhwBnAW+tqn9sZfsBP6uqJe35hcCRVbWiPX8U8CXg4Kq6tO3XS4GtgDsm2R9JkqRZyxFNSRuD/YBfV9WJQwVVtRKYl+TMobIkH0lyaHu8NslfJbkEeFmSC5P8TZKLgDcl2TbJV5Jc2f72busdk+RTrf4PkxzRml8ObN9GJ48F/hC4dChktj59s6quG2M/3kAXli9t9auqvlxVP+nhGEmSJM0ajmhK2hg8DbhqA9a7q6r2AUhyOLBNVT2nPf88cHxVXZLkN4Fzgae09Z5MF263Am5I8jHgKOBpA6OVH9iAPj0N+PQG7IckSdKcYtCUNJudNsbz/YGdutmrAGydZKv2+Kyquhu4O8lP6aboTqskhwGHAczbetvp3rwkSdKUcuqspI3BGmC3Ecrv4b7/Tj1k2PLbx3j+IGCvqlrS/rarqlvbsrsH6q1n5C/dRuvTWCa8TlWdVFVLq2rpvC0WTnIzkiRJGzeDpqSNwTeABUn+ZKggye7APLpRyQVJFgLPm0Sb5wFvHGhvyTj1b6WbSjvk88CzkiwbaOMFSXYeo42PAIckeebAOq9K8uhJ9FuSJGnWM2hKmnFVVcBLgN9pP2+yBjgG+A/gi8Aq4HPANZNo9ghgaZJVSa4HDh+nD/8FfDvJdUmOrao7gd8D/jTJ91sbhwI/HaONnwAHA8e1nzf5LvDbwC2T6LckSdKsl+7znSRppixYtEMtOuSDM92NOWXt8mXjV5IkSZOW5KqqWjpePUc0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVfzZ7oDkrSp23m7hazw5zgkSdIc4oimJEmSJKlXBk1JkiRJUq8MmpIkSZKkXhk0JUmSJEm9MmhKkiRJknpl0JQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF7Nn+kOSNKmbvWP17H4qLNmuhuSJGmWWrt82Ux34X4c0ZQkSZIk9cqgKUmSJEnqlUFTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlbbSSrE+ycuDvqHHq75vkWVPcp7OTbDOV25AkSZrt/B1NSRuzO6tqySTq7wvcBnxnoiskmV9V90ygXoBU1e9Ooj+SJEmbJEc0Jc06SdYmeVeSq5OsTvLkJIuBw4G3tNHP306ybZKvJLmy/e3d1j8myUlJzgM+k+TQJF9Lck6SG5L8dau3OMl3k3wUuBp4XNv2I5I8NMlZSa5Ncl2Sg9o6uyW5KMlVSc5NsmhGDpIkSdIMckRT0sZs8yQrB57/bVWd1h7/vKp2TfJ64Miq+v+SnAjcVlXHAST5PHB8VV2S5DeBc4GntPV3A/apqjuTHArsATwNuAO4MslZwM+BJwGvqarXtzaH+vIC4D+qalkrX5hkM+DDwIur6mctfL4P+OO+D4wkSdLGzKApaWM21tTZr7b/XgX8wSh19gd2GgiHWyfZqj0+o6ruHKh7flX9F0CSrwL7AP8X+FFVXTZC26uB45K8Hzizqr6V5Gl0YfX8ts15wE0jdSzJYcBhAPO23naU7kuSJM1OBk1Js9Xd7b/rGf3fsgcBew0LlEOjkrcPq1ujPB9er1tY9b0kuwG/C/xtm4Z7OrCmqvYar/NVdRJwEsCCRTsM37YkSdKs5jWakuaSW4GtBp6fB7xx6EmSsW4s9DtJHp5kc+AA4NtjbSjJY4A7qur/AMcBuwI3ANsm2avV2SzJUzdoTyRJkmYxg6akjdnmw37eZPk49f8ReMnQzYCAI4ClSVYluZ7uZkGjuQT4LLAS+EpVrRhnWzsDV7RrSI8G3ltVvwIOBN6f5NrW1pT+3IokSdLGKFXO2JK0aWs3A1paVW8cr+5UWLBoh1p0yAdnYtOSJGkOWLt82bRtK8lVVbV0vHqOaEqSJEmSeuXNgCRt8qrqFOCUGe6GJEnSnOGIpiRJkiSpVwZNSZIkSVKvDJqSJEmSpF55jaYkzbCdt1vIimm8W5wkSdJUc0RTkiRJktQrg6YkSZIkqVcGTUmSJElSrwyakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpV6mqme6DJG3SktwK3DDT/dCUeQTw85nuhKaM53du8/zObZ7fDfP4qtp2vErzp6MnkqQx3VBVS2e6E5oaSVZ4fucuz+/c5vmd2zy/U8ups5IkSZKkXhk0JUmSJEm9MmhK0sw7aaY7oCnl+Z3bPL9zm+d3bvP8TiFvBiRJkiRJ6pUjmpIkSZKkXhk0JWkaJHlBkhuS/EuSo0ZYviDJaW355UkWT38v9UBM4Bw/O8nVSe5JcuBM9FEbbgLn961Jrk+yKskFSR4/E/3UhpnA+T08yeokK5NckmSnmeinNsx453eg3oFJKol3ou2BQVOSpliSecAJwAuBnYBXjPAh5bXAzVX1ROB44P3T20s9EBM8x/8KHAp8fnp7pwdqguf3GmBpVe0CfBn4u+ntpTbUBM/v56tq56paQnduPzDN3dQGmuD5JclWwBHA5dPbw7nLoClJU28P4F+q6odV9SvgVODFw+q8GPh0e/xl4HlJMo191AMz7jmuqrVVtQr475nooB6QiZzfb1bVHe3pZcBjp7mP2nATOb+3DDx9KOBNTmaPifw/GOA9dF8i3DWdnZvLDJqSNPW2A/5t4Pm/t7IR61TVPcA64DempXfqw0TOsWavyZ7f1wL/NKU9Up8mdH6TvCHJD+jCyBHT1Dc9cOOe3yTPAB5XVWdOZ8fmOoOmJE29kUYmh38bPpE62nh5/ua2CZ/fJK8ClgLHTmmP1KcJnd+qOqGqtgf+HHjnlPdKfRnz/CZ5EN0lK382bT3aRBg0JWnq/TvwuIHnjwX+Y7Q6SeYDC4FfTEvv1IeJnGPNXhM6v0n2B44GXlRVd09T3/TATfb9eypwwJT2SH0a7/xuBTwNuDDJWmBP4AxvCPTAGTQlaepdCeyQ5AlJHgwcDJwxrM4ZwCHt8YHAN8ofOp5NJnKONXuNe37b1LuP04XMn85AH7XhJnJ+dxh4ugz4/jT2Tw/MmOe3qtZV1SOqanFVLaa7xvpFVbViZro7dxg0JWmKtWsu3wicC3wX+GJVrUny7iQvatX+AfiNJP8CvBUY9fbr2vhM5Bwn2T3JvwMvAz6eZM3M9ViTMcH38LHAlsCX2k9g+EXDLDHB8/vGJGuSrKT7N/qQUZrTRmaC51dTIH5hLkmSJEnqkyOakiRJkqReGTQlSZIkSb0yaEqSJEmSemXQlCRJkiT1yqApSZIkSeqVQVOSJEmS1CuDpiRJkiSpVwZNSZIkSVKv/h8XPklcf3CrtAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 2.1 Select the top n features based on feature importance from XG Boost\n",
    "\n",
    "np.random.seed(10)\n",
    "\n",
    "# define the model\n",
    "model = xgb.XGBRegressor(random_state=42)\n",
    "# fit the model\n",
    "model.fit(predictors, Target)\n",
    "\n",
    "# get importance\n",
    "features = predictors\n",
    "importances = model.feature_importances_\n",
    "indices = np.argsort(importances)\n",
    "\n",
    "feat_importances = pd.Series(model.feature_importances_, index=predictors.columns)\n",
    "feat_importances.nlargest(30).plot(kind='barh')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "#### As top drivers are same as random forest, not considering this method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['LatestDegreeCGPA',\n",
       " 'YearsOfExperince',\n",
       " 'CurrentCTC',\n",
       " 'EmpScore',\n",
       " 'HighestDegree_Ordinal',\n",
       " 'Enterprise',\n",
       " 'MidSized',\n",
       " 'Finance',\n",
       " 'Computer Science',\n",
       " 'Electronics',\n",
       " 'Current_less_Grad_Year',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded']"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#3.Select the features identified by Lasso regression\n",
    "\n",
    "np.random.seed(10)\n",
    "\n",
    "estimator = LassoCV(cv=5, normalize = True)\n",
    "\n",
    "sfm = SelectFromModel(estimator, prefit=False, norm_order=1, max_features=None)\n",
    "\n",
    "sfm.fit(predictors, Target)\n",
    "\n",
    "feature_idx = sfm.get_support()\n",
    "Lasso_features = predictors.columns[feature_idx].tolist()\n",
    "Lasso_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Current_less_Grad_Year',\n",
       " 'Electronics',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded',\n",
       " 'EmpScore',\n",
       " 'Enterprise',\n",
       " 'LatestDegreeCGPA',\n",
       " 'MidSized',\n",
       " 'TotalLeavesTaken',\n",
       " 'Finance',\n",
       " 'GraduatingInstitute_Ordinal',\n",
       " 'YearsOfExperince',\n",
       " 'Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded',\n",
       " 'CurrentCTC',\n",
       " 'Computer Science',\n",
       " 'ExpectedCTC',\n",
       " 'HighestDegree_Ordinal']"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Combining the features from random forest and lasso regression\n",
    "combined_feature_list =  Lasso_features + rf_top_features\n",
    "combined_feature_list = list(set(combined_feature_list))\n",
    "combined_feature_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors2 = df4[['Current_less_Grad_Year',\n",
    "                     'Electronics',\n",
    "                     'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
    "                     'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded',\n",
    "                     'EmpScore',\n",
    "                     'Enterprise',\n",
    "                     'LatestDegreeCGPA',\n",
    "                     'MidSized',\n",
    "                     'TotalLeavesTaken',\n",
    "                     'Finance',\n",
    "                     'GraduatingInstitute_Ordinal',\n",
    "                     'YearsOfExperince',\n",
    "                     'Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded',\n",
    "                     'CurrentCTC',\n",
    "                     'Computer Science',\n",
    "                     'ExpectedCTC',\n",
    "                     'HighestDegree_Ordinal']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train and Test Model for predicting 'FitmentPercent'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean MAE: 10.258 (0.223)\n"
     ]
    }
   ],
   "source": [
    "# Base Model 1: Default XG Boost with all Features\n",
    "\n",
    "# define model\n",
    "model_xgb_d1 = xgb.XGBRegressor(random_state=42)\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_xgb_d1, predictors, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean MAE: 10.253 (0.188)\n"
     ]
    }
   ],
   "source": [
    "# Base Model 2: Default XG Boost with selected Features\n",
    "\n",
    "# define model\n",
    "model_xgb_d2 = xgb.XGBRegressor(random_state=42)\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_xgb_d2, predictors2, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean MAE: 9.848 (0.203)\n"
     ]
    }
   ],
   "source": [
    "# Base Model 3: Default CAT Boost with all Features\n",
    "\n",
    "# define model\n",
    "model_cat_d1 = CatBoostRegressor(random_state=42)\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_cat_d1, predictors, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean MAE: 9.849 (0.212)\n"
     ]
    }
   ],
   "source": [
    "# Base Model 4: Default CAT Boost with selected Features\n",
    "\n",
    "# define model\n",
    "model_cat_d2 = CatBoostRegressor(iterations=2,\n",
    "                          learning_rate=1,\n",
    "                          depth=2,random_state=42)\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_cat_d2, predictors2, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean MAE: 9.772 (0.198)\n"
     ]
    }
   ],
   "source": [
    "# Base Model 5: Default LGB with all Features\n",
    "\n",
    "# define model\n",
    "model_lgb_d1 = lgb.LGBMRegressor(seed=42)\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_lgb_d1, predictors, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean MAE: 9.763 (0.213)\n"
     ]
    }
   ],
   "source": [
    "# Base Model 6: Default LGB with selected Features\n",
    "\n",
    "# define model\n",
    "model_lgb_d2 = lgb.LGBMRegressor(seed=42)\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=3, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_lgb_d2, predictors2, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "## There is not much diference between the three models and so will try to do parameter tuning on all"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "random_state=1\n",
    "n_iter=100\n",
    "\n",
    "train_data, test_data, train_targets, test_targets = train_test_split(predictors2, Target, \n",
    "                                                                      test_size=0.10, shuffle=True,\n",
    "                                                                      random_state=random_state)\n",
    "\n",
    "num_folds=10\n",
    "kf = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████| 100/100 [1:55:02<00:00, 82.20s/trial, best loss: 9.689993922514576]\n",
      "{'colsample_bytree': 0.6891542432417372, 'eta': 0.07812229461246026, 'max_depth': 3, 'subsample': 0.7907139653437618}\n",
      "Mean MAE: 9.690 (0.203)\n"
     ]
    }
   ],
   "source": [
    "## XG Boost\n",
    "\n",
    "def gb_mae_cv(params, random_state=random_state, cv=kf, X=predictors2, y=Target):\n",
    "    # the function gets a set of variable parameters in \"param\"\n",
    "    params = {'colsample_bytree': params['colsample_bytree'],            \n",
    "              'max_depth': int(params['max_depth']), \n",
    "              'subsample': params['subsample'], \n",
    "              'eta': params['eta']}\n",
    "    \n",
    "    # we use this params to create a new XGB Regressor\n",
    "    model = xgb.XGBRegressor(random_state=random_state, **params)\n",
    "    \n",
    "    # and then conduct the cross validation with the same folds as before\n",
    "    score = -cross_val_score(model, X, y, cv=cv, scoring=\"neg_mean_absolute_error\", n_jobs=-1).mean()\n",
    "\n",
    "    return score\n",
    "\n",
    "# possible values of parameters\n",
    "space={'colsample_bytree': hp.uniform('colsample_bytree', .3, 1),\n",
    "       'max_depth' : hp.randint('max_depth', 2, 30),\n",
    "       'subsample' : hp.uniform('subsample', .3, 1),\n",
    "       'eta': hp.uniform('eta', .01, .7)\n",
    "      }\n",
    "\n",
    "# trials will contain logging information\n",
    "trials = Trials()\n",
    "\n",
    "best_xgb =fmin(fn=gb_mae_cv, # function to optimize\n",
    "          space=space, \n",
    "          algo=tpe.suggest, # optimization algorithm, hyperotp will select its parameters automatically\n",
    "          max_evals=n_iter, # maximum number of iterations\n",
    "          trials=trials, # logging\n",
    "          rstate=np.random.RandomState(random_state) # fixing random state for the reproducibility\n",
    "         )\n",
    "\n",
    "print(best_xgb)\n",
    "# computing the score on the best model\n",
    "model_xgb = xgb.XGBRegressor(random_state=random_state, colsample_bytree = best_xgb['colsample_bytree'],\n",
    "                      max_depth=int(best_xgb['max_depth']),subsample= best_xgb['subsample'],eta=best_xgb['eta'])\n",
    "\n",
    "\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_xgb, predictors2, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 100/100 [10:01:58<00:00, 420.27s/trial, best loss: 9.848853942262561]\n",
      "{'bagging_temperature': 0.09057019926571874, 'border_count': 114, 'l2_leaf_reg': 14, 'learning_rate': 0.19777348457492086, 'max_depth': 7, 'min_data_in_leaf': 11, 'n_estimators': 932}\n",
      "Mean MAE: 10.282 (0.207)\n"
     ]
    }
   ],
   "source": [
    "# CAT Boost\n",
    "\n",
    "def gb_mae_cv(params, random_state=random_state, cv=kf, X=predictors, y=Target):\n",
    "    # the function gets a set of variable parameters in \"param\"\n",
    "    params = {'n_estimators': int(params['n_estimators']),\n",
    "              'min_data_in_leaf': int(params['min_data_in_leaf']),             \n",
    "              'max_depth': int(params['max_depth']), \n",
    "              'learning_rate': params['learning_rate'],\n",
    "              'bagging_temperature': params['bagging_temperature'],\n",
    "              'border_count': params['border_count'],\n",
    "              'l2_leaf_reg': params['l2_leaf_reg']              \n",
    "             }\n",
    "    \n",
    "    # we use this params to create a new CATB Regressor\n",
    "    model = CatBoostRegressor(random_state=random_state)\n",
    "    \n",
    "    # and then conduct the cross validation with the same folds as before\n",
    "    score = -cross_val_score(model, X, y, cv=cv, scoring=\"neg_mean_absolute_error\", n_jobs=-1).mean()\n",
    "\n",
    "    return score\n",
    "\n",
    "# possible values of parameters\n",
    "space={'n_estimators': hp.randint('n_estimators', 800, 2000),\n",
    "       'min_data_in_leaf' : hp.randint('min_data_in_leaf', 1, 20),\n",
    "       'max_depth' : hp.randint('max_depth', 5, 10),\n",
    "       'learning_rate': hp.uniform('learning_rate', .01, .5),\n",
    "       'bagging_temperature' : hp.uniform('bagging_temperature', 0, 1),\n",
    "       'border_count' : hp.randint('border_count', 1, 254),\n",
    "       'l2_leaf_reg' : hp.randint('l2_leaf_reg', 2, 30),\n",
    "       }\n",
    "\n",
    "# trials will contain logging information\n",
    "trials = Trials()\n",
    "\n",
    "best_cat =fmin(fn=gb_mae_cv, # function to optimize\n",
    "          space=space, \n",
    "          algo=tpe.suggest, # optimization algorithm, hyperotp will select its parameters automatically\n",
    "          max_evals=n_iter, # maximum number of iterations\n",
    "          trials=trials, # logging\n",
    "          rstate=np.random.RandomState(random_state) # fixing random state for the reproducibility\n",
    "         )\n",
    "\n",
    "print(best_cat)\n",
    "# computing the score on the test set\n",
    "model_cat = CatBoostRegressor(random_state=random_state,n_estimators=int(best_cat['n_estimators']),\n",
    "                          min_data_in_leaf=int(best_cat['min_data_in_leaf']),bagging_temperature = best_cat['bagging_temperature'],\n",
    "                          border_count=int(best_cat['border_count']),l2_leaf_reg=int(best_cat['l2_leaf_reg']),\n",
    "                          max_depth=int(best_cat['max_depth']),learning_rate=best_cat['learning_rate'])\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_cat, predictors, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reducing the number of iterations as it is not helping us to improve accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "random_state=1\n",
    "n_iter=50\n",
    "\n",
    "train_data, test_data, train_targets, test_targets = train_test_split(predictors2, Target, \n",
    "                                                                      test_size=0.10, shuffle=True,\n",
    "                                                                      random_state=random_state)\n",
    "\n",
