model-building.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('./marketdata2.csv').drop(['Unnamed: 0'], axis=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>marital</th>\n",
       "      <th>segment</th>\n",
       "      <th>SocialMedia</th>\n",
       "      <th>income</th>\n",
       "      <th>creditRating</th>\n",
       "      <th>aveSpend</th>\n",
       "      <th>touchpoints</th>\n",
       "      <th>nTouchpoints</th>\n",
       "      <th>recent_touchpoint</th>\n",
       "      <th>...</th>\n",
       "      <th>N</th>\n",
       "      <th>Y</th>\n",
       "      <th>C</th>\n",
       "      <th>B</th>\n",
       "      <th>D</th>\n",
       "      <th>A</th>\n",
       "      <th>F</th>\n",
       "      <th>E</th>\n",
       "      <th>New</th>\n",
       "      <th>G</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>38</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>5007.50</td>\n",
       "      <td>emailDestination mailDiscount</td>\n",
       "      <td>2</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>39</td>\n",
       "      <td>divorced</td>\n",
       "      <td>P3</td>\n",
       "      <td>U</td>\n",
       "      <td>100000.0</td>\n",
       "      <td>B</td>\n",
       "      <td>19.74</td>\n",
       "      <td>emailDestination mailGift</td>\n",
       "      <td>2</td>\n",
       "      <td>mailGift</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>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>41</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>D</td>\n",
       "      <td>320.10</td>\n",
       "      <td>emailDiscount mailGift emailDiscount</td>\n",
       "      <td>3</td>\n",
       "      <td>emailDiscount</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>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>60</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>152000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>39.47</td>\n",
       "      <td>mailDiscount</td>\n",
       "      <td>1</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>47</td>\n",
       "      <td>married</td>\n",
       "      <td>P2</td>\n",
       "      <td>Y</td>\n",
       "      <td>60000.0</td>\n",
       "      <td>A</td>\n",
       "      <td>19.89</td>\n",
       "      <td>mailDestination emailDestination emailDestination</td>\n",
       "      <td>3</td>\n",
       "      <td>emailDestination</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>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14984</th>\n",
       "      <td>32</td>\n",
       "      <td>married</td>\n",
       "      <td>P1</td>\n",
       "      <td>U</td>\n",
       "      <td>33000.0</td>\n",
       "      <td>E</td>\n",
       "      <td>4280.63</td>\n",
       "      <td>emailDestination emailDestination emailDestina...</td>\n",
       "      <td>7</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14985</th>\n",
       "      <td>53</td>\n",
       "      <td>divorced</td>\n",
       "      <td>P3</td>\n",
       "      <td>N</td>\n",
       "      <td>61000.0</td>\n",
       "      <td>F</td>\n",
       "      <td>1332.29</td>\n",
       "      <td>smsDestination mailDiscount emailDestination e...</td>\n",
       "      <td>6</td>\n",
       "      <td>mailDiscount</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>14986</th>\n",
       "      <td>30</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>44000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>4010.00</td>\n",
       "      <td>emailDestination emailDestination mailDiscount...</td>\n",
       "      <td>4</td>\n",
       "      <td>mailDiscount</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14987</th>\n",
       "      <td>52</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>71000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>507.46</td>\n",
       "      <td>emailDestination mailGift emailDiscount smsDes...</td>\n",
       "      <td>8</td>\n",
       "      <td>emailDestination</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14988</th>\n",
       "      <td>38</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>512.67</td>\n",
       "      <td>emailDestination mailDiscount</td>\n",
       "      <td>2</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>14989 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       age   marital segment SocialMedia    income creditRating  aveSpend  \\\n",
       "0       38    single      P4           U   51000.0            C   5007.50   \n",
       "1       39  divorced      P3           U  100000.0            B     19.74   \n",
       "2       41   married      P4           N   51000.0            D    320.10   \n",
       "3       60   married      P4           U  152000.0            C     39.47   \n",
       "4       47   married      P2           Y   60000.0            A     19.89   \n",
       "...    ...       ...     ...         ...       ...          ...       ...   \n",
       "14984   32   married      P1           U   33000.0            E   4280.63   \n",
       "14985   53  divorced      P3           N   61000.0            F   1332.29   \n",
       "14986   30    single      P4           N   44000.0            C   4010.00   \n",
       "14987   52   married      P4           N   71000.0            C    507.46   \n",
       "14988   38    single      P4           U   65000.0            C    512.67   \n",
       "\n",
       "                                             touchpoints  nTouchpoints  \\\n",
       "0                          emailDestination mailDiscount             2   \n",
       "1                              emailDestination mailGift             2   \n",
       "2                   emailDiscount mailGift emailDiscount             3   \n",
       "3                                           mailDiscount             1   \n",
       "4      mailDestination emailDestination emailDestination             3   \n",
       "...                                                  ...           ...   \n",
       "14984  emailDestination emailDestination emailDestina...             7   \n",
       "14985  smsDestination mailDiscount emailDestination e...             6   \n",
       "14986  emailDestination emailDestination mailDiscount...             4   \n",
       "14987  emailDestination mailGift emailDiscount smsDes...             8   \n",
       "14988                      emailDestination mailDiscount             2   \n",
       "\n",
       "      recent_touchpoint  ...  N  Y  C  B  D  A  F  E  New  G  \n",
       "0          mailDiscount  ...  0  0  1  0  0  0  0  0    0  0  \n",
       "1              mailGift  ...  0  0  0  1  0  0  0  0    0  0  \n",
       "2         emailDiscount  ...  1  0  0  0  1  0  0  0    0  0  \n",
       "3          mailDiscount  ...  0  0  1  0  0  0  0  0    0  0  \n",
       "4      emailDestination  ...  0  1  0  0  0  1  0  0    0  0  \n",
       "...                 ...  ... .. .. .. .. .. .. .. ..  ... ..  \n",
       "14984      mailDiscount  ...  0  0  0  0  0  0  0  1    0  0  \n",
       "14985      mailDiscount  ...  1  0  0  0  0  0  1  0    0  0  \n",
       "14986      mailDiscount  ...  1  0  1  0  0  0  0  0    0  0  \n",
       "14987  emailDestination  ...  1  0  1  0  0  0  0  0    0  0  \n",
       "14988      mailDiscount  ...  0  0  1  0  0  0  0  0    0  0  \n",
       "\n",
       "[14989 rows x 30 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check distribution of all numerical variables excluding one-hot encoded variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AGE Distribution\n",
      "count    14989.000000\n",
      "mean        39.223964\n",
      "std          9.276033\n",
      "min         20.000000\n",
      "25%         32.000000\n",
      "50%         37.000000\n",
      "75%         46.000000\n",
      "max         60.000000\n",
      "Name: age, dtype: float64\n",
      "-----\n",
      "INCOME Distribution\n",
      "count     14989.000000\n",
      "mean      71590.041216\n",
      "std       36821.917752\n",
      "min       10000.000000\n",
      "25%       45000.000000\n",
      "50%       64000.000000\n",
      "75%       90000.000000\n",
      "max      249000.000000\n",
      "Name: income, dtype: float64\n",
      "-----\n",
      "AVESPEND Distribution\n",
      "count    14989.000000\n",
      "mean       714.399585\n",
      "std       2164.764542\n",
      "min          0.000000\n",
      "25%         20.420000\n",
      "50%         91.470000\n",
      "75%        426.230000\n",
      "max      17889.740000\n",
      "Name: aveSpend, dtype: float64\n",
      "-----\n"
     ]
    }
   ],
   "source": [
    "vars_list = ['age','income','aveSpend']\n",
    "\n",
    "for var in vars_list:\n",
    "    print(var.upper() + ' Distribution')\n",
    "    print(df[var].describe())\n",
    "    print('-----')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that the max-min range of all 3 variables are very different from one another. When we do further analysis, like multivariate linear regression, for example, the attributed income will intrinsically influence the result more due to its larger value. Hence, it is important to standardize and normalize the data to bring all variables to the same range.\n",
    "\n",
    "I use a Robust Scaler (QuantileTransformer): similar to normalization but it instead uses the interquartile range, so that it is robust to outliers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import QuantileTransformer\n",
    "\n",
    "qt = QuantileTransformer(output_distribution='normal')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['nAveSpend'] = qt.fit_transform(df[['aveSpend']].values.reshape(-1,1))\n",
    "df['nIncome'] = qt.fit_transform(df[['income']].values.reshape(-1,1))\n",
    "df['nAge'] = qt.fit_transform(df[['age']].values.reshape(-1,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>marital</th>\n",
       "      <th>segment</th>\n",
       "      <th>SocialMedia</th>\n",
       "      <th>income</th>\n",
       "      <th>creditRating</th>\n",
       "      <th>aveSpend</th>\n",
       "      <th>touchpoints</th>\n",
       "      <th>nTouchpoints</th>\n",
       "      <th>recent_touchpoint</th>\n",
       "      <th>...</th>\n",
       "      <th>B</th>\n",
       "      <th>D</th>\n",
       "      <th>A</th>\n",
       "      <th>F</th>\n",
       "      <th>E</th>\n",
       "      <th>New</th>\n",
       "      <th>G</th>\n",
       "      <th>nAveSpend</th>\n",
       "      <th>nIncome</th>\n",
       "      <th>nAge</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>38</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>5007.50</td>\n",
       "      <td>emailDestination mailDiscount</td>\n",
       "      <td>2</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.798751</td>\n",
       "      <td>-0.406079</td>\n",
       "      <td>0.037646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>39</td>\n",
       "      <td>divorced</td>\n",
       "      <td>P3</td>\n",
       "      <td>U</td>\n",
       "      <td>100000.0</td>\n",
       "      <td>B</td>\n",
       "      <td>19.74</td>\n",
       "      <td>emailDestination mailGift</td>\n",
       "      <td>2</td>\n",
       "      <td>mailGift</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>-0.694311</td>\n",
       "      <td>0.912774</td>\n",
       "      <td>0.123259</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>41</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>D</td>\n",
       "      <td>320.10</td>\n",
       "      <td>emailDiscount mailGift emailDiscount</td>\n",
       "      <td>3</td>\n",
       "      <td>emailDiscount</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>0.559895</td>\n",
       "      <td>-0.406079</td>\n",
       "      <td>0.283522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>60</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>152000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>39.47</td>\n",
       "      <td>mailDiscount</td>\n",
       "      <td>1</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-0.362850</td>\n",
       "      <td>1.831224</td>\n",
       "      <td>5.199338</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>47</td>\n",
       "      <td>married</td>\n",
       "      <td>P2</td>\n",
       "      <td>Y</td>\n",
       "      <td>60000.0</td>\n",
       "      <td>A</td>\n",
       "      <td>19.89</td>\n",
       "      <td>mailDestination emailDestination emailDestination</td>\n",
       "      <td>3</td>\n",
       "      <td>emailDestination</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>-0.687284</td>\n",
       "      <td>-0.106841</td>\n",
       "      <td>0.744694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14984</th>\n",
       "      <td>32</td>\n",
       "      <td>married</td>\n",
       "      <td>P1</td>\n",
       "      <td>U</td>\n",
       "      <td>33000.0</td>\n",
       "      <td>E</td>\n",
       "      <td>4280.63</td>\n",
       "      <td>emailDestination emailDestination emailDestina...</td>\n",
       "      <td>7</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.748868</td>\n",
       "      <td>-1.340200</td>\n",
       "      <td>-0.659602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14985</th>\n",
       "      <td>53</td>\n",
       "      <td>divorced</td>\n",
       "      <td>P3</td>\n",
       "      <td>N</td>\n",
       "      <td>61000.0</td>\n",
       "      <td>F</td>\n",
       "      <td>1332.29</td>\n",
       "      <td>smsDestination mailDiscount emailDestination e...</td>\n",
       "      <td>6</td>\n",
       "      <td>mailDiscount</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>1.280981</td>\n",
       "      <td>-0.051460</td>\n",
       "      <td>1.269664</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14986</th>\n",
       "      <td>30</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>44000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>4010.00</td>\n",
       "      <td>emailDestination emailDestination mailDiscount...</td>\n",
       "      <td>4</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.694923</td>\n",
       "      <td>-0.758004</td>\n",
       "      <td>-0.985610</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14987</th>\n",
       "      <td>52</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>71000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>507.46</td>\n",
       "      <td>emailDestination mailGift emailDiscount smsDes...</td>\n",
       "      <td>8</td>\n",
       "      <td>emailDestination</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.779682</td>\n",
       "      <td>0.240671</td>\n",
       "      <td>1.169401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14988</th>\n",
       "      <td>38</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>512.67</td>\n",
       "      <td>emailDestination mailDiscount</td>\n",
       "      <td>2</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.788882</td>\n",
       "      <td>0.047692</td>\n",
       "      <td>0.037646</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>14989 rows × 33 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       age   marital segment SocialMedia    income creditRating  aveSpend  \\\n",
       "0       38    single      P4           U   51000.0            C   5007.50   \n",
       "1       39  divorced      P3           U  100000.0            B     19.74   \n",
       "2       41   married      P4           N   51000.0            D    320.10   \n",
       "3       60   married      P4           U  152000.0            C     39.47   \n",
       "4       47   married      P2           Y   60000.0            A     19.89   \n",
       "...    ...       ...     ...         ...       ...          ...       ...   \n",
       "14984   32   married      P1           U   33000.0            E   4280.63   \n",
       "14985   53  divorced      P3           N   61000.0            F   1332.29   \n",
       "14986   30    single      P4           N   44000.0            C   4010.00   \n",
       "14987   52   married      P4           N   71000.0            C    507.46   \n",
       "14988   38    single      P4           U   65000.0            C    512.67   \n",
       "\n",
       "                                             touchpoints  nTouchpoints  \\\n",
       "0                          emailDestination mailDiscount             2   \n",
       "1                              emailDestination mailGift             2   \n",
       "2                   emailDiscount mailGift emailDiscount             3   \n",
       "3                                           mailDiscount             1   \n",
       "4      mailDestination emailDestination emailDestination             3   \n",
       "...                                                  ...           ...   \n",
       "14984  emailDestination emailDestination emailDestina...             7   \n",
       "14985  smsDestination mailDiscount emailDestination e...             6   \n",
       "14986  emailDestination emailDestination mailDiscount...             4   \n",
       "14987  emailDestination mailGift emailDiscount smsDes...             8   \n",
       "14988                      emailDestination mailDiscount             2   \n",
       "\n",
       "      recent_touchpoint  ...  B  D  A  F  E  New  G  nAveSpend   nIncome  \\\n",
       "0          mailDiscount  ...  0  0  0  0  0    0  0   1.798751 -0.406079   \n",
       "1              mailGift  ...  1  0  0  0  0    0  0  -0.694311  0.912774   \n",
       "2         emailDiscount  ...  0  1  0  0  0    0  0   0.559895 -0.406079   \n",