    "num_folds=10\n",
    "kf = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████| 50/50 [2:10:43<00:00, 59.53s/trial, best loss: 9.734579263937245]\n",
      "{'learning_rate': 0.3486272160433078, 'max_depth': 3, 'min_data_in_leaf': 26, 'n_estimators': 184, 'num_leaves': 37}\n",
      "Mean MAE: 9.735 (0.203)\n"
     ]
    }
   ],
   "source": [
    "# Light GBM\n",
    "\n",
    "def gb_mae_cv(params, random_state=random_state, cv=kf, X=predictors2, y=Target):\n",
    "    # the function gets a set of variable parameters in \"param\"\n",
    "    params = {'n_estimators': int(params['n_estimators']),\n",
    "              'min_data_in_leaf': int(params['min_data_in_leaf']),             \n",
    "              'max_depth': int(params['max_depth']), \n",
    "              'num_leaves': int(params['num_leaves']), \n",
    "              'learning_rate': params['learning_rate']}\n",
    "    \n",
    "    # we use this params to create a new LGBM Regressor\n",
    "    model = lgb.LGBMRegressor(random_state=random_state,boosting = 'dart', **params)\n",
    "    \n",
    "    # and then conduct the cross validation with the same folds as before\n",
    "    score = -cross_val_score(model, X, y, cv=cv, scoring=\"neg_mean_absolute_error\", n_jobs=-1).mean()\n",
    "\n",
    "    return score\n",
    "\n",
    "# possible values of parameters\n",
    "space={'n_estimators': hp.randint('n_estimators', 90, 500),\n",
    "       'min_data_in_leaf' : hp.randint('min_data_in_leaf', 10, 200),\n",
    "       'max_depth' : hp.randint('max_depth', -1, 50),\n",
    "       'num_leaves' : hp.randint('num_leaves', 25, 100),\n",
    "       'learning_rate': hp.uniform('learning_rate', .05, .5)\n",
    "      }\n",
    "\n",
    "# trials will contain logging information\n",
    "trials = Trials()\n",
    "\n",
    "best_lgb=fmin(fn=gb_mae_cv, # function to optimize\n",
    "          space=space, \n",
    "          algo=tpe.suggest, # optimization algorithm, hyperotp will select its parameters automatically\n",
    "          max_evals=n_iter, # maximum number of iterations\n",
    "          trials=trials, # logging\n",
    "          rstate=np.random.RandomState(random_state) # fixing random state for the reproducibility\n",
    "         )\n",
    "\n",
    "print(best_lgb)\n",
    "# computing the score on the test set\n",
    "model_lgb = lgb.LGBMRegressor(random_state=random_state,boosting = 'dart', n_estimators=int(best_lgb['n_estimators']),\n",
    "                          min_data_in_leaf=int(best_lgb['min_data_in_leaf']),\n",
    "                          max_depth=int(best_lgb['max_depth']),num_leaves=int(best_lgb['num_leaves']),learning_rate=best_lgb['learning_rate'])\n",
    "\n",
    "\n",
    "\n",
    "# define model evaluation method\n",
    "cv = RepeatedKFold(n_splits=10, n_repeats=5, random_state=1)\n",
    "# evaluate model\n",
    "scores = cross_val_score(model_lgb, predictors2, Target, scoring='neg_mean_absolute_error', cv=cv, n_jobs=-1)\n",
    "# force scores to be positive\n",
    "scores = abs(scores)\n",
    "print('Mean MAE: %.3f (%.3f)' % (scores.mean(), scores.std()) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train and Test Model for predicting 'Bias Influential Factor'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "Target2 = df4['BiasInfluentialFactor']\n",
    "# encode string class values as integers\n",
    "label_encoder = LabelEncoder()\n",
    "label_encoder = label_encoder.fit(Target2)\n",
    "label_encoded_y = label_encoder.transform(Target2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,\n",
       "              importance_type='gain', interaction_constraints='',\n",
       "              learning_rate=0.300000012, max_delta_step=0, max_depth=6,\n",
       "              min_child_weight=1, missing=nan, monotone_constraints='()',\n",
       "              n_estimators=100, n_jobs=0, num_parallel_tree=1,\n",
       "              objective='multi:softprob', random_state=42, reg_alpha=0,\n",
       "              reg_lambda=1, scale_pos_weight=None, subsample=1,\n",
       "              tree_method='exact', validate_parameters=1, verbosity=None)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Fit XG Boost Model with all non- discriminatory fields\n",
    "\n",
    "# define model\n",
    "model_xgb_d1_class = xgb.XGBClassifier(random_state =42, n_estimators=100)\n",
    "# define model evaluation method\n",
    "model_xgb_d1_class.fit(predictors, label_encoded_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Make Predictions on test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>YearsOfExperince</th>\n",
       "      <th>GraduationYear</th>\n",
       "      <th>CurrentCTC</th>\n",
       "      <th>ExpectedCTC</th>\n",
       "      <th>MartialStatus</th>\n",
       "      <th>EmpScore</th>\n",
       "      <th>CurrentDesignation</th>\n",
       "      <th>CurrentCompanyType</th>\n",
       "      <th>DepartmentInCompany</th>\n",
       "      <th>TotalLeavesTaken</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5664</td>\n",
       "      <td>Theo</td>\n",
       "      <td>Native</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>2016</td>\n",
       "      <td>27</td>\n",
       "      <td>34</td>\n",
       "      <td>Single</td>\n",
       "      <td>4</td>\n",
       "      <td>SDE</td>\n",
       "      <td>Enterprise</td>\n",
       "      <td>Finance</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23568</td>\n",
       "      <td>Vernie</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>9</td>\n",
       "      <td>12</td>\n",
       "      <td>2009</td>\n",
       "      <td>29</td>\n",
       "      <td>33</td>\n",
       "      <td>Married</td>\n",
       "      <td>4</td>\n",
       "      <td>EM</td>\n",
       "      <td>Enterprise</td>\n",
       "      <td>Engineering</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21490</td>\n",
       "      <td>Wendell</td>\n",
       "      <td>English</td>\n",
       "      <td>28</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>M.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>9</td>\n",
       "      <td>2</td>\n",
       "      <td>2019</td>\n",
       "      <td>18</td>\n",
       "      <td>25</td>\n",
       "      <td>Married</td>\n",
       "      <td>3</td>\n",
       "      <td>SDE</td>\n",
       "      <td>Startup</td>\n",
       "      <td>Engineering</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8363</td>\n",
       "      <td>West</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>37</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>MS</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>8</td>\n",
       "      <td>11</td>\n",
       "      <td>2010</td>\n",
       "      <td>7</td>\n",
       "      <td>12</td>\n",
       "      <td>Single</td>\n",
       "      <td>5</td>\n",
       "      <td>DS</td>\n",
       "      <td>Enterprise</td>\n",
       "      <td>Product</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6165</td>\n",
       "      <td>Alford</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>7</td>\n",
       "      <td>6</td>\n",
       "      <td>2015</td>\n",
       "      <td>22</td>\n",
       "      <td>32</td>\n",
       "      <td>Married</td>\n",
       "      <td>4</td>\n",
       "      <td>SDE</td>\n",
       "      <td>Startup</td>\n",
       "      <td>Engineering</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0   5664     Theo                  Native   35    Male   \n",
       "1  23568   Vernie                 English   35    Male   \n",
       "2  21490  Wendell                 English   28    Male   \n",
       "3   8363     West                   Hindi   37    Male   \n",
       "4   6165   Alford                   Hindi   36  Female   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289           PhD   Information Technology   \n",
       "1                 JR85289        B.Tech   Information Technology   \n",
       "2                 JR87525        M.Tech  Artificial Intelligence   \n",
       "3                 JR87525            MS         Computer Science   \n",
       "4                 JR85289           PhD  Artificial Intelligence   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  YearsOfExperince  GraduationYear  \\\n",
       "0              Tier 2                 5                 5            2016   \n",
       "1              Tier 3                 9                12            2009   \n",
       "2              Tier 2                 9                 2            2019   \n",
       "3              Tier 2                 8                11            2010   \n",
       "4              Tier 2                 7                 6            2015   \n",
       "\n",
       "   CurrentCTC  ExpectedCTC MartialStatus  EmpScore CurrentDesignation  \\\n",
       "0          27           34        Single         4                SDE   \n",
       "1          29           33       Married         4                 EM   \n",
       "2          18           25       Married         3                SDE   \n",
       "3           7           12        Single         5                 DS   \n",
       "4          22           32       Married         4                SDE   \n",
       "\n",
       "  CurrentCompanyType DepartmentInCompany  TotalLeavesTaken  \n",
       "0         Enterprise             Finance                18  \n",
       "1         Enterprise         Engineering                 8  \n",
       "2            Startup         Engineering                18  \n",
       "3         Enterprise             Product                17  \n",
       "4            Startup         Engineering                 9  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Input the dataset\n",
    "df_test = pd.read_csv('Test.csv')\n",
    "\n",
    "df_test.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 8745 entries, 0 to 8744\n",
      "Data columns (total 20 columns):\n",
      " #   Column                   Non-Null Count  Dtype \n",
      "---  ------                   --------------  ----- \n",
      " 0   EmpID                    8745 non-null   int64 \n",
      " 1   EmpName                  8745 non-null   object\n",
      " 2   LanguageOfCommunication  8745 non-null   object\n",
      " 3   Age                      8745 non-null   int64 \n",
      " 4   Gender                   8745 non-null   object\n",
      " 5   JobProfileIDApplyingFor  8745 non-null   object\n",
      " 6   HighestDegree            8745 non-null   object\n",
      " 7   DegreeBranch             8745 non-null   object\n",
      " 8   GraduatingInstitute      8745 non-null   object\n",
      " 9   LatestDegreeCGPA         8745 non-null   int64 \n",
      " 10  YearsOfExperince         8745 non-null   int64 \n",
      " 11  GraduationYear           8745 non-null   int64 \n",
      " 12  CurrentCTC               8745 non-null   int64 \n",
      " 13  ExpectedCTC              8745 non-null   int64 \n",
      " 14  MartialStatus            8745 non-null   object\n",
      " 15  EmpScore                 8745 non-null   int64 \n",
      " 16  CurrentDesignation       8745 non-null   object\n",
      " 17  CurrentCompanyType       8745 non-null   object\n",
      " 18  DepartmentInCompany      8745 non-null   object\n",
      " 19  TotalLeavesTaken         8745 non-null   int64 \n",
      "dtypes: int64(9), object(11)\n",
      "memory usage: 1.3+ MB\n"
     ]
    }
   ],
   "source": [
    "#Check number of rows and columns, type of each columns\n",
    "df_test.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 8745 entries, 0 to 8744\n",
      "Data columns (total 20 columns):\n",
      " #   Column                   Non-Null Count  Dtype \n",
      "---  ------                   --------------  ----- \n",
      " 0   EmpID                    8745 non-null   object\n",
      " 1   EmpName                  8745 non-null   object\n",
      " 2   LanguageOfCommunication  8745 non-null   object\n",
      " 3   Age                      8745 non-null   int64 \n",
      " 4   Gender                   8745 non-null   object\n",
      " 5   JobProfileIDApplyingFor  8745 non-null   object\n",
      " 6   HighestDegree            8745 non-null   object\n",
      " 7   DegreeBranch             8745 non-null   object\n",
      " 8   GraduatingInstitute      8745 non-null   object\n",
      " 9   LatestDegreeCGPA         8745 non-null   int64 \n",
      " 10  YearsOfExperince         8745 non-null   int64 \n",
      " 11  GraduationYear           8745 non-null   int64 \n",
      " 12  CurrentCTC               8745 non-null   int64 \n",
      " 13  ExpectedCTC              8745 non-null   int64 \n",
      " 14  MartialStatus            8745 non-null   object\n",
      " 15  EmpScore                 8745 non-null   int64 \n",
      " 16  CurrentDesignation       8745 non-null   object\n",
      " 17  CurrentCompanyType       8745 non-null   object\n",
      " 18  DepartmentInCompany      8745 non-null   object\n",
      " 19  TotalLeavesTaken         8745 non-null   int64 \n",
      "dtypes: int64(8), object(12)\n",
      "memory usage: 1.3+ MB\n"
     ]
    }
   ],
   "source": [
    "# Ensure Correct Column Types\n",
    "String_cols = ['EmpID', 'EmpName', 'LanguageOfCommunication', 'Gender',\n",
    "               'JobProfileIDApplyingFor', 'HighestDegree', 'DegreeBranch',\n",
    "               'GraduatingInstitute',  'MartialStatus',\n",
    "               'CurrentDesignation', 'CurrentCompanyType',\n",
    "               'DepartmentInCompany']\n",
    "\n",
    "Numeric_cols = ['Age', 'YearsOfExperince',\n",
    "               'GraduationYear', 'CurrentCTC', 'ExpectedCTC',\n",
    "               'EmpScore', 'TotalLeavesTaken']\n",
    "\n",
    "df_test[String_cols] = df_test[String_cols].astype(str)\n",
    "df_test[Numeric_cols] = df_test[Numeric_cols].apply(pd.to_numeric, errors='coerce')\n",
    "df_test.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Ordinal Encoding for string variables with logical order\n",
    "\n",
    "# HighestDegree\n",
    "\n",
    "# create a list of our conditions\n",
    "conditions = [\n",
    "    (df_test['HighestDegree'] == \"B.Tech\") | (df_test['HighestDegree'] == \"BCA\") ,\n",
    "    (df_test['HighestDegree'] == \"M.Tech\") | (df_test['HighestDegree'] == \"MCA\") | (df_test['HighestDegree'] == \"MS\"),\n",
    "    (df_test['HighestDegree'] == \"Dual M.Tech\") | (df_test['HighestDegree'] == \"Dual MBA\") | (df_test['HighestDegree'] == \"PhD\"),\n",
    "    ]\n",
    "\n",
    "# create a list of the values we want to assign for each condition\n",
    "values = [1,2,3]\n",
    "\n",
    "# create a new column and use np.select to assign values to it using our lists as arguments\n",
    "df_test['HighestDegree_Ordinal'] = np.select(conditions, values)\n",
    "\n",
    "\n",
    "# GraduatingInstitute\n",
    "\n",
    "# create a list of our conditions\n",
    "conditions = [\n",
    "    (df_test['GraduatingInstitute'] == \"Tier 3\"),\n",
    "    (df_test['GraduatingInstitute'] == \"Tier 2\"),\n",
    "    (df_test['GraduatingInstitute'] == \"Tier 1\")]\n",
    "\n",
    "# create a list of the values we want to assign for each condition\n",
    "values = [1,2,3]\n",
    "\n",
    "# create a new column and use np.select to assign values to it using our lists as arguments\n",
    "df_test['GraduatingInstitute_Ordinal'] = np.select(conditions, values)\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8745\n",
      "8745\n",
      "8745\n",
      "8745\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>Design</th>\n",
       "      <th>Engineering</th>\n",
       "      <th>Finance</th>\n",
       "      <th>Product</th>\n",
       "      <th>Artificial Intelligence</th>\n",
       "      <th>Computer Science</th>\n",
       "      <th>Electrical</th>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <th>Electronics</th>\n",
       "      <th>Information Technology</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5664</td>\n",
       "      <td>Theo</td>\n",
       "      <td>Native</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23568</td>\n",
       "      <td>Vernie</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21490</td>\n",
       "      <td>Wendell</td>\n",
       "      <td>English</td>\n",
       "      <td>28</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>M.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8363</td>\n",
       "      <td>West</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>37</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>MS</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6165</td>\n",
       "      <td>Alford</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0   5664     Theo                  Native   35    Male   \n",
       "1  23568   Vernie                 English   35    Male   \n",
       "2  21490  Wendell                 English   28    Male   \n",
       "3   8363     West                   Hindi   37    Male   \n",
       "4   6165   Alford                   Hindi   36  Female   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289           PhD   Information Technology   \n",
       "1                 JR85289        B.Tech   Information Technology   \n",
       "2                 JR87525        M.Tech  Artificial Intelligence   \n",
       "3                 JR87525            MS         Computer Science   \n",
       "4                 JR85289           PhD  Artificial Intelligence   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  Design  Engineering  Finance  \\\n",
       "0              Tier 2                 5  ...       0            0        1   \n",
       "1              Tier 3                 9  ...       0            1        0   \n",
       "2              Tier 2                 9  ...       0            1        0   \n",
       "3              Tier 2                 8  ...       0            0        0   \n",
       "4              Tier 2                 7  ...       0            1        0   \n",
       "\n",
       "   Product Artificial Intelligence  Computer Science Electrical  \\\n",
       "0        0                       0                 0          0   \n",
       "1        0                       0                 0          0   \n",
       "2        0                       1                 0          0   \n",
       "3        1                       0                 1          0   \n",
       "4        0                       1                 0          0   \n",
       "\n",
       "  Electrical and Electronics Electronics  Information Technology  \n",
       "0                          0           0                       1  \n",
       "1                          0           0                       1  \n",
       "2                          0           0                       0  \n",
       "3                          0           0                       0  \n",
       "4                          0           0                       0  \n",
       "\n",
       "[5 rows x 36 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# One Hot Encoding\n",
    "\n",
    "# CurrentCompanyType has only 3 types of values\n",
    "\n",
    "# Get one hot encoding \n",
    "one_hot = pd.get_dummies(df_test['CurrentCompanyType'])\n",
    "\n",
    "# Join the encoded df_test\n",
    "df_test2 = df_test.join(one_hot)\n",
    "print(len(df_test))\n",
    "print(len(df_test2))\n",
    "\n",
    "# DepartmentInCompany has only 5 types of values\n",
    "\n",
    "# Get one hot encoding \n",
    "one_hot2 = pd.get_dummies(df_test['DepartmentInCompany'])\n",
    "\n",
    "# Join the encoded df_test\n",
    "df_test3 = df_test2.join(one_hot2)\n",
    "print(len(df_test3))\n",
    "\n",
    "# DepartmentInCompany has only 5 types of values\n",
    "\n",
    "# Get one hot encoding \n",
    "one_hot3 = pd.get_dummies(df_test['DegreeBranch'])\n",
    "\n",
    "# Join the encoded df_test\n",
    "df_test4 = df_test3.join(one_hot3)\n",
    "print(len(df_test4))\n",
    "\n",
    "df_test4.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>Engineering</th>\n",
       "      <th>Finance</th>\n",
       "      <th>Product</th>\n",
       "      <th>Artificial Intelligence</th>\n",
       "      <th>Computer Science</th>\n",
       "      <th>Electrical</th>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <th>Electronics</th>\n",
       "      <th>Information Technology</th>\n",
       "      <th>Current_less_Grad_Year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5664</td>\n",
       "      <td>Theo</td>\n",
       "      <td>Native</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23568</td>\n",
       "      <td>Vernie</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21490</td>\n",
       "      <td>Wendell</td>\n",
       "      <td>English</td>\n",
       "      <td>28</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>M.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8363</td>\n",
       "      <td>West</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>37</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>MS</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6165</td>\n",
       "      <td>Alford</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 37 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0   5664     Theo                  Native   35    Male   \n",
       "1  23568   Vernie                 English   35    Male   \n",
       "2  21490  Wendell                 English   28    Male   \n",
       "3   8363     West                   Hindi   37    Male   \n",
       "4   6165   Alford                   Hindi   36  Female   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289           PhD   Information Technology   \n",
       "1                 JR85289        B.Tech   Information Technology   \n",
       "2                 JR87525        M.Tech  Artificial Intelligence   \n",
       "3                 JR87525            MS         Computer Science   \n",
       "4                 JR85289           PhD  Artificial Intelligence   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  Engineering  Finance  Product  \\\n",
       "0              Tier 2                 5  ...            0        1        0   \n",
       "1              Tier 3                 9  ...            1        0        0   \n",
       "2              Tier 2                 9  ...            1        0        0   \n",
       "3              Tier 2                 8  ...            0        0        1   \n",
       "4              Tier 2                 7  ...            1        0        0   \n",
       "\n",
       "   Artificial Intelligence Computer Science  Electrical  \\\n",
       "0                        0                0           0   \n",
       "1                        0                0           0   \n",
       "2                        1                0           0   \n",
       "3                        0                1           0   \n",
       "4                        1                0           0   \n",
       "\n",
       "  Electrical and Electronics Electronics Information Technology  \\\n",
       "0                          0           0                      1   \n",
       "1                          0           0                      1   \n",
       "2                          0           0                      0   \n",
       "3                          0           0                      0   \n",
       "4                          0           0                      0   \n",
       "\n",
       "   Current_less_Grad_Year  \n",
       "0                       5  \n",
       "1                      12  \n",
       "2                       2  \n",
       "3                      11  \n",
       "4                       6  \n",
       "\n",
       "[5 rows x 37 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Current Year - Graduation Year\n",
    "\n",
    "df_test4['Current_less_Grad_Year'] = 2021 - df_test4['GraduationYear']\n",
    "df_test4.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ibaruah001\\AppData\\Roaming\\Python\\Python37\\site-packages\\category_encoders\\utils.py:21: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead\n",
      "  elif pd.api.types.is_categorical(cols):\n",
      "C:\\Users\\ibaruah001\\AppData\\Roaming\\Python\\Python37\\site-packages\\category_encoders\\utils.py:21: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead\n",
      "  elif pd.api.types.is_categorical(cols):\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>Artificial Intelligence</th>\n",
       "      <th>Computer Science</th>\n",
       "      <th>Electrical</th>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <th>Electronics</th>\n",
       "      <th>Information Technology</th>\n",
       "      <th>Current_less_Grad_Year</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5664</td>\n",
       "      <td>Theo</td>\n",
       "      <td>Native</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>JR85289SDEFinance</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>26.104478</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23568</td>\n",
       "      <td>Vernie</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>12</td>\n",
       "      <td>JR85289EMEngineering</td>\n",
       "      <td>16.707317</td>\n",
       "      <td>23.837398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21490</td>\n",
       "      <td>Wendell</td>\n",
       "      <td>English</td>\n",
       "      <td>28</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>M.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>JR87525SDEEngineering</td>\n",
       "      <td>17.331230</td>\n",
       "      <td>24.249211</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8363</td>\n",
       "      <td>West</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>37</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>MS</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>JR87525DSProduct</td>\n",
       "      <td>17.851852</td>\n",
       "      <td>25.518519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6165</td>\n",
       "      <td>Alford</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>JR85289SDEEngineering</td>\n",
       "      <td>18.176349</td>\n",
       "      <td>25.188797</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0   5664     Theo                  Native   35    Male   \n",
       "1  23568   Vernie                 English   35    Male   \n",
       "2  21490  Wendell                 English   28    Male   \n",
       "3   8363     West                   Hindi   37    Male   \n",
       "4   6165   Alford                   Hindi   36  Female   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289           PhD   Information Technology   \n",
       "1                 JR85289        B.Tech   Information Technology   \n",
       "2                 JR87525        M.Tech  Artificial Intelligence   \n",
       "3                 JR87525            MS         Computer Science   \n",
       "4                 JR85289           PhD  Artificial Intelligence   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  Artificial Intelligence  \\\n",
       "0              Tier 2                 5  ...                        0   \n",
       "1              Tier 3                 9  ...                        0   \n",
       "2              Tier 2                 9  ...                        1   \n",
       "3              Tier 2                 8  ...                        0   \n",
       "4              Tier 2                 7  ...                        1   \n",
       "\n",
       "   Computer Science  Electrical  Electrical and Electronics Electronics  \\\n",
       "0                 0           0                           0           0   \n",
       "1                 0           0                           0           0   \n",
       "2                 0           0                           0           0   \n",
       "3                 1           0                           0           0   \n",
       "4                 0           0                           0           0   \n",
       "\n",
       "   Information Technology Current_less_Grad_Year  \\\n",
       "0                       1                      5   \n",
       "1                       1                     12   \n",
       "2                       0                      2   \n",
       "3                       0                     11   \n",
       "4                       0                      6   \n",
       "\n",
       "  Job_Title_DegreeBrach_CurrentDesignation  \\\n",
       "0                        JR85289SDEFinance   \n",
       "1                     JR85289EMEngineering   \n",
       "2                    JR87525SDEEngineering   \n",
       "3                         JR87525DSProduct   \n",
       "4                    JR85289SDEEngineering   \n",
       "\n",
       "  Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded  \\\n",
       "0                                          19.000000             \n",
       "1                                          16.707317             \n",
       "2                                          17.331230             \n",
       "3                                          17.851852             \n",
       "4                                          18.176349             \n",
       "\n",
       "   Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded  \n",
       "0                                          26.104478              \n",
       "1                                          23.837398              \n",
       "2                                          24.249211              \n",
       "3                                          25.518519              \n",
       "4                                          25.188797              \n",
       "\n",
       "[5 rows x 40 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Target Encoding\n",
    "\n",
    "\n",
    "\n",
    "df_test4['Job_Title_DegreeBrach_CurrentDesignation'] = df_test4['JobProfileIDApplyingFor']+df_test4['CurrentDesignation'] + df_test4['DepartmentInCompany'] \n",
    "encoder = TargetEncoder()\n",
    "df_test4['Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded'] = encoder.fit_transform(df_test4['Job_Title_DegreeBrach_CurrentDesignation'], df_test4['CurrentCTC'])\n",
    "df_test4['Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded'] = encoder.fit_transform(df_test4['Job_Title_DegreeBrach_CurrentDesignation'], df_test4['ExpectedCTC'])\n",
    "\n",
    "df_test4.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>EmpName</th>\n",
       "      <th>LanguageOfCommunication</th>\n",
       "      <th>Age</th>\n",
       "      <th>Gender</th>\n",
       "      <th>JobProfileIDApplyingFor</th>\n",
       "      <th>HighestDegree</th>\n",
       "      <th>DegreeBranch</th>\n",
       "      <th>GraduatingInstitute</th>\n",
       "      <th>LatestDegreeCGPA</th>\n",
       "      <th>...</th>\n",
       "      <th>Electrical</th>\n",
       "      <th>Electrical and Electronics</th>\n",
       "      <th>Electronics</th>\n",
       "      <th>Information Technology</th>\n",
       "      <th>Current_less_Grad_Year</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Expected_Y_Encoded</th>\n",
       "      <th>Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5664</td>\n",
       "      <td>Theo</td>\n",
       "      <td>Native</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>JR85289SDEFinance</td>\n",
       "      <td>19.000000</td>\n",
       "      <td>26.104478</td>\n",
       "      <td>76.354434</td>\n",
       "      <td>76.354434</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23568</td>\n",
       "      <td>Vernie</td>\n",