       "3          mailDiscount  ...  0  0  0  0  0    0  0  -0.362850  1.831224   \n",
       "4      emailDestination  ...  0  0  1  0  0    0  0  -0.687284 -0.106841   \n",
       "...                 ...  ... .. .. .. .. ..  ... ..        ...       ...   \n",
       "14984      mailDiscount  ...  0  0  0  0  1    0  0   1.748868 -1.340200   \n",
       "14985      mailDiscount  ...  0  0  0  1  0    0  0   1.280981 -0.051460   \n",
       "14986      mailDiscount  ...  0  0  0  0  0    0  0   1.694923 -0.758004   \n",
       "14987  emailDestination  ...  0  0  0  0  0    0  0   0.779682  0.240671   \n",
       "14988      mailDiscount  ...  0  0  0  0  0    0  0   0.788882  0.047692   \n",
       "\n",
       "           nAge  \n",
       "0      0.037646  \n",
       "1      0.123259  \n",
       "2      0.283522  \n",
       "3      5.199338  \n",
       "4      0.744694  \n",
       "...         ...  \n",
       "14984 -0.659602  \n",
       "14985  1.269664  \n",
       "14986 -0.985610  \n",
       "14987  1.169401  \n",
       "14988  0.037646  \n",
       "\n",
       "[14989 rows x 33 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['age', 'marital', 'segment', 'SocialMedia', 'income', 'creditRating',\n",
       "       'aveSpend', 'touchpoints', 'nTouchpoints', 'recent_touchpoint',\n",
       "       'num_creditRating', 'single', 'divorced', 'married', 'unknown', 'P4',\n",
       "       'P3', 'P2', 'P1', 'U', 'N', 'Y', 'C', 'B', 'D', 'A', 'F', 'E', 'New',\n",
       "       'G', 'nAveSpend', 'nIncome', 'nAge'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check distribution of scaled variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot for NAGE: total counts = 14989\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot for NINCOME: total counts = 14989\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlUAAAE1CAYAAADOGhV1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3deXgU54Em8Lf6VutqXa37FkiAECABErKNbQzBxshxHCbx2DPEmzHxbMZhl33iZJLxxCSebCaejNlJspONvZkcG2xjT/CBsTH4AmNxSoAQh4TQfbakllotdavP2j9kFGSBWhLdXX28v+fRA+r+VPWqEM1LfdVfCaIoiiAiIiKiWyKTOgARERFRKGCpIiIiIvIClioiIiIiL2CpIiIiIvIClioiIiIiL1BIuXO3242xsTEolUoIgiBlFCIiIqIZiaIIh8OByMhIyGTTz0tJWqrGxsbQ2NgoZQQiIiKiOVm4cCGio6OnPS5pqVIqlQAmwqlUKo/j6+vrUVxc7OtYYY3H2D94nH2Px9g/eJx9j8fYP2ZznO12OxobGyf7y+dJWqquTfmpVCqo1epZfc1sx9H88Rj7B4+z7/EY+wePs+/xGPvHbI/zzS5Z4oXqRERERF7AUkVERETkBSxVRERERF7AUkVERETkBSxVRERERF7AUkVERETkBSxVRERERF7AUkVERETkBSxVRERERF4g6YrqRES+oonUwWC0eBwXoVEgWuv5NllERJ6wVBFRSHK6gdoGg8dxpYV6lioi8gpO/xERERF5AUsVERERkRewVBERERF5AUsVERERkRewVBERERF5AUsVERERkRewVBERERF5AUsVERERkRewVBERERF5AUsVERERkRewVBERERF5AUsVERERkRewVBERERF5AUsVERERkRewVBERERF5AUsVERERkRfMqlTt27cPmzZtwoYNG7B79+6bjvv444+xbt06r4UjIiIiChYKTwP6+vqwa9cu7N27FyqVCg8//DDKy8tRUFAwZdzAwAB++tOf+iwoERERUSDzeKaquroaFRUV0Ol00Gq12LhxIw4cODBt3NNPP40nn3zSJyGJiK5ntthhMFpm/FAoNVLHJKIw4/FMlcFgQFJS0uTner0edXV1U8b84Q9/wOLFi7Fs2bJ5haivr5/12Jqamnntg2aPx9g/eJznTxGhw/HznTOOqVxRgLb2No/bSo11oaNl2FvRwhJ/ln2Px9g/bvU4eyxVoihOe0wQhMnfNzY24uDBg/jd736H3t7eeYUoLi6GWq32OK6mpgZlZWXz2gfNDo+xf/A43xqD0YJsk9zjuOysbI9jUlP00MfneyNWWOLPsu/xGPvHbI6zzWab8USQx1KVnJyM06dPT35uMBig1+snPz9w4AD6+/vx5S9/GQ6HAwaDAY888gheeuml2XwPRERe43K70TtggXFkHKcu2tFhsEAuE1C5NBVJcVqp4xFRiPNYqiorK/GLX/wCRqMREREROHjwIJ599tnJ57dv347t27cDADo7O7F161YWKiLyO9OoDQeOt2JgeBwAoFHJERetxqDJhtc+vILlC5KwanEKlAquJENEvjGrM1U7duzA1q1b4XA4sGXLFpSUlGDbtm3Yvn07li5d6o+cREQ31dxtwgen2iFAwPrVWUhPisLC9Ah0G50Ytztx7HwPzjT242qXCXeVZiAzOVrqyEQUgjyWKgCoqqpCVVXVlMdefPHFaeMyMjLw4YcfeicZEZEHbreIExd6UNvQj6S4CNxbkYOYSBWAP1/7qVEpcHdZJhZmxeHjmk68fbQZD9yRj3R9lJTRiSgE8Tw4EQUlt1vEwRNtqG3ox5K8BDx0V8FkobqR9KQobLlnAWKj1Hj3eCtMozY/piWicMBSRURBxy2K+KimA1e7TLitJA13lWZAIff8cqZWyrGpMhcQgf3VLbA7XH5IS0ThgqWKiIKKKIp45VADLrcNYdXiZCxfmOT5i66ji1bj3jXZGDbbcPBEG9zu6cvGEBHNB0sVEQWVl95rwAenOrCsIBGrFiXPaxsZ+misXZ6Otl4z/vTRFS8nJKJwxVJFREHj7aPNeOVQA24rScNty9KmLEQ8V8X5iViSl4ADx9tw6uL8Fi4mIroeSxURBYUzDQa8+MZ5lC9JwWP3L76lQnXN7cvSkKGPwr/tOYOhkXEvpCSicMZSRUQBr7t/FM/9v9PITI7G/3ikFDLZrRcqAFDIZXjiwaWwjjvxv/ac4fVVRHRLWKqIKKBZxh34p9+egCAIePrr5dBqlF7dflpSFP7mi8WovWzAvqPNXt02EYWXWS3+SUTkL2aLHdZxJ4CJtah+8dpZdPePTZyhEgQYjBbYvLwUwn1rclB72YDfvX0RJQWJyE2L9er2iSg88EwVEQUU67gTtQ0G1DYY8Os3zqOuaQC3L0uD5brHnS63V/cpCAK+9ZXliIlU4l/+WMP1q4hoXliqiCggdQ+MouZSH4qy41Ccn+jz/cVGqfHfvlqKjj4zXj7Y4PP9EVHoYakiooBjs7vw/sl2REeqcMfydL/tt7RIjw2rs7D3oytobB/y236JKDSwVBFRwDl8phOjVgc2rM6CSin3677/5oFixMVo8G97zsDh5DQgEc0eSxURBZTj9T240jGM1YtTkJIQ6ff9R0Yo8eRfLEd7rxmvHGr0+/6JKHixVBFRwOgdHMP/e/cSUhMiUVqklyzHykXJuGdVJv7zwyto6hiWLAcRBReWKiIKGC++UQ8AWL86CzIvrJh+Kx5/oBi6KDX+bc8ZuLgoKBHNAksVEQWE81cHcPJiL+6/LRcxkSqp4yBKq8K2B4vR2jOCj063Sx2HiIIASxURSU4URfx23wUkxmqwflWW1HEm3VaShsKsOOw+cNnrC44SUehhqSIiyR09140rHcN49N5Ffn+330wEQcDXNi/GgGkc+z7hLWyIaGYsVUQkKYfTjT+8cxE5qTG4e2Wm1HGmWZqfiJWLkvGfHzRiZMwudRwiCmAsVUQkqXePtaB30ILHNi+GXCbtxek389j9i2G1OfHaB1xigYhujqWKiCQzZnXglYONWLYgEaWF0i2h4El2agzWrczC20db0Ge0SB2HiAIUSxURSeatI1dhttjx2OYlECReQsGTR+8tgkwA/njgktRRiChAsVQRkSScLjcOHG9FWZEeBRk6qeN4lKiLwP235+FIbScMPFtFRDfAUkVEkjhxoRfGERs2VeZKHWXWNt8+kfWd6haJkxBRIGKpIiJJvFvdgqS4CJQtSpY6yqzp47QoL07FwRPtXLeKiKZhqSIivzBb7DAYLTAYLai70o9zVwZwe0kaBoetk48bjJaALyubb8+F2WLHJ2c6pY5CRAFGIXUAIgoP1nEnahsMACYW+5QJQGy0evKxawqz46SINyOzxQ7ruBMAkBynRXpSFF7/+CqW5idOXmAfoVEgWiv97XWISDosVUTkV06XG5dbjchLj0WkRil1nFm5vhACQEGmDodrO/FOdStSEyMBAKWFepYqojDH6T8i8qumjmHYHC4syUuUOsq8FWbpoFbKUdc0IHUUIgogLFVE5Ff1zYOIi1YjPSlS6ijzplTIUZQTj+auYYxZHVLHIaIAwVJFRH7TP2RBn9GCJXkJAb/YpydL8xPgFidKIhERwFJFRH50qdUIuUwIyIvR5yo2So3slGhcbBmE2y1KHYeIAgAvVCciv3C53WjqNCEnLQYaVeC89Dhdbo8rpN9smYfFuQl491grOgxmrAyi9baIyDcC55WNiELa5dYhWG1OLMwMrLNUNofL4wXnNzuzlp0SDbVSjoa2IV9EI6Igw+k/IvKLExd6oVLIkJUSLXUUr5HLZSjI1KGl2wSrzSl1HCKSGEsVEfmcw+lCbYMBeemxUMhD62WnMCsOTpc4bRFTIgo/ofXqRkQB6fQlA6w2JxZk6qSO4nUpCVrERKpwvL5H6ihEJDGWKiLyuSNnOhGlVSJDHzpTf9cIgoCFWXG41GLEoMkqdRwikhBLFRH5lNXmxMmLfVhZlAyZLLjXprqZwqw4iAAO1/Imy0ThjKWKiHzqxIVe2B0ulC9JkTqKz+ii1chLj8VHNSxVROGMpYqIfOrImU4kxmpQEILXU11vTXEqWntG0NJtkjoKEUmEpYqIfMZsseNMgwF3rMiALMhvS+PJ6sXJkMsEnq0iCmMsVUTkM9V1PXC6RKxdni51FJ+L0qqwclEyDtd28rY1RGGKpYqIfObImU6kJUYiPyNW6ih+sXZFOowj47jUapQ6ChFJgKWKiHzCODKO81cHsHZFBoQQn/q7ZtXiFKgUMhw92yV1FCKSAEsVEfnE0bNdEMWJszfhIkKtwMrFyfi0rhsuTgEShR2WKiLyiSNnu5CbFoPM5NBb8HMmty9Lx5DZhostg1JHISI/Y6kiIq/rHRxDQ9sQ1q7IkDqK361alAyVUs4pQKIwxFJFRF73yWeFIhze9fd5GrUCqxYno7quh1OARGGGpYqIvO7ImS4syomHPl4rdRRJ3LEsHcOjNlxoHpA6ChH50axK1b59+7Bp0yZs2LABu3fvnvb8oUOHUFVVhfvvvx9///d/D7vd7vWgRBQc2npH0NozgjvC8CzVNWWL9FCr5Dh6tlvqKETkRx5LVV9fH3bt2oWXXnoJb775Jvbs2YOmpqbJ5y0WC370ox/ht7/9Lfbv3w+bzYbXX3/dp6GJKHB9cqYLMgG4fXma1FH8yulyw2C0wGC0YGTUjpKCRBw914WegdHJx80W/oeTKJQpPA2orq5GRUUFdLqJ+3Zt3LgRBw4cwJNPPgkA0Gq1+PDDD6FUKmGxWDA4OIiYmBjfpiaigCSKIo6c6UJJQRLiojVSx/Erm8OFuqY/T/clxGpgtjjw9tGWyXdAlhbqEa1VSRWRiHzMY6kyGAxISkqa/Fyv16Ourm7KGKVSicOHD+M73/kO9Ho9br/99jmFqK+vn/XYmpqaOW2b5o7H2D9C8Th3DdrRMziGVQWqad+fIkKHtnbP98VLT1Cirb3NC2MKPI7x7v6mjhHcIuQy4OzlTrhtE9eWpca60NEy7DFTsAnFn+VAw2PsH7d6nD2WKlGc/u6VG62OfOedd+LEiRN4/vnnsXPnTvzrv/7rrEMUFxdDrVZ7HFdTU4OysrJZb5fmjsfYP4LpOJstdljHnTOOidAoEK1V4cyb9VDI+/GXmysQ9bkzMgajBdkmucf9RUVFITsr+5bHAJjVGG/t70Zj8vra0NFnRmZGFmQyAakpeujj8z1mCibB9LMcrHiM/WM2x9lms814IshjqUpOTsbp06cnPzcYDNDr9ZOfDw8Po76+fvLsVFVVFXbs2OExPBEFB+u4E7UNhhnHlBbqEalR4pOzXSgrSp5WqMJVQYYOVzqG0dU/GnaLoBKFI48XqldWVuLYsWMwGo2wWq04ePAg1q5dO/m8KIp46qmn0N098S6Xd999F6Wlpb5LTEQB6ULLIIwj42F1WxpPslKioVTI0NQZelN+RDSdx1KVnJyMHTt2YOvWrXjwwQexefNmlJSUYNu2bTh//jzi4uLw7LPP4oknnsADDzyA1tZWPPXUU/7ITkQB5MiZLqhVcqxenCJ1lIChkMuQmxaDq10mLgRKFAY8Tv8BE1N6VVVVUx578cUXJ3+/fv16rF+/3rvJiChoOF1ufHquG+WLU6BRz+plJWwUZOjQ2D6MLoMZqxYlSx2HiHyIK6oT0S271GKE2WLn1N8NZCZHQ6WQoanTJHUUIvIxlioiumUnLvYiMkKJ0iK958FhZmIKMBbNXSY4XW6p4xCRD7FUEdEtcbrcqG0woHJpKpQKz0smhKOCDB1sDhcutRiljkJEPsRSRUS3pLVnBDa7i1N/M8hMjoJKKcOpS31SRyEiH2KpIqJb0tQxjJhIFZYWJHkeHKbkn00B1jYY4HByCpAoVLFUEdG82R0utPaMYNWiZMhl0++0QH9WkKGD1ebE2caZF1IlouDFUkVE89bcPbH+0uolXJvKk8zkKESoFTh6rlvqKETkIyxVRDRvTR3DiNIqkZ8eK3WUgCeXyVBaqMfx+h5OARKFKJYqIpoXm92Fjr5RFKTrbniTdZqurEgPy7gT5670Sx2FiHyApYqI5qWlxwS3KCI/g2epZmtxbgK0GgWq6zgFSBSKWKqIaF6udpoQpVUiOV4rdZSgoVTIsGpRCo7X98LFhUCJQg5LFRHNmc3uQnuvGfmc+puzypJUmC121F8dlDoKEXkZSxURzdm1qb8CTv3NWWmRHmqVHJ+e5xQgUahhqSKiObvaaUJUBKf+5kOjUqCsSI/j53vgdotSxyEiL2KpIqI5sTlcaO8zIz+DU3/zVbk0DUNmGy618l6ARKFEIXUAIgourd0muN2c+psPp8sNg9GCnNQYKOQCPjjVjiRdxLRxERoForUqCRIS0a1gqSKiOWm6wdTftbIwE5vD5etoAc/mcKGuaQAAkKGPxrH6HizInH7Gr7RQz1JFFIRYqoho1q5N/S3NT5hSBK4vCzdTmB3n63hBJT89Fq09IzAMWXltGlGI4DVVRDRrrd0jn0396aSOEvRy0mIgE4CrncNSRyEiL2GpIqJZa+42IZLv+vMKjUqBdH0UrnaZIIp8FyBRKGCpIqJZcbrcaO81Izc1hu/685L8dB1GxuwYNI1LHYWIvICliohmpdMwCqfLjdy0GKmjhIzctBgIAK52maSOQkRewFJFRLPS0m2CUiFDelKU1FFChlajRGpiJK528boqolDAUkVEHomiiNaeEWSlREMu58uGN+VnxGJoxIahEU4BEgU7vjoSkUd9Rgss407kpnHBT2/LS594JyWnAImCH0sVEXnU0j0CQQCyU6KljhJyoiKUSInXslQRhQCWKiLyqKXbhLTEKGhUXC/YF/IyYjEwbIVp1CZ1FCK6BSxVRDSjPuMYhsw2vuvPh/LTJ6ZVebaKKLixVBHRjM409gMAr6fyoZhINZJ0EWhmqSIKaixVRDSjs439SIjVICaSN/j1pbz0WPQZLRi12KWOQkTzxFJFRDdlGrWhqXOYZ6n8ID+DU4BEwY5XnRKFKbPFDuu4c8YxR891QRSBPF5P5XNx0RrEx2g4BUgUxFiqiMKUddyJ2gbDjGM+ru1EXLQaiboIP6UKb/npsTh1qQ+mURv0vGk1UdDh9B8R3ZDD6UZHnxnF+Qm8gbKfXJsCvPbmACIKLixVRHRDnQYznC4RS/MTpY4SNuJjNIiNUqHmcp/UUYhoHliqiOiGWrpHoFLIUJCpkzpK2BAEAfnpOlxuHcLIGN8FSBRsWKqIaBr35A2UY6DgDZT9Kj89Fm5RxMkLPVJHIaI54qslEU3TN2iB1ebkKuoSSIqLQEKsBp/WsVQRBRuWKiKapqXbBJkAZKewVPmbIAgoLdTjbGM/LOMOqeMQ0RywVBHRNC09I0hLioJaJZc6SlgqK0qG0+XGyYu8YJ0omLBUEdEUQ+ZxDJttXEVdQvkZsYiPUaO6rlvqKEQ0ByxVRDRFS/cIAPB6KgnJBAFrlqah5rIB47aZV70nosDBUkVEU7R0m5Coi0C0ljdQllJlSSrsDhdqPKx6T0SBg6WKiCZZxh3oHbTwLFUAWJKbgJhIFacAiYIISxURTWrtuTb1x+uppCaXy1BRnIpTF3thd7ikjkNEs8BSRUSTmrtMiNYqkRirkToKYWIK0Gpz4SzvBUgUFFiqiAgAYHO40GEYRV66jjdQDhAlBUmI1CjwKacAiYICSxURAQDaekbgdovIT+fUX6BQKmQoL07FyQu9cLrcUschIg9YqogIwMTUn1ajQEqCVuoodJ3KpakYtTpQ1zQgdRQi8oCliojgcLrR1mtGXlosp/4CzIpCPSLUcr4LkCgIzKpU7du3D5s2bcKGDRuwe/fuac+///77+OIXv4gHHngA3/zmN2EymbwelIh8p6PPDKfLjTxO/QUclVKOlYtScKK+Fy63KHUcIpqBx1LV19eHXbt24aWXXsKbb76JPXv2oKmpafL50dFR7Ny5Ey+88ALeeustFBYW4he/+IVPQxORdzV3maBWypGWFCV1FLqBypJUDI/acLFlUOooRDQDj6WquroaFRUV0Ol00Gq12LhxIw4cODD5vMPhwM6dO5GcnAwAKCwsRE9Pj+8SE5FXudxutPaMIDctBnIZp/4CUVlRMlQK2ZQpQLPFDoPRMuOH2WKXMDVR+FF4GmAwGJCUlDT5uV6vR11d3eTncXFxWL9+PQBgfHwcL7zwAv76r/96TiHq6+tnPbampmZO26a54zH2D6mPsyJCh7b2ThiGHbA5XIhW2dHW3jZtXHqC8oaP+2KMd/dX4Of9ee8YpMa60NEyPOWxvBQVDte0oTTTDpkgQBGhw/HznTNup2JpBpzW4RnHeIPUP8vhgMfYP271OHssVaI4fQ7/Rheyms1mfPOb30RRURG+9KUvzSlEcXEx1Gq1x3E1NTUoKyub07ZpbniM/SMQjrPBaEG2SY7m/g4oFVaUFudBIZ9+8joqKgrZWdkzbstbY7y9LX/uz5u5U1P00MfnT3lsBB14/qVaRCfmoSgnfvLPb67b8bZA+FkOdTzG/jGb42yz2WY8EeRx+i85ORkDA39+K6/BYIBer58yxmAw4JFHHkFRURF+/OMfe9okEQUItyiipXsE2SkxNyxUFDhWL06BQi5wIVCiAObxVbSyshLHjh2D0WiE1WrFwYMHsXbt2snnXS4X/vZv/xb33Xcf/uEf/oFvxyYKIr0DY7DanHzXXxCIjFBi+UI9qs/33HAGgYik53H6Lzk5GTt27MDWrVvhcDiwZcsWlJSUYNu2bdi+fTt6e3tx8eJFuFwuvPfeewAmpvN4xooo8F3tMkEuE5CdGi11FJqFyqWpOH2pD1e7TIjRqqSOQ0Sf47FUAUBVVRWqqqqmPPbiiy8CAJYuXYrLly97PxkR+ZTbLaKpcxjZqTFQKWa+NocCQ3lxKmT/eQ7Vdd24tyJH6jhE9Dm8iIIoTF3pHIZl3ImCDE79BYuYSBWW5ieguo5TgESBiKWKKEydutgLhVxAdmqM1FFoDipL0tDVP4rugTGpoxDR57BUEYUhl1tEzWUDp/6C0JriVAgCcPpSn9RRiOhzWKqIwlD91QGMjNlRkKGTOgrNUVyMBoty4lHbYJA6ChF9DksVURg6eq4baqUc2Smc+gtGt5WkodMwimGzTeooRHQdliqiMONyuVFd141lCxKhVPAlIBitWZoGALja5ftb0BDR7PEVlSjM1DVNTP2tWpQidRSap6S4COSmxaC5yyR1FCK6DksVUZg5eq4bEWo5lhYkSB2FbkFZUTIMQ1aMjNmljkJEn2GpIgojTpcbx853o3xJKpR8119QKyuauAcrz1YRBQ6WKqIwcu5KP8wWB25fliZ1FLpF+jgtEmI1vK6KKICwVBGFkcO1nYjUKFD62VkOCm4FGTr0DlpgtnAKkCgQsFQRhYlxuxPH63tw27J0Tv2FiILMiXXGmjp5toooELBUEYWJkxd6YbW5cFdphtRRyEt0UWro4yLQ1MFSRRQIWKqIwsRHNZ1IjNVgSR7f9RdKCjJ1MAxZMTzKhUCJpKaQOgAR+Z5p1IYzDQY8eGc+ZDJB6jjkgdPlhsFomXGMzeECMHFdVXVdD5o6hrFyUbI/4hHRTbBUEYWBo2e74HKLuKssU+ooNAs2hwt1TQMzjinMjgMARGtVSE2IRFMnSxWR1Dj9RxQGPq7tRE5qDHJSea+/ULQgU4dB0zgGTeNSRyEKayxVRCGuZ2AMl9uGeIF6CMvPiIUAvguQSGosVUQh7vCZTggCsHYFS1Wo0mqUSNdH4UrHEERRlDoOUdhiqSIKYaIo4uOaDhTnJSIpLkLqOORDCzJ1MI3aMTBslToKUdhiqSIKYU2dw+jqH8OdnPoLeXnpsZAJAq5wzSoiybBUEYWwD093QKmQ4baSVKmjkI9pVApkpUTjSscwpwCJJMJSRRSi7A4XPq7pxJriVERpVVLHIT9YmKXDqNWBrv5RqaMQhSWWKqIQdaK+F6NWBzaUZ0kdhfwkNy0WKoUMDW1DUkchCkssVUQh6uDJNujjIlBSkCR1FPIThVyGgkwdmjpNcDhdUschCjssVUQhyGC04NyVfqxflcXb0oSZwuw4OF1uNHeZpI5CFHZYqohC0Aen2gEA96zi1F+4SU2IREykCpc5BUjkdyxVRCHG7Rbx/ql2LFuQBH28Vuo45GeCIKAwKw6dhlEYR3jbGiJ/YqkiCjF1Tf0wDFmxYTXPUoWrazdbPl7fI3ESovDCUkUUYg6daEdUhBIVxVybKlzFRqmRmhCJ6vM9XLOKyI9YqohCiNlix7H6HtxVmgGVUi51HJJQYXYcegbGuMI6kR+xVBGFkMO1nXA43dhQni11FJJYQYYOCrkMH53ukDoKUdhQSB2AiLxDFEW8e6wVBRmxSIqLgMFomXG8zcF1jEKZWiXHisIkfFzbiceqlkDNM5dEPsdSRRQiLrYY0d5rxre+shzWcSdqGwwzjr92MTOFrrtWZODUxT4cPdvF5TWI/IDTf0Qh4sCxVkRqFFi7PF3qKBQgCrPjkKGPwjvVLVJHIQoLLFVEIcA0asPRc924e2UmNGqegKYJgiBgU2UuGtuH0cQL1ol8jqWKKAR8cKodTpcb967JkToKBZh1KzOhVsl5torID1iqiIKc2y3iwLE2LMlLQHZKjNRxKMBERihxV2kGDp/pwqjFLnUcopDGUkUU5M5d6UfP4BjPUtFNbarMhd3hwodcXoHIp1iqiILcu8daEROpwm0lXEGdbiwvPRaF2XF4p7qVK6wT+RBLFVEQGzRZceJCL9avyoJSwXWI6OY2Veaiq38UdU0DUkchClksVURB7ODxNrjdIjau4QrqNLPbl6UhWqvC/k95wTqRr7BUEQUpp8uNA8dbUVqkR1pilNRxKMCplHJsrMjGifoe9AyMSR2HKCSxVBEFqWPne2AcsWHzbblSR6EgUXVHHmQyGV4/3CR1FKKQxFJFFKT2f9qClAQtSouSpY5CQSI+RoN1KzPxwcl2DJttUschCjksVURBqKXbhAvNg7hvTS7kMkHqOBREHrq7AA6XG28fbZY6ClHI4f0siILQ/k9boJ8JVxgAABX9SURBVFLIsKGcN8mlm3O63DAYLVMeU8plWLFQj31Hm7F2RTo0KgUiNApEa1USpSQKHSxVREFm1GLHx7WduLM0g/8Q0oxsDtcNl1DITYtBbYMBL73XgOULk1BaqOfPEpEXcPqPKMi8f6oDNrsL9/MCdZqnlIRIpCVG4tyVfrjcXAyUyFtYqoiCiNst4p1PW7AoJx75GTqp41AQKy3UY9TqwJWOIamjEIUMliqiIFLbYEDP4BjPUtEty0qJRkKsBrUNBrh56xoir5hVqdq3bx82bdqEDRs2YPfu3Tcd993vfhd79+71WjgimurNI1ehi1ZjQaYOBqPlph82h0vqqBTgBEFAWZEeQyM21F42SB2HKCR4vFC9r68Pu3btwt69e6FSqfDwww+jvLwcBQUFU8Y888wzOHbsGMrLy30amChctXSbcLaxH1++u8Dj/dsKs+P8lIqCWX6GDnEX+7DvaDPuXZMDGZfnILolHs9UVVdXo6KiAjqdDlqtFhs3bsSBAwemjNm3bx/uuece3HfffT4LShTu3jh8FRqVHHeuyJA6CoUImSCgbFEyOg2jOHGhR+o4REHP45kqg8GApKSkyc/1ej3q6uqmjHn88ccBADU1NfMKUV9fP+ux890HzR6PsX/M5TiPWFz4uKYHqxZEYcQ0gLb2zhnHpyco0dbedstjvLkt/++vIOC+v0A85ipRREKMCv/x5lkobd0QhLmfreJrhu/xGPvHrR5nj6VKvMEFjPP5SzeT4uJiqNVqj+NqampQVlbm1X3TVDzG/jHX4/y7ty8AAB7fUgGZICDbJJ9xfFRUFLKzsm95jDe35e/9AQi47y9Qj3lqihO/2XcBLk06yotTPY6/Hl8zfI/H2D9mc5xtNtuMJ4I8Tv8lJydjYODP128YDAbo9fo5xCSiW2EZd+DA8TasKUlDSkKk1HEoBJUXpyAlQYtXDjXc8D/SRDQ7HktVZWUljh07BqPRCKvVioMHD2Lt2rX+yEZEAN4/2Y4xqwNfujNf6igUouQyGb5yz0I0dZpw+lKf1HGIgtaszlTt2LEDW7duxYMPPojNmzejpKQE27Ztw/nz5/2RkShsuVxuvPlJMxbnxqMwO17qOBTC7l6ZCX08z1YR3YpZ3fuvqqoKVVVVUx578cUXp43753/+Z++kIiIAQHVdDwxGCx5/oFjqKBTiFHIZvnLPAvzytXOobTCgrChZ6khEQYcrqhMFKJdbxMuHGpCZHIXVS1KkjkNhYN3KLCTFReDlgzxbRTQfLFVEAerTc13o6DPjL79QBDkXZSQ/UCpk+It1C9DQNoSzjf1SxyEKOixVRAHI5Rbx0nsNyE6Jxm0laVLHoTCyfnUWEmM1PFtFNA8sVUQB6MiZTnT1j+IvNxbx1iHkV0qFHFvWLcClVqPH2yER0VQsVUQBxuVy45WDDchNi8GaOS7ESOQNG8qzER+jwSuHGqSOQhRUWKqIAszHtZ3oHhjDIzxLRX7idLlhMFomP4bNNmysyEb91UF8cqYTBqMFZotd6phEAW9WSyoQkX84XW68cqgB+RmxKOc7/shPbA7XtKm+mEgVItQK7H7vMr64Nh9lRcmI1qokSkgUHHimiiiAvH+yHb2DFjyyscjr99gkmguFXIayIj26+sfQ0TcqdRyioMBSRRQgxqwO/PHAJSzKiceqRVx4kaRXnJeAaK0Kx+p74OY7AYk84vQfkR+YLXZYx52TnysidDAYLVPGvPZhI0ZG7dj+lQXoH7LedFs2h8tnOYmuJ5fLUL4kGe+f6sCpi32ouiNP6khEAY2lisgPrONO1DYYJj9va+9Etkk++fmw2YZDJ9pRlBMP48g4jCPjN91WYXacT7MSXW9BVhzONPbjjcNNuHdNDpQKTnAQ3Qz/dhAFgE/ruqFQyFBezIvTKbDIBAEVxakwDFlx8ESb1HGIAhpLFZHEOvrMaO0ZwRfKsxGpUUodh2ia7JRoLMzS4ZVDDbDanJ6/gChMsVQRScjtFnH0XDdiIlW4qzRD6jhENyQIArasW4Bhsw1vHrkqdRyigMVSRSSh81cHYBwZx20labxWhQJafroOa5am4k8fXoFhyOL5C4jCEF/FiSRiGrXheH0vslKikZsWI3UcIo/+5oFiuEXghdfPSx2FKCCxVBFJQBRFfFTTCUEA7irN4EKfFPCcLjcEAA/ckYcTF3rx3rFWGIyWyeVBeCsbIi6pQCSJ9n47uvqtuKs0g7f+oKBw7VY2iboIJMRq8Lv9F2FzutDT/eflQUoL9fx5prDGM1VEfma22HGp3Yr0pCgszo2XOg7RnMhlAu5ckYFRqwMnL/RJHYcooLBUEfmRKIo4XNsJEcDdZZz2o+CUmhiJxbnxqGvqh2mMSywQXcNSReRHl1qNaOs1ozAjArFRaqnjEM3bmqWp0KgUqGuxwuV2Sx2HKCCwVBH5ycCwFUfOdCE9KQq5ybzuhIKbRqXAXaUZMFlcOHa+V+o4RAGBpYrIDyzjDhw43gq1So4vlGdx2o9CQl56LLL1Kpy70o/W7hGp4xBJjqWKyMdEUcTv9l/EyJgdG8uzoeWtaCiELM6KQKJOg/dPt894I3CicMBSReRjb33SjJrLBqwpTkVaUpTUcYi8Si4TsLE8By6XiBfeOA+Xi9dXUfhiqSLyoQvNg/jtvgtYsTAJyxcmSR2HyCd00WrcVZqBKx3DeOlgg9RxiCTDUkXkI229I3j2P04gJUGLr1ct4XVUFNIKs+Nwx7I0vPp+Iz483S51HCJJcEV1oltktthhHZ+6Vs+gyYqf/P4UFHIB27+yAnI5//9Coe+v7lsE05gdP99zFgkxEVjGs7MUZvhKT3SLrONO1DYYJj+qz3fjJ78/hbFxB+6tyEF7nxlOXmdCYUAhl+H7j61Ghj4K//P3J9HWw3cEUnhhqSLyIofTjf1HWzAyZsemylwk6iKkjkTkV5ERSjzz+BpoVArsfPEYBk1WqSMR+Q1LFZGXOJwu7P+0Bb1GCzaszkI63+lHYSopLgLPPF6BsXEHnnnhGIbNNqkjEfkFSxWRF9gdLuz7pAXd/aNYvyoL+Rk6qSMRSSovPRb/8F/K0TNowfd/dZRrWFFYYKkiukWWcQfe+qQZfcYxfKEiG4XZcVJHIgoIyxYkYee2CvQPWfG9/30UA8OcCqTQxlJFdAtMozb8bHcN+oes2LgmBwU8Q0U0xdL8RPzwG2swZLbhe/9+FAajRepIRD7DUkU0T139o3jq55+gq38MmypzkJcWK3UkooC0ODcBzz6xBuYxO777y0/Q0m2SOhKRT7BUEc3D+asD+Pa/HcHYuANP/VUZslNjpI5EJDmnyw2D0XLDj7hoDZ76q5VwuUV895efoOZyn9RxibyOi38SzdGHpzvwi1fPICUhEs88XgGZIKC2wSB1LCLJ2Rwu1DUNzDjmgTvy8OHpDvzoNyfwt19aivsqc/2Ujsj3WKqIZsnucOH371zEW0eaUVKQiO89thpREUpeI0I0B1FaFb79aBn+71sX8O9/qsOVjmFsWbcAis/ddSBCo0C0ViVRSqL5YakimoX23hH8yx9r0Nozgs235+LrVcVQKjh7TjQfgkzA7cvSAAE4dLId9VcH8YWKbERFKCfHlBbqWaoo6LBUEc1AFEW8U92K/3irHhEaBZ55vAIrFyVLHYso6MlkAtYuT0dKvBYf1XTi1fcbsWF1FjKTo6WORjRvLFVEN9HWM4L/83od6q8OorRIj//+1RWIi9FIHYsopCzMikOiLgIHjrfirU+asXJRMlYu0ksdi2heWKqIPscy7sBL7zVg39FmaNUKfPPLJdhYkQOZTJA6GlFIio/R4C/WLcDhM104fakPV7uGERetgT5eK3U0ojlhqSL6jMvlxgenO/DHdy9heNSGL5Rn46/vW4TYKLXU0YhCnlIhx/pVWSjI0OFwbSf++fencF9lDr52/2JoNUrPGyAKACxVFPZEUcTJC734/TuX0NFnRmFWHJ7+ejkWZvF2M0T+lpMag7QvFKK524R3j7Xi6LluPHhnPu6/LZfligIeSxWFJbPFDuu4Ew3tQ3j94yZc6RhGcrwW//XLJSgr1EMQBAwMW+B2e96WzeHyfWCiMKJSyvHIF4qwqTIXu9+7jD+8cwl7P2rCA2vzUXV7LqL4rkAKUCxVFJbONvZj94HL6OofhVajwJ0r0rEoNwEyQcCZxn4AQGF2HBrahjxuizdQJvKNhVlx+OG2NWhsH8KeQ4146b3L+M8PGlFRnIp1qzKxfKEecl7rSAGEpYrChsstouZyH948fBV1TQPQahS4fVkaluQlTFt4kIikde2WNwCgi1LjiS8txX1rcnDkbCdOXOjFkbNd0EWrsXZFOlYvSsHivASuHUeSY6mikDc0Mo6DJ9vw3vE29A9ZER+jxsMbChEbpWKZIgpQN7vlzeLcBBRmxaG1ZwQ9g2N4t7oVbx1pRoRagRWFSSgtTMbS/ASkJkZCEHgWi/yLpYpCjsstoqljCGcb+3GmsR+XW41wuUUsW5CIv3mgGOVLUmA0jfN+fURBSi6XIT9Dh7+4ZyFiIlU4d6Ufpy714dTFPlTX9QAA4mPUWJKXiEU58ViQqUNOWgw0Kv6TR77FnzAKKm63CLPFjkHTOIwj49f9aoVxZOL3PQNjsIw7AQBZydHYWJGN20rSkJIQCQAwmsZ5cTlRiNCoFSgvTkV5cSpEUUSnYRT1zYO4cHUQ9c0D+ORsFwBAJgBZKTHIz4jFggwd8jN1yE2LhVopl/g7oFAyq1K1b98+/OpXv4LD4cBjjz2GRx99dMrzly5dwtNPP43R0VGsXLkSP/zhD6FQsK/R3ImiCOPIOLr7x9A9MIru/jH0Gscmy9PQyDicLnHa1+mi1IiP0SA+