       "      <td>English</td>\n",
       "      <td>35</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>B.Tech</td>\n",
       "      <td>Information Technology</td>\n",
       "      <td>Tier 3</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>12</td>\n",
       "      <td>JR85289EMEngineering</td>\n",
       "      <td>16.707317</td>\n",
       "      <td>23.837398</td>\n",
       "      <td>73.981538</td>\n",
       "      <td>73.981538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21490</td>\n",
       "      <td>Wendell</td>\n",
       "      <td>English</td>\n",
       "      <td>28</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>M.Tech</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>JR87525SDEEngineering</td>\n",
       "      <td>17.331230</td>\n",
       "      <td>24.249211</td>\n",
       "      <td>75.790534</td>\n",
       "      <td>75.790534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8363</td>\n",
       "      <td>West</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>37</td>\n",
       "      <td>Male</td>\n",
       "      <td>JR87525</td>\n",
       "      <td>MS</td>\n",
       "      <td>Computer Science</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>JR87525DSProduct</td>\n",
       "      <td>17.851852</td>\n",
       "      <td>25.518519</td>\n",
       "      <td>75.434902</td>\n",
       "      <td>75.434902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6165</td>\n",
       "      <td>Alford</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>36</td>\n",
       "      <td>Female</td>\n",
       "      <td>JR85289</td>\n",
       "      <td>PhD</td>\n",
       "      <td>Artificial Intelligence</td>\n",
       "      <td>Tier 2</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>JR85289SDEEngineering</td>\n",
       "      <td>18.176349</td>\n",
       "      <td>25.188797</td>\n",
       "      <td>76.249896</td>\n",
       "      <td>76.249896</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID  EmpName LanguageOfCommunication  Age  Gender  \\\n",
       "0   5664     Theo                  Native   35    Male   \n",
       "1  23568   Vernie                 English   35    Male   \n",
       "2  21490  Wendell                 English   28    Male   \n",
       "3   8363     West                   Hindi   37    Male   \n",
       "4   6165   Alford                   Hindi   36  Female   \n",
       "\n",
       "  JobProfileIDApplyingFor HighestDegree             DegreeBranch  \\\n",
       "0                 JR85289           PhD   Information Technology   \n",
       "1                 JR85289        B.Tech   Information Technology   \n",
       "2                 JR87525        M.Tech  Artificial Intelligence   \n",
       "3                 JR87525            MS         Computer Science   \n",
       "4                 JR85289           PhD  Artificial Intelligence   \n",
       "\n",
       "  GraduatingInstitute  LatestDegreeCGPA  ...  Electrical  \\\n",
       "0              Tier 2                 5  ...           0   \n",
       "1              Tier 3                 9  ...           0   \n",
       "2              Tier 2                 9  ...           0   \n",
       "3              Tier 2                 8  ...           0   \n",
       "4              Tier 2                 7  ...           0   \n",
       "\n",
       "   Electrical and Electronics  Electronics  Information Technology  \\\n",
       "0                           0            0                       1   \n",
       "1                           0            0                       1   \n",
       "2                           0            0                       0   \n",
       "3                           0            0                       0   \n",
       "4                           0            0                       0   \n",
       "\n",
       "  Current_less_Grad_Year  Job_Title_DegreeBrach_CurrentDesignation  \\\n",
       "0                      5                         JR85289SDEFinance   \n",
       "1                     12                      JR85289EMEngineering   \n",
       "2                      2                     JR87525SDEEngineering   \n",
       "3                     11                          JR87525DSProduct   \n",
       "4                      6                     JR85289SDEEngineering   \n",
       "\n",
       "  Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded  \\\n",
       "0                                          19.000000             \n",
       "1                                          16.707317             \n",
       "2                                          17.331230             \n",
       "3                                          17.851852             \n",
       "4                                          18.176349             \n",
       "\n",
       "  Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded  \\\n",
       "0                                          26.104478              \n",
       "1                                          23.837398              \n",
       "2                                          24.249211              \n",
       "3                                          25.518519              \n",
       "4                                          25.188797              \n",
       "\n",
       "  Job_Title_DegreeBrach_CurrentDesignation_Expected_Y_Encoded  \\\n",
       "0                                          76.354434            \n",
       "1                                          73.981538            \n",
       "2                                          75.790534            \n",
       "3                                          75.434902            \n",
       "4                                          76.249896            \n",
       "\n",
       "   Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded  \n",
       "0                                          76.354434   \n",
       "1                                          73.981538   \n",
       "2                                          75.790534   \n",
       "3                                          75.434902   \n",
       "4                                          76.249896   \n",
       "\n",
       "[5 rows x 42 columns]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Target encoidng for fit percent\n",
    "\n",
    "df4_temp = df4.groupby(['Job_Title_DegreeBrach_CurrentDesignation'])['FitmentPercent'].mean().reset_index()\n",
    "df4_temp.columns = ['Job_Title_DegreeBrach_CurrentDesignation','Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded']\n",
    "df_test4 = pd.merge(df_test4, df4_temp,  how='left', on =['Job_Title_DegreeBrach_CurrentDesignation'])\n",
    "\n",
    "df_test4.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictors3 = df_test4[['Current_less_Grad_Year',\n",
    "                     'Electronics',\n",
    "                     'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
    "                     'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded',\n",
    "                     'EmpScore',\n",
    "                     'Enterprise',\n",
    "                     'LatestDegreeCGPA',\n",
    "                     'MidSized',\n",
    "                     'TotalLeavesTaken',\n",
    "                     'Finance',\n",
    "                     'GraduatingInstitute_Ordinal',\n",
    "                     'YearsOfExperince',\n",
    "                     'Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded',\n",
    "                     'CurrentCTC',\n",
    "                     'Computer Science',\n",
    "                     'ExpectedCTC',\n",
    "                     'HighestDegree_Ordinal']]\n",
    "\n",
    "\n",
    "predictors4 = df_test4[['LatestDegreeCGPA', 'YearsOfExperince','CurrentCTC', 'ExpectedCTC', \n",
    "                   'EmpScore','TotalLeavesTaken', 'HighestDegree_Ordinal',\n",
    "                   'GraduatingInstitute_Ordinal', 'Enterprise', 'MidSized', 'Startup',\n",
    "                   'Customer Success', 'Design', 'Engineering', 'Finance', 'Product',\n",
    "                   'Artificial Intelligence', 'Computer Science', 'Electrical',\n",
    "                   'Electrical and Electronics', 'Electronics', 'Information Technology',\n",
    "                   'Current_less_Grad_Year',\n",
    "                   'Job_Title_DegreeBrach_CurrentDesignation_Currect_CTC_Encoded',\n",
    "                   'Job_Title_DegreeBrach_CurrentDesignation_Expected_CTC_Encoded',\n",
    "                   'Job_Title_DegreeBrach_CurrentDesignation_Y_Encoded']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bias Influential Factor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>BiasInfluentialFactor_Temp</th>\n",
       "      <th>BiasInfluentialFactor</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>DegreeBranch</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>DegreeBranch</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5</td>\n",
       "      <td>HighestDegree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9</td>\n",
       "      <td>nan</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>Gender</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   BiasInfluentialFactor_Temp BiasInfluentialFactor\n",
       "0                           1          DegreeBranch\n",
       "1                           1          DegreeBranch\n",
       "2                           5         HighestDegree\n",
       "3                           9                   nan\n",
       "4                           4                Gender"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_xgb_d1_class = pd.DataFrame(model_xgb_d1_class.predict(predictors4))\n",
    "model_xgb_d1_class.columns = ['BiasInfluentialFactor_Temp']\n",
    "model_xgb_d1_class['BiasInfluentialFactor'] = label_encoder.inverse_transform(model_xgb_d1_class['BiasInfluentialFactor_Temp'])\n",
    "model_xgb_d1_class.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fit Percent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\tlearn: 14.1234268\ttotal: 84.4ms\tremaining: 1m 18s\n",
      "1:\tlearn: 13.5658369\ttotal: 100ms\tremaining: 46.5s\n",
      "2:\tlearn: 13.1776948\ttotal: 111ms\tremaining: 34.4s\n",
      "3:\tlearn: 12.9093884\ttotal: 133ms\tremaining: 30.8s\n",
      "4:\tlearn: 12.7116732\ttotal: 141ms\tremaining: 26.1s\n",
      "5:\tlearn: 12.5730239\ttotal: 154ms\tremaining: 23.8s\n",
      "6:\tlearn: 12.4790981\ttotal: 166ms\tremaining: 21.9s\n",
      "7:\tlearn: 12.4069096\ttotal: 180ms\tremaining: 20.8s\n",
      "8:\tlearn: 12.3551258\ttotal: 194ms\tremaining: 19.9s\n",
      "9:\tlearn: 12.3188745\ttotal: 201ms\tremaining: 18.5s\n",
      "10:\tlearn: 12.2946707\ttotal: 209ms\tremaining: 17.5s\n",
      "11:\tlearn: 12.2754243\ttotal: 219ms\tremaining: 16.8s\n",
      "12:\tlearn: 12.2554640\ttotal: 230ms\tremaining: 16.3s\n",
      "13:\tlearn: 12.2399625\ttotal: 240ms\tremaining: 15.7s\n",
      "14:\tlearn: 12.2238573\ttotal: 248ms\tremaining: 15.2s\n",
      "15:\tlearn: 12.2159453\ttotal: 258ms\tremaining: 14.8s\n",
      "16:\tlearn: 12.1965776\ttotal: 277ms\tremaining: 14.9s\n",
      "17:\tlearn: 12.1889846\ttotal: 285ms\tremaining: 14.5s\n",
      "18:\tlearn: 12.1740690\ttotal: 294ms\tremaining: 14.1s\n",
      "19:\tlearn: 12.1646209\ttotal: 303ms\tremaining: 13.8s\n",
      "20:\tlearn: 12.1567682\ttotal: 315ms\tremaining: 13.7s\n",
      "21:\tlearn: 12.1456894\ttotal: 327ms\tremaining: 13.5s\n",
      "22:\tlearn: 12.1361739\ttotal: 336ms\tremaining: 13.3s\n",
      "23:\tlearn: 12.1275282\ttotal: 344ms\tremaining: 13s\n",
      "24:\tlearn: 12.1090740\ttotal: 362ms\tremaining: 13.1s\n",
      "25:\tlearn: 12.1001071\ttotal: 370ms\tremaining: 12.9s\n",
      "26:\tlearn: 12.0886966\ttotal: 380ms\tremaining: 12.7s\n",
      "27:\tlearn: 12.0797631\ttotal: 388ms\tremaining: 12.5s\n",
      "28:\tlearn: 12.0691268\ttotal: 403ms\tremaining: 12.6s\n",
      "29:\tlearn: 12.0637966\ttotal: 412ms\tremaining: 12.4s\n",
      "30:\tlearn: 12.0560626\ttotal: 421ms\tremaining: 12.2s\n",
      "31:\tlearn: 12.0487638\ttotal: 427ms\tremaining: 12s\n",
      "32:\tlearn: 12.0381873\ttotal: 448ms\tremaining: 12.2s\n",
      "33:\tlearn: 12.0296051\ttotal: 457ms\tremaining: 12.1s\n",
      "34:\tlearn: 12.0101089\ttotal: 469ms\tremaining: 12s\n",
      "35:\tlearn: 12.0053927\ttotal: 484ms\tremaining: 12s\n",
      "36:\tlearn: 11.9916225\ttotal: 515ms\tremaining: 12.5s\n",
      "37:\tlearn: 11.9792702\ttotal: 525ms\tremaining: 12.3s\n",
      "38:\tlearn: 11.9722936\ttotal: 536ms\tremaining: 12.3s\n",
      "39:\tlearn: 11.9633003\ttotal: 555ms\tremaining: 12.4s\n",
      "40:\tlearn: 11.9594077\ttotal: 568ms\tremaining: 12.3s\n",
      "41:\tlearn: 11.9457349\ttotal: 579ms\tremaining: 12.3s\n",
      "42:\tlearn: 11.9367828\ttotal: 587ms\tremaining: 12.1s\n",
      "43:\tlearn: 11.9253315\ttotal: 596ms\tremaining: 12s\n",
      "44:\tlearn: 11.9201796\ttotal: 605ms\tremaining: 11.9s\n",
      "45:\tlearn: 11.9134674\ttotal: 613ms\tremaining: 11.8s\n",
      "46:\tlearn: 11.8995469\ttotal: 621ms\tremaining: 11.7s\n",
      "47:\tlearn: 11.8920514\ttotal: 632ms\tremaining: 11.6s\n",
      "48:\tlearn: 11.8835364\ttotal: 641ms\tremaining: 11.6s\n",
      "49:\tlearn: 11.8788988\ttotal: 648ms\tremaining: 11.4s\n",
      "50:\tlearn: 11.8752311\ttotal: 654ms\tremaining: 11.3s\n",
      "51:\tlearn: 11.8695536\ttotal: 660ms\tremaining: 11.2s\n",
      "52:\tlearn: 11.8658687\ttotal: 668ms\tremaining: 11.1s\n",
      "53:\tlearn: 11.8573585\ttotal: 676ms\tremaining: 11s\n",
      "54:\tlearn: 11.8488219\ttotal: 685ms\tremaining: 10.9s\n",
      "55:\tlearn: 11.8432736\ttotal: 701ms\tremaining: 11s\n",
      "56:\tlearn: 11.8300458\ttotal: 708ms\tremaining: 10.9s\n",
      "57:\tlearn: 11.8226281\ttotal: 717ms\tremaining: 10.8s\n",
      "58:\tlearn: 11.8148199\ttotal: 725ms\tremaining: 10.7s\n",
      "59:\tlearn: 11.8054156\ttotal: 737ms\tremaining: 10.7s\n",
      "60:\tlearn: 11.7967912\ttotal: 744ms\tremaining: 10.6s\n",
      "61:\tlearn: 11.7896721\ttotal: 753ms\tremaining: 10.6s\n",
      "62:\tlearn: 11.7827239\ttotal: 783ms\tremaining: 10.8s\n",
      "63:\tlearn: 11.7773211\ttotal: 795ms\tremaining: 10.8s\n",
      "64:\tlearn: 11.7695929\ttotal: 806ms\tremaining: 10.8s\n",
      "65:\tlearn: 11.7671638\ttotal: 816ms\tremaining: 10.7s\n",
      "66:\tlearn: 11.7596725\ttotal: 833ms\tremaining: 10.8s\n",
      "67:\tlearn: 11.7542015\ttotal: 868ms\tremaining: 11s\n",
      "68:\tlearn: 11.7485329\ttotal: 893ms\tremaining: 11.2s\n",
      "69:\tlearn: 11.7413151\ttotal: 941ms\tremaining: 11.6s\n",
      "70:\tlearn: 11.7282690\ttotal: 947ms\tremaining: 11.5s\n",
      "71:\tlearn: 11.7233847\ttotal: 956ms\tremaining: 11.4s\n",
      "72:\tlearn: 11.7176783\ttotal: 968ms\tremaining: 11.4s\n",
      "73:\tlearn: 11.7150765\ttotal: 997ms\tremaining: 11.6s\n",
      "74:\tlearn: 11.7087446\ttotal: 1.01s\tremaining: 11.5s\n",
      "75:\tlearn: 11.7031960\ttotal: 1.01s\tremaining: 11.4s\n",
      "76:\tlearn: 11.6931760\ttotal: 1.02s\tremaining: 11.4s\n",
      "77:\tlearn: 11.6820673\ttotal: 1.03s\tremaining: 11.3s\n",
      "78:\tlearn: 11.6722763\ttotal: 1.04s\tremaining: 11.2s\n",
      "79:\tlearn: 11.6649804\ttotal: 1.05s\tremaining: 11.1s\n",
      "80:\tlearn: 11.6584587\ttotal: 1.05s\tremaining: 11.1s\n",
      "81:\tlearn: 11.6555510\ttotal: 1.06s\tremaining: 11s\n",
      "82:\tlearn: 11.6480409\ttotal: 1.07s\tremaining: 11s\n",
      "83:\tlearn: 11.6374351\ttotal: 1.08s\tremaining: 10.9s\n",
      "84:\tlearn: 11.6303498\ttotal: 1.08s\tremaining: 10.8s\n",
      "85:\tlearn: 11.6289837\ttotal: 1.09s\tremaining: 10.7s\n",
      "86:\tlearn: 11.6219114\ttotal: 1.1s\tremaining: 10.7s\n",
      "87:\tlearn: 11.6138903\ttotal: 1.11s\tremaining: 10.7s\n",
      "88:\tlearn: 11.6018565\ttotal: 1.12s\tremaining: 10.6s\n",
      "89:\tlearn: 11.5989486\ttotal: 1.13s\tremaining: 10.6s\n",
      "90:\tlearn: 11.5919713\ttotal: 1.14s\tremaining: 10.5s\n",
      "91:\tlearn: 11.5866089\ttotal: 1.14s\tremaining: 10.4s\n",
      "92:\tlearn: 11.5830113\ttotal: 1.15s\tremaining: 10.4s\n",
      "93:\tlearn: 11.5772410\ttotal: 1.16s\tremaining: 10.4s\n",
      "94:\tlearn: 11.5694368\ttotal: 1.17s\tremaining: 10.3s\n",
      "95:\tlearn: 11.5595255\ttotal: 1.18s\tremaining: 10.2s\n",
      "96:\tlearn: 11.5485960\ttotal: 1.19s\tremaining: 10.2s\n",
      "97:\tlearn: 11.5347390\ttotal: 1.21s\tremaining: 10.3s\n",
      "98:\tlearn: 11.5247120\ttotal: 1.22s\tremaining: 10.2s\n",
      "99:\tlearn: 11.5182706\ttotal: 1.22s\tremaining: 10.2s\n",
      "100:\tlearn: 11.5084125\ttotal: 1.23s\tremaining: 10.1s\n",
      "101:\tlearn: 11.4972490\ttotal: 1.25s\tremaining: 10.2s\n",
      "102:\tlearn: 11.4894625\ttotal: 1.25s\tremaining: 10.1s\n",
      "103:\tlearn: 11.4868705\ttotal: 1.26s\tremaining: 10s\n",
      "104:\tlearn: 11.4743423\ttotal: 1.27s\tremaining: 9.98s\n",
      "105:\tlearn: 11.4699916\ttotal: 1.27s\tremaining: 9.93s\n",
      "106:\tlearn: 11.4666909\ttotal: 1.29s\tremaining: 9.96s\n",
      "107:\tlearn: 11.4586006\ttotal: 1.3s\tremaining: 9.93s\n",
      "108:\tlearn: 11.4480747\ttotal: 1.31s\tremaining: 9.87s\n",
      "109:\tlearn: 11.4404861\ttotal: 1.31s\tremaining: 9.83s\n",
      "110:\tlearn: 11.4330701\ttotal: 1.32s\tremaining: 9.78s\n",
      "111:\tlearn: 11.4247487\ttotal: 1.33s\tremaining: 9.75s\n",
      "112:\tlearn: 11.4109698\ttotal: 1.34s\tremaining: 9.71s\n",
      "113:\tlearn: 11.4064827\ttotal: 1.35s\tremaining: 9.66s\n",
      "114:\tlearn: 11.3966169\ttotal: 1.35s\tremaining: 9.62s\n",
      "115:\tlearn: 11.3878563\ttotal: 1.36s\tremaining: 9.6s\n",
      "116:\tlearn: 11.3799500\ttotal: 1.37s\tremaining: 9.55s\n",
      "117:\tlearn: 11.3716997\ttotal: 1.38s\tremaining: 9.52s\n",
      "118:\tlearn: 11.3621175\ttotal: 1.39s\tremaining: 9.49s\n",
      "119:\tlearn: 11.3526434\ttotal: 1.4s\tremaining: 9.47s\n",
      "120:\tlearn: 11.3462856\ttotal: 1.41s\tremaining: 9.44s\n",
      "121:\tlearn: 11.3405004\ttotal: 1.42s\tremaining: 9.41s\n",
      "122:\tlearn: 11.3291057\ttotal: 1.42s\tremaining: 9.37s\n",
      "123:\tlearn: 11.3225617\ttotal: 1.43s\tremaining: 9.33s\n",
      "124:\tlearn: 11.3127676\ttotal: 1.44s\tremaining: 9.31s\n",
      "125:\tlearn: 11.3019832\ttotal: 1.45s\tremaining: 9.28s\n",
      "126:\tlearn: 11.2942370\ttotal: 1.46s\tremaining: 9.26s\n",
      "127:\tlearn: 11.2849232\ttotal: 1.47s\tremaining: 9.22s\n",
      "128:\tlearn: 11.2727509\ttotal: 1.48s\tremaining: 9.22s\n",
      "129:\tlearn: 11.2644570\ttotal: 1.49s\tremaining: 9.19s\n",
      "130:\tlearn: 11.2581532\ttotal: 1.5s\tremaining: 9.16s\n",
      "131:\tlearn: 11.2491816\ttotal: 1.5s\tremaining: 9.13s\n",
      "132:\tlearn: 11.2429383\ttotal: 1.51s\tremaining: 9.09s\n",
      "133:\tlearn: 11.2338574\ttotal: 1.52s\tremaining: 9.05s\n",
      "134:\tlearn: 11.2268594\ttotal: 1.53s\tremaining: 9.02s\n",
      "135:\tlearn: 11.2175078\ttotal: 1.54s\tremaining: 8.99s\n",
      "136:\tlearn: 11.2075968\ttotal: 1.54s\tremaining: 8.97s\n",
      "137:\tlearn: 11.1976078\ttotal: 1.55s\tremaining: 8.93s\n",
      "138:\tlearn: 11.1877227\ttotal: 1.56s\tremaining: 8.9s\n",
      "139:\tlearn: 11.1801925\ttotal: 1.57s\tremaining: 8.87s\n",
      "140:\tlearn: 11.1714907\ttotal: 1.58s\tremaining: 8.87s\n",
      "141:\tlearn: 11.1625537\ttotal: 1.59s\tremaining: 8.84s\n",
      "142:\tlearn: 11.1532152\ttotal: 1.6s\tremaining: 8.82s\n",
      "143:\tlearn: 11.1422041\ttotal: 1.61s\tremaining: 8.79s\n",
      "144:\tlearn: 11.1361534\ttotal: 1.61s\tremaining: 8.75s\n",
      "145:\tlearn: 11.1285315\ttotal: 1.63s\tremaining: 8.76s\n",
      "146:\tlearn: 11.1207790\ttotal: 1.63s\tremaining: 8.73s\n",
      "147:\tlearn: 11.1122770\ttotal: 1.65s\tremaining: 8.73s\n",
      "148:\tlearn: 11.1062446\ttotal: 1.66s\tremaining: 8.71s\n",
      "149:\tlearn: 11.0978539\ttotal: 1.69s\tremaining: 8.81s\n",
      "150:\tlearn: 11.0921760\ttotal: 1.7s\tremaining: 8.78s\n",
      "151:\tlearn: 11.0870710\ttotal: 1.7s\tremaining: 8.74s\n",
      "152:\tlearn: 11.0806101\ttotal: 1.72s\tremaining: 8.73s\n",
      "153:\tlearn: 11.0756885\ttotal: 1.73s\tremaining: 8.74s\n",
      "154:\tlearn: 11.0690035\ttotal: 1.74s\tremaining: 8.74s\n",
      "155:\tlearn: 11.0608057\ttotal: 1.77s\tremaining: 8.79s\n",
      "156:\tlearn: 11.0549045\ttotal: 1.78s\tremaining: 8.79s\n",
      "157:\tlearn: 11.0457335\ttotal: 1.79s\tremaining: 8.79s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "158:\tlearn: 11.0376422\ttotal: 1.8s\tremaining: 8.77s\n",
      "159:\tlearn: 11.0316432\ttotal: 1.83s\tremaining: 8.82s\n",
      "160:\tlearn: 11.0272887\ttotal: 1.84s\tremaining: 8.8s\n",
      "161:\tlearn: 11.0180580\ttotal: 1.85s\tremaining: 8.78s\n",
      "162:\tlearn: 11.0114183\ttotal: 1.86s\tremaining: 8.76s\n",
      "163:\tlearn: 11.0021825\ttotal: 1.91s\tremaining: 8.97s\n",
      "164:\tlearn: 10.9916056\ttotal: 1.93s\tremaining: 8.95s\n",
      "165:\tlearn: 10.9863606\ttotal: 1.93s\tremaining: 8.92s\n",
      "166:\tlearn: 10.9793927\ttotal: 1.94s\tremaining: 8.89s\n",
      "167:\tlearn: 10.9741555\ttotal: 1.98s\tremaining: 9.02s\n",
      "168:\tlearn: 10.9669771\ttotal: 1.99s\tremaining: 8.98s\n",
      "169:\tlearn: 10.9586088\ttotal: 2s\tremaining: 8.95s\n",
      "170:\tlearn: 10.9474141\ttotal: 2.04s\tremaining: 9.07s\n",
      "171:\tlearn: 10.9434597\ttotal: 2.05s\tremaining: 9.05s\n",
      "172:\tlearn: 10.9301098\ttotal: 2.06s\tremaining: 9.03s\n",
      "173:\tlearn: 10.9197348\ttotal: 2.07s\tremaining: 9.01s\n",
      "174:\tlearn: 10.9135842\ttotal: 2.09s\tremaining: 9.05s\n",
      "175:\tlearn: 10.9092094\ttotal: 2.1s\tremaining: 9.02s\n",
      "176:\tlearn: 10.9034975\ttotal: 2.11s\tremaining: 8.99s\n",
      "177:\tlearn: 10.8967142\ttotal: 2.12s\tremaining: 8.96s\n",
      "178:\tlearn: 10.8864841\ttotal: 2.13s\tremaining: 8.96s\n",
      "179:\tlearn: 10.8813616\ttotal: 2.13s\tremaining: 8.92s\n",
      "180:\tlearn: 10.8747154\ttotal: 2.14s\tremaining: 8.89s\n",
      "181:\tlearn: 10.8680158\ttotal: 2.15s\tremaining: 8.87s\n",
      "182:\tlearn: 10.8599230\ttotal: 2.16s\tremaining: 8.85s\n",
      "183:\tlearn: 10.8521847\ttotal: 2.17s\tremaining: 8.83s\n",
      "184:\tlearn: 10.8476186\ttotal: 2.19s\tremaining: 8.85s\n",
      "185:\tlearn: 10.8404329\ttotal: 2.21s\tremaining: 8.85s\n",
      "186:\tlearn: 10.8392315\ttotal: 2.21s\tremaining: 8.81s\n",
      "187:\tlearn: 10.8296268\ttotal: 2.22s\tremaining: 8.79s\n",
      "188:\tlearn: 10.8218674\ttotal: 2.23s\tremaining: 8.77s\n",
      "189:\tlearn: 10.8141838\ttotal: 2.24s\tremaining: 8.76s\n",
      "190:\tlearn: 10.8076358\ttotal: 2.25s\tremaining: 8.74s\n",
      "191:\tlearn: 10.7950191\ttotal: 2.26s\tremaining: 8.71s\n",
      "192:\tlearn: 10.7879158\ttotal: 2.27s\tremaining: 8.68s\n",
      "193:\tlearn: 10.7807320\ttotal: 2.27s\tremaining: 8.65s\n",
      "194:\tlearn: 10.7690928\ttotal: 2.28s\tremaining: 8.63s\n",
      "195:\tlearn: 10.7614912\ttotal: 2.3s\tremaining: 8.63s\n",
      "196:\tlearn: 10.7581593\ttotal: 2.31s\tremaining: 8.61s\n",
      "197:\tlearn: 10.7518248\ttotal: 2.34s\tremaining: 8.68s\n",
      "198:\tlearn: 10.7461154\ttotal: 2.35s\tremaining: 8.68s\n",
      "199:\tlearn: 10.7398018\ttotal: 2.36s\tremaining: 8.65s\n",
      "200:\tlearn: 10.7286017\ttotal: 2.37s\tremaining: 8.63s\n",
      "201:\tlearn: 10.7184721\ttotal: 2.38s\tremaining: 8.6s\n",
      "202:\tlearn: 10.7106968\ttotal: 2.39s\tremaining: 8.59s\n",
      "203:\tlearn: 10.7053018\ttotal: 2.4s\tremaining: 8.57s\n",
      "204:\tlearn: 10.7001795\ttotal: 2.41s\tremaining: 8.55s\n",
      "205:\tlearn: 10.6937826\ttotal: 2.42s\tremaining: 8.52s\n",
      "206:\tlearn: 10.6872831\ttotal: 2.43s\tremaining: 8.51s\n",
      "207:\tlearn: 10.6810782\ttotal: 2.44s\tremaining: 8.49s\n",
      "208:\tlearn: 10.6735176\ttotal: 2.44s\tremaining: 8.46s\n",
      "209:\tlearn: 10.6681279\ttotal: 2.45s\tremaining: 8.43s\n",
      "210:\tlearn: 10.6634047\ttotal: 2.46s\tremaining: 8.41s\n",
      "211:\tlearn: 10.6542296\ttotal: 2.47s\tremaining: 8.39s\n",
      "212:\tlearn: 10.6442236\ttotal: 2.47s\tremaining: 8.35s\n",
      "213:\tlearn: 10.6352379\ttotal: 2.48s\tremaining: 8.32s\n",
      "214:\tlearn: 10.6281508\ttotal: 2.49s\tremaining: 8.3s\n",
      "215:\tlearn: 10.6207026\ttotal: 2.5s\tremaining: 8.28s\n",
      "216:\tlearn: 10.6151479\ttotal: 2.5s\tremaining: 8.25s\n",
      "217:\tlearn: 10.6115559\ttotal: 2.51s\tremaining: 8.22s\n",
      "218:\tlearn: 10.6057080\ttotal: 2.52s\tremaining: 8.2s\n",
      "219:\tlearn: 10.5997830\ttotal: 2.53s\tremaining: 8.18s\n",
      "220:\tlearn: 10.5887283\ttotal: 2.54s\tremaining: 8.15s\n",
      "221:\tlearn: 10.5816602\ttotal: 2.54s\tremaining: 8.13s\n",
      "222:\tlearn: 10.5735409\ttotal: 2.55s\tremaining: 8.11s\n",
      "223:\tlearn: 10.5678818\ttotal: 2.56s\tremaining: 8.1s\n",
      "224:\tlearn: 10.5634662\ttotal: 2.58s\tremaining: 8.09s\n",
      "225:\tlearn: 10.5569823\ttotal: 2.58s\tremaining: 8.07s\n",
      "226:\tlearn: 10.5509732\ttotal: 2.59s\tremaining: 8.04s\n",
      "227:\tlearn: 10.5446380\ttotal: 2.6s\tremaining: 8.02s\n",
      "228:\tlearn: 10.5385576\ttotal: 2.61s\tremaining: 8s\n",
      "229:\tlearn: 10.5325460\ttotal: 2.62s\tremaining: 8s\n",
      "230:\tlearn: 10.5235018\ttotal: 2.63s\tremaining: 7.98s\n",
      "231:\tlearn: 10.5185302\ttotal: 2.64s\tremaining: 7.96s\n",
      "232:\tlearn: 10.5112387\ttotal: 2.65s\tremaining: 7.94s\n",
      "233:\tlearn: 10.5070354\ttotal: 2.65s\tremaining: 7.91s\n",
      "234:\tlearn: 10.5005783\ttotal: 2.66s\tremaining: 7.88s\n",
      "235:\tlearn: 10.4893672\ttotal: 2.67s\tremaining: 7.86s\n",
      "236:\tlearn: 10.4823707\ttotal: 2.67s\tremaining: 7.85s\n",
      "237:\tlearn: 10.4732973\ttotal: 2.68s\tremaining: 7.82s\n",
      "238:\tlearn: 10.4662473\ttotal: 2.7s\tremaining: 7.83s\n",
      "239:\tlearn: 10.4603084\ttotal: 2.71s\tremaining: 7.8s\n",
      "240:\tlearn: 10.4529282\ttotal: 2.71s\tremaining: 7.78s\n",
      "241:\tlearn: 10.4469899\ttotal: 2.72s\tremaining: 7.76s\n",
      "242:\tlearn: 10.4410856\ttotal: 2.73s\tremaining: 7.75s\n",
      "243:\tlearn: 10.4350047\ttotal: 2.75s\tremaining: 7.74s\n",
      "244:\tlearn: 10.4272163\ttotal: 2.75s\tremaining: 7.72s\n",
      "245:\tlearn: 10.4205358\ttotal: 2.77s\tremaining: 7.72s\n",
      "246:\tlearn: 10.4161264\ttotal: 2.78s\tremaining: 7.71s\n",
      "247:\tlearn: 10.4081124\ttotal: 2.79s\tremaining: 7.68s\n",
      "248:\tlearn: 10.4031439\ttotal: 2.79s\tremaining: 7.67s\n",
      "249:\tlearn: 10.3920862\ttotal: 2.8s\tremaining: 7.65s\n",
      "250:\tlearn: 10.3887129\ttotal: 2.82s\tremaining: 7.64s\n",
      "251:\tlearn: 10.3827649\ttotal: 2.82s\tremaining: 7.62s\n",
      "252:\tlearn: 10.3732421\ttotal: 2.83s\tremaining: 7.6s\n",
      "253:\tlearn: 10.3665098\ttotal: 2.84s\tremaining: 7.58s\n",
      "254:\tlearn: 10.3536529\ttotal: 2.85s\tremaining: 7.56s\n",
      "255:\tlearn: 10.3468065\ttotal: 2.85s\tremaining: 7.54s\n",
      "256:\tlearn: 10.3371800\ttotal: 2.87s\tremaining: 7.53s\n",
      "257:\tlearn: 10.3307663\ttotal: 2.87s\tremaining: 7.51s\n",
      "258:\tlearn: 10.3228159\ttotal: 2.88s\tremaining: 7.5s\n",
      "259:\tlearn: 10.3176141\ttotal: 2.89s\tremaining: 7.48s\n",
      "260:\tlearn: 10.3117020\ttotal: 2.9s\tremaining: 7.47s\n",
      "261:\tlearn: 10.3056378\ttotal: 2.91s\tremaining: 7.45s\n",
      "262:\tlearn: 10.3017032\ttotal: 2.92s\tremaining: 7.43s\n",
      "263:\tlearn: 10.2916159\ttotal: 2.93s\tremaining: 7.42s\n",
      "264:\tlearn: 10.2796218\ttotal: 2.94s\tremaining: 7.41s\n",
      "265:\tlearn: 10.2717449\ttotal: 2.96s\tremaining: 7.41s\n",
      "266:\tlearn: 10.2638441\ttotal: 2.97s\tremaining: 7.39s\n",
      "267:\tlearn: 10.2581364\ttotal: 2.98s\tremaining: 7.38s\n",
      "268:\tlearn: 10.2519516\ttotal: 2.99s\tremaining: 7.36s\n",
      "269:\tlearn: 10.2392254\ttotal: 3s\tremaining: 7.35s\n",
      "270:\tlearn: 10.2335962\ttotal: 3s\tremaining: 7.33s\n",
      "271:\tlearn: 10.2257649\ttotal: 3.01s\tremaining: 7.31s\n",
      "272:\tlearn: 10.2213811\ttotal: 3.02s\tremaining: 7.3s\n",
      "273:\tlearn: 10.2109647\ttotal: 3.04s\tremaining: 7.3s\n",
      "274:\tlearn: 10.2074662\ttotal: 3.05s\tremaining: 7.28s\n",
      "275:\tlearn: 10.1995878\ttotal: 3.06s\tremaining: 7.26s\n",
      "276:\tlearn: 10.1944314\ttotal: 3.06s\tremaining: 7.24s\n",
      "277:\tlearn: 10.1870480\ttotal: 3.08s\tremaining: 7.24s\n",
      "278:\tlearn: 10.1842223\ttotal: 3.09s\tremaining: 7.24s\n",
      "279:\tlearn: 10.1779233\ttotal: 3.1s\tremaining: 7.22s\n",
      "280:\tlearn: 10.1736169\ttotal: 3.11s\tremaining: 7.2s\n",
      "281:\tlearn: 10.1661300\ttotal: 3.11s\tremaining: 7.18s\n",
      "282:\tlearn: 10.1597539\ttotal: 3.12s\tremaining: 7.16s\n",
      "283:\tlearn: 10.1552573\ttotal: 3.13s\tremaining: 7.15s\n",
      "284:\tlearn: 10.1492211\ttotal: 3.14s\tremaining: 7.13s\n",
      "285:\tlearn: 10.1414021\ttotal: 3.15s\tremaining: 7.12s\n",
      "286:\tlearn: 10.1379347\ttotal: 3.16s\tremaining: 7.1s\n",
      "287:\tlearn: 10.1347490\ttotal: 3.17s\tremaining: 7.1s\n",
      "288:\tlearn: 10.1295603\ttotal: 3.18s\tremaining: 7.08s\n",
      "289:\tlearn: 10.1239695\ttotal: 3.19s\tremaining: 7.06s\n",
      "290:\tlearn: 10.1186784\ttotal: 3.2s\tremaining: 7.04s\n",
      "291:\tlearn: 10.1113238\ttotal: 3.21s\tremaining: 7.03s\n",
      "292:\tlearn: 10.1070393\ttotal: 3.22s\tremaining: 7.03s\n",
      "293:\tlearn: 10.0988332\ttotal: 3.23s\tremaining: 7.01s\n",
      "294:\tlearn: 10.0918936\ttotal: 3.24s\tremaining: 7s\n",
      "295:\tlearn: 10.0855866\ttotal: 3.25s\tremaining: 6.98s\n",
      "296:\tlearn: 10.0818116\ttotal: 3.26s\tremaining: 6.97s\n",
      "297:\tlearn: 10.0733962\ttotal: 3.27s\tremaining: 6.95s\n",
      "298:\tlearn: 10.0687284\ttotal: 3.27s\tremaining: 6.93s\n",
      "299:\tlearn: 10.0617684\ttotal: 3.28s\tremaining: 6.91s\n",
      "300:\tlearn: 10.0578535\ttotal: 3.29s\tremaining: 6.89s\n",
      "301:\tlearn: 10.0534608\ttotal: 3.3s\tremaining: 6.89s\n",
      "302:\tlearn: 10.0458947\ttotal: 3.31s\tremaining: 6.87s\n",
      "303:\tlearn: 10.0415434\ttotal: 3.31s\tremaining: 6.85s\n",
      "304:\tlearn: 10.0338371\ttotal: 3.32s\tremaining: 6.83s\n",
      "305:\tlearn: 10.0284064\ttotal: 3.34s\tremaining: 6.83s\n",
      "306:\tlearn: 10.0273153\ttotal: 3.35s\tremaining: 6.82s\n",
      "307:\tlearn: 10.0201866\ttotal: 3.36s\tremaining: 6.81s\n",
      "308:\tlearn: 10.0154810\ttotal: 3.37s\tremaining: 6.79s\n",
      "309:\tlearn: 10.0123948\ttotal: 3.37s\tremaining: 6.77s\n",
      "310:\tlearn: 10.0077251\ttotal: 3.38s\tremaining: 6.75s\n",
      "311:\tlearn: 9.9980143\ttotal: 3.39s\tremaining: 6.74s\n",
      "312:\tlearn: 9.9973910\ttotal: 3.4s\tremaining: 6.73s\n",
      "313:\tlearn: 9.9917677\ttotal: 3.42s\tremaining: 6.73s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "314:\tlearn: 9.9866681\ttotal: 3.42s\tremaining: 6.71s\n",
      "315:\tlearn: 9.9807800\ttotal: 3.43s\tremaining: 6.69s\n",
      "316:\tlearn: 9.9732520\ttotal: 3.44s\tremaining: 6.68s\n",
      "317:\tlearn: 9.9674585\ttotal: 3.45s\tremaining: 6.66s\n",
      "318:\tlearn: 9.9630715\ttotal: 3.46s\tremaining: 6.65s\n",
      "319:\tlearn: 9.9586184\ttotal: 3.46s\tremaining: 6.63s\n",
      "320:\tlearn: 9.9539808\ttotal: 3.47s\tremaining: 6.61s\n",
      "321:\tlearn: 9.9511286\ttotal: 3.48s\tremaining: 6.59s\n",
      "322:\tlearn: 9.9479605\ttotal: 3.49s\tremaining: 6.58s\n",
      "323:\tlearn: 9.9425874\ttotal: 3.5s\tremaining: 6.56s\n",
      "324:\tlearn: 9.9350892\ttotal: 3.5s\tremaining: 6.55s\n",
      "325:\tlearn: 9.9279205\ttotal: 3.51s\tremaining: 6.53s\n",
      "326:\tlearn: 9.9155925\ttotal: 3.52s\tremaining: 6.52s\n",
      "327:\tlearn: 9.9085440\ttotal: 3.53s\tremaining: 6.51s\n",