VoNMfTRUShky9NGIUE/8HHYPjKF7YGxyPC8uJwp+1193dY1aKUdZoR5lhXqIoohhsw1tfSNo7R5Ba88ITl7oxQenOgAAMkFAcoIWKfFaZOijkJMWi+Q4LXQxasRFaxCtVXIKkebEY/Pp6+vDrl27sHfvXqhUKjz88MMoLy9HQUHB5JinnnoK//RP/4Tly5fj+9//Pl599VU88sgjPg1OwcnhdGHU6oDRNHGWqX/Yiv4hC7oHxtAzMIaewTHY7H8+i6SQy5CSoEVibASW5iciKkKJqAgldNHqiY8oNWKj1FOujbI5XLjQPCjFt0dEfnSz664+rzA7Dm73xJmqO5aLGLU60D9khWHIAuPIOFp7R3D+6sC0/7DJZQJiIlWIiVQh9rP/uCXFRSAuWoO4mInXn4nXIg0iNQoWMPJcqqqrq1FRUQGdTgcA2LhxIw4cOIAnn3wSANDV1YXx8XEsX74cAPDQQw/h5z//+axKlShO/ADb7fZZB7bZbLMeC0xcX3PiQg9GzHa4IUIUJ/br/uzXiY9rj137PeAWRbhdIlwQIbpFuN0iXO6J593uz/7iCdd+EfD5v0vXPheEieenPjZ1/ORfRGFykxPbvPbYZ89Pfv75/X32dZ/f5pQhgjDl66fs77o8fX1mtJouTdvHte/f5XJPHBv3xLF1ut0Q3SJcronn3W4RDpcbNrsTNrsL43Ynxq/9anPB5Z5+lkkuExAfG4HUOA2Kc2ORqItAQmwEEmM10EWrIbsuiN3pRmP7EAA3zKNWmEet6Pjc9vLSYyHDzNN7LqfDK2Pmuy2VQpj2Nb7c33zHBPP+RPf0Y+zL/QXiMfDH/q7/WQ6K708AYrVyxGqjUJAeNTkmJy0G55sGYBl3wmpzwGpzwWpzYtzuhNXmhHnMimHTGM5dcUC84euYDDGRSkRqlYhQK6FVKxChVkCtkkMmCJDLBchkAhSyiV/lchnkMgEyeC5iPX2j6Bxp8DjOH9yYeJ2/9uESr/0ecLmv/fsw9d9Ml1uceM51/fiJ5wUIkMsAQS5MHCdBgEwuQC6bOD6CMPFvhOzacfvscZls4kyjXCaDIBMg/+y58iXJt3SPVk8d41pfudZfPk8Qb/bMZ37961/DYrFgx44dAIDXXnsNdXV1ePbZZwEAZ86cwXPPPYeXX34ZANDW1oZvfOMbeO+99zyGN5vNaGxs9DiOiIiIKFAsXLgQ0dHR0x73eKbqRp3r+lOcnp6fSWRkJBYuXAilkvPWREREFNhEUYTD4UBkZOQNn/dYqpKTk3H69OnJzw0GA/R6/ZTnBwb+PKfd398/5fmZyGSyGzY9IiIiokCk0dx8etHjyoeVlZU4duwYjEYjrFYrDh48iLVr104+n56eDrVajZqaGgDAG2+8MeV5IiIionDg8ZoqYGJJhV//+tdwOBzYsmULtm3bhm3btmH79u1YunQpLl++jKeffhpjY2NYvHgxfvKTn0Cl4g0viYiIKHzMqlQRERER0cx44zMiIiIiL2CpIiIiIvIClioiIiIiL2CpIiIiIvIClioiIiIiL2CpIiIiIvIClioiIiIiL2CpIiIiIvIClioiCkgnTpzA5s2bpY5BRDRrLFVEREREXqCQOgARhbcTJ05g165dyMzMxJUrV2C32/GDH/wAgiBMjtm7dy8OHToEmUyGtrY2KJVK/PSnP8XChQvR39+PZ555Bs3NzZDJZHj44YexdetW9Pb2YufOnejq6oIoinjwwQfx+OOPo7OzE1/72tdQUVGBs2fPwul04jvf+Q727NmD5uZmFBcX4/nnn4dMJkNtbS1+9rOfwWq1QhAEfOtb38Ldd98t4dEiokDGM1VEJLm6ujp8/etfxxtvvIEtW7bgl7/85bQxp06dwj/+4z/i7bffRmlpKX7zm98AAH74wx8iJycHBw4cwJ49e/Dqq6+ira0N3/72t1FeXo59+/bh5ZdfxltvvYX9+/cDADo7O7Fu3Trs378fFRUV+PGPf4znn38e+/fvx+nTp3H27FmYTCZ873vfw3PPPYfXX38dv/rVr7Bz5050d3f79dgQUfBgqSIiyaWlpWHRokUAgMWLF8NkMk0bs2TJEqSkpEwbU11dja9+9asAgOjoaLz99ttISkpCbW0tHn300cnHH3roIRw5cgQAoFQqsW7dOgBAVlYWVqxYgaioKKjVauj1ephMJpw9exb9/f34u7/7O3zxi1/EN77xDQiCgIaGBt8eDCIKWpz+IyLJaTSayd8LggBRFGc9RqFQTJkq7OjogE6nm7YNt9sNp9MJYKJUXf81SqVy2v5cLhfy8/Px2muvTT7W19eH+Pj4uX57RBQmeKaKiILamjVr8Kc//QkAYDab8bWvfQ1tbW1YtmwZdu/ePfn4G2+8gcrKyllvd/ny5Whra8OpU6cAAJcuXcLGjRthMBi8/00QUUjgmSoiCmo/+MEPsHPnTlRVVUEURTzxxBMoLi7Gz372M/zoRz/C3r17YbfbUVVVhYceeghdXV2z2m58fDx+/vOf47nnnoPNZoMoinjuueeQnp7u4++IiIKVIN7oPDsRERERzQmn/4iIiIi8gKWKiIiIyAtYqoiIiIi8gKWKiIiIyAtYqoiIiIi8gKWKiIiIyAtYqoiIiIi8gKWKiIiIyAv+P4ad6kCzhPH3AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot for NAVESPEND: total counts = 14989\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "sns.set(style=\"whitegrid\")\n",
    "\n",
    "df_explore = df[['nAge', 'nIncome', 'nAveSpend']]\n",
    "for cat in df_explore:\n",
    "    cat_num = df_explore[cat]\n",
    "    print(\"Plot for %s: total counts = %d\" % (cat.upper(), len(cat_num)))\n",
    "    f, ax = plt.subplots(figsize=(10,5))\n",
    "    plot = sns.distplot(cat_num)\n",
    "    plot.set_xticklabels(plot.get_xticklabels(), rotation=45)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now our numerical variables follow a normal distribution. We will start labeling our data using the most recent touchpoint. Lets see how many possible labels are there in our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['mailDiscount', 'mailGift', 'emailDiscount', 'emailDestination',\n",
       "       'smsDiscount', 'smsDestination', 'emailGift', 'smsGift',\n",
       "       'mailDestination'], dtype=object)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['recent_touchpoint'].unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I assume these are the latest touchpoint each customer went through before making a purchase."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: 'mailDiscount',\n",
       " 1: 'mailGift',\n",
       " 2: 'emailDiscount',\n",
       " 3: 'emailDestination',\n",
       " 4: 'smsDiscount',\n",
       " 5: 'smsDestination',\n",
       " 6: 'emailGift',\n",
       " 7: 'smsGift',\n",
       " 8: 'mailDestination'}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#create a mapping from labels to a unique integer and vice versa for labelling and prediction later\n",
    "labels = df['recent_touchpoint'].unique()\n",
    "i = 0\n",
    "idx2class = {} \n",
    "class2idx = {}\n",
    "for tp in labels:\n",
    "    idx2class[i] = tp\n",
    "    class2idx[tp] = i\n",
    "    i += 1\n",
    "idx2class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'mailDiscount': 0,\n",
       " 'mailGift': 1,\n",
       " 'emailDiscount': 2,\n",
       " 'emailDestination': 3,\n",
       " 'smsDiscount': 4,\n",
       " 'smsDestination': 5,\n",
       " 'emailGift': 6,\n",
       " 'smsGift': 7,\n",
       " 'mailDestination': 8}"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class2idx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['label'] = df['recent_touchpoint'].replace(class2idx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": false
   },
   "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>marital</th>\n",
       "      <th>segment</th>\n",
       "      <th>SocialMedia</th>\n",
       "      <th>income</th>\n",
       "      <th>creditRating</th>\n",
       "      <th>aveSpend</th>\n",
       "      <th>touchpoints</th>\n",
       "      <th>nTouchpoints</th>\n",
       "      <th>recent_touchpoint</th>\n",
       "      <th>...</th>\n",
       "      <th>D</th>\n",
       "      <th>A</th>\n",
       "      <th>F</th>\n",
       "      <th>E</th>\n",
       "      <th>New</th>\n",
       "      <th>G</th>\n",
       "      <th>nAveSpend</th>\n",
       "      <th>nIncome</th>\n",
       "      <th>nAge</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>38</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>5007.50</td>\n",
       "      <td>emailDestination mailDiscount</td>\n",
       "      <td>2</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>1.798751</td>\n",
       "      <td>-0.406079</td>\n",
       "      <td>0.037646</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>39</td>\n",
       "      <td>divorced</td>\n",
       "      <td>P3</td>\n",
       "      <td>U</td>\n",
       "      <td>100000.0</td>\n",
       "      <td>B</td>\n",
       "      <td>19.74</td>\n",
       "      <td>emailDestination mailGift</td>\n",
       "      <td>2</td>\n",
       "      <td>mailGift</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>0</td>\n",
       "      <td>-0.694311</td>\n",
       "      <td>0.912774</td>\n",
       "      <td>0.123259</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>41</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>D</td>\n",
       "      <td>320.10</td>\n",
       "      <td>emailDiscount mailGift emailDiscount</td>\n",
       "      <td>3</td>\n",
       "      <td>emailDiscount</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>0.559895</td>\n",
       "      <td>-0.406079</td>\n",
       "      <td>0.283522</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>60</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>152000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>39.47</td>\n",
       "      <td>mailDiscount</td>\n",
       "      <td>1</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>-0.362850</td>\n",
       "      <td>1.831224</td>\n",
       "      <td>5.199338</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>47</td>\n",
       "      <td>married</td>\n",
       "      <td>P2</td>\n",
       "      <td>Y</td>\n",
       "      <td>60000.0</td>\n",
       "      <td>A</td>\n",
       "      <td>19.89</td>\n",
       "      <td>mailDestination emailDestination emailDestination</td>\n",
       "      <td>3</td>\n",
       "      <td>emailDestination</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>-0.687284</td>\n",
       "      <td>-0.106841</td>\n",
       "      <td>0.744694</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14984</th>\n",
       "      <td>32</td>\n",
       "      <td>married</td>\n",
       "      <td>P1</td>\n",
       "      <td>U</td>\n",
       "      <td>33000.0</td>\n",
       "      <td>E</td>\n",
       "      <td>4280.63</td>\n",
       "      <td>emailDestination emailDestination emailDestina...</td>\n",
       "      <td>7</td>\n",
       "      <td>mailDiscount</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>1.748868</td>\n",
       "      <td>-1.340200</td>\n",
       "      <td>-0.659602</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14985</th>\n",
       "      <td>53</td>\n",
       "      <td>divorced</td>\n",
       "      <td>P3</td>\n",
       "      <td>N</td>\n",
       "      <td>61000.0</td>\n",
       "      <td>F</td>\n",
       "      <td>1332.29</td>\n",
       "      <td>smsDestination mailDiscount emailDestination e...</td>\n",
       "      <td>6</td>\n",
       "      <td>mailDiscount</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>1.280981</td>\n",
       "      <td>-0.051460</td>\n",
       "      <td>1.269664</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14986</th>\n",
       "      <td>30</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>44000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>4010.00</td>\n",
       "      <td>emailDestination emailDestination mailDiscount...</td>\n",
       "      <td>4</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>1.694923</td>\n",
       "      <td>-0.758004</td>\n",
       "      <td>-0.985610</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14987</th>\n",
       "      <td>52</td>\n",
       "      <td>married</td>\n",
       "      <td>P4</td>\n",
       "      <td>N</td>\n",
       "      <td>71000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>507.46</td>\n",
       "      <td>emailDestination mailGift emailDiscount smsDes...</td>\n",
       "      <td>8</td>\n",
       "      <td>emailDestination</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>0</td>\n",
       "      <td>0.779682</td>\n",
       "      <td>0.240671</td>\n",
       "      <td>1.169401</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14988</th>\n",
       "      <td>38</td>\n",
       "      <td>single</td>\n",
       "      <td>P4</td>\n",
       "      <td>U</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>C</td>\n",
       "      <td>512.67</td>\n",
       "      <td>emailDestination mailDiscount</td>\n",
       "      <td>2</td>\n",
       "      <td>mailDiscount</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>0</td>\n",
       "      <td>0.788882</td>\n",
       "      <td>0.047692</td>\n",
       "      <td>0.037646</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>14989 rows × 34 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       age   marital segment SocialMedia    income creditRating  aveSpend  \\\n",
       "0       38    single      P4           U   51000.0            C   5007.50   \n",
       "1       39  divorced      P3           U  100000.0            B     19.74   \n",
       "2       41   married      P4           N   51000.0            D    320.10   \n",
       "3       60   married      P4           U  152000.0            C     39.47   \n",
       "4       47   married      P2           Y   60000.0            A     19.89   \n",
       "...    ...       ...     ...         ...       ...          ...       ...   \n",
       "14984   32   married      P1           U   33000.0            E   4280.63   \n",
       "14985   53  divorced      P3           N   61000.0            F   1332.29   \n",
       "14986   30    single      P4           N   44000.0            C   4010.00   \n",
       "14987   52   married      P4           N   71000.0            C    507.46   \n",
       "14988   38    single      P4           U   65000.0            C    512.67   \n",
       "\n",
       "                                             touchpoints  nTouchpoints  \\\n",
       "0                          emailDestination mailDiscount             2   \n",
       "1                              emailDestination mailGift             2   \n",
       "2                   emailDiscount mailGift emailDiscount             3   \n",
       "3                                           mailDiscount             1   \n",
       "4      mailDestination emailDestination emailDestination             3   \n",
       "...                                                  ...           ...   \n",
       "14984  emailDestination emailDestination emailDestina...             7   \n",
       "14985  smsDestination mailDiscount emailDestination e...             6   \n",
       "14986  emailDestination emailDestination mailDiscount...             4   \n",
       "14987  emailDestination mailGift emailDiscount smsDes...             8   \n",
       "14988                      emailDestination mailDiscount             2   \n",
       "\n",
       "      recent_touchpoint  ...  D  A  F  E  New  G  nAveSpend   nIncome  \\\n",
       "0          mailDiscount  ...  0  0  0  0    0  0   1.798751 -0.406079   \n",
       "1              mailGift  ...  0  0  0  0    0  0  -0.694311  0.912774   \n",
       "2         emailDiscount  ...  1  0  0  0    0  0   0.559895 -0.406079   \n",
       "3          mailDiscount  ...  0  0  0  0    0  0  -0.362850  1.831224   \n",
       "4      emailDestination  ...  0  1  0  0    0  0  -0.687284 -0.106841   \n",
       "...                 ...  ... .. .. .. ..  ... ..        ...       ...   \n",
       "14984      mailDiscount  ...  0  0  0  1    0  0   1.748868 -1.340200   \n",
       "14985      mailDiscount  ...  0  0  1  0    0  0   1.280981 -0.051460   \n",
       "14986      mailDiscount  ...  0  0  0  0    0  0   1.694923 -0.758004   \n",
       "14987  emailDestination  ...  0  0  0  0    0  0   0.779682  0.240671   \n",
       "14988      mailDiscount  ...  0  0  0  0    0  0   0.788882  0.047692   \n",
       "\n",
       "           nAge  label  \n",
       "0      0.037646      0  \n",
       "1      0.123259      1  \n",
       "2      0.283522      2  \n",
       "3      5.199338      0  \n",
       "4      0.744694      3  \n",
       "...         ...    ...  \n",
       "14984 -0.659602      0  \n",
       "14985  1.269664      0  \n",
       "14986 -0.985610      0  \n",
       "14987  1.169401      3  \n",
       "14988  0.037646      0  \n",
       "\n",
       "[14989 rows x 34 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Select columns we want to use for our model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['age', 'marital', 'segment', 'SocialMedia', 'income', 'creditRating',\n",
       "       'aveSpend', 'touchpoints', 'nTouchpoints', 'recent_touchpoint',\n",
       "       'num_creditRating', 'single', 'divorced', 'married', 'unknown', 'P4',\n",
       "       'P3', 'P2', 'P1', 'U', 'N', 'Y', 'C', 'B', 'D', 'A', 'F', 'E', 'New',\n",
       "       'G', 'nAveSpend', 'nIncome', 'nAge', 'label'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When a dataset is imbalanced, i.e., the distribution of target variables is not uniform, we should make sure that the training-validation split is stratified. This ensures that the distribution of the target variable is preserved in both the training and validation datasets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def stratified_split(df, target, val_percent=0.2):\n",
    "    '''\n",
    "    Function to split a dataframe into train and validation sets, while preserving the ratio of the labels in the target variable\n",
    "    Inputs:\n",
    "    - df, the dataframe\n",
    "    - target, the target variable\n",
    "    - val_percent, the percentage of validation samples, default 0.2\n",
    "    Outputs:\n",
    "    - train_idxs, the indices of the training dataset\n",
    "    - val_idxs, the indices of the validation dataset\n",
    "    '''\n",
    "    classes=list(df[target].unique())\n",
    "    train_idxs, val_idxs = [], []\n",
    "    for c in classes:\n",
    "        idx=list(df[df[target]==c].index)\n",
    "        np.random.shuffle(idx)\n",
    "        val_size=int(len(idx)*val_percent)\n",
    "        val_idxs+=idx[:val_size]\n",
    "        train_idxs+=idx[val_size:]\n",
    "    return train_idxs, val_idxs\n",
    "\n",
    "\n",
    "train_idxs, val_idxs = stratified_split(df, 'label', val_percent=0.25)\n",
    "\n",
    "val_idxs, test_idxs = stratified_split(df[df.index.isin(val_idxs)], 'label', val_percent=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_stratified(df, col):\n",
    "    '''\n",
    "    Analyzes the ratio of different classes in a categorical variable within a dataframe\n",
    "    Inputs:\n",
    "    - dataframe\n",
    "    - categorical column to be analyzed\n",
    "    Returns: None\n",
    "    '''\n",
    "    classes=list(df[col].unique())\n",
    "    \n",
    "    for c in classes:\n",
    "        print(f'Proportion of records with {c}: {len(df[df[col]==c])*1./len(df):0.2} ({len(df[df[col]==c])} / {len(df)})')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------- STRATIFIED SAMPLING REPORT ----------\n",
      "-------- Label proportions in FULL data --------\n",
      "Proportion of records with 0: 0.12 (1837 / 14989)\n",
      "Proportion of records with 1: 0.043 (643 / 14989)\n",
      "Proportion of records with 2: 0.22 (3229 / 14989)\n",
      "Proportion of records with 3: 0.37 (5542 / 14989)\n",
      "Proportion of records with 4: 0.077 (1149 / 14989)\n",
      "Proportion of records with 5: 0.039 (585 / 14989)\n",