      "328:\tlearn: 9.9042332\ttotal: 3.54s\tremaining: 6.49s\n",
      "329:\tlearn: 9.8998310\ttotal: 3.55s\tremaining: 6.48s\n",
      "330:\tlearn: 9.8970579\ttotal: 3.56s\tremaining: 6.46s\n",
      "331:\tlearn: 9.8890640\ttotal: 3.56s\tremaining: 6.44s\n",
      "332:\tlearn: 9.8813302\ttotal: 3.57s\tremaining: 6.43s\n",
      "333:\tlearn: 9.8752434\ttotal: 3.58s\tremaining: 6.41s\n",
      "334:\tlearn: 9.8706107\ttotal: 3.59s\tremaining: 6.4s\n",
      "335:\tlearn: 9.8679149\ttotal: 3.6s\tremaining: 6.38s\n",
      "336:\tlearn: 9.8617376\ttotal: 3.61s\tremaining: 6.37s\n",
      "337:\tlearn: 9.8565520\ttotal: 3.62s\tremaining: 6.36s\n",
      "338:\tlearn: 9.8512577\ttotal: 3.63s\tremaining: 6.34s\n",
      "339:\tlearn: 9.8453385\ttotal: 3.63s\tremaining: 6.33s\n",
      "340:\tlearn: 9.8432238\ttotal: 3.64s\tremaining: 6.32s\n",
      "341:\tlearn: 9.8371766\ttotal: 3.65s\tremaining: 6.3s\n",
      "342:\tlearn: 9.8298371\ttotal: 3.67s\tremaining: 6.29s\n",
      "343:\tlearn: 9.8237252\ttotal: 3.68s\tremaining: 6.29s\n",
      "344:\tlearn: 9.8160078\ttotal: 3.69s\tremaining: 6.28s\n",
      "345:\tlearn: 9.8136865\ttotal: 3.7s\tremaining: 6.26s\n",
      "346:\tlearn: 9.8097269\ttotal: 3.7s\tremaining: 6.25s\n",
      "347:\tlearn: 9.8021133\ttotal: 3.71s\tremaining: 6.23s\n",
      "348:\tlearn: 9.7977897\ttotal: 3.72s\tremaining: 6.22s\n",
      "349:\tlearn: 9.7913506\ttotal: 3.73s\tremaining: 6.2s\n",
      "350:\tlearn: 9.7871947\ttotal: 3.74s\tremaining: 6.19s\n",
      "351:\tlearn: 9.7784090\ttotal: 3.75s\tremaining: 6.17s\n",
      "352:\tlearn: 9.7738161\ttotal: 3.75s\tremaining: 6.16s\n",
      "353:\tlearn: 9.7644541\ttotal: 3.76s\tremaining: 6.14s\n",
      "354:\tlearn: 9.7576666\ttotal: 3.77s\tremaining: 6.12s\n",
      "355:\tlearn: 9.7533694\ttotal: 3.77s\tremaining: 6.11s\n",
      "356:\tlearn: 9.7479488\ttotal: 3.79s\tremaining: 6.1s\n",
      "357:\tlearn: 9.7399287\ttotal: 3.79s\tremaining: 6.09s\n",
      "358:\tlearn: 9.7340382\ttotal: 3.81s\tremaining: 6.08s\n",
      "359:\tlearn: 9.7277475\ttotal: 3.82s\tremaining: 6.06s\n",
      "360:\tlearn: 9.7229913\ttotal: 3.83s\tremaining: 6.06s\n",
      "361:\tlearn: 9.7185517\ttotal: 3.83s\tremaining: 6.04s\n",
      "362:\tlearn: 9.7145197\ttotal: 3.84s\tremaining: 6.02s\n",
      "363:\tlearn: 9.7086374\ttotal: 3.85s\tremaining: 6.01s\n",
      "364:\tlearn: 9.7012266\ttotal: 3.86s\tremaining: 5.99s\n",
      "365:\tlearn: 9.6928926\ttotal: 3.87s\tremaining: 5.98s\n",
      "366:\tlearn: 9.6897872\ttotal: 3.87s\tremaining: 5.96s\n",
      "367:\tlearn: 9.6837477\ttotal: 3.88s\tremaining: 5.95s\n",
      "368:\tlearn: 9.6802454\ttotal: 3.89s\tremaining: 5.93s\n",
      "369:\tlearn: 9.6726796\ttotal: 3.9s\tremaining: 5.93s\n",
      "370:\tlearn: 9.6682769\ttotal: 3.91s\tremaining: 5.91s\n",
      "371:\tlearn: 9.6632602\ttotal: 3.92s\tremaining: 5.89s\n",
      "372:\tlearn: 9.6595153\ttotal: 3.92s\tremaining: 5.88s\n",
      "373:\tlearn: 9.6525739\ttotal: 3.93s\tremaining: 5.87s\n",
      "374:\tlearn: 9.6496899\ttotal: 3.94s\tremaining: 5.85s\n",
      "375:\tlearn: 9.6453226\ttotal: 3.94s\tremaining: 5.83s\n",
      "376:\tlearn: 9.6388482\ttotal: 3.95s\tremaining: 5.82s\n",
      "377:\tlearn: 9.6334790\ttotal: 3.96s\tremaining: 5.81s\n",
      "378:\tlearn: 9.6291160\ttotal: 3.97s\tremaining: 5.79s\n",
      "379:\tlearn: 9.6221213\ttotal: 3.98s\tremaining: 5.78s\n",
      "380:\tlearn: 9.6167989\ttotal: 3.98s\tremaining: 5.76s\n",
      "381:\tlearn: 9.6102620\ttotal: 3.99s\tremaining: 5.75s\n",
      "382:\tlearn: 9.6046993\ttotal: 4.01s\tremaining: 5.74s\n",
      "383:\tlearn: 9.5989646\ttotal: 4.01s\tremaining: 5.73s\n",
      "384:\tlearn: 9.5913103\ttotal: 4.02s\tremaining: 5.71s\n",
      "385:\tlearn: 9.5838383\ttotal: 4.03s\tremaining: 5.7s\n",
      "386:\tlearn: 9.5786846\ttotal: 4.04s\tremaining: 5.69s\n",
      "387:\tlearn: 9.5747424\ttotal: 4.05s\tremaining: 5.68s\n",
      "388:\tlearn: 9.5714702\ttotal: 4.06s\tremaining: 5.67s\n",
      "389:\tlearn: 9.5649166\ttotal: 4.07s\tremaining: 5.66s\n",
      "390:\tlearn: 9.5614573\ttotal: 4.1s\tremaining: 5.67s\n",
      "391:\tlearn: 9.5557450\ttotal: 4.11s\tremaining: 5.66s\n",
      "392:\tlearn: 9.5523381\ttotal: 4.11s\tremaining: 5.64s\n",
      "393:\tlearn: 9.5489412\ttotal: 4.12s\tremaining: 5.63s\n",
      "394:\tlearn: 9.5441231\ttotal: 4.13s\tremaining: 5.62s\n",
      "395:\tlearn: 9.5351652\ttotal: 4.14s\tremaining: 5.6s\n",
      "396:\tlearn: 9.5286923\ttotal: 4.15s\tremaining: 5.59s\n",
      "397:\tlearn: 9.5238533\ttotal: 4.15s\tremaining: 5.57s\n",
      "398:\tlearn: 9.5199859\ttotal: 4.16s\tremaining: 5.56s\n",
      "399:\tlearn: 9.5106825\ttotal: 4.17s\tremaining: 5.55s\n",
      "400:\tlearn: 9.5057136\ttotal: 4.18s\tremaining: 5.53s\n",
      "401:\tlearn: 9.4999765\ttotal: 4.19s\tremaining: 5.52s\n",
      "402:\tlearn: 9.4964415\ttotal: 4.2s\tremaining: 5.51s\n",
      "403:\tlearn: 9.4932173\ttotal: 4.21s\tremaining: 5.5s\n",
      "404:\tlearn: 9.4887864\ttotal: 4.22s\tremaining: 5.49s\n",
      "405:\tlearn: 9.4859486\ttotal: 4.22s\tremaining: 5.47s\n",
      "406:\tlearn: 9.4788986\ttotal: 4.23s\tremaining: 5.46s\n",
      "407:\tlearn: 9.4757793\ttotal: 4.24s\tremaining: 5.45s\n",
      "408:\tlearn: 9.4699799\ttotal: 4.25s\tremaining: 5.43s\n",
      "409:\tlearn: 9.4653658\ttotal: 4.26s\tremaining: 5.42s\n",
      "410:\tlearn: 9.4595423\ttotal: 4.28s\tremaining: 5.42s\n",
      "411:\tlearn: 9.4556191\ttotal: 4.29s\tremaining: 5.41s\n",
      "412:\tlearn: 9.4496781\ttotal: 4.29s\tremaining: 5.4s\n",
      "413:\tlearn: 9.4444079\ttotal: 4.3s\tremaining: 5.38s\n",
      "414:\tlearn: 9.4410300\ttotal: 4.31s\tremaining: 5.37s\n",
      "415:\tlearn: 9.4376859\ttotal: 4.32s\tremaining: 5.35s\n",
      "416:\tlearn: 9.4328464\ttotal: 4.33s\tremaining: 5.34s\n",
      "417:\tlearn: 9.4268381\ttotal: 4.33s\tremaining: 5.33s\n",
      "418:\tlearn: 9.4225730\ttotal: 4.35s\tremaining: 5.32s\n",
      "419:\tlearn: 9.4191343\ttotal: 4.35s\tremaining: 5.31s\n",
      "420:\tlearn: 9.4166743\ttotal: 4.36s\tremaining: 5.29s\n",
      "421:\tlearn: 9.4132074\ttotal: 4.37s\tremaining: 5.28s\n",
      "422:\tlearn: 9.4082328\ttotal: 4.38s\tremaining: 5.27s\n",
      "423:\tlearn: 9.4043858\ttotal: 4.39s\tremaining: 5.26s\n",
      "424:\tlearn: 9.3996693\ttotal: 4.39s\tremaining: 5.24s\n",
      "425:\tlearn: 9.3941943\ttotal: 4.4s\tremaining: 5.23s\n",
      "426:\tlearn: 9.3845052\ttotal: 4.41s\tremaining: 5.22s\n",
      "427:\tlearn: 9.3803539\ttotal: 4.42s\tremaining: 5.21s\n",
      "428:\tlearn: 9.3784059\ttotal: 4.43s\tremaining: 5.19s\n",
      "429:\tlearn: 9.3700641\ttotal: 4.44s\tremaining: 5.18s\n",
      "430:\tlearn: 9.3656202\ttotal: 4.45s\tremaining: 5.17s\n",
      "431:\tlearn: 9.3584248\ttotal: 4.46s\tremaining: 5.16s\n",
      "432:\tlearn: 9.3552851\ttotal: 4.46s\tremaining: 5.14s\n",
      "433:\tlearn: 9.3521139\ttotal: 4.47s\tremaining: 5.13s\n",
      "434:\tlearn: 9.3469589\ttotal: 4.5s\tremaining: 5.14s\n",
      "435:\tlearn: 9.3392863\ttotal: 4.52s\tremaining: 5.14s\n",
      "436:\tlearn: 9.3329331\ttotal: 4.53s\tremaining: 5.13s\n",
      "437:\tlearn: 9.3273806\ttotal: 4.53s\tremaining: 5.11s\n",
      "438:\tlearn: 9.3232154\ttotal: 4.54s\tremaining: 5.1s\n",
      "439:\tlearn: 9.3178622\ttotal: 4.57s\tremaining: 5.11s\n",
      "440:\tlearn: 9.3142636\ttotal: 4.57s\tremaining: 5.09s\n",
      "441:\tlearn: 9.3095366\ttotal: 4.58s\tremaining: 5.08s\n",
      "442:\tlearn: 9.3072154\ttotal: 4.59s\tremaining: 5.06s\n",
      "443:\tlearn: 9.3016415\ttotal: 4.59s\tremaining: 5.05s\n",
      "444:\tlearn: 9.2967939\ttotal: 4.6s\tremaining: 5.03s\n",
      "445:\tlearn: 9.2944446\ttotal: 4.61s\tremaining: 5.02s\n",
      "446:\tlearn: 9.2884870\ttotal: 4.61s\tremaining: 5s\n",
      "447:\tlearn: 9.2825422\ttotal: 4.62s\tremaining: 4.99s\n",
      "448:\tlearn: 9.2809320\ttotal: 4.63s\tremaining: 4.98s\n",
      "449:\tlearn: 9.2764069\ttotal: 4.64s\tremaining: 4.97s\n",
      "450:\tlearn: 9.2724964\ttotal: 4.64s\tremaining: 4.95s\n",
      "451:\tlearn: 9.2667885\ttotal: 4.65s\tremaining: 4.94s\n",
      "452:\tlearn: 9.2627865\ttotal: 4.66s\tremaining: 4.92s\n",
      "453:\tlearn: 9.2587872\ttotal: 4.66s\tremaining: 4.91s\n",
      "454:\tlearn: 9.2552966\ttotal: 4.67s\tremaining: 4.9s\n",
      "455:\tlearn: 9.2502474\ttotal: 4.68s\tremaining: 4.88s\n",
      "456:\tlearn: 9.2442380\ttotal: 4.69s\tremaining: 4.87s\n",
      "457:\tlearn: 9.2410114\ttotal: 4.69s\tremaining: 4.86s\n",
      "458:\tlearn: 9.2385160\ttotal: 4.71s\tremaining: 4.85s\n",
      "459:\tlearn: 9.2324293\ttotal: 4.71s\tremaining: 4.83s\n",
      "460:\tlearn: 9.2270824\ttotal: 4.72s\tremaining: 4.82s\n",
      "461:\tlearn: 9.2206146\ttotal: 4.73s\tremaining: 4.81s\n",
      "462:\tlearn: 9.2171097\ttotal: 4.74s\tremaining: 4.8s\n",
      "463:\tlearn: 9.2120434\ttotal: 4.74s\tremaining: 4.78s\n",
      "464:\tlearn: 9.2097098\ttotal: 4.75s\tremaining: 4.77s\n",
      "465:\tlearn: 9.2040223\ttotal: 4.76s\tremaining: 4.76s\n",
      "466:\tlearn: 9.2001944\ttotal: 4.76s\tremaining: 4.74s\n",
      "467:\tlearn: 9.1959739\ttotal: 4.77s\tremaining: 4.73s\n",
      "468:\tlearn: 9.1924733\ttotal: 4.78s\tremaining: 4.72s\n",
      "469:\tlearn: 9.1849285\ttotal: 4.79s\tremaining: 4.7s\n",
      "470:\tlearn: 9.1792069\ttotal: 4.79s\tremaining: 4.69s\n",
      "471:\tlearn: 9.1756036\ttotal: 4.8s\tremaining: 4.68s\n",
      "472:\tlearn: 9.1731494\ttotal: 4.8s\tremaining: 4.66s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "473:\tlearn: 9.1703070\ttotal: 4.81s\tremaining: 4.65s\n",
      "474:\tlearn: 9.1661579\ttotal: 4.83s\tremaining: 4.64s\n",
      "475:\tlearn: 9.1621109\ttotal: 4.83s\tremaining: 4.63s\n",
      "476:\tlearn: 9.1580151\ttotal: 4.85s\tremaining: 4.63s\n",
      "477:\tlearn: 9.1550422\ttotal: 4.87s\tremaining: 4.62s\n",
      "478:\tlearn: 9.1519972\ttotal: 4.87s\tremaining: 4.61s\n",
      "479:\tlearn: 9.1491499\ttotal: 4.88s\tremaining: 4.6s\n",
      "480:\tlearn: 9.1438382\ttotal: 4.89s\tremaining: 4.59s\n",
      "481:\tlearn: 9.1407554\ttotal: 4.9s\tremaining: 4.57s\n",
      "482:\tlearn: 9.1336372\ttotal: 4.91s\tremaining: 4.56s\n",
      "483:\tlearn: 9.1315558\ttotal: 4.92s\tremaining: 4.55s\n",
      "484:\tlearn: 9.1254190\ttotal: 4.92s\tremaining: 4.54s\n",
      "485:\tlearn: 9.1220294\ttotal: 4.94s\tremaining: 4.53s\n",
      "486:\tlearn: 9.1203883\ttotal: 4.95s\tremaining: 4.52s\n",
      "487:\tlearn: 9.1154069\ttotal: 4.96s\tremaining: 4.51s\n",
      "488:\tlearn: 9.1123527\ttotal: 4.96s\tremaining: 4.5s\n",
      "489:\tlearn: 9.1103554\ttotal: 4.97s\tremaining: 4.49s\n",
      "490:\tlearn: 9.1039902\ttotal: 4.98s\tremaining: 4.47s\n",
      "491:\tlearn: 9.1001281\ttotal: 4.99s\tremaining: 4.47s\n",
      "492:\tlearn: 9.0945978\ttotal: 5s\tremaining: 4.45s\n",
      "493:\tlearn: 9.0908420\ttotal: 5.01s\tremaining: 4.45s\n",
      "494:\tlearn: 9.0856512\ttotal: 5.02s\tremaining: 4.43s\n",
      "495:\tlearn: 9.0805077\ttotal: 5.04s\tremaining: 4.43s\n",
      "496:\tlearn: 9.0755092\ttotal: 5.04s\tremaining: 4.42s\n",
      "497:\tlearn: 9.0719306\ttotal: 5.05s\tremaining: 4.4s\n",
      "498:\tlearn: 9.0687671\ttotal: 5.06s\tremaining: 4.39s\n",
      "499:\tlearn: 9.0649860\ttotal: 5.06s\tremaining: 4.38s\n",
      "500:\tlearn: 9.0621724\ttotal: 5.08s\tremaining: 4.37s\n",
      "501:\tlearn: 9.0595418\ttotal: 5.08s\tremaining: 4.35s\n",
      "502:\tlearn: 9.0564364\ttotal: 5.09s\tremaining: 4.34s\n",
      "503:\tlearn: 9.0526514\ttotal: 5.1s\tremaining: 4.33s\n",
      "504:\tlearn: 9.0464865\ttotal: 5.12s\tremaining: 4.33s\n",
      "505:\tlearn: 9.0415899\ttotal: 5.13s\tremaining: 4.32s\n",
      "506:\tlearn: 9.0392643\ttotal: 5.14s\tremaining: 4.3s\n",
      "507:\tlearn: 9.0348589\ttotal: 5.14s\tremaining: 4.29s\n",
      "508:\tlearn: 9.0306802\ttotal: 5.15s\tremaining: 4.28s\n",
      "509:\tlearn: 9.0256795\ttotal: 5.17s\tremaining: 4.28s\n",
      "510:\tlearn: 9.0230631\ttotal: 5.18s\tremaining: 4.26s\n",
      "511:\tlearn: 9.0205142\ttotal: 5.19s\tremaining: 4.25s\n",
      "512:\tlearn: 9.0152274\ttotal: 5.19s\tremaining: 4.24s\n",
      "513:\tlearn: 9.0110746\ttotal: 5.21s\tremaining: 4.24s\n",
      "514:\tlearn: 9.0095089\ttotal: 5.23s\tremaining: 4.23s\n",
      "515:\tlearn: 9.0072245\ttotal: 5.24s\tremaining: 4.22s\n",
      "516:\tlearn: 9.0039932\ttotal: 5.24s\tremaining: 4.21s\n",
      "517:\tlearn: 9.0003492\ttotal: 5.25s\tremaining: 4.2s\n",
      "518:\tlearn: 8.9939949\ttotal: 5.26s\tremaining: 4.19s\n",
      "519:\tlearn: 8.9891373\ttotal: 5.28s\tremaining: 4.18s\n",
      "520:\tlearn: 8.9846410\ttotal: 5.29s\tremaining: 4.17s\n",
      "521:\tlearn: 8.9789515\ttotal: 5.3s\tremaining: 4.16s\n",
      "522:\tlearn: 8.9761355\ttotal: 5.3s\tremaining: 4.15s\n",
      "523:\tlearn: 8.9736231\ttotal: 5.31s\tremaining: 4.14s\n",
      "524:\tlearn: 8.9681449\ttotal: 5.32s\tremaining: 4.12s\n",
      "525:\tlearn: 8.9636320\ttotal: 5.33s\tremaining: 4.11s\n",
      "526:\tlearn: 8.9593347\ttotal: 5.34s\tremaining: 4.1s\n",
      "527:\tlearn: 8.9563199\ttotal: 5.34s\tremaining: 4.09s\n",
      "528:\tlearn: 8.9528389\ttotal: 5.35s\tremaining: 4.08s\n",
      "529:\tlearn: 8.9488815\ttotal: 5.36s\tremaining: 4.06s\n",
      "530:\tlearn: 8.9438493\ttotal: 5.37s\tremaining: 4.05s\n",
      "531:\tlearn: 8.9379039\ttotal: 5.38s\tremaining: 4.04s\n",
      "532:\tlearn: 8.9328512\ttotal: 5.38s\tremaining: 4.03s\n",
      "533:\tlearn: 8.9300445\ttotal: 5.4s\tremaining: 4.02s\n",
      "534:\tlearn: 8.9267673\ttotal: 5.41s\tremaining: 4.01s\n",
      "535:\tlearn: 8.9230075\ttotal: 5.42s\tremaining: 4s\n",
      "536:\tlearn: 8.9163529\ttotal: 5.42s\tremaining: 3.99s\n",
      "537:\tlearn: 8.9122788\ttotal: 5.43s\tremaining: 3.98s\n",
      "538:\tlearn: 8.9083250\ttotal: 5.44s\tremaining: 3.96s\n",
      "539:\tlearn: 8.9032088\ttotal: 5.45s\tremaining: 3.96s\n",
      "540:\tlearn: 8.8987874\ttotal: 5.46s\tremaining: 3.95s\n",
      "541:\tlearn: 8.8925127\ttotal: 5.47s\tremaining: 3.94s\n",
      "542:\tlearn: 8.8873902\ttotal: 5.48s\tremaining: 3.92s\n",
      "543:\tlearn: 8.8827664\ttotal: 5.5s\tremaining: 3.92s\n",
      "544:\tlearn: 8.8774940\ttotal: 5.51s\tremaining: 3.91s\n",
      "545:\tlearn: 8.8739580\ttotal: 5.52s\tremaining: 3.9s\n",
      "546:\tlearn: 8.8699088\ttotal: 5.53s\tremaining: 3.89s\n",
      "547:\tlearn: 8.8677234\ttotal: 5.54s\tremaining: 3.88s\n",
      "548:\tlearn: 8.8643154\ttotal: 5.55s\tremaining: 3.87s\n",
      "549:\tlearn: 8.8598805\ttotal: 5.55s\tremaining: 3.86s\n",
      "550:\tlearn: 8.8511322\ttotal: 5.56s\tremaining: 3.85s\n",
      "551:\tlearn: 8.8431777\ttotal: 5.57s\tremaining: 3.84s\n",
      "552:\tlearn: 8.8347219\ttotal: 5.58s\tremaining: 3.83s\n",
      "553:\tlearn: 8.8300163\ttotal: 5.6s\tremaining: 3.82s\n",
      "554:\tlearn: 8.8266750\ttotal: 5.61s\tremaining: 3.81s\n",
      "555:\tlearn: 8.8245159\ttotal: 5.62s\tremaining: 3.8s\n",
      "556:\tlearn: 8.8214029\ttotal: 5.63s\tremaining: 3.79s\n",
      "557:\tlearn: 8.8181866\ttotal: 5.64s\tremaining: 3.78s\n",
      "558:\tlearn: 8.8133354\ttotal: 5.64s\tremaining: 3.77s\n",
      "559:\tlearn: 8.8092076\ttotal: 5.65s\tremaining: 3.75s\n",
      "560:\tlearn: 8.8057163\ttotal: 5.67s\tremaining: 3.75s\n",
      "561:\tlearn: 8.8000195\ttotal: 5.67s\tremaining: 3.73s\n",
      "562:\tlearn: 8.7956215\ttotal: 5.68s\tremaining: 3.72s\n",
      "563:\tlearn: 8.7900657\ttotal: 5.69s\tremaining: 3.71s\n",
      "564:\tlearn: 8.7844662\ttotal: 5.7s\tremaining: 3.7s\n",
      "565:\tlearn: 8.7744229\ttotal: 5.71s\tremaining: 3.69s\n",
      "566:\tlearn: 8.7694485\ttotal: 5.72s\tremaining: 3.68s\n",
      "567:\tlearn: 8.7648534\ttotal: 5.73s\tremaining: 3.67s\n",
      "568:\tlearn: 8.7577750\ttotal: 5.75s\tremaining: 3.67s\n",
      "569:\tlearn: 8.7507439\ttotal: 5.77s\tremaining: 3.66s\n",
      "570:\tlearn: 8.7474456\ttotal: 5.78s\tremaining: 3.66s\n",
      "571:\tlearn: 8.7452785\ttotal: 5.79s\tremaining: 3.64s\n",
      "572:\tlearn: 8.7409792\ttotal: 5.8s\tremaining: 3.63s\n",
      "573:\tlearn: 8.7373424\ttotal: 5.81s\tremaining: 3.63s\n",
      "574:\tlearn: 8.7324346\ttotal: 5.83s\tremaining: 3.62s\n",
      "575:\tlearn: 8.7267667\ttotal: 5.84s\tremaining: 3.61s\n",
      "576:\tlearn: 8.7229572\ttotal: 5.85s\tremaining: 3.6s\n",
      "577:\tlearn: 8.7158378\ttotal: 5.85s\tremaining: 3.58s\n",
      "578:\tlearn: 8.7077250\ttotal: 5.86s\tremaining: 3.57s\n",
      "579:\tlearn: 8.7017835\ttotal: 5.87s\tremaining: 3.56s\n",
      "580:\tlearn: 8.6988596\ttotal: 5.88s\tremaining: 3.55s\n",
      "581:\tlearn: 8.6949923\ttotal: 5.88s\tremaining: 3.54s\n",
      "582:\tlearn: 8.6910667\ttotal: 5.91s\tremaining: 3.54s\n",
      "583:\tlearn: 8.6877749\ttotal: 5.93s\tremaining: 3.53s\n",
      "584:\tlearn: 8.6840617\ttotal: 5.94s\tremaining: 3.52s\n",
      "585:\tlearn: 8.6798840\ttotal: 5.95s\tremaining: 3.51s\n",
      "586:\tlearn: 8.6755857\ttotal: 5.96s\tremaining: 3.5s\n",
      "587:\tlearn: 8.6727612\ttotal: 5.96s\tremaining: 3.49s\n",
      "588:\tlearn: 8.6662215\ttotal: 5.97s\tremaining: 3.48s\n",
      "589:\tlearn: 8.6622316\ttotal: 5.98s\tremaining: 3.46s\n",
      "590:\tlearn: 8.6594355\ttotal: 5.98s\tremaining: 3.45s\n",
      "591:\tlearn: 8.6556854\ttotal: 5.99s\tremaining: 3.44s\n",
      "592:\tlearn: 8.6529157\ttotal: 6s\tremaining: 3.43s\n",
      "593:\tlearn: 8.6483933\ttotal: 6.01s\tremaining: 3.42s\n",
      "594:\tlearn: 8.6433291\ttotal: 6.01s\tremaining: 3.41s\n",
      "595:\tlearn: 8.6402432\ttotal: 6.02s\tremaining: 3.39s\n",
      "596:\tlearn: 8.6366401\ttotal: 6.04s\tremaining: 3.39s\n",
      "597:\tlearn: 8.6325519\ttotal: 6.05s\tremaining: 3.38s\n",
      "598:\tlearn: 8.6268026\ttotal: 6.06s\tremaining: 3.37s\n",
      "599:\tlearn: 8.6216739\ttotal: 6.07s\tremaining: 3.36s\n",
      "600:\tlearn: 8.6192689\ttotal: 6.08s\tremaining: 3.35s\n",
      "601:\tlearn: 8.6137642\ttotal: 6.1s\tremaining: 3.34s\n",
      "602:\tlearn: 8.6101534\ttotal: 6.1s\tremaining: 3.33s\n",
      "603:\tlearn: 8.6077196\ttotal: 6.11s\tremaining: 3.32s\n",
      "604:\tlearn: 8.6043403\ttotal: 6.12s\tremaining: 3.31s\n",
      "605:\tlearn: 8.6023765\ttotal: 6.13s\tremaining: 3.29s\n",
      "606:\tlearn: 8.5970510\ttotal: 6.13s\tremaining: 3.28s\n",
      "607:\tlearn: 8.5909212\ttotal: 6.14s\tremaining: 3.27s\n",
      "608:\tlearn: 8.5870917\ttotal: 6.15s\tremaining: 3.26s\n",
      "609:\tlearn: 8.5824302\ttotal: 6.16s\tremaining: 3.25s\n",
      "610:\tlearn: 8.5803122\ttotal: 6.17s\tremaining: 3.24s\n",
      "611:\tlearn: 8.5745863\ttotal: 6.17s\tremaining: 3.23s\n",
      "612:\tlearn: 8.5694601\ttotal: 6.18s\tremaining: 3.22s\n",
      "613:\tlearn: 8.5626962\ttotal: 6.19s\tremaining: 3.2s\n",
      "614:\tlearn: 8.5578109\ttotal: 6.21s\tremaining: 3.2s\n",
      "615:\tlearn: 8.5536172\ttotal: 6.21s\tremaining: 3.19s\n",
      "616:\tlearn: 8.5471633\ttotal: 6.22s\tremaining: 3.18s\n",
      "617:\tlearn: 8.5424822\ttotal: 6.23s\tremaining: 3.17s\n",
      "618:\tlearn: 8.5387538\ttotal: 6.24s\tremaining: 3.15s\n",
      "619:\tlearn: 8.5323416\ttotal: 6.25s\tremaining: 3.14s\n",
      "620:\tlearn: 8.5292634\ttotal: 6.25s\tremaining: 3.13s\n",
      "621:\tlearn: 8.5277088\ttotal: 6.26s\tremaining: 3.12s\n",
      "622:\tlearn: 8.5230071\ttotal: 6.27s\tremaining: 3.11s\n",
      "623:\tlearn: 8.5186171\ttotal: 6.28s\tremaining: 3.1s\n",
      "624:\tlearn: 8.5156898\ttotal: 6.3s\tremaining: 3.09s\n",
      "625:\tlearn: 8.5107597\ttotal: 6.31s\tremaining: 3.08s\n",
      "626:\tlearn: 8.5081637\ttotal: 6.32s\tremaining: 3.07s\n",
      "627:\tlearn: 8.5009159\ttotal: 6.32s\tremaining: 3.06s\n",
      "628:\tlearn: 8.4988016\ttotal: 6.33s\tremaining: 3.05s\n",
      "629:\tlearn: 8.4967620\ttotal: 6.34s\tremaining: 3.04s\n",
      "630:\tlearn: 8.4887145\ttotal: 6.35s\tremaining: 3.03s\n",
      "631:\tlearn: 8.4828755\ttotal: 6.36s\tremaining: 3.02s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "632:\tlearn: 8.4798044\ttotal: 6.37s\tremaining: 3.01s\n",
      "633:\tlearn: 8.4714032\ttotal: 6.38s\tremaining: 3s\n",
      "634:\tlearn: 8.4652604\ttotal: 6.38s\tremaining: 2.98s\n",
      "635:\tlearn: 8.4603997\ttotal: 6.39s\tremaining: 2.98s\n",
      "636:\tlearn: 8.4577884\ttotal: 6.4s\tremaining: 2.96s\n",
      "637:\tlearn: 8.4553167\ttotal: 6.41s\tremaining: 2.95s\n",
      "638:\tlearn: 8.4498246\ttotal: 6.42s\tremaining: 2.94s\n",
      "639:\tlearn: 8.4465806\ttotal: 6.43s\tremaining: 2.93s\n",
      "640:\tlearn: 8.4439240\ttotal: 6.43s\tremaining: 2.92s\n",
      "641:\tlearn: 8.4397593\ttotal: 6.44s\tremaining: 2.91s\n",
      "642:\tlearn: 8.4335196\ttotal: 6.45s\tremaining: 2.9s\n",
      "643:\tlearn: 8.4274846\ttotal: 6.45s\tremaining: 2.89s\n",
      "644:\tlearn: 8.4237082\ttotal: 6.46s\tremaining: 2.88s\n",
      "645:\tlearn: 8.4216674\ttotal: 6.47s\tremaining: 2.87s\n",
      "646:\tlearn: 8.4153697\ttotal: 6.48s\tremaining: 2.85s\n",
      "647:\tlearn: 8.4120193\ttotal: 6.49s\tremaining: 2.84s\n",
      "648:\tlearn: 8.4085835\ttotal: 6.49s\tremaining: 2.83s\n",
      "649:\tlearn: 8.4047450\ttotal: 6.5s\tremaining: 2.82s\n",
      "650:\tlearn: 8.4001491\ttotal: 6.51s\tremaining: 2.81s\n",
      "651:\tlearn: 8.3948627\ttotal: 6.52s\tremaining: 2.8s\n",
      "652:\tlearn: 8.3888940\ttotal: 6.52s\tremaining: 2.79s\n",
      "653:\tlearn: 8.3833876\ttotal: 6.53s\tremaining: 2.78s\n",
      "654:\tlearn: 8.3794843\ttotal: 6.54s\tremaining: 2.76s\n",
      "655:\tlearn: 8.3750381\ttotal: 6.54s\tremaining: 2.75s\n",
      "656:\tlearn: 8.3704198\ttotal: 6.55s\tremaining: 2.74s\n",
      "657:\tlearn: 8.3686835\ttotal: 6.56s\tremaining: 2.73s\n",
      "658:\tlearn: 8.3645606\ttotal: 6.57s\tremaining: 2.72s\n",
      "659:\tlearn: 8.3624764\ttotal: 6.58s\tremaining: 2.71s\n",
      "660:\tlearn: 8.3595329\ttotal: 6.59s\tremaining: 2.7s\n",
      "661:\tlearn: 8.3536529\ttotal: 6.6s\tremaining: 2.69s\n",
      "662:\tlearn: 8.3503132\ttotal: 6.61s\tremaining: 2.68s\n",
      "663:\tlearn: 8.3459817\ttotal: 6.61s\tremaining: 2.67s\n",
      "664:\tlearn: 8.3426125\ttotal: 6.62s\tremaining: 2.66s\n",
      "665:\tlearn: 8.3397262\ttotal: 6.63s\tremaining: 2.65s\n",
      "666:\tlearn: 8.3355911\ttotal: 6.64s\tremaining: 2.64s\n",
      "667:\tlearn: 8.3305989\ttotal: 6.64s\tremaining: 2.63s\n",
      "668:\tlearn: 8.3269915\ttotal: 6.65s\tremaining: 2.61s\n",
      "669:\tlearn: 8.3236725\ttotal: 6.66s\tremaining: 2.6s\n",
      "670:\tlearn: 8.3188584\ttotal: 6.67s\tremaining: 2.59s\n",
      "671:\tlearn: 8.3140435\ttotal: 6.68s\tremaining: 2.58s\n",
      "672:\tlearn: 8.3098722\ttotal: 6.69s\tremaining: 2.57s\n",
      "673:\tlearn: 8.3051934\ttotal: 6.7s\tremaining: 2.56s\n",
      "674:\tlearn: 8.3027613\ttotal: 6.7s\tremaining: 2.55s\n",
      "675:\tlearn: 8.2956423\ttotal: 6.71s\tremaining: 2.54s\n",
      "676:\tlearn: 8.2916028\ttotal: 6.72s\tremaining: 2.53s\n",
      "677:\tlearn: 8.2881042\ttotal: 6.74s\tremaining: 2.52s\n",
      "678:\tlearn: 8.2836928\ttotal: 6.74s\tremaining: 2.51s\n",
      "679:\tlearn: 8.2808750\ttotal: 6.75s\tremaining: 2.5s\n",
      "680:\tlearn: 8.2778807\ttotal: 6.76s\tremaining: 2.49s\n",
      "681:\tlearn: 8.2747510\ttotal: 6.77s\tremaining: 2.48s\n",
      "682:\tlearn: 8.2715973\ttotal: 6.79s\tremaining: 2.48s\n",
      "683:\tlearn: 8.2690204\ttotal: 6.8s\tremaining: 2.47s\n",
      "684:\tlearn: 8.2647137\ttotal: 6.83s\tremaining: 2.46s\n",
      "685:\tlearn: 8.2617786\ttotal: 6.84s\tremaining: 2.45s\n",
      "686:\tlearn: 8.2582138\ttotal: 6.85s\tremaining: 2.44s\n",
      "687:\tlearn: 8.2541903\ttotal: 6.86s\tremaining: 2.43s\n",
      "688:\tlearn: 8.2515752\ttotal: 6.88s\tremaining: 2.42s\n",
      "689:\tlearn: 8.2483839\ttotal: 6.89s\tremaining: 2.42s\n",
      "690:\tlearn: 8.2437724\ttotal: 6.89s\tremaining: 2.4s\n",
      "691:\tlearn: 8.2404626\ttotal: 6.9s\tremaining: 2.39s\n",
      "692:\tlearn: 8.2373589\ttotal: 6.92s\tremaining: 2.39s\n",
      "693:\tlearn: 8.2341745\ttotal: 6.93s\tremaining: 2.38s\n",
      "694:\tlearn: 8.2316106\ttotal: 6.94s\tremaining: 2.37s\n",
      "695:\tlearn: 8.2270010\ttotal: 6.94s\tremaining: 2.35s\n",
      "696:\tlearn: 8.2255593\ttotal: 6.96s\tremaining: 2.35s\n",
      "697:\tlearn: 8.2212242\ttotal: 6.97s\tremaining: 2.34s\n",
      "698:\tlearn: 8.2167379\ttotal: 6.98s\tremaining: 2.33s\n",
      "699:\tlearn: 8.2143553\ttotal: 6.99s\tremaining: 2.32s\n",
      "700:\tlearn: 8.2100617\ttotal: 7s\tremaining: 2.31s\n",
      "701:\tlearn: 8.2068788\ttotal: 7.01s\tremaining: 2.3s\n",
      "702:\tlearn: 8.2045610\ttotal: 7.02s\tremaining: 2.29s\n",
      "703:\tlearn: 8.2018471\ttotal: 7.02s\tremaining: 2.27s\n",
      "704:\tlearn: 8.1989792\ttotal: 7.03s\tremaining: 2.26s\n",
      "705:\tlearn: 8.1956272\ttotal: 7.04s\tremaining: 2.25s\n",
      "706:\tlearn: 8.1912283\ttotal: 7.05s\tremaining: 2.24s\n",
      "707:\tlearn: 8.1891755\ttotal: 7.06s\tremaining: 2.23s\n",
      "708:\tlearn: 8.1860045\ttotal: 7.07s\tremaining: 2.22s\n",
      "709:\tlearn: 8.1832435\ttotal: 7.07s\tremaining: 2.21s\n",
      "710:\tlearn: 8.1788418\ttotal: 7.08s\tremaining: 2.2s\n",
      "711:\tlearn: 8.1755244\ttotal: 7.09s\tremaining: 2.19s\n",
      "712:\tlearn: 8.1725911\ttotal: 7.1s\tremaining: 2.18s\n",
      "713:\tlearn: 8.1663908\ttotal: 7.11s\tremaining: 2.17s\n",
      "714:\tlearn: 8.1607202\ttotal: 7.11s\tremaining: 2.16s\n",
      "715:\tlearn: 8.1581005\ttotal: 7.12s\tremaining: 2.15s\n",
      "716:\tlearn: 8.1539084\ttotal: 7.13s\tremaining: 2.14s\n",
      "717:\tlearn: 8.1493325\ttotal: 7.14s\tremaining: 2.13s\n",
      "718:\tlearn: 8.1477970\ttotal: 7.14s\tremaining: 2.12s\n",
      "719:\tlearn: 8.1450116\ttotal: 7.15s\tremaining: 2.1s\n",
      "720:\tlearn: 8.1421790\ttotal: 7.17s\tremaining: 2.1s\n",
      "721:\tlearn: 8.1368959\ttotal: 7.17s\tremaining: 2.09s\n",
      "722:\tlearn: 8.1354740\ttotal: 7.18s\tremaining: 2.08s\n",
      "723:\tlearn: 8.1313759\ttotal: 7.19s\tremaining: 2.07s\n",
      "724:\tlearn: 8.1290117\ttotal: 7.2s\tremaining: 2.06s\n",
      "725:\tlearn: 8.1262809\ttotal: 7.21s\tremaining: 2.04s\n",
      "726:\tlearn: 8.1239648\ttotal: 7.21s\tremaining: 2.03s\n",
      "727:\tlearn: 8.1194426\ttotal: 7.22s\tremaining: 2.02s\n",
      "728:\tlearn: 8.1159876\ttotal: 7.23s\tremaining: 2.01s\n",
      "729:\tlearn: 8.1140805\ttotal: 7.24s\tremaining: 2s\n",
      "730:\tlearn: 8.1097159\ttotal: 7.25s\tremaining: 1.99s\n",
      "731:\tlearn: 8.1051945\ttotal: 7.29s\tremaining: 1.99s\n",
      "732:\tlearn: 8.1033533\ttotal: 7.29s\tremaining: 1.98s\n",
      "733:\tlearn: 8.1006723\ttotal: 7.3s\tremaining: 1.97s\n",
      "734:\tlearn: 8.0977355\ttotal: 7.31s\tremaining: 1.96s\n",
      "735:\tlearn: 8.0946820\ttotal: 7.32s\tremaining: 1.95s\n",
      "736:\tlearn: 8.0903926\ttotal: 7.33s\tremaining: 1.94s\n",
      "737:\tlearn: 8.0866642\ttotal: 7.33s\tremaining: 1.93s\n",
      "738:\tlearn: 8.0846825\ttotal: 7.34s\tremaining: 1.92s\n",
      "739:\tlearn: 8.0796578\ttotal: 7.35s\tremaining: 1.91s\n",
      "740:\tlearn: 8.0755698\ttotal: 7.38s\tremaining: 1.9s\n",
      "741:\tlearn: 8.0731168\ttotal: 7.39s\tremaining: 1.89s\n",
      "742:\tlearn: 8.0697118\ttotal: 7.39s\tremaining: 1.88s\n",
      "743:\tlearn: 8.0665762\ttotal: 7.4s\tremaining: 1.87s\n",
      "744:\tlearn: 8.0635638\ttotal: 7.41s\tremaining: 1.86s\n",
      "745:\tlearn: 8.0592811\ttotal: 7.42s\tremaining: 1.85s\n",
      "746:\tlearn: 8.0560077\ttotal: 7.43s\tremaining: 1.84s\n",
      "747:\tlearn: 8.0537498\ttotal: 7.45s\tremaining: 1.83s\n",
      "748:\tlearn: 8.0478773\ttotal: 7.46s\tremaining: 1.82s\n",
      "749:\tlearn: 8.0455133\ttotal: 7.47s\tremaining: 1.81s\n",
      "750:\tlearn: 8.0419591\ttotal: 7.48s\tremaining: 1.8s\n",
      "751:\tlearn: 8.0404461\ttotal: 7.49s\tremaining: 1.79s\n",
      "752:\tlearn: 8.0389259\ttotal: 7.5s\tremaining: 1.78s\n",
      "753:\tlearn: 8.0331423\ttotal: 7.5s\tremaining: 1.77s\n",
      "754:\tlearn: 8.0304819\ttotal: 7.52s\tremaining: 1.76s\n",
      "755:\tlearn: 8.0292447\ttotal: 7.52s\tremaining: 1.75s\n",
      "756:\tlearn: 8.0260877\ttotal: 7.53s\tremaining: 1.74s\n",
      "757:\tlearn: 8.0197305\ttotal: 7.54s\tremaining: 1.73s\n",
      "758:\tlearn: 8.0166647\ttotal: 7.55s\tremaining: 1.72s\n",
      "759:\tlearn: 8.0096606\ttotal: 7.56s\tremaining: 1.71s\n",
      "760:\tlearn: 8.0074116\ttotal: 7.57s\tremaining: 1.7s\n",
      "761:\tlearn: 8.0032682\ttotal: 7.58s\tremaining: 1.69s\n",
      "762:\tlearn: 8.0007120\ttotal: 7.58s\tremaining: 1.68s\n",
      "763:\tlearn: 7.9991095\ttotal: 7.59s\tremaining: 1.67s\n",
      "764:\tlearn: 7.9968349\ttotal: 7.6s\tremaining: 1.66s\n",
      "765:\tlearn: 7.9930029\ttotal: 7.61s\tremaining: 1.65s\n",
      "766:\tlearn: 7.9904980\ttotal: 7.62s\tremaining: 1.64s\n",
      "767:\tlearn: 7.9879004\ttotal: 7.63s\tremaining: 1.63s\n",
      "768:\tlearn: 7.9834119\ttotal: 7.63s\tremaining: 1.62s\n",
      "769:\tlearn: 7.9778536\ttotal: 7.64s\tremaining: 1.61s\n",
      "770:\tlearn: 7.9720182\ttotal: 7.65s\tremaining: 1.6s\n",
      "771:\tlearn: 7.9695145\ttotal: 7.66s\tremaining: 1.59s\n",
      "772:\tlearn: 7.9677405\ttotal: 7.67s\tremaining: 1.58s\n",
      "773:\tlearn: 7.9643198\ttotal: 7.67s\tremaining: 1.57s\n",
      "774:\tlearn: 7.9615842\ttotal: 7.68s\tremaining: 1.56s\n",
      "775:\tlearn: 7.9571152\ttotal: 7.69s\tremaining: 1.54s\n",
      "776:\tlearn: 7.9526515\ttotal: 7.7s\tremaining: 1.54s\n",
      "777:\tlearn: 7.9492951\ttotal: 7.71s\tremaining: 1.52s\n",
      "778:\tlearn: 7.9461576\ttotal: 7.72s\tremaining: 1.52s\n",
      "779:\tlearn: 7.9421406\ttotal: 7.73s\tremaining: 1.5s\n",
      "780:\tlearn: 7.9369489\ttotal: 7.74s\tremaining: 1.5s\n",
      "781:\tlearn: 7.9353568\ttotal: 7.75s\tremaining: 1.49s\n",
      "782:\tlearn: 7.9271754\ttotal: 7.76s\tremaining: 1.48s\n",
      "783:\tlearn: 7.9232724\ttotal: 7.77s\tremaining: 1.47s\n",
      "784:\tlearn: 7.9211503\ttotal: 7.77s\tremaining: 1.46s\n",
      "785:\tlearn: 7.9196679\ttotal: 7.78s\tremaining: 1.44s\n",
      "786:\tlearn: 7.9139172\ttotal: 7.8s\tremaining: 1.44s\n",
      "787:\tlearn: 7.9124453\ttotal: 7.8s\tremaining: 1.43s\n",
      "788:\tlearn: 7.9095647\ttotal: 7.82s\tremaining: 1.42s\n",
      "789:\tlearn: 7.9059174\ttotal: 7.83s\tremaining: 1.41s\n",
      "790:\tlearn: 7.9027465\ttotal: 7.85s\tremaining: 1.4s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "791:\tlearn: 7.9015952\ttotal: 7.85s\tremaining: 1.39s\n",
      "792:\tlearn: 7.8974260\ttotal: 7.86s\tremaining: 1.38s\n",
      "793:\tlearn: 7.8915307\ttotal: 7.87s\tremaining: 1.37s\n",
      "794:\tlearn: 7.8892523\ttotal: 7.87s\tremaining: 1.36s\n",
      "795:\tlearn: 7.8852738\ttotal: 7.89s\tremaining: 1.35s\n",
      "796:\tlearn: 7.8806836\ttotal: 7.89s\tremaining: 1.34s\n",
      "797:\tlearn: 7.8777468\ttotal: 7.9s\tremaining: 1.33s\n",
      "798:\tlearn: 7.8736391\ttotal: 7.91s\tremaining: 1.32s\n",
      "799:\tlearn: 7.8688762\ttotal: 7.93s\tremaining: 1.31s\n",
      "800:\tlearn: 7.8667452\ttotal: 7.93s\tremaining: 1.3s\n",
      "801:\tlearn: 7.8630616\ttotal: 7.94s\tremaining: 1.29s\n",
      "802:\tlearn: 7.8608954\ttotal: 7.95s\tremaining: 1.28s\n",
      "803:\tlearn: 7.8582706\ttotal: 7.96s\tremaining: 1.27s\n",
      "804:\tlearn: 7.8534288\ttotal: 7.96s\tremaining: 1.26s\n",
      "805:\tlearn: 7.8498584\ttotal: 7.97s\tremaining: 1.25s\n",
      "806:\tlearn: 7.8458052\ttotal: 7.98s\tremaining: 1.24s\n",
      "807:\tlearn: 7.8427159\ttotal: 7.99s\tremaining: 1.23s\n",
      "808:\tlearn: 7.8376283\ttotal: 8s\tremaining: 1.22s\n",
      "809:\tlearn: 7.8346193\ttotal: 8s\tremaining: 1.21s\n",
      "810:\tlearn: 7.8302836\ttotal: 8.01s\tremaining: 1.19s\n",
      "811:\tlearn: 7.8270462\ttotal: 8.01s\tremaining: 1.18s\n",
      "812:\tlearn: 7.8240604\ttotal: 8.02s\tremaining: 1.17s\n",