      "Proportion of records with 6: 0.048 (722 / 14989)\n",
      "Proportion of records with 7: 0.04 (598 / 14989)\n",
      "Proportion of records with 8: 0.046 (684 / 14989)\n",
      "-------- Label proportions in TRAIN data --------\n",
      "Proportion of records with 1: 0.043 (483 / 11245)\n",
      "Proportion of records with 2: 0.22 (2422 / 11245)\n",
      "Proportion of records with 0: 0.12 (1378 / 11245)\n",
      "Proportion of records with 3: 0.37 (4157 / 11245)\n",
      "Proportion of records with 4: 0.077 (862 / 11245)\n",
      "Proportion of records with 7: 0.04 (449 / 11245)\n",
      "Proportion of records with 6: 0.048 (542 / 11245)\n",
      "Proportion of records with 8: 0.046 (513 / 11245)\n",
      "Proportion of records with 5: 0.039 (439 / 11245)\n",
      "------ Label proportions in VALIDATION data -----\n",
      "Proportion of records with 2: 0.22 (404 / 1875)\n",
      "Proportion of records with 3: 0.37 (693 / 1875)\n",
      "Proportion of records with 1: 0.043 (80 / 1875)\n",
      "Proportion of records with 4: 0.077 (144 / 1875)\n",
      "Proportion of records with 7: 0.04 (75 / 1875)\n",
      "Proportion of records with 0: 0.12 (230 / 1875)\n",
      "Proportion of records with 6: 0.048 (90 / 1875)\n",
      "Proportion of records with 8: 0.046 (86 / 1875)\n",
      "Proportion of records with 5: 0.039 (73 / 1875)\n",
      "-------- Label proportions in TEST data ---------\n",
      "Proportion of records with 0: 0.12 (229 / 1869)\n",
      "Proportion of records with 5: 0.039 (73 / 1869)\n",
      "Proportion of records with 6: 0.048 (90 / 1869)\n",
      "Proportion of records with 4: 0.077 (143 / 1869)\n",
      "Proportion of records with 2: 0.22 (403 / 1869)\n",
      "Proportion of records with 3: 0.37 (692 / 1869)\n",
      "Proportion of records with 7: 0.04 (74 / 1869)\n",
      "Proportion of records with 1: 0.043 (80 / 1869)\n",
      "Proportion of records with 8: 0.045 (85 / 1869)\n"
     ]
    }
   ],
   "source": [
    "print('---------- STRATIFIED SAMPLING REPORT ----------')\n",
    "print('-------- Label proportions in FULL data --------')\n",
    "test_stratified(df, 'label')\n",
    "print('-------- Label proportions in TRAIN data --------')\n",
    "test_stratified(df[df.index.isin(train_idxs)], 'label')\n",
    "print('------ Label proportions in VALIDATION data -----')\n",
    "test_stratified(df[df.index.isin(val_idxs)], 'label')\n",
    "print('-------- Label proportions in TEST data ---------')\n",
    "test_stratified(df[df.index.isin(test_idxs)], 'label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Retrieved Training Data\n",
      "Retrieved Validation Data\n",
      "Retrieved Test Data\n"
     ]
    }
   ],
   "source": [
    "train_df = df[df.index.isin(train_idxs)]\n",
    "X_train = train_df[['nTouchpoints', 'single', 'divorced', 'married', 'unknown', 'P4',\n",
    "       'P3', 'P2', 'P1', 'U', 'N', 'Y', 'C', 'B', 'D', 'A', 'F', 'E', 'New',\n",
    "       'G', 'nAveSpend', 'nIncome', 'nAge']].values\n",
    "Y_train = train_df[['label']].values\n",
    "print('Retrieved Training Data')\n",
    "val_df = df[df.index.isin(val_idxs)]\n",
    "X_val = val_df[['nTouchpoints', 'single', 'divorced', 'married', 'unknown', 'P4',\n",
    "       'P3', 'P2', 'P1', 'U', 'N', 'Y', 'C', 'B', 'D', 'A', 'F', 'E', 'New',\n",
    "       'G', 'nAveSpend', 'nIncome', 'nAge']].values\n",
    "Y_val = val_df[['label']].values\n",
    "print('Retrieved Validation Data')\n",
    "test_df = df[df.index.isin(test_idxs)]\n",
    "X_test = test_df[['nTouchpoints', 'single', 'divorced', 'married', 'unknown', 'P4',\n",
    "       'P3', 'P2', 'P1', 'U', 'N', 'Y', 'C', 'B', 'D', 'A', 'F', 'E', 'New',\n",
    "       'G', 'nAveSpend', 'nIncome', 'nAge']].values\n",
    "Y_test = test_df[['label']].values\n",
    "print('Retrieved Test Data')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Building"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### For simplicity, I will explore Logistic regression(OneVsRest) as the baseline model and progressively use more complicated ensemble models for our prediction task such as XGBoost.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import accuracy_score, f1_score\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "#store data, all in numpy arrays\n",
    "training_data = {'X_train':X_train,'Y_train':Y_train,\n",
    "                'X_val': X_val,'Y_val':Y_val,\n",
    "                'X_test': X_test,'Y_test':Y_test}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 100 out of 100 | elapsed:    1.3s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,\n",
       "                       criterion='gini', max_depth=None, max_features='auto',\n",
       "                       max_leaf_nodes=None, max_samples=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "                       n_jobs=None, oob_score=False, random_state=27, verbose=1,\n",
       "                       warm_start=False)"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "clf = RandomForestClassifier(n_jobs=None,random_state=27,\n",
    "                       verbose=1)\n",
    "clf.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 100 out of 100 | elapsed:    0.1s finished\n"
     ]
    }
   ],
   "source": [
    "predicted_labels = clf.predict(training_data['X_test'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7667201712145533"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(training_data['Y_test'], predicted_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 40 candidates, totalling 200 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:    1.1s\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:    2.8s\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:    4.4s\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:    8.5s\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:   13.1s\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed:   16.0s\n",
      "[Parallel(n_jobs=-1)]: Done  56 tasks      | elapsed:   19.5s\n",
      "[Parallel(n_jobs=-1)]: Done  69 tasks      | elapsed:   26.2s\n",
      "[Parallel(n_jobs=-1)]: Done  82 tasks      | elapsed:   32.2s\n",
      "[Parallel(n_jobs=-1)]: Done  97 tasks      | elapsed:   37.3s\n",
      "[Parallel(n_jobs=-1)]: Done 112 tasks      | elapsed:   46.3s\n",
      "[Parallel(n_jobs=-1)]: Done 129 tasks      | elapsed:   52.1s\n",
      "[Parallel(n_jobs=-1)]: Done 146 tasks      | elapsed:  1.0min\n",
      "[Parallel(n_jobs=-1)]: Done 165 tasks      | elapsed:  1.2min\n",
      "[Parallel(n_jobs=-1)]: Done 184 tasks      | elapsed:  1.3min\n",
      "[Parallel(n_jobs=-1)]: Done 200 out of 200 | elapsed:  1.5min finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 250 out of 250 | elapsed:    2.1s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                              class_weight=None,\n",
       "                                              criterion='gini', max_depth=None,\n",
       "                                              max_features='auto',\n",
       "                                              max_leaf_nodes=None,\n",
       "                                              max_samples=None,\n",
       "                                              min_impurity_decrease=0.0,\n",
       "                                              min_impurity_split=None,\n",
       "                                              min_samples_leaf=1,\n",
       "                                              min_samples_split=2,\n",
       "                                              min_weight_fraction_leaf=0.0,\n",
       "                                              n_estimators=100, n_jobs=None,\n",
       "                                              oob_score=False, random_state=27,\n",
       "                                              verbose=1, warm_start=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'criterion': ['gini'], 'max_depth': [8, 9, 10, 11, 12],\n",
       "                         'max_features': ['auto'],\n",
       "                         'n_estimators': range(100, 500, 50)},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring='f1_micro', verbose=10)"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "params = {\n",
    "    'n_estimators'      : range(100,500,50),\n",
    "    'max_depth'         : [8, 9, 10, 11, 12],\n",
    "    'max_features': ['auto'],\n",
    "    'criterion' :['gini']\n",
    "}\n",
    "#metrics to consider: f1_micro, f1_macro, roc_auc_ovr\n",
    "gsearch1 = GridSearchCV(estimator = clf, param_grid = params, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=5)\n",
    "gsearch1.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.5012894619831036params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 100}\",\n",
       "  1: \"mean:0.5019119608714984params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 150}\",\n",
       "  2: \"mean:0.5002223210315696params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 200}\",\n",
       "  3: \"mean:0.5024455313472653params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 250}\",\n",
       "  4: \"mean:0.500578034682081params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 300}\",\n",
       "  5: \"mean:0.50102267674522params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 350}\",\n",
       "  6: \"mean:0.5020898176967541params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 400}\",\n",
       "  7: \"mean:0.5022676745220098params{'criterion': 'gini', 'max_depth': 8, 'max_features': 'auto', 'n_estimators': 450}\",\n",
       "  8: \"mean:0.5226322810137839params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 100}\",\n",
       "  9: \"mean:0.5232547799021787params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 150}\",\n",
       "  10: \"mean:0.5232547799021787params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 200}\",\n",
       "  11: \"mean:0.5250333481547355params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 250}\",\n",
       "  12: \"mean:0.525300133392619params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 300}\",\n",
       "  13: \"mean:0.5245887060915962params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 350}\",\n",
       "  14: \"mean:0.5277901289461983params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 400}\",\n",
       "  15: \"mean:0.5272565584704313params{'criterion': 'gini', 'max_depth': 9, 'max_features': 'auto', 'n_estimators': 450}\",\n",
       "  16: \"mean:0.5453090262338817params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 100}\",\n",
       "  17: \"mean:0.5469986660738105params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 150}\",\n",
       "  18: \"mean:0.5469097376611828params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 200}\",\n",
       "  19: \"mean:0.5469097376611828params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 250}\",\n",
       "  20: \"mean:0.5469097376611828params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 300}\",\n",
       "  21: \"mean:0.5477990217874611params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 350}\",\n",
       "  22: \"mean:0.5476211649622054params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 400}\",\n",
       "  23: \"mean:0.5482436638506003params{'criterion': 'gini', 'max_depth': 10, 'max_features': 'auto', 'n_estimators': 450}\",\n",
       "  24: \"mean:0.574744330813695params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 100}\",\n",
       "  25: \"mean:0.5764339706536239params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 150}\",\n",
       "  26: \"mean:0.5734993330369053params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 200}\",\n",
       "  27: \"mean:0.5732325477990218params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 250}\",\n",
       "  28: \"mean:0.5726989773232548params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 300}\",\n",
       "  29: \"mean:0.5727879057358826params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 350}\",\n",
       "  30: \"mean:0.5724321920853713params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 400}\",\n",
       "  31: \"mean:0.5715429079590929params{'criterion': 'gini', 'max_depth': 11, 'max_features': 'auto', 'n_estimators': 450}\",\n",
       "  32: \"mean:0.5936860827034238params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 100}\",\n",
       "  33: \"mean:0.597332147621165params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 150}\",\n",
       "  34: \"mean:0.5992885726989773params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 200}\",\n",
       "  35: \"mean:0.5996442863494886params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 250}\",\n",
       "  36: \"mean:0.5974210760337929params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 300}\",\n",
       "  37: \"mean:0.5967985771453979params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 350}\",\n",
       "  38: \"mean:0.5977767896843041params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 400}\",\n",
       "  39: \"mean:0.5991107158737217params{'criterion': 'gini', 'max_depth': 12, 'max_features': 'auto', 'n_estimators': 450}\"},\n",
       " {'best_mean': 0.5996442863494886,\n",
       "  'best_param': {'criterion': 'gini',\n",
       "   'max_depth': 12,\n",
       "   'max_features': 'auto',\n",
       "   'n_estimators': 250}})"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:    2.6s\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:    5.5s\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:    6.2s\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:   11.9s\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:   15.3s\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed:   19.3s\n",
      "[Parallel(n_jobs=-1)]: Done  56 tasks      | elapsed:   22.7s\n",
      "[Parallel(n_jobs=-1)]: Done  69 tasks      | elapsed:   29.0s\n",
      "[Parallel(n_jobs=-1)]: Done  82 tasks      | elapsed:   35.6s\n",
      "[Parallel(n_jobs=-1)]: Done  97 tasks      | elapsed:   44.6s\n",
      "[Parallel(n_jobs=-1)]: Done 112 tasks      | elapsed:   49.9s\n",
      "[Parallel(n_jobs=-1)]: Done 129 tasks      | elapsed:  1.1min\n",
      "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:  1.3min finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    3.0s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                              class_weight=None,\n",
       "                                              criterion='gini', max_depth=12,\n",
       "                                              max_features='auto',\n",
       "                                              max_leaf_nodes=None,\n",
       "                                              max_samples=None,\n",
       "                                              min_impurity_decrease=0.0,\n",
       "                                              min_impurity_split=None,\n",
       "                                              min_samples_leaf=1,\n",
       "                                              min_samples_split=2,\n",
       "                                              min_weight_fraction_leaf=0.0,\n",
       "                                              n_estimators=250, n_jobs=None,\n",
       "                                              oob_score=False, random_state=27,\n",
       "                                              verbose=1, warm_start=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'max_depth': [11, 12, 13],\n",
       "                         'n_estimators': range(200, 300, 10)},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring='f1_micro', verbose=10)"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf2 = gsearch1.best_estimator_\n",
    "\n",
    "params1 = {\n",
    "    'n_estimators'      : range(200,300,10),\n",
    "    'max_depth'         : [11, 12,13]\n",
    "}\n",
    "#metrics to consider: f1_micro, f1_macro, roc_auc_ovr\n",
    "gsearch2 = GridSearchCV(estimator = clf2, param_grid = params1, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=5)\n",
    "gsearch2.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.5734993330369053params{'max_depth': 11, 'n_estimators': 200}\",\n",
       "  1: \"mean:0.5741218319253002params{'max_depth': 11, 'n_estimators': 210}\",\n",
       "  2: \"mean:0.5721654068474878params{'max_depth': 11, 'n_estimators': 220}\",\n",
       "  3: \"mean:0.5726989773232548params{'max_depth': 11, 'n_estimators': 230}\",\n",
       "  4: \"mean:0.5734104046242775params{'max_depth': 11, 'n_estimators': 240}\",\n",
       "  5: \"mean:0.5732325477990218params{'max_depth': 11, 'n_estimators': 250}\",\n",
       "  6: \"mean:0.5727879057358827params{'max_depth': 11, 'n_estimators': 260}\",\n",
       "  7: \"mean:0.572610048910627params{'max_depth': 11, 'n_estimators': 270}\",\n",
       "  8: \"mean:0.5721654068474878params{'max_depth': 11, 'n_estimators': 280}\",\n",
       "  9: \"mean:0.57207647843486params{'max_depth': 11, 'n_estimators': 290}\",\n",
       "  10: \"mean:0.5992885726989773params{'max_depth': 12, 'n_estimators': 200}\",\n",
       "  11: \"mean:0.5998221431747444params{'max_depth': 12, 'n_estimators': 210}\",\n",
       "  12: \"mean:0.598399288572699params{'max_depth': 12, 'n_estimators': 220}\",\n",
       "  13: \"mean:0.5985771453979547params{'max_depth': 12, 'n_estimators': 230}\",\n",
       "  14: \"mean:0.5991996442863494params{'max_depth': 12, 'n_estimators': 240}\",\n",
       "  15: \"mean:0.5996442863494886params{'max_depth': 12, 'n_estimators': 250}\",\n",
       "  16: \"mean:0.5982214317474435params{'max_depth': 12, 'n_estimators': 260}\",\n",