      "813:\tlearn: 7.8210832\ttotal: 8.03s\tremaining: 1.16s\n",
      "814:\tlearn: 7.8163222\ttotal: 8.04s\tremaining: 1.15s\n",
      "815:\tlearn: 7.8119643\ttotal: 8.05s\tremaining: 1.14s\n",
      "816:\tlearn: 7.8070397\ttotal: 8.06s\tremaining: 1.13s\n",
      "817:\tlearn: 7.8023774\ttotal: 8.07s\tremaining: 1.12s\n",
      "818:\tlearn: 7.8000355\ttotal: 8.08s\tremaining: 1.11s\n",
      "819:\tlearn: 7.7981131\ttotal: 8.08s\tremaining: 1.1s\n",
      "820:\tlearn: 7.7959250\ttotal: 8.09s\tremaining: 1.09s\n",
      "821:\tlearn: 7.7907141\ttotal: 8.1s\tremaining: 1.08s\n",
      "822:\tlearn: 7.7868136\ttotal: 8.1s\tremaining: 1.07s\n",
      "823:\tlearn: 7.7828677\ttotal: 8.11s\tremaining: 1.06s\n",
      "824:\tlearn: 7.7790400\ttotal: 8.12s\tremaining: 1.05s\n",
      "825:\tlearn: 7.7754392\ttotal: 8.13s\tremaining: 1.04s\n",
      "826:\tlearn: 7.7739582\ttotal: 8.13s\tremaining: 1.03s\n",
      "827:\tlearn: 7.7709443\ttotal: 8.15s\tremaining: 1.02s\n",
      "828:\tlearn: 7.7673160\ttotal: 8.15s\tremaining: 1.01s\n",
      "829:\tlearn: 7.7641072\ttotal: 8.16s\tremaining: 1s\n",
      "830:\tlearn: 7.7609999\ttotal: 8.17s\tremaining: 993ms\n",
      "831:\tlearn: 7.7595578\ttotal: 8.18s\tremaining: 983ms\n",
      "832:\tlearn: 7.7582539\ttotal: 8.19s\tremaining: 973ms\n",
      "833:\tlearn: 7.7571985\ttotal: 8.19s\tremaining: 963ms\n",
      "834:\tlearn: 7.7538992\ttotal: 8.2s\tremaining: 953ms\n",
      "835:\tlearn: 7.7516871\ttotal: 8.21s\tremaining: 943ms\n",
      "836:\tlearn: 7.7470965\ttotal: 8.22s\tremaining: 933ms\n",
      "837:\tlearn: 7.7437719\ttotal: 8.23s\tremaining: 924ms\n",
      "838:\tlearn: 7.7420636\ttotal: 8.24s\tremaining: 914ms\n",
      "839:\tlearn: 7.7400451\ttotal: 8.25s\tremaining: 904ms\n",
      "840:\tlearn: 7.7381887\ttotal: 8.26s\tremaining: 894ms\n",
      "841:\tlearn: 7.7345058\ttotal: 8.27s\tremaining: 884ms\n",
      "842:\tlearn: 7.7306142\ttotal: 8.28s\tremaining: 874ms\n",
      "843:\tlearn: 7.7276511\ttotal: 8.28s\tremaining: 864ms\n",
      "844:\tlearn: 7.7255544\ttotal: 8.29s\tremaining: 854ms\n",
      "845:\tlearn: 7.7212265\ttotal: 8.31s\tremaining: 845ms\n",
      "846:\tlearn: 7.7167015\ttotal: 8.32s\tremaining: 835ms\n",
      "847:\tlearn: 7.7139369\ttotal: 8.33s\tremaining: 825ms\n",
      "848:\tlearn: 7.7101690\ttotal: 8.34s\tremaining: 815ms\n",
      "849:\tlearn: 7.7060849\ttotal: 8.37s\tremaining: 808ms\n",
      "850:\tlearn: 7.7017704\ttotal: 8.38s\tremaining: 798ms\n",
      "851:\tlearn: 7.6988098\ttotal: 8.43s\tremaining: 791ms\n",
      "852:\tlearn: 7.6982926\ttotal: 8.44s\tremaining: 781ms\n",
      "853:\tlearn: 7.6953119\ttotal: 8.44s\tremaining: 771ms\n",
      "854:\tlearn: 7.6899909\ttotal: 8.45s\tremaining: 761ms\n",
      "855:\tlearn: 7.6863334\ttotal: 8.46s\tremaining: 751ms\n",
      "856:\tlearn: 7.6837697\ttotal: 8.47s\tremaining: 741ms\n",
      "857:\tlearn: 7.6780774\ttotal: 8.48s\tremaining: 731ms\n",
      "858:\tlearn: 7.6759084\ttotal: 8.48s\tremaining: 721ms\n",
      "859:\tlearn: 7.6723885\ttotal: 8.49s\tremaining: 711ms\n",
      "860:\tlearn: 7.6696759\ttotal: 8.51s\tremaining: 702ms\n",
      "861:\tlearn: 7.6661175\ttotal: 8.52s\tremaining: 692ms\n",
      "862:\tlearn: 7.6644219\ttotal: 8.53s\tremaining: 682ms\n",
      "863:\tlearn: 7.6622548\ttotal: 8.54s\tremaining: 672ms\n",
      "864:\tlearn: 7.6580900\ttotal: 8.55s\tremaining: 662ms\n",
      "865:\tlearn: 7.6550488\ttotal: 8.56s\tremaining: 652ms\n",
      "866:\tlearn: 7.6520509\ttotal: 8.56s\tremaining: 642ms\n",
      "867:\tlearn: 7.6480946\ttotal: 8.57s\tremaining: 632ms\n",
      "868:\tlearn: 7.6445792\ttotal: 8.58s\tremaining: 622ms\n",
      "869:\tlearn: 7.6385360\ttotal: 8.59s\tremaining: 612ms\n",
      "870:\tlearn: 7.6303278\ttotal: 8.6s\tremaining: 602ms\n",
      "871:\tlearn: 7.6263729\ttotal: 8.61s\tremaining: 592ms\n",
      "872:\tlearn: 7.6229590\ttotal: 8.62s\tremaining: 582ms\n",
      "873:\tlearn: 7.6195842\ttotal: 8.63s\tremaining: 572ms\n",
      "874:\tlearn: 7.6169880\ttotal: 8.63s\tremaining: 562ms\n",
      "875:\tlearn: 7.6152290\ttotal: 8.64s\tremaining: 552ms\n",
      "876:\tlearn: 7.6104714\ttotal: 8.65s\tremaining: 543ms\n",
      "877:\tlearn: 7.6078751\ttotal: 8.66s\tremaining: 532ms\n",
      "878:\tlearn: 7.6042435\ttotal: 8.66s\tremaining: 522ms\n",
      "879:\tlearn: 7.6011166\ttotal: 8.67s\tremaining: 512ms\n",
      "880:\tlearn: 7.5983164\ttotal: 8.68s\tremaining: 502ms\n",
      "881:\tlearn: 7.5975026\ttotal: 8.68s\tremaining: 492ms\n",
      "882:\tlearn: 7.5929841\ttotal: 8.69s\tremaining: 482ms\n",
      "883:\tlearn: 7.5906263\ttotal: 8.7s\tremaining: 472ms\n",
      "884:\tlearn: 7.5873401\ttotal: 8.71s\tremaining: 463ms\n",
      "885:\tlearn: 7.5833934\ttotal: 8.72s\tremaining: 453ms\n",
      "886:\tlearn: 7.5798004\ttotal: 8.72s\tremaining: 443ms\n",
      "887:\tlearn: 7.5763052\ttotal: 8.73s\tremaining: 433ms\n",
      "888:\tlearn: 7.5744666\ttotal: 8.74s\tremaining: 423ms\n",
      "889:\tlearn: 7.5705174\ttotal: 8.74s\tremaining: 413ms\n",
      "890:\tlearn: 7.5672364\ttotal: 8.75s\tremaining: 403ms\n",
      "891:\tlearn: 7.5637232\ttotal: 8.76s\tremaining: 393ms\n",
      "892:\tlearn: 7.5612869\ttotal: 8.77s\tremaining: 383ms\n",
      "893:\tlearn: 7.5570351\ttotal: 8.78s\tremaining: 373ms\n",
      "894:\tlearn: 7.5546300\ttotal: 8.78s\tremaining: 363ms\n",
      "895:\tlearn: 7.5518131\ttotal: 8.79s\tremaining: 353ms\n",
      "896:\tlearn: 7.5508350\ttotal: 8.8s\tremaining: 343ms\n",
      "897:\tlearn: 7.5491054\ttotal: 8.81s\tremaining: 334ms\n",
      "898:\tlearn: 7.5455393\ttotal: 8.82s\tremaining: 324ms\n",
      "899:\tlearn: 7.5419805\ttotal: 8.83s\tremaining: 314ms\n",
      "900:\tlearn: 7.5382265\ttotal: 8.84s\tremaining: 304ms\n",
      "901:\tlearn: 7.5358410\ttotal: 8.84s\tremaining: 294ms\n",
      "902:\tlearn: 7.5330890\ttotal: 8.85s\tremaining: 284ms\n",
      "903:\tlearn: 7.5298171\ttotal: 8.86s\tremaining: 274ms\n",
      "904:\tlearn: 7.5289958\ttotal: 8.86s\tremaining: 264ms\n",
      "905:\tlearn: 7.5273479\ttotal: 8.88s\tremaining: 255ms\n",
      "906:\tlearn: 7.5244265\ttotal: 8.88s\tremaining: 245ms\n",
      "907:\tlearn: 7.5206568\ttotal: 8.89s\tremaining: 235ms\n",
      "908:\tlearn: 7.5179278\ttotal: 8.9s\tremaining: 225ms\n",
      "909:\tlearn: 7.5158292\ttotal: 8.91s\tremaining: 215ms\n",
      "910:\tlearn: 7.5130584\ttotal: 8.92s\tremaining: 206ms\n",
      "911:\tlearn: 7.5088473\ttotal: 8.93s\tremaining: 196ms\n",
      "912:\tlearn: 7.5065337\ttotal: 8.94s\tremaining: 186ms\n",
      "913:\tlearn: 7.5046294\ttotal: 8.95s\tremaining: 176ms\n",
      "914:\tlearn: 7.5018966\ttotal: 8.96s\tremaining: 166ms\n",
      "915:\tlearn: 7.4998831\ttotal: 8.97s\tremaining: 157ms\n",
      "916:\tlearn: 7.4953738\ttotal: 8.97s\tremaining: 147ms\n",
      "917:\tlearn: 7.4917315\ttotal: 8.98s\tremaining: 137ms\n",
      "918:\tlearn: 7.4902159\ttotal: 8.99s\tremaining: 127ms\n",
      "919:\tlearn: 7.4881519\ttotal: 9s\tremaining: 117ms\n",
      "920:\tlearn: 7.4839501\ttotal: 9.01s\tremaining: 108ms\n",
      "921:\tlearn: 7.4821224\ttotal: 9.02s\tremaining: 97.9ms\n",
      "922:\tlearn: 7.4771872\ttotal: 9.03s\tremaining: 88ms\n",
      "923:\tlearn: 7.4748385\ttotal: 9.03s\tremaining: 78.2ms\n",
      "924:\tlearn: 7.4695537\ttotal: 9.04s\tremaining: 68.4ms\n",
      "925:\tlearn: 7.4671940\ttotal: 9.05s\tremaining: 58.7ms\n",
      "926:\tlearn: 7.4629664\ttotal: 9.06s\tremaining: 48.9ms\n",
      "927:\tlearn: 7.4587184\ttotal: 9.07s\tremaining: 39.1ms\n",
      "928:\tlearn: 7.4541413\ttotal: 9.08s\tremaining: 29.3ms\n",
      "929:\tlearn: 7.4511682\ttotal: 9.09s\tremaining: 19.5ms\n",
      "930:\tlearn: 7.4486106\ttotal: 9.1s\tremaining: 9.77ms\n",
      "931:\tlearn: 7.4469043\ttotal: 9.1s\tremaining: 0us\n",
      "[LightGBM] [Warning] boosting is set=dart, boosting_type=gbdt will be ignored. Current value: boosting=dart\n",
      "[LightGBM] [Warning] min_data_in_leaf is set=26, min_child_samples=20 will be ignored. Current value: min_data_in_leaf=26\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>BiasInfluentialFactor</th>\n",
       "      <th>Tune_XGBoost_Pred</th>\n",
       "      <th>Tune_catoost_Pred</th>\n",
       "      <th>Tune_lgboost_Pred</th>\n",
       "      <th>Ensemble</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5664</td>\n",
       "      <td>DegreeBranch</td>\n",
       "      <td>85.046638</td>\n",
       "      <td>81.834640</td>\n",
       "      <td>84.872841</td>\n",
       "      <td>83.918040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23568</td>\n",
       "      <td>DegreeBranch</td>\n",
       "      <td>82.543213</td>\n",
       "      <td>78.959032</td>\n",
       "      <td>83.996590</td>\n",
       "      <td>81.832945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21490</td>\n",
       "      <td>HighestDegree</td>\n",
       "      <td>78.630707</td>\n",
       "      <td>78.201040</td>\n",
       "      <td>78.705865</td>\n",
       "      <td>78.512537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8363</td>\n",
       "      <td>nan</td>\n",
       "      <td>63.221035</td>\n",
       "      <td>52.945330</td>\n",
       "      <td>62.621525</td>\n",
       "      <td>59.595963</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6165</td>\n",
       "      <td>Gender</td>\n",
       "      <td>84.470673</td>\n",
       "      <td>87.459835</td>\n",
       "      <td>85.167257</td>\n",
       "      <td>85.699255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>17679</td>\n",
       "      <td>nan</td>\n",
       "      <td>63.043877</td>\n",
       "      <td>66.995853</td>\n",
       "      <td>63.192774</td>\n",
       "      <td>64.410835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>9715</td>\n",
       "      <td>DegreeBranch</td>\n",
       "      <td>79.625107</td>\n",
       "      <td>77.391229</td>\n",
       "      <td>79.194648</td>\n",
       "      <td>78.736995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>12591</td>\n",
       "      <td>nan</td>\n",
       "      <td>59.519886</td>\n",
       "      <td>54.654342</td>\n",
       "      <td>59.446620</td>\n",
       "      <td>57.873616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>6140</td>\n",
       "      <td>YearsOfExperince</td>\n",
       "      <td>62.684628</td>\n",
       "      <td>67.307872</td>\n",
       "      <td>61.022800</td>\n",
       "      <td>63.671767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>21018</td>\n",
       "      <td>CurrentCompanyType</td>\n",
       "      <td>66.655975</td>\n",
       "      <td>68.969488</td>\n",
       "      <td>67.049320</td>\n",
       "      <td>67.558261</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   EmpID BiasInfluentialFactor  Tune_XGBoost_Pred  Tune_catoost_Pred  \\\n",
       "0   5664          DegreeBranch          85.046638          81.834640   \n",
       "1  23568          DegreeBranch          82.543213          78.959032   \n",
       "2  21490         HighestDegree          78.630707          78.201040   \n",
       "3   8363                   nan          63.221035          52.945330   \n",
       "4   6165                Gender          84.470673          87.459835   \n",
       "5  17679                   nan          63.043877          66.995853   \n",
       "6   9715          DegreeBranch          79.625107          77.391229   \n",
       "7  12591                   nan          59.519886          54.654342   \n",
       "8   6140      YearsOfExperince          62.684628          67.307872   \n",
       "9  21018    CurrentCompanyType          66.655975          68.969488   \n",
       "\n",
       "   Tune_lgboost_Pred   Ensemble  \n",
       "0          84.872841  83.918040  \n",
       "1          83.996590  81.832945  \n",
       "2          78.705865  78.512537  \n",
       "3          62.621525  59.595963  \n",
       "4          85.167257  85.699255  \n",
       "5          63.192774  64.410835  \n",
       "6          79.194648  78.736995  \n",
       "7          59.446620  57.873616  \n",
       "8          61.022800  63.671767  \n",
       "9          67.049320  67.558261  "
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# Prediction for tuned XG Boost\n",
    "\n",
    "model_xgb.fit(predictors2,Target)\n",
    "model_xgb_pred = pd.DataFrame(model_xgb.predict(predictors3))\n",
    "model_xgb_pred.columns = ['Tune_XGBoost_Pred']\n",
    "\n",
    "\n",
    "\n",
    "# Prediction for tuned Cat Boost\n",
    "\n",
    "model_cat.fit(predictors2,Target)\n",
    "model_cat_pred = pd.DataFrame(model_cat.predict(predictors3))\n",
    "model_cat_pred.columns = ['Tune_catoost_Pred']\n",
    "\n",
    "# Prediction for tuned lgb Boost\n",
    "\n",
    "model_lgb.fit(predictors2,Target)\n",
    "model_lgb_pred = pd.DataFrame(model_lgb.predict(predictors3))\n",
    "model_lgb_pred.columns = ['Tune_lgboost_Pred']\n",
    "\n",
    "\n",
    "Model_Pred = pd.concat([df_test4['EmpID'],model_xgb_d1_class['BiasInfluentialFactor'],model_xgb_pred,model_cat_pred,model_lgb_pred],axis=1)\n",
    "Model_Pred['Ensemble'] = Model_Pred[['Tune_XGBoost_Pred', 'Tune_catoost_Pred','Tune_lgboost_Pred']].mean(axis=1)\n",
    "\n",
    "Model_Pred.head(n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EmpID</th>\n",
       "      <th>BiasInfluentialFactor</th>\n",
       "      <th>FitmentPercent</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5664</td>\n",
       "      <td>DegreeBranch</td>\n",
       "      <td>83.918040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>23568</td>\n",
       "      <td>DegreeBranch</td>\n",
       "      <td>81.832945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>21490</td>\n",
       "      <td>HighestDegree</td>\n",
       "      <td>78.512537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8363</td>\n",
       "      <td>nan</td>\n",
       "      <td>59.595963</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6165</td>\n",
       "      <td>Gender</td>\n",
       "      <td>85.699255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>9731</td>\n",
       "      <td>nan</td>\n",
       "      <td>65.798386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>16330</td>\n",
       "      <td>YearsOfExperince</td>\n",
       "      <td>78.460854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>18456</td>\n",
       "      <td>nan</td>\n",
       "      <td>73.426304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>22486</td>\n",
       "      <td>Gender</td>\n",
       "      <td>68.323516</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>4900</td>\n",
       "      <td>DegreeBranch</td>\n",
       "      <td>78.313497</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    EmpID BiasInfluentialFactor  FitmentPercent\n",
       "0    5664          DegreeBranch       83.918040\n",
       "1   23568          DegreeBranch       81.832945\n",
       "2   21490         HighestDegree       78.512537\n",
       "3    8363                   nan       59.595963\n",
       "4    6165                Gender       85.699255\n",
       "..    ...                   ...             ...\n",
       "95   9731                   nan       65.798386\n",
       "96  16330      YearsOfExperince       78.460854\n",
       "97  18456                   nan       73.426304\n",
       "98  22486                Gender       68.323516\n",
       "99   4900          DegreeBranch       78.313497\n",
       "\n",
       "[100 rows x 3 columns]"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Model_Pred_F = Model_Pred[['EmpID','BiasInfluentialFactor','Ensemble']]\n",
    "Model_Pred_F.columns = ['EmpID','BiasInfluentialFactor','FitmentPercent']\n",
    "Model_Pred_F.head(n=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Model_Pred_F.to_csv('Predictions_06062021_IB_F.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}