       "  17: \"mean:0.598399288572699params{'max_depth': 12, 'n_estimators': 270}\",\n",
       "  18: \"mean:0.5979546465095599params{'max_depth': 12, 'n_estimators': 280}\",\n",
       "  19: \"mean:0.5968875055580258params{'max_depth': 12, 'n_estimators': 290}\",\n",
       "  20: \"mean:0.625255669186305params{'max_depth': 13, 'n_estimators': 200}\",\n",
       "  21: \"mean:0.6270342374388617params{'max_depth': 13, 'n_estimators': 210}\",\n",
       "  22: \"mean:0.6255224544241885params{'max_depth': 13, 'n_estimators': 220}\",\n",
       "  23: \"mean:0.6247220987105381params{'max_depth': 13, 'n_estimators': 230}\",\n",
       "  24: \"mean:0.6250778123610493params{'max_depth': 13, 'n_estimators': 240}\",\n",
       "  25: \"mean:0.6249888839484214params{'max_depth': 13, 'n_estimators': 250}\",\n",
       "  26: \"mean:0.6227656736327256params{'max_depth': 13, 'n_estimators': 260}\",\n",
       "  27: \"mean:0.6239217429968875params{'max_depth': 13, 'n_estimators': 270}\",\n",
       "  28: \"mean:0.6234771009337484params{'max_depth': 13, 'n_estimators': 280}\",\n",
       "  29: \"mean:0.6232992441084926params{'max_depth': 13, 'n_estimators': 290}\"},\n",
       " {'best_mean': 0.6270342374388617,\n",
       "  'best_param': {'max_depth': 13, 'n_estimators': 210}})"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 12 candidates, totalling 60 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:    4.7s\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:    9.1s\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:    9.7s\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:   17.3s\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:   21.7s\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed:   26.1s\n",
      "[Parallel(n_jobs=-1)]: Done  52 out of  60 | elapsed:   30.5s remaining:    4.7s\n",
      "[Parallel(n_jobs=-1)]: Done  60 out of  60 | elapsed:   33.8s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    3.2s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                              class_weight=None,\n",
       "                                              criterion='gini', max_depth=13,\n",
       "                                              max_features='auto',\n",
       "                                              max_leaf_nodes=None,\n",
       "                                              max_samples=None,\n",
       "                                              min_impurity_decrease=0.0,\n",
       "                                              min_impurity_split=None,\n",
       "                                              min_samples_leaf=1,\n",
       "                                              min_samples_split=2,\n",
       "                                              min_weight_fraction_leaf=0.0,\n",
       "                                              n_estimators=210, n_jobs=None,\n",
       "                                              oob_score=False, random_state=27,\n",
       "                                              verbose=1, warm_start=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'max_depth': [13, 14, 15],\n",
       "                         'n_estimators': range(200, 220, 5)},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring='f1_micro', verbose=10)"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf3 = gsearch2.best_estimator_\n",
    "\n",
    "params2 = {\n",
    "    'n_estimators'      : range(200,220,5),\n",
    "    'max_depth'         : [13,14,15]\n",
    "}\n",
    "#metrics to consider: f1_micro, f1_macro, roc_auc_ovr\n",
    "gsearch3 = GridSearchCV(estimator = clf3, param_grid = params2, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=5)\n",
    "gsearch3.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.625255669186305params{'max_depth': 13, 'n_estimators': 200}\",\n",
       "  1: \"mean:0.6259670964873277params{'max_depth': 13, 'n_estimators': 205}\",\n",
       "  2: \"mean:0.6270342374388617params{'max_depth': 13, 'n_estimators': 210}\",\n",
       "  3: \"mean:0.6265895953757226params{'max_depth': 13, 'n_estimators': 215}\",\n",
       "  4: \"mean:0.6510449088483771params{'max_depth': 14, 'n_estimators': 200}\",\n",
       "  5: \"mean:0.6509559804357492params{'max_depth': 14, 'n_estimators': 205}\",\n",
       "  6: \"mean:0.6508670520231215params{'max_depth': 14, 'n_estimators': 210}\",\n",
       "  7: \"mean:0.6507781236104936params{'max_depth': 14, 'n_estimators': 215}\",\n",
       "  8: \"mean:0.669364161849711params{'max_depth': 15, 'n_estimators': 200}\",\n",
       "  9: \"mean:0.6689195197865718params{'max_depth': 15, 'n_estimators': 205}\",\n",
       "  10: \"mean:0.6701645175633615params{'max_depth': 15, 'n_estimators': 210}\",\n",
       "  11: \"mean:0.669897732325478params{'max_depth': 15, 'n_estimators': 215}\"},\n",
       " {'best_mean': 0.6701645175633615,\n",
       "  'best_param': {'max_depth': 15, 'n_estimators': 210}})"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 6 candidates, totalling 30 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:    4.6s\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:    9.6s\n",
      "[Parallel(n_jobs=-1)]: Done  19 out of  30 | elapsed:   14.9s remaining:    8.6s\n",
      "[Parallel(n_jobs=-1)]: Done  23 out of  30 | elapsed:   15.4s remaining:    4.7s\n",
      "[Parallel(n_jobs=-1)]: Done  27 out of  30 | elapsed:   19.0s remaining:    2.1s\n",
      "[Parallel(n_jobs=-1)]: Done  30 out of  30 | elapsed:   19.3s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    2.7s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                              class_weight=None,\n",
       "                                              criterion='gini', max_depth=15,\n",
       "                                              max_features='auto',\n",
       "                                              max_leaf_nodes=None,\n",
       "                                              max_samples=None,\n",
       "                                              min_impurity_decrease=0.0,\n",
       "                                              min_impurity_split=None,\n",
       "                                              min_samples_leaf=1,\n",
       "                                              min_samples_split=2,\n",
       "                                              min_weight_fraction_leaf=0.0,\n",
       "                                              n_estimators=210, n_jobs=None,\n",
       "                                              oob_score=False, random_state=27,\n",
       "                                              verbose=1, warm_start=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'max_depth': range(14, 20)}, pre_dispatch='2*n_jobs',\n",
       "             refit=True, return_train_score=False, scoring='f1_micro',\n",
       "             verbose=10)"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf4 = gsearch3.best_estimator_\n",
    "\n",
    "params3 = {\n",
    "    'max_depth'         : range(14,20,1)\n",
    "}\n",
    "#metrics to consider: f1_micro, f1_macro, roc_auc_ovr\n",
    "gsearch4 = GridSearchCV(estimator = clf4, param_grid = params3, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=5)\n",
    "gsearch4.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.6508670520231215params{'max_depth': 14}\",\n",
       "  1: \"mean:0.6701645175633615params{'max_depth': 15}\",\n",
       "  2: \"mean:0.6927523343708314params{'max_depth': 16}\",\n",
       "  3: \"mean:0.7013783903957315params{'max_depth': 17}\",\n",
       "  4: \"mean:0.7136505113383727params{'max_depth': 18}\",\n",
       "  5: \"mean:0.7153401511783015params{'max_depth': 19}\"},\n",
       " {'best_mean': 0.7153401511783015, 'best_param': {'max_depth': 19}})"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 16 candidates, totalling 80 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:    4.7s\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:   10.8s\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:   11.2s\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:   20.8s\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:   25.8s\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed:   30.6s\n",
      "[Parallel(n_jobs=-1)]: Done  56 tasks      | elapsed:   35.8s\n",
      "[Parallel(n_jobs=-1)]: Done  74 out of  80 | elapsed:   50.1s remaining:    4.1s\n",
      "[Parallel(n_jobs=-1)]: Done  80 out of  80 | elapsed:   50.5s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    2.7s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                              class_weight=None,\n",
       "                                              criterion='gini', max_depth=19,\n",
       "                                              max_features='auto',\n",
       "                                              max_leaf_nodes=None,\n",
       "                                              max_samples=None,\n",
       "                                              min_impurity_decrease=0.0,\n",
       "                                              min_impurity_split=None,\n",
       "                                              min_samples_leaf=1,\n",
       "                                              min_samples_split=2,\n",
       "                                              min_weight_fraction_leaf=0.0,\n",
       "                                              n_estimators=210, n_jobs=None,\n",
       "                                              oob_score=False, random_state=27,\n",
       "                                              verbose=1, warm_start=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'max_depth': range(19, 50, 2)},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring='f1_micro', verbose=10)"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf5 = gsearch4.best_estimator_\n",
    "\n",
    "params4 = {\n",
    "    'max_depth'         : range(19,50,2)\n",
    "}\n",
    "#metrics to consider: f1_micro, f1_macro, roc_auc_ovr\n",
    "gsearch5 = GridSearchCV(estimator = clf5, param_grid = params4, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=5)\n",
    "gsearch5.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.7153401511783015params{'max_depth': 19}\",\n",
       "  1: \"mean:0.7201422854602045params{'max_depth': 21}\",\n",
       "  2: \"mean:0.7244108492663406params{'max_depth': 23}\",\n",
       "  3: \"mean:0.7262783459315252params{'max_depth': 25}\",\n",
       "  4: \"mean:0.7250333481547354params{'max_depth': 27}\",\n",
       "  5: \"mean:0.7256558470431304params{'max_depth': 29}\",\n",
       "  6: \"mean:0.7244108492663405params{'max_depth': 31}\",\n",
       "  7: \"mean:0.7256558470431302params{'max_depth': 33}\",\n",
       "  8: \"mean:0.7251222765673633params{'max_depth': 35}\",\n",
       "  9: \"mean:0.7247665629168519params{'max_depth': 37}\",\n",
       "  10: \"mean:0.7247665629168519params{'max_depth': 39}\",\n",
       "  11: \"mean:0.7247665629168519params{'max_depth': 41}\",\n",
       "  12: \"mean:0.7247665629168519params{'max_depth': 43}\",\n",
       "  13: \"mean:0.7247665629168519params{'max_depth': 45}\",\n",
       "  14: \"mean:0.7247665629168519params{'max_depth': 47}\",\n",
       "  15: \"mean:0.7247665629168519params{'max_depth': 49}\"},\n",
       " {'best_mean': 0.7262783459315252, 'best_param': {'max_depth': 25}})"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 3 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 out of  15 | elapsed:    4.6s remaining:   29.6s\n",
      "[Parallel(n_jobs=-1)]: Done   4 out of  15 | elapsed:    4.6s remaining:   12.6s\n",
      "[Parallel(n_jobs=-1)]: Done   6 out of  15 | elapsed:    4.6s remaining:    6.9s\n",
      "[Parallel(n_jobs=-1)]: Done   8 out of  15 | elapsed:    4.7s remaining:    4.2s\n",
      "[Parallel(n_jobs=-1)]: Done  10 out of  15 | elapsed:    8.6s remaining:    4.3s\n",
      "[Parallel(n_jobs=-1)]: Done  12 out of  15 | elapsed:    8.6s remaining:    2.2s\n",
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed:    8.7s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    3.1s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=RandomForestClassifier(bootstrap=True, ccp_alpha=0.0,\n",
       "                                              class_weight=None,\n",
       "                                              criterion='gini', max_depth=25,\n",
       "                                              max_features='auto',\n",
       "                                              max_leaf_nodes=None,\n",
       "                                              max_samples=None,\n",
       "                                              min_impurity_decrease=0.0,\n",
       "                                              min_impurity_split=None,\n",
       "                                              min_samples_leaf=1,\n",
       "                                              min_samples_split=2,\n",
       "                                              min_weight_fraction_leaf=0.0,\n",
       "                                              n_estimators=210, n_jobs=None,\n",
       "                                              oob_score=False, random_state=27,\n",
       "                                              verbose=1, warm_start=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'max_depth': [24, 25, 26]}, pre_dispatch='2*n_jobs',\n",
       "             refit=True, return_train_score=False, scoring='f1_micro',\n",
       "             verbose=10)"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf6 = gsearch5.best_estimator_\n",
    "\n",
    "params5 = {\n",
    "    'max_depth'         : [24,25,26]\n",
    "}\n",
    "#metrics to consider: f1_micro, f1_macro, roc_auc_ovr\n",
    "gsearch6 = GridSearchCV(estimator = clf6, param_grid = params5, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=5)\n",
    "gsearch6.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.7234326367274344params{'max_depth': 24}\",\n",
       "  1: \"mean:0.7262783459315252params{'max_depth': 25}\",\n",
       "  2: \"mean:0.7222765673632725params{'max_depth': 26}\"},\n",
       " {'best_mean': 0.7262783459315252, 'best_param': {'max_depth': 25}})"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_clf = gsearch6.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    2.7s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    0.1s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Accuracy:1.0\n",
      "Train F1-Score(Micro):1.0\n",
      "------\n",
      "Test Accuracy:0.7768860353130016\n",
      "Test F1-Score(Micro):0.7768860353130017\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done 210 out of 210 | elapsed:    0.6s finished\n"
     ]
    }
   ],
   "source": [
    "final_clf.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))\n",
    "predicted_labels = final_clf.predict(training_data['X_test'])\n",
    "train_pred = final_clf.predict(training_data['X_train'])\n",
    "print('Train Accuracy:'+str(accuracy_score(training_data['Y_train'], train_pred)))\n",
    "print('Train F1-Score(Micro):'+str(f1_score(training_data['Y_train'], train_pred,average='micro')))\n",
    "print('------')\n",
    "print('Test Accuracy:'+str(accuracy_score(training_data['Y_test'], predicted_labels)))\n",
    "print('Test F1-Score(Micro):'+str(f1_score(training_data['Y_test'], predicted_labels,average='micro')))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Achieved F1-Score(Micro) of 0.776886 with RandomForestClassifier. There is overfitting in this model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAF3CAYAAACSQ46hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeVxU5f///+cwLGJWLm8gs91E376RSklITXMLNUBRM8O0LNdyCZe0ct93Lbc0S3PLHRQrNNMyldy+lZp7VuYGuKGiyDLn94c/5iOpYXKOMPm4327ebpxzZl7nGlnmOde5znXZDMMwBAAAAEu45XcDAAAA/s0IWwAAABYibAEAAFiIsAUAAGAhwhYAAICFCFsAAAAWcs/vBgAo+MqWLSt/f3+5uf3f57OAgAANHTr0lurt2LFDS5Ys0aBBg8xq4jXKli2rhIQEFS9e3LJzXM/ixYuVnp6uFi1a3NbzAii4CFsAbspnn31mWnA5ePCgEhMTTalV0Gzfvl1lypTJ72YAKEAIWwDy5Ndff9XQoUN19uxZZWVlqWXLlmratKkcDoeGDRumn3/+WampqTIMQ0OGDNH999+vDz/8UOfPn9e7776rRo0aafDgwVq5cqUkafPmzc7tiRMn6qefflJSUpLKli2rMWPGaOrUqVq9erUcDodKlSql/v37y8/P74btO3LkiF599VWFhITop59+UmZmpt555x0tXLhQhw4dUkBAgMaNG6djx46pZcuWqly5svbu3SvDMNSvXz8FBQUpIyNDI0aMUEJCgux2uwIDA/Xuu++qSJEiqlWrlgIDA7Vv3z5169ZNa9eu1caNG1WoUCGFhoaqX79+OnXqlJKTk1WqVClNmDBBJUqUUK1atRQZGamEhAQdP35c9evX1zvvvCNJWrJkiWbOnCk3NzcVK1ZMI0eOVMmSJbV27VpNnTpVGRkZKlSokHr16qWnnnrqtnyfAeSBAQC58Pf3N8LCwoyIiAjnv5MnTxoZGRlGgwYNjF27dhmGYRjnzp0z6tevb/z444/G//t//8/o3LmzkZWVZRiGYUybNs1o3769YRiGsXTpUqNdu3aGYRjGDz/8YLzwwgvOc129/eGHHxqhoaFGRkaGYRiGERMTY7z99tvO7QULFhht2rS5YZtPnTpl/Pnnn4a/v7+xZs0awzAMo1+/fkbNmjWN8+fPG2lpaUbVqlWN7du3Ox+3YsUKwzAM49tvvzWqVq1qpKenGx988IHRqVMnIz093cjKyjJ69+5t9O3b1zAMw6hZs6YxadIk53l79eplzJgxwzAMw5g1a5Yxbdo0wzAMw+FwGG3atDE++eQT5/NGjBhhGIZhnDhxwqhQoYJx+PBhY8+ePUZwcLBx7NgxwzAMY+bMmUbfvn2N3377zQgLCzNOnz5tGIZh7N+/36hataqRmpr6j7+fAG4verYA3JTrXUY8ePCgDh8+rPfee8+5Ly0tTbt371ZUVJTuvfdeLViwQH/++ac2b96su+666x+f98knn5S7+5U/VevWrdPOnTvVpEkTSZLD4dClS5dyreHh4aFatWpJkh566CE99dRTKlKkiCTJ19dXKSkp8vX11b333qvw8HBJUo0aNWS327Vv3z6tX79e0dHR8vDwkCS1bNlSb731lrN+UFDQdc/76quvatu2bZo5c6Z+//13HThwQE888YTzeO3atSVJfn5+KlGihFJSUrR161ZVq1ZNJUuWlCS99tprkqR58+YpKSnJuS1JNptNhw8fVrly5XL9PwCQfwhbAG5ZVlaW7rnnHi1fvty57+TJk7r77rv17bffaujQoWrdurVq166txx57TCtWrLimhs1mk3HVEq0ZGRk5jhcuXNj5tcPhUJs2bRQVFSVJSk9PV0pKSq7t9PDwkM1my7F9PXa7Pce2w+GQ3W6Xw+G4Zv/V7by6jVcbPXq0duzYoSZNmig4OFiZmZk5XquXl5fz6+z/B7vdnqOtaWlpOnr0qBwOh5555hlNmDDBeez48ePy9fX9u5cOoABg6gcAt+zRRx+Vl5eXM2wdP35cYWFh2rVrlzZu3KiaNWsqKipKFSpU0Jo1a5SVlSXpSqjJzMyUJBUvXlzHjh3TqVOnZBiG1qxZc8PzVatWTUuWLNGFCxckSR988IFznJMZTp8+rfXr10uS1q5dKw8PD/n7++vZZ5/VggULlJGRIYfDoXnz5qlq1arXrXH1a9uwYYNeffVVNWrUSCVKlNCmTZuc/wc3EhwcrISEBCUlJUmSFixYoNGjRyskJEQbN27Ur7/+Kkn67rvvFBERocuXL5v18gFYhJ4tALfM09NTU6ZM0dChQzVjxgxlZmaqa9euqlSpkooWLaoePXooPDxcdrtdQUFBzoHtTz31lCZMmKC33npLkydPVvPmzdWkSRP5+Pjoueeeu+H5XnzxRSUmJqpZs2ay2WwqWbKkRowYYdrryQ6OY8aMUaFChTR58mTZ7XZ17NhRI0eOVKNGjZSZmanAwED17dv3ujWqV6+uwYMHS5LeeustjRo1SlOmTJHdblfFihV1+PDhv21D2bJl1bNnT7Vp00aS5OPjo2HDhsnPz0+DBg1St27dZBiG3N3dNXXq1Bv2qgEoOGzG1X3aAHCHOnLkiMLDw/Xjjz/md1MA/MtwGREAAMBC9GwBAABYiJ4tAAAACxXIAfIOh0OpqanX3K4NAABQ0BiGoYyMDN1111051pDNViDDVmpqqvbv35/fzQAAALhp/v7+uvvuu6/ZXyDDVvaEg/7+/vL09Mzn1gAAANxYenq69u/ff8MJkwtk2Mq+dOjp6ZljhmUAAICC6kZDnxggDwAAYCHCFgAAgIUIWwAAABYibAEAAFiIsAUAAGAhwhYAAICFCFsAAAAWImwBAABYiLAFAABgIcIWAACAhQhbAADgphmZjgJZqyArkGsjAgCAgsnm7qYTYw+YUuu+7mVMqVPQ3VTPVlxcnBo0aKC6detq3rx51xz/+uuvFR4erhdeeEG9e/dWenq6JOnYsWNq0aKF6tWrp44dOyo1NdXc1gMAABRwuYatxMREjR8/XvPnz9fy5cu1cOFCHTx40Hn84sWLGjRokGbOnKkvvvhCly9fVkxMjCRp4MCBioqKUnx8vAICAjRlyhTrXgkAAEABlGvY2rRpk0JCQlS0aFEVLlxYoaGhio+Pdx4vXLiw1q5dq//85z+6ePGiTp06pXvuuUcZGRnaunWrQkNDJUmNGzfO8TwAAIA7Qa5jtpKSkuTj4+Pc9vX11Y4dO3I8xsPDQ999953eeecd+fr6qlq1ajpz5oyKFCkid/crp/Dx8VFiYuI/atyuXbv+0eMBAIC1KlWqZGq97du3m1qvIMo1bBmGcc0+m812zb4aNWpo8+bNGjdunAYMGKB33nnnpp73dwICAuTl5fWPngMAAFyH2eEtP1y+fPlvO4hyvYzo5+enkydPOreTkpLk6+vr3D579qw2bNjg3A4PD9e+fftUvHhxXbhwQVlZWZKk5OTkHM8DAAC4E+QatqpUqaKEhASdPn1aly5d0urVq1W9enXnccMw1LNnTx07dkyS9NVXX6lixYry8PBQUFCQvvzyS0lSbGxsjucBAADcCW6qZys6OlqtWrVSo0aNFBYWpsDAQLVt21Y7d+5UsWLFNHjwYLVv314RERH6/fff1bNnT0lS//79tWjRIjVo0EDbtm3T22+/bfkLAgAAKEhsxvUGZeWz7GufjNkCAKDgYVLTnHLLLSzXAwAAYCHCFgAAgIUIWwAAABYibAEAAFiIsAUAAGAhwhYAAICFCFsAAAAWImwBAABYiLAFAABgIcIWAACAhQhbAAAAFiJsAQAAWIiwBQAAYCHCFgAAgIUIWwAAABYibAEAAFiIsAUAAGAhwhYAAICFCFsAAAAWImwBAABYiLAFAABgIcIWAACAhQhbAAAAFiJsAQAAWIiwBQAAYCHCFgAAgIUIWwAAABYibAEAAFiIsAUAAGAhwhYAAICFCFsAAAAWImwBAABYiLAFAABgoZsKW3FxcWrQoIHq1q2refPmXXN8zZo1atiwoSIiIvTmm28qJSVFkhQbG6tq1aqpYcOGatiwocaPH29u6wEAAAo499wekJiYqPHjx2vZsmXy9PRU8+bNFRwcrMcff1ySdOHCBQ0YMEBLly6Vn5+fPvjgA02cOFF9+vTRzp071bt3b4WFhVn+QgAAAAqiXHu2Nm3apJCQEBUtWlSFCxdWaGio4uPjncczMjI0YMAA+fn5SZLKli2r48ePS5J27typ2NhYRUREqEePHs4eLwAAgDtFrmErKSlJPj4+zm1fX18lJiY6t4sVK6Y6depIktLS0jR9+nTnto+Pjzp37qzly5erZMmSGjRokNntBwAAKNByvYxoGMY1+2w22zX7zp8/rzfffFPlypVTZGSkJGny5MnO423atHGGsJu1a9euf/R4AABgrUqVKplab/v27abWK4hyDVt+fn7atm2bczspKUm+vr45HpOUlKQ33nhDISEheu+99yRdCV9Lly7Va6+9JulKaHN3z/V0OQQEBMjLy+sfPQcAALgOs8Nbfrh8+fLfdhDlehmxSpUqSkhI0OnTp3Xp0iWtXr1a1atXdx7PyspShw4dVL9+fb3//vvOXq/ChQtrxowZ+vnnnyVJc+fOVd26dfP6egAAAFzKTfVsRUdHq1WrVsrIyFDTpk0VGBiotm3bqkuXLjpx4oR2796trKwsrVq1StKVHqmhQ4dqwoQJGjBggNLS0vTII49o1KhRlr8gAACAgsRmXG9QVj7L7o7jMiIAAAXPibEHTKlzX/cyptTJb7nlFmaQBwAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQjcVtuLi4tSgQQPVrVtX8+bNu+b4mjVr1LBhQ0VEROjNN99USkqKJOnYsWNq0aKF6tWrp44dOyo1NdXc1gMAABRwuYatxMREjR8/XvPnz9fy5cu1cOFCHTx40Hn8woULGjBggKZPn64VK1aobNmymjhxoiRp4MCBioqKUnx8vAICAjRlyhTrXgkAAEABlGvY2rRpk0JCQlS0aFEVLlxYoaGhio+Pdx7PyMjQgAED5OfnJ0kqW7asjh8/royMDG3dulWhoaGSpMaNG+d4HgAAwJ3APbcHJCUlycfHx7nt6+urHTt2OLeLFSumOnXqSJLS0tI0ffp0tWzZUmfOnFGRIkXk7n7lFD4+PkpMTPxHjdu1a9c/ejwAALBWpUqVTK23fft2U+sVRLmGLcMwrtlns9mu2Xf+/Hm9+eabKleunCIjI68brK73vL8TEBAgLy+vf/QcAADgOswOb/nh8uXLf9tBlOtlRD8/P508edK5nZSUJF9f3xyPSUpKUlRUlMqVK6ehQ4dKkooXL64LFy4oKytLkpScnHzN8wAAAP7tcg1bVapUUUJCgk6fPq1Lly5p9erVql69uvN4VlaWOnTooPr16+v999939l55eHgoKChIX375pSQpNjY2x/MAAADuBLleRvTz81N0dLRatWqljIwMNW3aVIGBgWrbtq26dOmiEydOaPfu3crKytKqVaskXbn8N3ToUPXv31+9e/fW1KlTVbJkSY0bN87yFwQAAFCQ2IzrDcrKZ9nXPhmzBQBAwXNi7AFT6tzXvYwpdfJbbrmFGeQBAECBYWQ6CnS9W5HrZUQAAIDbxebupsQPfjCtnl/XENNq3Sp6tgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAMAdw8jMuu313E09IwAAQAFmc7cradJXptXz7VRfysr828fQswUAAGAhwhYAAICFCFsAAAAWImwBAABYiLAFAABgIcIWAACAhQhbAAAAFiJsAQAAWIiwBQAAYCHCFgAAgIUIWwAAABYibAEAAFiIsAUAAGChmwpbcXFxatCggerWrat58+bd8HG9evXSsmXLnNuxsbGqVq2aGjZsqIYNG2r8+PF5bzEAAIALcc/tAYmJiRo/fryWLVsmT09PNW/eXMHBwXr88cdzPKZ///5KSEhQcHCwc//OnTvVu3dvhYWFWdN6AACAAi7Xnq1NmzYpJCRERYsWVeHChRUaGqr4+Pgcj4mLi1Pt2rVVv379HPt37typ2NhYRUREqEePHkpJSTG39QAAAAVcrj1bSUlJ8vHxcW77+vpqx44dOR7Tpk0bSdL27dtz7Pfx8VG7du0UGBiocePGadCgQRo7duxNN27Xrl03/VgAAGC9SpUqmVrvr9nB7Pp/PYcV9XPLK7mGLcMwrtlns9lu6uSTJ092ft2mTRvVqVPnpp6XLSAgQF5eXv/oOQAAwHVYEX5u9zkCAgL+NnDlehnRz89PJ0+edG4nJSXJ19c31xOfP39es2bNcm4bhiF391yzHQAAwL9KrmGrSpUqSkhI0OnTp3Xp0iWtXr1a1atXz7Vw4cKFNWPGDP3888+SpLlz56pu3bp5bzEAAIALybWryc/PT9HR0WrVqpUyMjLUtGlTBQYGqm3bturSpYsqVKhw3efZ7XZNmDBBAwYMUFpamh555BGNGjXK9BcAAABQkN3Udb3w8HCFh4fn2Pfxxx9f87gRI0bk2A4KClJMTEwemgcAAODamEEeAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBCNxW24uLi1KBBA9WtW1fz5s274eN69eqlZcuWObePHTumFi1aqF69eurYsaNSU1Pz3mIAAAAXkmvYSkxM1Pjx4zV//nwtX75cCxcu1MGDB695TIcOHRQfH59j/8CBAxUVFaX4+HgFBARoypQp5rYeAACggMs1bG3atEkhISEqWrSoChcurNDQ0GtCVVxcnGrXrq369es792VkZGjr1q0KDQ2VJDVu3Pia5wEAAPzbuef2gKSkJPn4+Di3fX19tWPHjhyPadOmjSRp+/btzn1nzpxRkSJF5O5+5RQ+Pj5KTEw0pdEAAACuItewZRjGNftsNluuhW/1eVfbtWvXP3o8AACwVqVKlUytd3VHjRX1/3oOK+rnlldyDVt+fn7atm2bczspKUm+vr65nrh48eK6cOGCsrKyZLfblZycfFPPu1pAQIC8vLz+0XMAAIDrsCL83O5zBAQE/G3gynXMVpUqVZSQkKDTp0/r0qVLWr16tapXr57riT08PBQUFKQvv/xSkhQbG3tTzwMAAPg3yTVs+fn5KTo6Wq1atVKjRo0UFhamwMBAtW3bVjt37vzb5/bv31+LFi1SgwYNtG3bNr399tumNRwAAMAV5HoZUZLCw8MVHh6eY9/HH398zeNGjBiRY7tUqVKaM2dOHpoHAADg2phBHgAAwEIuEbaMzKwCWQsAACA3N3UZMb/Z3O1KnjrXlFo+HV8xpQ4AAMDNcImeLQAAAFdF2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQoQtAAAACxG2AAAALETYAgAAsBBhCwAAwEKELQAAAAsRtgAAACxE2AIAALAQYQsAAMBChC0AAAALEbYAAAAsRNgCAACwEGELAADAQjcVtuLi4tSgQQPVrVtX8+bNu+b4nj171KRJE4WGhur9999XZmamJCk2NlbVqlVTw4YN1bBhQ40fP97c1gMAABRw7rk9IDExUePHj9eyZcvk6emp5s2bKzg4WI8//rjzMT179tSQIUP05JNP6r333tOiRYsUFRWlnTt3qnfv3goLC7P0RQAAABRUufZsbdq0SSEhISpatKgKFy6s0NBQxcfHO48fPXpUaWlpevLJJyVJjRs3dh7fuXOnYmNjFRERoR49eiglJcWilwEAAFAw5dqzlZSUJB8fH+e2r6+vduzYccPjPj4+SkxMdH7drl07BQYGaty4cRo0aJDGjh17043btWuXJKlSpUo3/ZybsX37dlPrAQBwp7D6Pdns+n89hxX1s/PKjeQatgzDuGafzWa7qeOTJ0927mvTpo3q1KmT2+lyCAgIkJeX1z96zs2w4j8aAAD8c7fjPdnqcwQEBPxt4Mr1MqKfn59Onjzp3E5KSpKvr+8NjycnJ8vX11fnz5/XrFmznPsNw5C7e67ZDgAA4F8l17BVpUoVJSQk6PTp07p06ZJWr16t6tWrO4+XKlVKXl5ezi662NhYVa9eXYULF9aMGTP0888/S5Lmzp2runXrWvQyAAAACqZcu5r8/PwUHR2tVq1aKSMjQ02bNlVgYKDatm2rLl26qEKFChozZoz69Omj1NRUlS9fXq1atZLdbteECRM0YMAApaWl6ZFHHtGoUaNux2sCAAAoMG7qul54eLjCw8Nz7Pv444+dX5crV05Lliy55nlBQUGKiYnJYxMBAABcFzPIAwAAWIiwBQAAYCHCFgAAgIUIWwAAABYibAEAAFiIsAUAAGAhwhYAAICFCFsAAAAWImwBAABYiLAFAABgIcIWAACAhQhbAAAAFiJsAQAAWIiwBQAAYCHCFgAAgIUIWwAAABYibAEAAFiIsAUAAGAhwhYAAICFCFsAAAAWImwBAABYiLAFAABgIcLWv4AjM71A1ysIMrPMfU1m1ysI0rMyCmQtAHB17vndAOSdm7unNk8LM61ecPuVptUqKNztnpo0N9S0ep1eWZVjOyMrXR52T9PqX69eela6PE06x/Vqedo91CCmvyn1v4wceJ1zZsrTbt6fnL/Ws7r+lX1Z8rTbTap/bS2r698O6VkOedrN+RxvZi0gPxG2JBmZmbK5m/NfYWatO0lWZrrs7uYECTNr3SwPu6f6LqpnWr3BzeKv2edp91T9FRGm1P8qYoUpdf4JT7u7Xlj2oWn1vmjc5Zr6YUtnmVZ/ZZPXrtnnabcrbMlic+o3ffG69RsuufZ7fyuWNzXv5/Gf8LS76aVlh0yptbDxY6bUAfIbqUCSzd1diVNHmVLLr+M7ptS509jdPTVvljk9Ty1eW5X7gwAAuE3onwWAO0R6lqNA1wP+rejZug0cmelyM/Gyltn1ANwZPO1uarw0wbR6y5o8Y1qtgiIry5DdbitwteDaCFu3gZu7p37/sJFp9R7pEmtaLQDA/7HbbVq++KQptRq++B9T6sD1cRkRAADAQoQtAAAACxG2AAC4TRxZRoGuB2swZgsA4DIyswy5mzTo3MxaN8vNblPCZ8mm1XvmVR/TasE6NxW24uLiNHXqVGVkZOi1115TixYtchzfs2eP+vTpowsXLigoKEgDBw6Uu7u7jh07pp49e+rUqVN69NFHNWbMGN11112WvBAAwL+fu92mD2JOmFKra+R9ptQBcpPrZcTExESNHz9e8+fP1/Lly7Vw4UIdPHgwx2N69uypvn37atWqVTIMQ4sWLZIkDRw4UFFRUYqPj1dAQICmTJlizasAAAAooHLt2dq0aZNCQkJUtGhRSVJoaKji4+PVqVMnSdLRo0eVlpamJ598UpLUuHFjffjhh3rxxRe1detWTZ482bn/lVdeUc+ePXNtlGFcuQadnv5/i/1menr8w5d2fZcvX77u/kzPQpbWz/K615T6NzyHZ1FL62dlZchuN+d7cKNaHh7mvIYbfQ+8TKp/o3MUdre2viQVtZvzc3Tj+t4W1/cypf6NzlHU3Zyf0RvVv3IOc0Zf3Kj+ve7mDKW9cX1Tyv/tOe6xZ1la38ueaWl9u7u19eVhTv0bncORZcjNpMujN6qVWcicCW1v+J5s3p+K654j08u8IeuXL1925pXs/PJXNuNGR/5/06ZN08WLFxUdHS1JWrx4sXbs2KHBgwdLkn788UeNGjVKn3/+uSTpjz/+ULt27TRnzhw1bdpU69evlyRlZmbqySef1K5du3Jt+Pnz57V///6bfJkAAAD5z9/fX3ffffc1+3P9jHO9LGaz2XI9ntvz/s5dd90lf39/eXh43PRzAAAA8oNhGMrIyLjhuPRcw5afn5+2bdvm3E5KSpKvr2+O4ydP/t9su8nJyfL19VXx4sV14cIFZWVlyW63O/ffDDc3t+smQwAAgIKoUKEbD0fK9aJllSpVlJCQoNOnT+vSpUtavXq1qlev7jxeqlQpeXl5afv27ZKk2NhYVa9eXR4eHgoKCtKXX36ZYz8AAMCdJNcxW9KVqR+mTZumjIwMNW3aVG3btlXbtm3VpUsXVahQQXv37lWfPn2Umpqq8uXLa/jw4fL09NTRo0fVu3dvnTp1SiVLltS4ceN0773mDRQHAAAo6G4qbAEAAODWsFwPAACAhQhbAAAAFiJsAQAAWIiwBQAAYCHCFgDgtvvtt9/yuwnAbUPYAvCvdPHixfxuAv4iMzNTX375pVq2bKnGjRubUjMxMdGUOoCVCFu5uHjxovbu3SvDMPjjXQDwPbi9ypUrp//+97/Of+XLl1dISIi6deums2fP5rn+sWPH/vZfXjRs2DDH6hdWSUlJ0enTp53bW7ZsybGdV7/++quSkpIkSdOnT1eHDh00ceJEpaWlmXYOSdq/f7/i4+P17VsZwW0AACAASURBVLff6s8//zS19p9//qkxY8aoevXqeueddxQcHKxvvvnGlNodOnRwfv3pp5+aUvOv5s+f7/z6wIEDOY4NHTrUknOabf78+fr6668lSU2bNlXt2rX1/PPP648//sjnlhU8Vrzvu/Q8W+fOndPu3bsVEhKiGTNmaPfu3XrrrbdUunRpU+onJCSoX79+ysrK0oIFCxQREaExY8aoWrVqeapbrly5HGs+uru7y83NTenp6SpSpIi2bt2a16arVq1af7uupFl/6DIzM7Vhw4Zr3ngbNWpkSv2/ioyMVExMjCm1zp8/rxkzZujee+/VCy+8oK5du2r//v2qVKmShgwZIj8/vzzVP378uAYPHqwTJ06oTp06at++vex2uySpffv2mjZtmhkv47Y7efKkFi1apIMHD2rcuHF5qpX9c3r1nyGbzaakpCRlZmZqz549t1x7w4YNGjhwoOrUqaPo6Gh5enrmqa3Xs3v3brVr107Dhg1zrpAxfvx4LVu2TB9//LHKlSuXp/offfSRFixYILvdrsqVK+vIkSOqW7eutmzZIm9vb40ePTrPr+HUqVPq0qWLDhw4oIcfflg2m02//fabnnzySY0dO1b33HPPLdf++uuvtWDBAv3yyy+qW7eu6tWrp759+2rt2rV5bne2Ro0aKTY2VpK5fx+udnXdv57DzHOGh4erRo0aeu6551SpUiXT1gaeNm2aEhIS1L9/fz366KMKDw/XRx99pHXr1mn37t0aNmxYnurn9p719NNP56n+1dLT0/XJJ5/ot99+U79+/TRr1iy1a9fOtN9vq973c10bsSDr1q2bnn32WRmGobi4OLVq1Ur9+/fX3LlzTak/btw4zZ8/X23btpWvr6/mzp2rbt265fk/fe/evZKk/v37q2LFioqIiJDNZtOqVav0/fffm9F0zZkzR4ZhaPLkyXrwwQfVuHFj2e12xcXF6ciRI6acQ5K6d++uY8eOqXTp0jn+MJgRtipUqKDMzExJVxb5zH5Tzg6reXkjlqRevXqpdOnS2rt3r2bPnq2OHTsqIiJCX375pfr376+PPvooT/Xfe+89hYWFqWzZspo0aZI6dOigqVOnyt3d3ZRLH5MmTfrb4506dcrzOa7nP//5j95880298MILea711zfd1NRUjRw5Uhs2bNDgwYPzVLtatWpasWKFPvjgAzVt2lT9+vXT/fff7zx+9de3auTIkRo7dqyCg4Od+6KjoxUUFKQRI0Zo1qxZeaofFxenr776ShcvXlSdOnW0adMmeXt7q0WLFmrQoEEeW3/F4MGDValSJc2aNUseHh6SrryhTZw4UcOGDdOIESNuuXbnzp1Vr149LVy4UA8//LAkmRYgsl1dz6q+g6vrWtk/8emnn+r777/X3Llz9d577ykwMFC1atXK8/c6NjZWS5YscS6SbLfbVapUKUVFRSk8PDzP7f7www8lSWfPntXhw4dVsWJFubm56ccff5S/v78WLFiQ53NkGzRokIoXL67du3fLbrfr8OHDev/990354CFZ977v0mHr7NmzevXVVzVkyBA1btxYTZo00bx580yr73A45OPj49x+/PHHTastSTt27NDAgQOd26GhoZoyZYoptUuVKiVJ2rdvn4YPH+7c//rrr5s2ViK7fnx8vGn1rrZw4UINHjxYrVu31vPPPy8p56fYvDpy5IimTJmijIwMPffcc3rppZckSU2aNDElsJ89e1ZNmjSRJE2dOlXdu3dXz549NX78+DzXlqSsrCzNmjVLrVu3lpvb7R8RkP3GbJaEhAT16dNHVatW1YoVK1SkSJE81/T29lbXrl114sQJdezYUffcc48zuJvRu3vu3LkcQSvbs88+qzFjxuS5vru7u7y9veXt7a0HH3xQ3t7ekq68WWZ/nVf79u3ThAkTcuzz9PRUt27d1LBhwzzVXrFihWJiYhQVFaVSpUrphRdeUFZWVp5q/h2zg9ztPoePj48iIyNVpkwZJSQkaO7cudq0aVOew5bdbncGLUnq2LGjJMnNzc2UHqE5c+ZIktq2batJkyY5g/XRo0fVr1+/PNe/2i+//KKYmBitX79e3t7eGjlypCmBMZtV7/suHbYcDof27t2rNWvWaPbs2dq/f7+zJ8QM9913n9atWyebzaZz585p3rx5pnwazubt7a2lS5eqfv36cjgcWr58uYoWLWpa/Ww//PCDQkJCJEnfffed81KWGUqXLq2kpCT5+vqaVjNb+fLlNXPmTI0YMULr1q1Tnz59TP1D5+7urkOHDumxxx7TzJkznft3795tynnsdrsOHDigMmXKyGazaeTIkWrTpo2zizqvunbtqsTERHl7e6tt27Z5rvdPrF692rSf1YsXL2rEiBHO3qyqVauaUleSvv32Ww0aNEjVqlXTunXrTAlwV8vMzJTD4bgm7DocDmVkZOS5/tV1//p7a9bvgpeX13X322y2PId4f39/9erVSz169NC6desUExOjkydPql27dmrRooVq1KiRp/rSlTFUtWvXlnRlsHz212aG6tsR4qQrYeXQoUMqV66cKleurOnTp+f5UrR05efxwoULzp//0NBQSVeGUpjp2LFjzqAlXek9zuvYy7+y2WxKT093fk/OnDlj6vfHqvd9lw5b3bt31+DBg9WqVSs99NBDaty4sXr37m1a/UGDBmno0KE6fvy46tSpo5CQEA0aNMi0+qNHj9bgwYM1ZMgQubm5qUqVKho1apRp9SVpyJAh6tWrl3OAbalSpUw9R1pamurVqyd/f/8cn5Bmz55tSv1ChQppwIABWrt2rV577TWlpqaaUle6cpmvY8eO+uqrr+Tv7y9JWrNmjQYNGnTNJ/1b8e6776pDhw6Kjo5WWFiYPDw8NHXqVPXo0UMHDx7Mc33pymswa/zd9Vxv7N+FCxf08MMPm9Jtf3VvVlxcXI5P33nVpUsX7d69W0OHDtUzzzxjWt2rPf3005o0aZK6dOmSY/+UKVMUEBCQ5/q///67WrVqdc3XhmGYNrD5796ozHoTs9vtqlOnjurUqaPTp09r+fLlGjt2rClha9WqVSa08O/9XaBLTk427Tzly5fXxYsXdfbsWZ06dUonT55UWlqaChUqlKe64eHh6tWrl0aOHOkMXKmpqXrvvfcUERFhRtMlSf/73//Uq1cvZwfCypUrFRQUZFp9SWrVqpVat26t5ORkDR06VGvWrNFbb71lWn2r3vddeoD88uXLr+nm/vzzz/Xyyy/nU4tuzdmzZy3p0bpadvo3+zxbtmy57v7KlSvnuXZiYqIGDx6sP/74Q0899ZQ6deqk7777Ti+++GKea1+vfo8ePVSoUCHnDQtm1T906JCCgoLUo0cP52DjPXv26L///W+e6w8ZMkS///67s/15Gcx8PUePHs2x7ebmpnvuuce0UFSuXDm5u7vL19f3mrE3ee2VGDx4sLp3767ChQub0dTrunDhgtq1a6fk5GRVqFBBhmFo9+7dKl68uKZOnZrn37cb/X5lM+P3LCAg4Lo3g2QHiZ07d+b5HK7ur78Hf5U9bMMsqampWr16taZOnapjx45p165deaqXlZWlAQMGaOXKlc7xtQcPHlTDhg01YMAAcxqtK2P95s6d6/y5rVKliqKiouTubm6/zsGDB7V582ZlZWWpcuXKpvT+Wc0lw9acOXOUmpqqefPmqUWLFs79mZmZio2N1Zo1a/JU/3bdybdnzx5FR0crLS1NCxcu1CuvvKIJEybof//7nyn1pSt/JPr06aOjR49q3rx56t69u4YNG6YHHnjAlPrt2rVTzZo1VbNmTd13332m1Mz2xhtv6H//+5+CgoL01VdfSVKO8Wdm1A8ICFClSpVctr6V/z+3w+1+E7OCYRj64YcftGfPHrm5uSkgIMD0T/NW+jd8D/4tvv/+eyUkJCghIUEOh0PVq1dXjRo1TPt5SkxM1I4dOyRd6YUyc1hMtgsXLuj8+fM5biQw8zzp6enasGGDzp07l2O/WXfAP//88zmGedhsNhUqVEiPPfaYevXqdcu/Dy4Ztr755hvt3LlTS5YsUdOmTZ373d3d9fTTT193wOo/cbv++LRo0UKDBg1S9+7dFRsbq40bN2r8+PFasmSJKfWlK2/IrVu31pgxYxQTE6PFixdr+fLlpt1I8PPPP2v9+vVav369MjMzVb16ddWqVUtPPPFEnmuHhYVp5cqVkqSMjAw1atRIX3zxRZ7rUh9AQdShQwfVrFlTNWrUMP3D6+3w0Ucfafr06SpatKjz7nGzxs1le/XVV2UYxjXvw2Z90BwyZIgeeOABZ7ZYsWKFdu7cqVq1amnevHm3fIexS47Zql27tmrXrq0GDRo4x9qYKfubeL273goVKqTU1FRTznvp0qUcc4JVrVpVI0eOzHPdq505c0bVqlXTmDFjZLPZ1KxZM1Pv2HziiSf0xBNPqEWLFoqPj9dHH32kTz75JM/d3lLOu908PDxMv/uN+gAKko8++kgrVqzQ559/rg4dOmjVqlWWzVlohSVLlmjNmjUqXry4Zec4c+aMVqxYYVn97du3q0+fPs7tqKgoNW7cWMOHD8/TbAEuGbaynTx5Un379r1mQk2zBkx+88032r17t+rUqSPpyp1Nvr6+unjxosLDw/Xaa6/lqX7RokW1d+9e5yXLFStW6N57781rs3MoVKiQTpw44TzHtm3bTJ3cceDAgdq+fbvsdruefvpp9e/f35RxJNdj9R1B1AeQn8aMGaMTJ07ol19+UZs2bbR06VLt3bvX1Bu/rFSyZEnT38P+KiQkRJs2bVJISIglU964ubnp+++/17PPPivpyqVdT09PnTx5Mk+zHbjkZcRsoaGh6tmzp/PW+mwPPfSQKfWbN2+u6dOnOwcdX7hwQR06dNCsWbPUuHHjPKfrw4cPq1evXtq5c6cKFSrkvMPrscceM6P5kq7M5dW3b18dPnxYDz30kFJSUjRhwgQ9+eSTptTv3r279u/frzJlyig4OFiVK1fWo48+akrtvw7cTUxMlJ+fn2ld09QHUJA0atRIMTExioyMVGxsrDIzM50TLbuCvn37av/+/QoODs7xod7MCZZnzZqlESNGON/zs//e5XWS62wHDhxQr169dOzYMRmGoYcffljDhw9XfHy87r//fkVGRt5SXZfu2SpWrJiz18kKZ86cyXHXlZeXl1JSUuTu7m5KL8JDDz2kiRMnqnDhwnI4HDp16lSOOUrMEBgYqCVLluj3339XVlaWHnvsMVN7tsaOHSvpyvptCQkJ6tChgy5evGjKTPhW39JNfQAFSXZPTfb7S3p6er5MWHyr/Pz88rzMWW5mz56ttWvXWjK4X5LKlCmjZcuWKSUlRXa7XUWKFNGGDRvyPL2ES4etoKAgjRo1Ss8++2yOifkqVqxoSv3nn39er776qnPOkNWrV6t27dqKjY3NMcPsrZo9e7ZiYmIUExOjo0ePqkOHDnrttdecM5nnxcSJE9W5c2e9++671z1u1mDCQ4cOOe+e2bNnj5544glT5s6RrL8LivoACpJ69erp7bffVkpKimbNmqUVK1YoLCwsv5t10zp16qSLFy/q8OHD8vf3V1pamulTr/j6+lo+VZJ0ZbqMhQsXavHixbp8+bLWr1+fp3oufRkxKirqmn02m83UAeDr1q3Txo0bZbfbVaVKFdWoUUM//fSTHn300Txfmw4LC9OiRYucP4yXLl1Ss2bNFBcXl+d2r127VrVq1brhAqm32hX6V+Hh4c67Z5566imX+hQGAAXB1bOsr1+/Xps2bZJhGAoJCVHNmjUt68Uxm1WLOF+ta9eu2rlzpypWrJjjpiCzOhA2b96szz//XGvWrJGbm5sGDhzonJQ6L1w6bN0OBw4cUEpKSo45Q8xawTw0NFRffPGFc8K3zMxMRUZGmhK2sr3++uv69NNPTat3Pd99951++OEHZWZmKiQkxDm7MgAgd9lzO/717Tg5OVmZmZmmjUey2osvvqgpU6aobdu2io2N1cGDB9WtWzdT7x60qgNh1qxZWrhwoTw8PFS/fn3Vr19fr7/+utauXZunutlc8jLigAEDNGDAALVu3fq6Y6fMChcDBw7UunXr9OCDDzr32Ww205aiqVOnjvMypSTnZUozpaWl6fjx4ypZsqSpdbPNmDFDq1atUnh4uAzD0NSpU3XgwAF16NDBkvMBwL/NX9/QU1NTNXLkSOd6oa7CqkWcrxYZGan9+/dry5YtyszMVHBwcJ5X45CkcePGqVatWmrRooWCgoJks9nMXYvXtEq3UePGjSVdmb3cShs3blR8fHye16W6kW7duunrr7/W1q1b5e7urlatWpk+4P/06dOqVauWSpQoIS8vL9PvVFu+fLkWL17s/D9q1qyZGjduTNgCgFtw9XqhK1asMH3xdCtZtYjz1WJjYzVp0iTVqVNHDodDnTp1UseOHXNMcH4r1q9fr5UrV2r48OFKTk5W/fr1lZ6eblKr/wWXETdu3KiEhARlZWUpODhYzz33nGm133jjDU2aNEne3t6m1bxaZGTkDbtEzbJjxw6VKFHimv1mDa6+ehZz6conm4YNG5p6KRQA/u0uXryoESNGOHuzqlatmt9N+sdOnTqloUOHOsecBQcHq0+fPvL19TXtHA0bNtSsWbNUrFgxSVc6FFq1apXjfSiv9u7dq2XLlikuLk4lSpTQyy+/nGNpwFvhkj1b2T799FOtXLnSeQnrww8/1P79+03r8br33nv1wgsv6KmnnsoxXYJZA/FKlCihbdu2KTAw0NTpGK7Wq1cv57p5VggJCVHnzp2d18tjY2PzvFwSANxJru7NiouLM22h99utRIkSGjVqlPbu3St3d3eVLVvW9MmWHQ6HM2hJUvHixU09R2Zmpk6cOKHy5curTJkySk1N1bhx4/Ictly6Zys8PFyLFi1y9jxdvHhRTZs2NW0COKvv5AsJCblm9nszJ2eTpOjoaNWoUUOBgYE5Loea1bVrGIbmz5+vzZs3O++eeemll0xf5R0A/q3KlSsnd3d3+fr65ggOrjZB8caNG9WrVy/5+vrK4XDo3LlzmjBhggIDA007R48ePVSsWDHnZcPFixcrJSVFo0ePNqV+165ddezYMZUuXTrH9yKvnSwuHbasuoSVnJwsHx+fHLfjXs1VbsOVrtzl8ldm//Javco7APybHT169G+Pu8qcemFhYRozZozKlSsnSdq5c6f69++vZcuWmXaOtLQ0TZw4UT/88IPzA/6bb75p2ti2evXq6auvvjK9R86lux8qV66st99+29nTFBMTY8q0DH369NG0adP0yiuv5Fi5PJtZQeXSpUuaNGmSc8xZSEiIunbtauokcGbdtnojt2OVdwD4N3OVMJUbT09PZ9CSpAoVKph+Dg8PD1WsWFE9e/bU6dOntXbtWlMvu5YuXVrJycmmjjOTXLxny+FwaN68ec6EGxwcrKioqDxPPpbt559/1vbt2/XKK6+oQ4cO+uWXXzRw4EDVq1fPlPrvvvuuvL291axZM0nSokWLdP78edO6Q6UrM7zPnz9fFy9elGEYcjgcOnLkiGkTv9apU0eLFi2ydJV3AEDBN2DAAGVkZKhZs2ay2+364osvdOTIEbVq1UqSOXNUvvvuu3I4HBo5cqROnz6t4cOHy9vbW4MGDcpzbenKjXE//vij/P39c4ylzuuUTy7ds+Xm5qbKlSvL4XDI3d1dVatWNS1oSdLQoUPVs2dPrV69WoUKFVJsbKw6depkWtj65Zdfckz21q9fPzVo0MCU2tmio6NVu3Ztbd++XZGRkVq/fr3KlCljWv3bsco7AKDg+/XXXyVJY8aMybH/ww8/NG2Oyl27djmHChUvXlyjR49WeHh4nutma9++vWm1rubSYWvWrFmaM2eOatWqJYfDoZkzZ6pTp05q1KiRKfUdDoeefvppde/eXc8//7xKliyprKwsU2pLVwY/njt3Tvfcc48k6dy5c7Lb7abVl668hi5duigzM1Ply5dX8+bN1bx5c9PqP/LII4qKirJ0lXcAQME3Z84cy8/hcDiUlJTkvMx36tQpU5eJq1y5smm1rubSYevzzz9XbGys7r77bklS586dFRUVZVrY8vb21qeffqrNmzerX79++uyzz0y9Nvzaa6+padOmzkHsa9euNX2iVm9vb6Wnp+uRRx7RL7/8oqCgIF2+fNm0+rdjlXcAQMHlcDg0f/58Va5cWf7+/po9e7YWL16s8uXLq2/fvqZOzNqhQwdFRkaqUqVKMgxDO3bs0Pvvv29afau49JitZs2aac6cOfLy8pIkZWRk6OWXX9aSJUtMqZ+YmKjFixerSpUqqlixokaPHq2WLVvqvvvuM6W+JO3fv19bt26Vw+FQ5cqVVbZsWdNqS9LcuXO1du1ajRkzRi+99JIefvhhGYahTz75xJT6f/75Z47ljAzD0Ny5c9WyZUtT6gMACrbRo0fr0KFDev/995WYmKh27dpp4sSJOnDggPbs2aMRI0aYer7ExET99NNPcnd3V4UKFUwfzG4Flw5bPXr00KFDhxQWFia73a6vv/5anp6ezm7Agr5kTEREhBo2bKiwsDDLeodOnz6tlStX6ty5czp+/Lh27typatWq6Z133jGlfr169TRt2jQ9/PDD2rdvn/r06aPChQvrs88+M6U+AKBgCw8PV0xMjNzd3TV06FClpqZq2LBhkqT69eubOrH2uXPnFBcXp7Nnz+aYbqigD11x6cuIDzzwgB544AGdO3dOkhQUFCTpyjwcrmDs2LFauXKlWrVqpZIlSyoiIkKhoaGmXqps27atypYtq/vvv18lS5Y0fUHqYcOGqWPHjqpSpYpWr16tbt26mXYZFwBQ8Lm5uTknst6yZUuOQeYOh8PUc3Xt2lV33323ypQpY/pcWFZy6Z6tf5Nt27Zp2LBhOnTokH766SfT6jZp0kRLly41rd717N27V23atNHYsWNZqgcA7jDNmzfXuHHjlJqaqsjISG3YsEFFixbV3r179f7775v6HhQeHu6Sa++6dM/W0qVLNX78eOeSN9kTau7atSufW3ZzsrKytGHDBn3xxRfaunWrqlWrpvfee8/Uc9SpU0eLFy9WSEhIjjsd8zrDe7ly5ZyfKrLz+quvvirJ/CWHAAAFV3R0tF566SVduHBBnTp1UtGiRTV//nxNnjzZtLWEs/33v//V3r17c0ye6gpcumerVq1amjJlih5//PEc3YlmT59glWrVqumJJ55QRESEatasacli1KNGjdLnn3+eY+FOZngHAJgpPT1daWlpzqmMfv75Z91777165JFHTD1PZGSk9u7dqxIlSsjLy8tlVi1x6Z6t4sWLu1y6vdrKlStVtGhRS8+xevVqJSQk5FiE2kyuOlgRAGAeT09PeXp6Ki4uTgcPHlT79u21evVq08PWpEmTTK13u7hk2Mq+XvvAAw+oc+fOql27do7eLDNnk7VC+/btNW3aNDVu3Pi6A/zMTOgPPvigUlJSLAtbrjpYEQBgrjFjxujEiRP65Zdf1LZtWy1dulR79+5V796981z72LFjkuSy7zMuGbbWr18v6cqClB4eHtq4caPzmM1mK/Bhq2LFioqNjVXnzp0tP5fNZtMLL7ygMmXK5FjKyIxlEyTp5MmTmjlzpim1AACua8OGDYqJiVFkZKSKFCmimTNnKiIiwpSw9corr8hmszkvG2bjMqKFrl6oOXug3IULF7Rnzx5TFrq02u+//67ff/9df/75p/744w/VqFFDbm5u2rBhgx5//HFFRkaadi6r5xpz1cGKAABzZS+bkx2G0tPTTVtKZ+3atZKuTF6+YcMGpaSkOIeuuEJvl0uGrWzjx4/XTz/9pM8++0ypqakaP368qlatqrfeeiu/m/a3su/OaNmypZYvX67ixYtLklJSUkxvu1XrPGU7cOCAGjdurOLFiztn8pfMvRQKACj46tWrp7ffflspKSmaNWuWVqxYobCwMFPP0aNHDx0/flylS5fOsb+gz+/o0mHrm2++0fLlyyVdWaNv9uzZioyMLPBhK1tSUlKOAfLe3t5KTk7Oxxb9c61atcrvJgAACoB27drp+++/1/3336/jx4+rc+fOqlmzpqnn2Ldvn7766iuX6M26mkuHrczMTF2+fFmFCxd2bruS5557Tq1bt9bzzz8vh8Oh+Ph41a9fP7+b9Y9s2bLF+XVGRoa2b9+uoKAgUy+FAgAKvjfffFMRERGKjo62ZCojSSpdurSSk5NdYj3Eq7n0PFuffPKJlixZotq1a0uSvvvuOzVr1sylFkFetWqVtmzZIpvNpmeeecb5WlzV2bNnFR0dzaB5ALjDfPvtt1q5cqVzku6IiAjTVxV544039OOPP8rf3z9HoDPrpi+ruHTYkqSffvpJW7ZskYeHhypVqqTAwMD8btIdLT09XWFhYVq9enV+NwUAkA/S0tL07bffavr06Tpz5ozWrVtnWu2rr6ZczerxyXnl0pcRs+fbyl5c+Y8//tAff/xR4Kd++Ddp2bJljmV7jhw5oho1auRzqwAA+eHgwYP64osvFB8fr5IlS5o+rregh6obcemwlT3flnRlvNb27dv19NNPE7Zuo6vnCrPZbCpWrJgef/zxfGwRACA/hIeHy263KyIiQp999pnLjauykstfRrzamTNn1KNHD33yySf53RQAAO4o+/btU9myZXPs27Bhg6pVq5ZPLSo4XLpn66/uvvtu/fnnn/ndDAAA7jjZQev06dNasmSJFi1apPT09BxXoe5ULh22WrdunWO80B9//EGCBgAgH2zevFmff/651qxZIzc3Nw0cOND0SU1dlUtfRkxISMixXaxYMZaNAQDgNpo1a5YWLlwoDw8P1a9fX/Xr19frr7/uXGIHLt6z9cwzz2jjxo1KSEhQVlaWgoODCVsAANxG48aNU61atdSiRQsFBQXJZrO53AzvVnPpsPXpp59q5cqVCg8Pl2EYYVphIAAAAhtJREFU+vDDD7V//361a9cuv5sGAMAdYf369Vq5cqWGDx+u5ORk1a9fX+np6fndrALFpS8jhoeHa9GiRfL29pYkXbx4UU2bNtWXX36Zzy0DAODOs3fvXi1btkxxcXEqUaKEXn75ZbVo0SK/m5XvXLpnyzAMZ9CSpEKFCslut+djiwAAuDNlZmbqxIkTKl++vMqUKaPU1FSNGzeOsCUXD1uVK1fW22+/7Vz0OCYmRk8//XQ+twoAgDtP9+7ddezYMZUuXdo5Zuv555/P51YVDC59GdHhcGjevHn64Ycf5HA4FBISoqioKHl4eOR30wAAuKPUq1dPX331FYPjr8Mlw9a0adPUvn37/G4GAAD4/7311lvq378/y/Rch0teRoyPjydsAQBQgKSlpalevXry9/eXp6enc//s2bPzsVUFg0uGLQAAULDQCXJjLnkZsVy5cte9JmwYhmw2m/bs2ZMPrQIAALiWS/ZslStXTrGxsfndDAAAgFy55XcDAAAA/s1csmerXr16kq5MoLZhwwadPXs2x/FGjRrlR7MAAACu4ZJjtrJ17dr1mgnUJGn48OH52CoAAID/45I9W9n27dun+Pj4/G4GAADADbn0mK3SpUsrKSkpv5sBAABwQy7ds8UEagAAoKBz6bDFBGoAAPx/7dpBCQAAAAKx/q1NIShsKe5xrLse5AEA1l0/WwAA68QWAECR2AIAKBJbAABFAUQDYabrbSy9AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(figsize=(10,5))\n",
    "plot = sns.barplot(x=features, y=final_clf.feature_importances_)\n",
    "ax.set_title('Feature Importance')\n",
    "plot.set_xticklabels(plot.get_xticklabels(),rotation='vertical')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "#allow logloss and classification error plots for each iteraetion of xgb model\n",
    "def plot_compare(metrics,eval_results,epochs):\n",
    "    for m in metrics:\n",
    "        test_score = eval_results['val'][m]\n",
    "        train_score = eval_results['train'][m]\n",
    "        rang = range(0, epochs)\n",
    "        plt.rcParams[\"figure.figsize\"] = [6,6]\n",
    "        plt.plot(rang, test_score,\"c\", label=\"Val\")\n",
    "        plt.plot(rang, train_score,\"orange\", label=\"Train\")\n",
    "        title_name = m + \" plot\"\n",
    "        plt.title(title_name)\n",
    "        plt.xlabel('Iterations')\n",
    "        plt.ylabel(m)\n",
    "        lgd = plt.legend()\n",
    "        plt.show()\n",
    "        \n",
    "def fitXgb(sk_model, training_data=training_data,epochs=300):\n",
    "    print('Fitting model...')\n",
    "    sk_model.fit(training_data['X_train'], training_data['Y_train'].reshape(training_data['Y_train'].shape[0],))\n",
    "    print('Fitting done!')\n",
    "    train = xgb.DMatrix(training_data['X_train'], label=training_data['Y_train'])\n",
    "    val = xgb.DMatrix(training_data['X_val'], label=training_data['Y_val'])\n",
    "    params = sk_model.get_xgb_params()\n",
    "    metrics = ['mlogloss','merror']\n",
    "    params['eval_metric'] = metrics\n",
    "    store = {}\n",
    "    evallist = [(val, 'val'),(train,'train')]\n",
    "    xgb_model = xgb.train(params, train, epochs, evallist,evals_result=store,verbose_eval=100)\n",
    "    print('-- Model Report --')\n",
    "    print('XGBoost Accuracy: '+str(accuracy_score(sk_model.predict(training_data['X_test']), training_data['Y_test'])))\n",
    "    print('XGBoost F1-Score (Micro): '+str(f1_score(sk_model.predict(training_data['X_test']),training_data['Y_test'],average='micro')))\n",
    "    plot_compare(metrics,store,epochs)\n",
    "    features = ['nTouchpoints', 'single', 'divorced', 'married', 'unknown', 'P4',\n",
    "       'P3', 'P2', 'P1', 'U', 'N', 'Y', 'C', 'B', 'D', 'A', 'F', 'E', 'New',\n",
    "       'G', 'nAveSpend', 'nIncome', 'nAge']\n",
    "    f, ax = plt.subplots(figsize=(10,5))\n",
    "    plot = sns.barplot(x=features, y=sk_model.feature_importances_)\n",
    "    ax.set_title('Feature Importance')\n",
    "    plot.set_xticklabels(plot.get_xticklabels(),rotation='vertical')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### XGBoost hyperparameter tuning with GridSearchCV\n",
    "\n",
    "#### Initial model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost.sklearn import XGBClassifier\n",
    "#initial model\n",
    "xgb1 = XGBClassifier(learning_rate=0.1,\n",
    "                    n_estimators=1000,\n",
    "                    max_depth=5,\n",
    "                    min_child_weight=1,\n",
    "                    gamma=0,\n",
    "                    subsample=0.8,\n",
    "                    colsample_bytree=0.8,\n",
    "                    objective='multi:softmax',\n",
    "                    nthread=4,\n",
    "                    num_class=9,\n",
    "                    seed=27)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting model...\n",
      "Fitting done!\n",
      "[0]\tval-mlogloss:2.07116\tval-merror:0.49547\ttrain-mlogloss:2.07036\ttrain-merror:0.49435\n",
      "[100]\tval-mlogloss:0.94984\tval-merror:0.28160\ttrain-mlogloss:0.84953\ttrain-merror:0.22979\n",
      "[200]\tval-mlogloss:0.67841\tval-merror:0.15947\ttrain-mlogloss:0.54003\ttrain-merror:0.08706\n",
      "[299]\tval-mlogloss:0.51569\tval-merror:0.09920\ttrain-mlogloss:0.36673\ttrain-merror:0.03691\n",
      "-- Model Report --\n",
      "XGBoost Accuracy: 0.9668271803103263\n",
      "XGBoost F1-Score (Micro): 0.9668271803103263\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAF3CAYAAACSQ46hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdfVzN9/8/8MfpdIlN+HYattlm0sfSTKFhGSKldGVmmrCJfOYqF8tGzDXTZBizzTQU0SjZpLmaobnoOwsJzeYidCEyKXU6798ffp1vx9WJ83rXOTzut5vbrfN+n/N8v05H5zzO6/16v14KSZIkEBEREZEszGq7AURERERPMoYtIiIiIhkxbBERERHJiGGLiIiISEYMW0REREQyYtgiIiIikpF5bTeAiIxfy5Yt4eDgADOz//t+5uTkhNmzZz9WvYyMDCQkJGDGjBmimniPli1bIi0tDQ0bNpTtGPezceNGlJWVITg4uEaPS0TGi2GLiKrlhx9+EBZcsrOzkZubK6SWsUlPT0eLFi1quxlEZEQYtojIIH/99Rdmz56N69evo6KiAgMHDkTfvn2h0WgwZ84c/PnnnyguLoYkSZg1axaaNGmCxYsX499//8Unn3wCf39/zJw5E1u3bgUAHDx4UHt7yZIlOHr0KPLy8tCyZUtERUVh+fLlSE1NhUajQdOmTTFt2jTY29s/sH0XL17EoEGD4ObmhqNHj0KtVuPjjz9GfHw8zp49CycnJyxcuBCXLl3CwIED0b59e2RlZUGSJEydOhWurq4oLy/HvHnzkJaWBqVSCWdnZ3zyySeoV68eunXrBmdnZ5w6dQrjxo3Drl27sH//flhbW8PT0xNTp07F1atXkZ+fj6ZNm2LRokVo1KgRunXrhoCAAKSlpeHy5cvw8vLCxx9/DABISEjAqlWrYGZmhgYNGmD+/Plo3Lgxdu3aheXLl6O8vBzW1taIiIjAG2+8USOvMxEZQCIi0sPBwUHy8fGR+vTpo/1XUFAglZeXS97e3tLx48clSZKkGzduSF5eXtIff/wh/e///q80atQoqaKiQpIkSVqxYoU0fPhwSZIk6ccff5SGDRsmSZIk/f7771Lv3r21x6p6e/HixZKnp6dUXl4uSZIkbd68WRo7dqz29vr166WhQ4c+sM1Xr16VLly4IDk4OEg7duyQJEmSpk6dKnXt2lX6999/pdLSUqlTp05Senq69n5btmyRJEmS9uzZI3Xq1EkqKyuTvvzyS2nkyJFSWVmZVFFRIU2aNEmKjIyUJEmSunbtKi1dulR73IiICOm7776TJEmSYmJipBUrVkiSJEkajUYaOnSotHLlSu3j5s2bJ0mSJF25ckVq3bq1dP78eenkyZNShw4dpEuXLkmSJEmrVq2SIiMjpb///lvy8fGRCgsLJUmSpNOnT0udOnWSiouLH/n1JKKaxZ4tIqqW+51GzM7Oxvnz5/Hpp59qt5WWliIzMxMDBgxA/fr1sX79ely4cAEHDx5E3bp1H/m4bdq0gbn5nbeq3bt349ixYwgKCgIAaDQalJSU6K1hYWGBbt26AQBefPFFvPHGG6hXrx4AQKVSoaioCCqVCvXr14evry8AoEuXLlAqlTh16hT27t2L8PBwWFhYAAAGDhyIjz76SFvf1dX1vscdNGgQjhw5glWrVuGff/7BmTNn8Prrr2v3d+/eHQBgb2+PRo0aoaioCIcPH0bnzp3RuHFjAMDgwYMBALGxscjLy9PeBgCFQoHz58/D0dFR7++AiGoPwxYRPbaKigo8++yzSEpK0m4rKCjAM888gz179mD27NkYMmQIunfvjldeeQVbtmy5p4ZCoYBUZYnW8vJynf116tTR/qzRaDB06FAMGDAAAFBWVoaioiK97bSwsIBCodC5fT9KpVLntkajgVKphEajuWd71XZWbWNVCxYsQEZGBoKCgtChQweo1Wqd52plZaX9ufL3oFQqddpaWlqKnJwcaDQavPnmm1i0aJF23+XLl6FSqR721InICHDqByJ6bC+//DKsrKy0Yevy5cvw8fHB8ePHsX//fnTt2hUDBgxA69atsWPHDlRUVAC4E2rUajUAoGHDhrh06RKuXr0KSZKwY8eOBx6vc+fOSEhIwM2bNwEAX375pXackwiFhYXYu3cvAGDXrl2wsLCAg4MD3nrrLaxfvx7l5eXQaDSIjY1Fp06d7luj6nPbt28fBg0aBH9/fzRq1AgHDhzQ/g4epEOHDkhLS0NeXh4AYP369ViwYAHc3Nywf/9+/PXXXwCAX3/9FX369MHt27dFPX0ikgl7tojosVlaWmLZsmWYPXs2vvvuO6jVaowZMwYuLi6wtbXFhAkT4OvrC6VSCVdXV+3A9jfeeAOLFi3CRx99hK+++gr9+/dHUFAQ7Ozs8Pbbbz/weO+88w5yc3PRr18/KBQKNG7cGPPmzRP2fCqDY1RUFKytrfHVV19BqVRixIgRmD9/Pvz9/aFWq+Hs7IzIyMj71nB3d8fMmTMBAB999BE+//xzLFu2DEqlEm3btsX58+cf2oaWLVti4sSJGDp0KADAzs4Oc+bMgb29PWbMmIFx48ZBkiSYm5tj+fLlD+xVIyLjoZCq9mkTET2lLl68CF9fX/zxxx+13RQiesLwNCIRERGRjNizRURERCQj9mwRERERycgoB8hrNBoUFxffc7k2ERERkbGRJAnl5eWoW7euzhqylYwybBUXF+P06dO13QwiIiKianNwcMAzzzxzz3ajDFuVEw46ODjA0tKylltDRERE9GBlZWU4ffr0AydMNsqwVXnq0NLSUmeGZSIiIiJj9aChTxwgT0RERCSjaoWt5ORkeHt7o0ePHoiNjX3g/SIiIrBp06Z7tmdmZsLJyenxW0lERERkovSGrdzcXERHRyMuLg5JSUmIj49Hdnb2PfcJCwtDSkrKPY8vKSnBjBkz7llcloiIiOhpoDdsHThwAG5ubrC1tUWdOnXg6el5T6hKTk5G9+7d4eXldc/j582bh8GDBwtrMBEREZEp0TtAPi8vD3Z2dtrbKpUKGRkZOvepXDA1PT1dZ/vOnTtRWlqKXr16PVbjjh8//liPIyIiIjIWesPW/Vbzqc5Eo/n5+Vi+fDliYmIeq2EA4OTkxKsRiYiIyKjdvn37oR1Eek8j2tvbo6CgQHs7Ly8PKpVK74H37NmD69evIzg4GH5+fgAAPz8/3Lx5szrtJiIiInoi6A1bHTt2RFpaGgoLC1FSUoLU1FS4u7vrLfzOO+9gx44dSEpKQlJSEgAgKSkJ9erVM7zVRERERCaiWj1b4eHhCAkJgb+/P3x8fODs7IzQ0FAcO3asJtpIREREZLIU0v0GZdWyynOfHLNFRERExk5fbuEM8lQtFeoyo6xFRERk7IxybUQyPkpzS2xb6S2klteHPwupQ0REZArYs0VEREQkI4YtIiIiIhkxbBERERHJiGGLiIiISEYMW0REREQyYtgiIiIikhHDFhEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGTFsEREREcmIYYuIiIhIRgxbRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFREREJCOGLSIiIiIZMWwRERERyYhhi4iIiEhGDFtEREREMmLYIiIiIpIRwxYRERGRjBi2iIiIiGTEsEVEREQkI4YtIiIiIhlVK2wlJyfD29sbPXr0QGxs7APvFxERgU2bNmlvp6enIygoCH5+fhg0aBBycnIMbzERERGRCdEbtnJzcxEdHY24uDgkJSUhPj4e2dnZ99wnLCwMKSkpOtsnTpyI2bNnIykpCb6+vpg1a5bY1hMREREZOb1h68CBA3Bzc4OtrS3q1KkDT0/Pe0JVcnIyunfvDi8vL+22srIyjBkzBo6OjgCAli1b4vLly4KbT0RERGTczPXdIS8vD3Z2dtrbKpUKGRkZOvcZOnQogDunDStZWlrCz88PAKDRaLB06VJ4eHgIaTQRERGRqdAbtiRJumebQqGo9gHKysowadIkqNVqDB8+/JEad/z48Ue6P8nHxcVFaL2qwZyIiOhJpjds2dvb48iRI9rbeXl5UKlU1SpeXFyMESNGwNbWFsuXL4eFhcUjNc7JyQlWVlaP9BgyDaLDGxERUW25ffv2QzuI9I7Z6tixI9LS0lBYWIiSkhKkpqbC3d29WgefOHEimjVrhi+//BKWlpbVbzURERHRE6JaPVvh4eEICQlBeXk5+vbtC2dnZ4SGhmL06NFo3br1fR+XmZmJnTt34tVXX4W/vz+AO+O9vv32W7HPgIiIiMiIKaT7DcqqZZXdcTyNaFy2rfQWUsfrw5+F1CEiIjIG+nILZ5AnIiIikhHDFhEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGTFsEREREcmIYYuIiIhIRgxbRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFREREJCOGLSIiIiIZMWwRERERyYhhi4iIiEhGDFtEREREMmLYIiIiIpIRwxYRERGRjBi2iIiIiGTEsEVEREQkI4YtIiIiIhkxbBERERHJiGGLiIiISEYMW0REREQyYtgiIiIikhHDFhEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGVUrbCUnJ8Pb2xs9evRAbGzsA+8XERGBTZs2aW9funQJwcHB6NWrF0aMGIHi4mLDW0xERERkQvSGrdzcXERHRyMuLg5JSUmIj49Hdnb2PfcJCwtDSkqKzvbp06djwIABSElJgZOTE5YtWya29URERERGTm/YOnDgANzc3GBra4s6derA09PznlCVnJyM7t27w8vLS7utvLwchw8fhqenJwAgMDDwnscRERERPenM9d0hLy8PdnZ22tsqlQoZGRk69xk6dCgAID09Xbvt2rVrqFevHszN7xzCzs4Oubm5QhpNREREZCr0hi1Jku7ZplAo9BZ+3MdVdfz48Ue6P8nHxcVFaL2qwZyIiOhJpjds2dvb48iRI9rbeXl5UKlUegs3bNgQN2/eREVFBZRKJfLz86v1uKqcnJxgZWX1SI8h0yA6vBEREdWW27dvP7SDSO+YrY4dOyItLQ2FhYUoKSlBamoq3N3d9R7YwsICrq6u+PnnnwEAiYmJ1XocERER0ZNEb9iyt7dHeHg4QkJC4O/vDx8fHzg7OyM0NBTHjh176GOnTZuGDRs2wNvbG0eOHMHYsWOFNZyIiIjIFCik+w2uqmWV3XE8jWhctq30FlLH68OfhdQhIiIyBvpyC2eQJyIiIpIRwxYRERGRjBi2iIiIiGTEsEVEREQkI4YtIiIiIhkxbBERERHJiGGLiIiISEYMW0REREQyYtgiIiIikhHDFhEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGTFsEREREcmIYYuIiIhIRgxbRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFRERE1SapNUZZy5iZ13YDiIiIyHQozM1w5YszQmo9N76FkDrGjj1bRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFREREJCOGLSIiIiIZMWwRERERyYhhi4iIiEhG1QpbycnJ8Pb2Ro8ePRAbG3vP/pMnTyIoKAienp6YPHky1Go1AODixYsIDg6Gn58fBg4ciJycHLGtJyIiIjJyesNWbm4uoqOjERcXh6SkJMTHxyM7O1vnPhMnTkRkZCS2b98OSZKwYcMGAMCXX36J3r17IykpCT179kR0dLQ8z4KIiIjISOkNWwcOHICbmxtsbW1Rp04deHp6IiUlRbs/JycHpaWlaNOmDQAgMDBQu1+j0eDmzZsAgJKSElhbW8vxHIiIiIiMlt61EfPy8mBnZ6e9rVKpkJGR8cD9dnZ2yM3NBQCMGTMG/fv3x5o1a1BeXo74+PhHatzx48cf6f4kHxcXF6H10tPThdYjIqKawc+DR6c3bEmSdM82hUJRrf0RERGYMWMGPDw8sH37dowcORJbtmzRefzDODk5wcrKqlr3JdMi+o+ViIhM05PweXD79u2HdhDpPY1ob2+PgoIC7e28vDyoVKoH7s/Pz4dKpUJhYSHOnj0LDw8PAICnpyfy8/Nx7dq1x3oiRERERKZIb9jq2LEj0tLSUFhYiJKSEqSmpsLd3V27v2nTprCystJ2AyYmJsLd3R0NGjSAlZUVjhw5AuBON2HdunXRsGFDmZ4KERERkfHRexrR3t4e4eHhCAkJQXl5Ofr27QtnZ2eEhoZi9OjRaN26NaKiojBlyhQUFxejVatWCAkJgUKhwNKlSzFz5kyUlpaibt26WLJkSU08JyIiIiKjoZDuN+iqllWe++SYLeOybaW3kDpeH/4spA4REdWOK1+cEVLnufEthNSpbfpyC2eQJyIiIpIRwxYRERGRjBi2iIiIiGTEsEVEREQkI4YtIiIiIhkxbBERERHJiGGLiIiISEYMW0REREQyYtgiIiIikhHDFhEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGTFsEREREcmIYYuIiIhIRgxbRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFREREJCOGLSIiIiIZMWwRERERyYhhi4iI6AmiUUtGWetpZl7bDagOSV0BhbnS6GoREREZGzNzBbKX5Aqp9eooeyF1nnYmEbYU5krkL18rpJbdiPeF1CEiIiKqDp5GJCIiIpIRwxYRERGRjBi2iIjIZKgrxA3YFlmL6GFMYswWERERAJgrFfhy8xUhtcYEPCekDpE+1erZSk5Ohre3N3r06IHY2Nh79p88eRJBQUHw9PTE5MmToVarAQB5eXkYNmwY/P390b9/f1y8eFFs64mIiIiMnN6wlZubi+joaMTFxSEpKQnx8fHIzs7Wuc/EiRMRGRmJ7du3Q5IkbNiwAQDw8ccfo2vXrkhMTISfnx+ioqLkeRZERERERkpv2Dpw4ADc3Nxga2uLOnXqwNPTEykpKdr9OTk5KC0tRZs2bQAAgYGBSElJQWFhIbKystC/f38AQFBQEMaOHSvT0yAiIiIyTnrHbOXl5cHOzk57W6VSISMj44H77ezskJubiwsXLqBJkyaYM2cODh48iCZNmiAyMvKRGnf8+HEAgIuLyyM9Tp/09HSh9Z4GfA2IyBjwvUg/uX9HfA0end6wJUn3Xq2hUCj07ler1cjMzMSoUaMwefJkbNy4EZMmTcKaNWuq3TgnJydYWVlV+/7VJfo/Cj06vgZEZAz4XqSf3L+jJ+E1uH37traD6H70nka0t7dHQUGB9nZeXh5UKtUD9+fn50OlUsHOzg5169ZF165dAQA+Pj46PWJERERETwO9Yatjx45IS0tDYWEhSkpKkJqaCnd3d+3+pk2bwsrKStsNmJiYCHd3d7z44ouwt7fHr7/+CgDYvXs3XnvtNZmeBhEREZFxqlbPVnh4OEJCQuDv7w8fHx84OzsjNDQUx44dAwBERUVh7ty58PLyQklJCUJCQgAAS5cuxXfffQcfHx+sXr0ac+bMkffZEBERERmZak1q6uvrC19fX51t3377rfZnR0dHJCQk3PO4V1555ZHGaBERERE9abhcDxEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGTFsEREREcmIYYuIiIhIRgxbRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFRERUQzQVklHXI3lUa21EIiIiMpyZUoG0H/KF1XtzkJ2wWiQf9mwRERERyYhhi4iIiEhGDFtEREREMmLYIiIiIpIRwxYRERGRjBi2iIiIiGTEsEVEREQkI4YtIiIiIhkxbBERERHJiGGLiIiISEYMW0REREQyYtiqARp1mVHXIyIiIvlwIeoaYGZuiX8W+wur99LoRGG1iIiISF7s2SIiIiKjIak1Rl3vcbBni4joKVFWoYGlUtx3bNH1iABAYW6G3C9/F1bPfoybsFqPi2GLiOgpYak0Q+CPacLqbQp6855tIgMYwxw9KRi2iIhIGEulGd7ddFZIrfjAV4TUIapt1frKkJycDG9vb/To0QOxsbH37D958iSCgoLg6emJyZMnQ61W6+zPzMyEk5OTmBYTERERmRC9YSs3NxfR0dGIi4tDUlIS4uPjkZ2drXOfiRMnIjIyEtu3b4ckSdiwYYN2X0lJCWbMmIHy8nLxrSciIiIycnrD1oEDB+Dm5gZbW1vUqVMHnp6eSElJ0e7PyclBaWkp2rRpAwAIDAzU2T9v3jwMHjxYfMuJiJ4wZRUVRl2PiB6P3jFbeXl5sLOz095WqVTIyMh44H47Ozvk5uYCAHbu3InS0lL06tXrsRp3/PhxAICLi8tjPf5B0tPThdbTR3T7AdN/DjXdfiJT4OLiAr+EFP13rKakvr10/tZq4r1I7vcKueu/9lprWFtbCqtfWlqGEyeOaW/zNaj5+ncfo3Wr12BpYy2sdllJKY5nnnjoffSGLUmS7tmmUCj07s/Pz8fy5csRExNTjaben5OTE6ysrB778Q8ixwtZ00z9OZh6++npVFZRAUul0uhqPYzcf2tPYv2kjQXC6vu98z9P5O/IlOrf7xh5S7cJq60a6QUnJydtB9H96A1b9vb2OHLkiPZ2Xl4eVCqVzv6Cgv/7j5mfnw+VSoU9e/bg+vXrCA4O1u7z8/NDbGws6tWr98hPhoiotlkqlfBJ2Cik1ta+7wipQ0TGT++YrY4dOyItLQ2FhYUoKSlBamoq3N3dtfubNm0KKysrbRddYmIi3N3d8c4772DHjh1ISkpCUlISACApKckog5Z019WTxlKLiIiITF+1erbCw8MREhKC8vJy9O3bF87OzggNDcXo0aPRunVrREVFYcqUKSguLkarVq0QEhJSE20XRmFujtzlnwupZT/iYyF1iIiI6MlQrUlNfX194evrq7Pt22+/1f7s6OiIhISEh9Y4derUYzSPiIiIyLRxHYQnhEZdZpS1iIiInnZcrucJYWZuiYMrfITU6jB8q5A6RERExJ4tIiIiIlkxbBERERHJiGGLiIiISEYMW0REREQyYtgiIiIikhHDFhEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGTFsEREREcmIYYuIiIhIRgxbRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFREREJCOGLSIiIiIZMWwRERERyYhhi4iIiEhGDFtEREREMmLYIiIiIpIRwxYRERGRjBi2iIiIiGTEsEVEREQkI4YtIiIiIhkxbBERERHJiGGLiIiISEbVClvJycnw9vZGjx49EBsbe8/+kydPIigoCJ6enpg8eTLUajUAID09HUFBQfDz88OgQYOQk5MjtvVERERERk5v2MrNzUV0dDTi4uKQlJSE+Ph4ZGdn69xn4sSJiIyMxPbt2yFJEjZs2KDdPnv2bCQlJcHX1xezZs2S51kQERERGSm9YevAgQNwc3ODra0t6tSpA09PT6SkpGj35+TkoLS0FG3atAEABAYGIiUlBWVlZRgzZgwcHR0BAC1btsTly5dlehpERERExslc3x3y8vJgZ2enva1SqZCRkfHA/XZ2dsjNzYWlpSX8/PwAABqNBkuXLoWHh8cjNe748eMAABcXl0d6nD7p6ek6t02tfk0co6brE5kCU/s7u/sYfC+q+fp3H4OvQc3Xv/sYctSvzCsPojdsSZJ0zzaFQlHt/WVlZZg0aRLUajWGDx+u73A6nJycYGVl9UiPqQ45ftE1Wb8mjmHq9YlMAd8rnvz6NXEM1q/9Yzg5OT00cOk9jWhvb4+CggLt7by8PKhUqgfuz8/P1+4vLi7G0KFDoVarsXz5clhYWDzWkyAylLqizKjrERHRk0tvz1bHjh2xZMkSFBYWwsbGBqmpqZg5c6Z2f9OmTWFlZYX09HS4uLggMTER7u7uAO4MkG/WrBlmzJih09tFdLcKdRmU5pay1TJXWmLpWk8h9QFg5PvbhdUiIqInm96wZW9vj/DwcISEhKC8vBx9+/aFs7MzQkNDMXr0aLRu3RpRUVGYMmUKiouL0apVK4SEhCAzMxM7d+7Eq6++Cn9/fwB3xnt9++23sj8pMj1Kc0vExogJQ8GDGYSIiMh46A1bAODr6wtfX1+dbVVDk6OjIxISEnT2t2rVCqdOnRLQRCLTUF5RBgulmN45kbWIiKh2VStsEZF+FkpLRG7oJaTWzH4p+u8kWFlFOSyV4sZViq5HRGSqGLaICABgqbSA9+Zpwur9HDBdWK3qKqtQw1Ip5m1NZC0ierrxnYTIRJRVlMFS0KlFkbWqf0yx4eV+9SyV5vD5MUZI/a1Bg4XUISJi2CIyEZZKS3ht6SOk1rY+W4TUeRSWSnP03rRYWL2fAkcLq0VEJKdqLURNRERERI+HYYuIiIhIRgxbRERERDJi2CIiIiKSEcMWERERkYwYtoiIiIhkxLBFREREJCOGLSIiIiIZMWwRERERyYhhi4iIiEhGDFtEREREMmLYIiIiIpIRwxYRERGRjBi2iIiIiGTEsEVEREQkI4YtIiIiIhkxbBERERHJiGGLiIiISEYMW0REREQyYtgiIiIikhHDFhEREZGMGLaIiIiIZMSwRURERCQjhi0iIiIiGTFsEREREcmoWmErOTkZ3t7e6NGjB2JjY+/Zf/LkSQQFBcHT0xOTJ0+GWq0GAFy6dAnBwcHo1asXRowYgeLiYrGtJyIiIjJyesNWbm4uoqOjERcXh6SkJMTHxyM7O1vnPhMnTkRkZCS2b98OSZKwYcMGAMD06dMxYMAApKSkwMnJCcuWLZPnWRAREREZKXN9dzhw4ADc3Nxga2sLAPD09ERKSgpGjhwJAMjJyUFpaSnatGkDAAgMDMTixYvxzjvv4PDhw/jqq6+0299//31MnDhRb6MkSQIAlJWVabepLS0e8and3+3bt++7XW1pLWv9Cqv6Quo/7BiwtJW1vpnM9S0s5K1vJaj+w45Rx1ze52CrFPP/6MH1bYTUf9AxbJVWstYHAFtzed8rbM31vm0aVL++ubjRHfc7Rn0xzX9gfQB4Vlkha30rpVrW+kpzMfUfeAwLmesD0FjJ+ztSW2vkrS/ureK+x1Bbif07q8wrlfnlbgrpQXv+vxUrVuDWrVsIDw8HAGzcuBEZGfKm+CUAACAASURBVBmYOXMmAOCPP/7A559/jnXr1gEAzp07h2HDhmHNmjXo27cv9u7dCwBQq9Vo06YNjh8/rrfh//77L06fPl3Np0lERERU+xwcHPDMM8/cs13vd5z7ZTGFQqF3v77HPUzdunXh4OAACwuLaj+GiIiIqDZIkoTy8nLUrVv3vvv1hi17e3scOXJEezsvLw8qlUpnf0FBgfZ2fn4+VCoVGjZsiJs3b6KiogJKpVK7vTrMzMzumwyJiIiIjJG19YOHI+k9admxY0ekpaWhsLAQJSUlSE1Nhbu7u3Z/06ZNYWVlhfT0dABAYmIi3N3dYWFhAVdXV/z8888624mIiIieJnrHbAF3pn5YsWIFysvL0bdvX4SGhiI0NBSjR49G69atkZWVhSlTpqC4uBitWrXC3LlzYWlpiZycHEyaNAlXr15F48aNsXDhQtSvL26gOBEREZGxq1bYIiIiIqLHwxnkiYiIiGTEsEVEREQkI4YtIiIiIhkxbBERERHJiGGLiIhq3N9//13bTSCqMQxbRPREunXrVm03ge6iVqvx888/Y+DAgQgMDBRWNzc3V1gtIjkwbOlx69YtZGVlQZIkvnkbAb4GNcvR0RH/+c9/tP9atWoFNzc3jBs3DtevXze4/qVLlx76zxB+fn46q1/IpaioCIWFhdrbhw4d0rktwl9//YW8vDwAwDfffIOwsDAsWbIEpaWlwo5x+vRppKSkYM+ePbhw4YKwugBw4cIFREVFwd3dHR9//DE6dOiAnTt3CqsfFham/fn7778XVrdSXFyc9uczZ87o7Js9e7bw48khLi4Ov/zyCwCgb9++6N69O3r27Ilz587VcsuMjxyf+yY9z9aNGzeQmZkJNzc3fPfdd8jMzMRHH32E5s2bC6mflpaGqVOnoqKiAuvXr0efPn0QFRWFzp07G1TX0dFRZ81Hc3NzmJmZoaysDPXq1cPhw4cNbTq6dev20HUlRb3RqdVq7Nu3754PXn9/fyH17xYQEIDNmzcLqfXvv//iu+++Q/369dG7d2+MGTMGp0+fhouLC2bNmgV7e3uD6l++fBkzZ87ElStX4OHhgeHDh0OpVAIAhg8fjhUrVoh4GjWuoKAAGzZsQHZ2NhYuXGhQrcr/p1XfhhQKBfLy8qBWq3Hy5MnHrr1v3z5Mnz4dHh4eCA8Ph6WlpUFtvZ/MzEwMGzYMc+bM0a6QER0djU2bNuHbb7+Fo6Ojwcf4+uuvsX79eiiVSrRv3x4XL15Ejx49cOjQIdjY2GDBggUG1b969SpGjx6NM2fOoFmzZlAoFPj777/Rpk0bfPHFF3j22Wcfu/Yvv/yC9evX48SJE+jRowd69eqFyMhI7Nq1y6A2383f3x+JiYkAxL5HVKpa8+76Io/n6+uLLl264O2334aLi4uwtYFXrFiBtLQ0TJs2DS+//DJ8fX3x9ddfY/fu3cjMzMScOXMMqq/vM6tdu3YG1a+qrKwMK1euxN9//42pU6ciJiYGw4YNE/b3Ldfnvt61EY3ZuHHj8NZbb0GSJCQnJyMkJATTpk3D2rVrhdRfuHAh4uLiEBoaCpVKhbVr12LcuHEG/9KzsrIAANOmTUPbtm3Rp08fKBQKbN++Hb/99puIpmPNmjWQJAlfffUVXnjhBQQGBkKpVCI5ORkXL14UcgwAGD9+PC5duoTmzZvrvDGICFutW7eGWq0GcGeRz8oP5cqwasgHMQBERESgefPmyMrKwurVqzFixAj06dMHP//8M6ZNm4avv/7aoPqffvopfHx80LJlSyxduhRhYWFYvnw5zM3NhZz2WLp06UP3jxw50uBj3M///M//4L///S969+5tcK27P3SLi4sxf/587Nu3DzNnzjSodufOnbFlyxZ8+eWX6Nu3L6ZOnYomTZpo91f9+XHNnz8fX3zxBTp06KDdFh4eDldXV8ybNw8xMTEGHyM5ORnbtm3DrVu34OHhgQMHDsDGxgbBwcHw9vY2uP7MmTPh4uKCmJgYWFhYALjzgbZkyRLMmTMH8+bNe+zao0aNQq9evRAfH49mzZoBgLAAUVXVmnL0H1StKWf/xPfff4/ffvsNa9euxaeffgpnZ2d069bN4Nc5MTERCQkJ2kWSlUolmjZtigEDBsDX19fgdi9evBgAcP36dZw/fx5t27aFmZkZ/vjjDzg4OGD9+vUGH6PSjBkz0LBhQ2RmZkKpVOL8+fOYPHmywV86Ksn1uW/SYev69esYNGgQZs2ahcDAQAQFBSE2NlZYfY1GAzs7O+3tV199VVhtAMjIyMD06dO1tz09PbFs2TIhtZs2bQoAOHXqFObOnavd/sEHHwgdK3Hq1CmkpKQIq1dVfHw8Zs6ciSFDhqBnz54AdL/BGurixYtYtmwZysvL8fbbb+Pdd98FAAQFBQkJ7NevX0dQUBAAYPny5Rg/fjwmTpyI6Ohog2sDQEVFBWJiYjBkyBCYmdX8iIDKD2ZR0tLSMGXKFHTq1AlbtmxBvXr1DK5pY2ODMWPG4MqVKxgxYgSeffZZbXAX0bt748YNnaBV6a233kJUVJTB9YE7Pd82NjawsbHBCy+8ABsbGwB3PjArfzbEqVOnsGjRIp1tlpaWGDduHPz8/AyqvWXLFmzevBkDBgxA06ZN0bt3b1RUVBhUUx85wlxN1bezs0NAQABatGiBtLQ0rF27FgcOHDA4bCmVSm3QAoARI0YAAMzMzIT0CK1ZswYAEBoaiqVLl2qDdU5ODqZOnWpw/apOnDiBzZs3Y+/evbCxscH8+fOFBMZKcn3um3TY0mg0yMrKwo4dO7B69WqcPn1a2xMiwnPPPYfdu3dDoVDgxo0biI2NFfJtuJKNjQ1+/PFHeHl5QaPRICkpCba2tsLqV/r999/h5uYGAPj111+1p7JEaN68OfLy8qBSqYTVrNSqVSusWrUK8+bNw+7duzFlyhShb3Tm5uY4e/YsXnnlFaxatUq7PTMzU8hxlEolzpw5gxYtWkChUGD+/PkYOnSotovaUGPGjEFubi5sbGwQGhpqcL1HkZqaKuz/6q1btzBv3jxtb1anTp2E1AWAPXv2YMaMGejcuTN2794tJMBVpVarodFo7gm7Go0G5eXlQo5Rtfbdf7si/p9aWVndd7tCoTA4xDs4OCAiIgITJkzA7t27sXnzZhQUFGDYsGEIDg5Gly5dDKpf6cyZM+jevTuAO4PlK38WFazlDnCVQkNDcfbsWTg6OqJ9+/b45ptvhJyK1mg0uHnzpvb/v6enJ4A7QylEunTpkjZoAXd6jw0de3k3hUKBsrIy7Wty7do1oa+PXJ/7Jh22xo8fj5kzZyIkJAQvvvgiAgMDMWnSJGH1Z8yYgdmzZ+Py5cvw8PCAm5sbZsyYIaz+ggULMHPmTMyaNQtmZmbo2LEjPv/8c2H1AWDWrFmIiIjQDq5t2rSp0GOUlpaiV69ecHBw0PmGtHr1aiH1ra2t8dlnn2HXrl0YPHgwiouLhdQF7pzmGzFiBLZt2wYHBwcAwI4dOzBjxox7vuk/jk8++QRhYWEIDw+Hj48PLCwssHz5ckyYMAHZ2dkG1wfuPAeRA43vdr+xfzdv3kSzZs2EdNtX7c1KTk7W+fZtqNGjRyMzMxOzZ8/Gm2++KaxuVe3atcPSpUsxevRone3Lli2Dk5OTkGP8888/CAkJuednSZKEDG5+2AeVqA8xpVIJDw8PeHh4oLCwEElJSfjiiy+Eha3t27cLqfMgDwtz+fn5wo7TqlUr3Lp1C9evX8fVq1dRUFCA0tJSWFtbG1TX19cXERERmD9/vjZwFRcX49NPP0WfPn1ENB0A8NprryEiIkLbgbB161a4uroKqw8AISEhGDJkCPLz8zF79mzs2LEDH330kbD6cn3um/QA+aSkpHu6udetW4f33nuvllr0eK5fvy5Lj1ZVlelf9HEOHTp03+3t27c3uHZubi5mzpyJc+fO4Y033sDIkSPx66+/4p133jG49v3qT5gwAdbW1toLFkTVP3v2LFxdXTFhwgTtYOOTJ0/iP//5j8H1Z82ahX/++UfbfkMGM99PTk6Ozm0zMzM8++yzwkKRo6MjzM3NoVKp7hl3Y2iPxMyZMzF+/HjUqVNHRFPv6+bNmxg2bBjy8/PRunVrSJKEzMxMNGzYEMuXLxfy9/agv7FKhv6tOTk53fdikMogcezYMYPqPwnu/ju4W+WwDVGKi4uRmpqK5cuX49KlSzh+/LhB9SoqKvDZZ59h69at2vG12dnZ8PPzw2effSam0bgz1m/t2rXa/7MdO3bEgAEDYG4utl8nOzsbBw8eREVFBdq3by+k909uJhm21qxZg+LiYsTGxiI4OFi7Xa1WIzExETt27DCofk1dyXfy5EmEh4ejtLQU8fHxeP/997Fo0SK89tprQuoDd94kpkyZgpycHMTGxmL8+PGYM2cOnn/+eSH1hw0bhq5du6Jr16547rnnhNSs9OGHH+K1116Dq6srtm3bBgA6489E1HdycoKLi4vJ1pfz91MTavpDTA6SJOH333/HyZMnYWZmBicnJ+Hf5uX0JLwGT4rffvsNaWlpSEtLg0ajgbu7O7p06SLs/1Nubi4yMjIA3OmFEjksptLNmzfx77//6lxIIPI4ZWVl2LdvH27cuKGzXdQV8D179tQZ5qFQKGBtbY1XXnkFERERj/33YJJha+fOnTh27BgSEhLQt29f7XZzc3O0a9fuvgNWH0VNvfkEBwdjxowZGD9+PBITE7F//35ER0cjISFBSH3gzgfykCFDEBUVhc2bN2Pjxo1ISkoSdiHBn3/+ib1792Lv3r1Qq9Vwd3dHt27d8Prrrxtc28fHB1u3bgUAlJeXw9/fHz/99JPBdVmfiIxRWFgYunbtii5dugj/8loTvv76a3zzzTewtbXVXj0u6mKUSoMGDYIkSfd8Dov6ojlr1iw8//zz2myxZcsWHDt2DN26dUNsbOxjX2FskmO2unfvju7du8Pb21s71kakyhfxfle9WVtbo7i4WMhxS0pKdOYE69SpE+bPn29w3aquXbuGzp07IyoqCgqFAv369RN6xebrr7+O119/HcHBwUhJScHXX3+NlStXGtztDehe7WZhYSH86jfWJyJj8vXXX2PLli1Yt24dwsLCsH37dtnmLJRDQkICduzYgYYNG8p2jGvXrmHLli2y1U9PT8eUKVO0twcMGIDAwEDMnTvXoNkCTDJsVSooKEBkZOQ9E2qKGiy5c+dOZGZmwsPDA8CdK5tUKhVu3boFX19fDB482KD6tra2yMrK0p6y3LJlC+rXr29os3VYW1vjypUr2mMcOXJE6OSO06dPR3p6OpRKJdq1a4dp06YJGa91P6Z8SfeTUJ+I5BUVFYUrV67gxIkTGDp0KH788UdkZWUJvfBLTo0bNxb+GXY3Nzc3HDhwAG5ubrJMeWNmZobffvsNb731FoA7p3YtLS1RUFBg0GwHJnkasZKnpycmTpyovbS+0osvviikfv/+/fHNN99oBx3fvHkTYWFhiImJQWBgoMHp+vz584iIiMCxY8dgbW2tvcLrlVdeEdF8AHfm8oqMjMT58+fx4osvoqioCIsWLUKbNm2E1B8/fjxOnz6NFi1aoEOHDmjfvj1efvllIbXvHribm5sLe3t7YV3TrE9ExsTf3x+bN29GQEAAEhMToVartRMtm4LIyEicPn0aHTp00PlSL3KC5ZiYGMybN0/7mV/5fmfoJNeVzpw5g4iICFy6dAmSJKFZs2aYO3cuUlJS0KRJEwQEBDxWXZPu2WrQoIG210kO165d07nqysrKCkVFRTA3NxfSi/Diiy9iyZIlqFOnDjQaDa5evaozR4kIzs7OSEhIwD///IOKigq88sorQnu2vvjiCwB31m5LS0tDWFgYbt26JWQmfLkv52Z9IjImlT01lZ8vZWVltTJh8eOyt7c3eJkzfVavXo1du3bJMrgfAFq0aIFNmzahqKgISqUS9erVw759+wyeXsKkw5arqys+//xzvPXWWzoT87Vt21ZI/Z49e2LQoEHaOUNSU1PRvXt3JCYm6sww+7hWr16NzZs3Y/PmzcjJyUFYWBgGDx6sncncEEuWLMGoUaPwySef3He/qMGEZ8+e1V49c/LkSbz++uvC5s6R+yoo1iciY9KrVy+MHTsWRUVFiImJwZYtW+Dj41Pbzaq2kSNH4tatWzh//jwcHBxQWloqfOoVlUol+1RJwJ3pMuLj47Fx40bcvn0be/fuNaieSZ9GHDBgwD3bFAqF0AHgu3fvxv79+6FUKtGxY0d06dIFR48excsvv2zwuWkfHx9s2LBB+5+xpKQE/fr1Q3JyssHt3rVrF7p16/bABVIftyv0br6+vtqrZ9544w2T+hZGRGQMqs6yvnfvXhw4cACSJMHNzQ1du3aVrRdHNLkWca5qzJgxOHbsGNq2batzUZCoDoSDBw9i3bp12LFjB8zMzDB9+nTtpNSGMOmwVRPOnDmDoqIinTlDRK1g7unpiZ9++kk74ZtarUZAQICQsFXpgw8+wPfffy+s3v38+uuv+P3336FWq+Hm5qadXZmIiPSrnNvx7o/j/Px8qNVqYeOR5PbOO+9g2bJlCA0NRWJiIrKzszFu3DihVw/K1YEQExOD+Ph4WFhYwMvLC15eXvjggw+wa9cug+pWMsnTiJ999hk+++wzDBky5L5jp0SFi+nTp2P37t144YUXtNsUCoWwpWg8PDy0pykBaE9TilRaWorLly+jcePGQutW+u6777B9+3b4+vpCkiQsX74cZ86cQVhYmCzHIyJ60tz9gV5cXIz58+dr1ws1FXIt4lxVQEAATp8+jUOHDkGtVqNDhw4Gr8YBAAsXLkS3bt0QHBwMV1dXKBQKsWvxCqtUgwIDAwHcmb1cTvv370dKSorB61I9yLhx4/DLL7/g8OHDMDc3R0hIiPAB/4WFhejWrRsaNWoEKysr4VeqJSUlYePGjdrfUb9+/RAYGMiwRUT0GKquF7plyxbhi6fLSa5FnKtKTEzE0qVL4eHhAY1Gg5EjR2LEiBE6E5w/jr1792Lr1q2YO3cu8vPz4eXlhbKyMkGtfgJOI+7fvx9paWmoqKhAhw4d8Pbbbwur/eGHH2Lp0qWwsbERVrOqgICAB3aJipKRkYFGjRrds13U4Oqqs5gDd77Z+Pn5CT0VSkT0pLt16xbmzZun7c3q1KlTbTfpkV29ehWzZ8/Wjjnr0KEDpkyZApVKJewYfn5+iImJQYMGDQDc6VAICQnR+RwyVFZWFjZt2oTk5GQ0atQI7733ns7SgI/DJHu2Kn3//ffYunWr9hTW4sWLcfr0aWE9XvXr10fv3r3xxhtv6EyXIGogXqNGjXDkyBE4OzsLnY6hqoiICO26eXJwc3PDqFGjtOfLExMTDV4uiYjoaVK1Nys5OVnYQu81rVGjRvj888+RlZUFc3NztGzZUvhkyxqNRhu0AKBhw4ZCj6FWq3HlyhW0atUKLVq0QHFxMRYuXGhw2DLpni1fX19s2LBB2/N069Yt9O3bV9gEcHJfyefm5nbP7PciJ2cDgPDwcHTp0gXOzs46p0NFde1KkoS4uDgcPHhQe/XMu+++K3yVdyKiJ5WjoyPMzc2hUql0goOpTVC8f/9+REREQKVSQaPR4MaNG1i0aBGcnZ2FHWPChAlo0KCB9rThxo0bUVRUhAULFgipP2bMGFy6dAnNmzfXeS0M7WQx6bAl1yms/Px82NnZ6VyOW5WpXIYL3LnK5W6i/3jlXuWdiOhJlpOT89D9pjKnno+PD6KiouDo6AgAOHbsGKZNm4ZNmzYJO0ZpaSmWLFmC33//XfsF/7///a+wsW29evXCtm3bhPfImXT3Q/v27TF27FhtT9PmzZuFTMswZcoUrFixAu+//77OyuWVRAWVkpISLF26VDvmzM3NDWPGjBE6CZyoy1YfpCZWeSciepKZSpjSx9LSUhu0AKB169bCj2FhYYG2bdti4sSJKCwsxK5du4Sedm3evDny8/OFjjMDTLxnS6PRIDY2VptwO3TogAEDBhg8+VilP//8E+np6Xj//fcRFhaGEydOYPr06ejVq5eQ+p988glsbGzQr18/AMCGDRvw77//CusOBe7M8B4XF4dbt25BkiRoNBpcvHhR2MSvHh4e2LBhg6yrvBMRkfH77LPPUF5ejn79+kGpVOKnn37CxYsXERISAkDMHJWffPIJNBoN5s+fj8LCQsydOxc2NjaYMWOGwbWBOxfG/fHHH3BwcNAZS23olE8m3bNlZmaG9u3bQ6PRwNzcHJ06dRIWtABg9uzZmDhxIlJTU2FtbY3ExESMHDlSWNg6ceKEzmRvU6dOhbe3t5DalcLDw9G9e3ekp6cjICAAe/fuRYsWLYTVr4lV3omIyPj99ddfAICoqCid7YsXLxY2R+Xx48e1Q4UaNmyIBQsWwNfX1+C6lYYPHy6sVlUmHbZiYmKwZs0adOvWDRqNBqtWrcLIkSPh7+8vpL5Go0G7du0wfvx49OzZE40bN0ZFRYWQ2sCdwY83btzAs88+CwC4ceMGlEqlsPrAnecwevRoqNVqtGrVCv3790f//v2F1X/ppZcwYMAAWVd5JyIi47dmzRrZj6HRaJCXl6c9zXf16lWhy8S1b99eWK2qTDpsrVu3DomJiXjmmWcAAKNGjcKAAQOEhS0bGxt8//33OHjwIKZOnYoffvhB6LnhwYMHo2/fvtpB7Lt27RI+UauNjQ3Kysrw0ksv4cSJE3B1dcXt27eF1a+JVd6JiMh4aTQaxMXFoX379nBwcMDq1auxceNGtGrVCpGRkUInZg0LC0NAQABcXFwgSRIyMjIwefJkYfXlYtJjtvr164c1a9bAysoKAFBeXo733nsPCQkJQurn5uZi48aN6NixI9q2bYsFCxZg4MCBeO6554TUB4DTp0/j8OHD0Gg0aN++PVq2bCmsNgCsXbsWu3btQlRUFN599100a9YMkiRh5cqVQupfuHBBZzkjSZKwdu1aDBw4UEh9IiIybgsWLMDZs2cxefJk5ObmYtiwYViyZAnOnDmDkydPYt68eUKPl5ubi6NHj8Lc3BytW7cWPphdDiYdtiZMmICzZ8/Cx8cHSqUSv/zyCywtLbXdgMa+ZEyfPn3g5+cHHx8f2XqHCgsLsXXrVty4cQOXL1/GsWPH0LlzZ3z88cdC6vfq1QsrVqxAs2bNcOrUKUyZMgV16tTBDz/8IKQ+EREZN19fX2zevBnm5uaYPXs2iouLMWfOHACAl5eX0Im1b9y4geTkZFy/fl1nuiFjH7pi0qcRn3/+eTz//PO4ceMGAMDV1RXAnXk4TMEXX3yBrVu3IiQkBI0bN0afPn3g6ekp9FRlaGgoWrZsiSZNmqBx48bCF6SeM2cORowYgY4dOyI1NRXjxo0TdhqXiIiMn5mZmXYi60OHDukMMtdoNEKPNWbMGDzzzDNo0aKF8Lmw5GTSPVtPkiNHjmDOnDk4e/Ysjh49KqxuUFAQfvzxR2H17icrKwtDhw7FF198waV6iIieMv3798fChQtRXFyMgIAA7Nu3D7a2tsjKysLkyZOFfgb5+vqa5Nq7Jt2z9eOPPyI6Olq75E3lhJrHjx+v5ZZVT0VFBfbt24effvoJhw8fRufOnfHpp58KPYaHhwc2btwINzc3nSsdDZ3h3dHRUfutojKvDxo0CID4JYeIiMh4hYeH491338XNmzcxcuRI2NraIi4uDl999ZWwtYQr/ec//0FWVpbO5KmmwKR7trp164Zly5bh1Vdf1elOFD19glw6d+6M119/HX369EHXrl1lWYz6888/x7p163QW7uQM70REJFJZWRlKS0u1Uxn9+eefqF+/Pl566SWhxwkICEBWVhYaNWoEKysrk1m1xKR7tho2bGhy6baqrVu3wtbWVtZjpKamIi0tTWcRapFMdbAiERGJY2lpCUtLSyQnJyM7OxvDhw9Hamqq8LC1dOlSofVqikmGrcrztc8//zxGjRqF7t276/RmiZxNVg7Dhw/HihUrEBgYeN8BfiIT+gsvvICioiLZwpapDlYkIiKxoqKicOXKFZw4cQKhoaH48ccfkZWVhUmTJhlc+9KlSwBgsp8zJhm29u7dC+DOgpQWFhbYv3+/dp9CoTD6sNW2bVskJiZi1KhRsh9LoVCgd+/eaNGihc5SRiKWTQCAgoICrFq1SkgtIiIyXfv27cPmzZsREBCAevXqYdWqVejTp4+QsPX+++9DoVBoTxtW4mlEGVVdqLlyoNzNmzdx8uRJIQtdyu2ff/7BP//8gwsXLuDcuXPo0qULzMzMsG/fPrz66qsICAgQdiy55xoz1cGKREQkVuWyOZVhqKysTNhSOrt27QJwZ/Lyffv2oaioSDt0xRR6u0wybFWKjo7G0aNH8cMPP6C4uBjR0dHo1KkTPvroo9pu2kNVXp0xcOBAJCUloWHDhgCAoqIi4W2Xa52nSmfOnEFgYCAaNmyonckfEHsqlIiIjF+vXr0wduxYFBUVISYmBlu2bIGPj4/QY0yYMAGXL19G8+bNdbYb+/yOJh22du7ciaSkJAB31uhbvXo1AgICjD5sVcrLy9MZIG9jY4P8/PxabNGjCwkJqe0mEBGRERg2bBh+++03NGnSBJcvX8aoUaPQtWtXocc4deoUtm3bZhK9WVWZdNhSq9W4ffs26tSpo71tSt5++20MGTIEPXv2hEajQUpKCry8vGq7WY/k0KFD2p/Ly8uRnp4OV1dXoadCiYjI+P33v/9Fnz59EB4eLstURgDQvHlz5Ofnm8R6iFWZ9DxbK1euREJCArp37w4A+PXXX9GvXz+TWgR5+/btOHToEBQKBd58803tczFV169fR3h4OAfNExE9Zfbs2YOtW7dqJ+nu06eP8FVFPvzwQ/zxxx9wcHDQCXSiLvqSi0mHLQA4evQoDh06BAsLC7i4uMDZ2bm2m/RUKysrg4+PD1JTU2u7KUREVAtKs5yS7AAAA6BJREFUS0uxZ88efPPNN7h27Rp2794trHbVsylVyT0+2VAmfRqxcr6tysWVz507h3Pnzhn91A9PkoEDB+os23Px4kV06dKllltFRES1ITs7Gz/99BNSUlLQuHFj4eN6jT1UPYhJh63K+baAO+O10tPT0a5dO4atGlR1rjCFQoEGDRrg1VdfrcUWERFRbfD19YVSqUSfPn3www8/mNy4KjmZ/GnEqq5du4YJEyZg5cqVtd0UIiKip8qpU6fQsmVLnW379u1D586da6lFxsOke7bu9swzz+DChQu13QwiIqKnTmXQKiwsREJCAjZs2ICysjKds1BPK5MOW0OGDNEZL3Tu3DkmaCIiolpw8OBBrFu3Djt27ICZmRmmT58ufFJTU2XSpxHT0tJ0bjdo0IDLxhAREdWgmJgYxMfHw8LCAl5eXvDy8sIHH3ygXWKHTLxn680338T+/fuRlpaGiooKdOjQgWGLiIioBi1cuBDdunVDcHAwXF1doVAoTG6Gd7mZdNj6/vvvsXXrVvj6+kKSJCxevBinT5/GsGHDartpRERET4W9e/di69atmDt3LvLz8+Hl5YWysrLabpZRMenTiL6+vtiwYQNsbGwAALdu3ULfvn3x888/13LLiIiInj5ZWVnYtGkTkpOT0ahRI7z33nsIDg6u7WbVOpPu2ZIkSRu0AMDa2hpKpbIWW0RERPR0UqvVuHLlClq1aoUWLVqguLgYCxcuZNiCiYet9u3bY+zYsdpFjzdv3ox27drVcquIiIiePuPHj8elS5fQvHlz7Zitnv+vvfu1USAK4gA8XAKNkJAtgQouT6JpAIGgBCQ1YEnwOLYABB0QWkAhUcuJ+5cLkFPk7bv7PrlrfnIyM5n3+po5VTsUPUZsmibW63Xs9/tomiaGw2GMx+Podru5owHAv5JSiu12azn+jiKLreVyGZPJJHcMAODDdDqN+XzumZ47ihwj1nWt2AKAFrlcLpFSisFgEL1e7+v7arXKmKodiiy2AIB20QR5rMgxYlVVd2fC1+s1Op1OHA6HDKkAAG4V2dmqqio2m03uGAAAv3rJHQAA4C8rsrOVUoqI9wNqu90uzufzj/+j0ShHLACAG0XubH2azWY3B9QiIhaLRcZUAADfiuxsfToej1HXde4YAAAPFb2z1e/343Q65Y4BAPBQ0Z0tB9QAgLYruthyQA0AaLuiF+QBANqu6J0tAIC2U2wBADyRYgsA4IkUWwAAT/QG2KRenJLwaXwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fitXgb(xgb1, training_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Tune max_depth and min_child_weight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getTrainScores(gs):\n",
    "    results = {}\n",
    "    runs = 0\n",
    "    for x,y in zip(list(gs.cv_results_['mean_test_score']), gs.cv_results_['params']):\n",
    "        results[runs] = 'mean:' + str(x) + 'params' + str(y)\n",
    "        runs += 1\n",
    "    best = {'best_mean': gs.best_score_, \"best_param\":gs.best_params_}\n",
    "    return results, best\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 12 candidates, totalling 60 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 tasks      | elapsed:  2.8min\n",
      "[Parallel(n_jobs=-1)]: Done   9 tasks      | elapsed:  5.7min\n",
      "[Parallel(n_jobs=-1)]: Done  16 tasks      | elapsed:  7.8min\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed: 15.3min\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed: 23.8min\n",
      "[Parallel(n_jobs=-1)]: Done  45 tasks      | elapsed: 30.8min\n",
      "[Parallel(n_jobs=-1)]: Done  52 out of  60 | elapsed: 38.7min remaining:  5.9min\n",
      "[Parallel(n_jobs=-1)]: Done  60 out of  60 | elapsed: 44.0min finished\n",
      "/usr/local/lib/python3.7/site-packages/sklearn/preprocessing/_label.py:235: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n",
      "/usr/local/lib/python3.7/site-packages/sklearn/preprocessing/_label.py:268: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score=nan,\n",
       "             estimator=XGBClassifier(base_score=None, booster=None,\n",
       "                                     colsample_bylevel=None,\n",
       "                                     colsample_bynode=None,\n",
       "                                     colsample_bytree=0.8, gamma=0, gpu_id=None,\n",
       "                                     importance_type='gain',\n",
       "                                     interaction_constraints=None,\n",
       "                                     learning_rate=0.1, max_delta_step=None,\n",
       "                                     max_depth=5, min_child_weight=1,\n",
       "                                     missing=nan, monotone_constraints=None,\n",
       "                                     n_estimators=1000,...\n",
       "                                     objective='multi:softmax',\n",
       "                                     random_state=None, reg_alpha=None,\n",
       "                                     reg_lambda=None, scale_pos_weight=None,\n",
       "                                     seed=27, subsample=0.8, tree_method=None,\n",
       "                                     validate_parameters=None, verbosity=None),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'max_depth': range(3, 10, 2),\n",
       "                         'min_child_weight': range(1, 6, 2)},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring='f1_micro', verbose=10)"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_test1 = {\n",
    " 'max_depth':range(3,10,2),\n",
    " 'min_child_weight':range(1,6,2)\n",
    "}\n",
    "#metrics to consider: f1_micro, f1_macro, roc_auc_ovr\n",
    "gsearch1 = GridSearchCV(estimator = xgb1, param_grid = param_test1, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=5)\n",
    "gsearch1.fit(X_train, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.8630502445531347params{'max_depth': 3, 'min_child_weight': 1}\",\n",
       "  1: \"mean:0.8412627834593153params{'max_depth': 3, 'min_child_weight': 3}\",\n",
       "  2: \"mean:0.8196531791907514params{'max_depth': 3, 'min_child_weight': 5}\",\n",
       "  3: \"mean:0.9425522454424188params{'max_depth': 5, 'min_child_weight': 1}\",\n",
       "  4: \"mean:0.9318808359270786params{'max_depth': 5, 'min_child_weight': 3}\",\n",
       "  5: \"mean:0.9207647843485993params{'max_depth': 5, 'min_child_weight': 5}\",\n",
       "  6: \"mean:0.9451311694086261params{'max_depth': 7, 'min_child_weight': 1}\",\n",
       "  7: \"mean:0.9397065362383282params{'max_depth': 7, 'min_child_weight': 3}\",\n",
       "  8: \"mean:0.9338372610048911params{'max_depth': 7, 'min_child_weight': 5}\",\n",
       "  9: \"mean:0.9458425967096489params{'max_depth': 9, 'min_child_weight': 1}\",\n",
       "  10: \"mean:0.9400622498888394params{'max_depth': 9, 'min_child_weight': 3}\",\n",
       "  11: \"mean:0.9368608270342375params{'max_depth': 9, 'min_child_weight': 5}\"},\n",
       " {'best_mean': 0.9458425967096489,\n",
       "  'best_param': {'max_depth': 9, 'min_child_weight': 1}})"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Through GridSearchCV, we see that the best parameters for max_depth and min_child_weight are 9 and 1 respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting model...\n",
      "Fitting done!\n",
      "[0]\tval-mlogloss:2.02634\tval-merror:0.39200\ttrain-mlogloss:2.01941\ttrain-merror:0.36078\n",
      "[100]\tval-mlogloss:0.42041\tval-merror:0.06987\ttrain-mlogloss:0.27151\ttrain-merror:0.01272\n",
      "[200]\tval-mlogloss:0.22706\tval-merror:0.02987\ttrain-mlogloss:0.09315\ttrain-merror:0.00044\n",
      "[299]\tval-mlogloss:0.16545\tval-merror:0.02507\ttrain-mlogloss:0.04507\ttrain-merror:0.00000\n",
      "-- Model Report --\n",
      "XGBoost Accuracy: 0.9684323167469235\n",
      "XGBoost F1-Score (Micro): 0.9684323167469235\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb2 = XGBClassifier(learning_rate=0.1,\n",
    "                    n_estimators=1000,\n",
    "                    max_depth=9,\n",
    "                    min_child_weight=1,\n",
    "                    gamma=0,\n",
    "                    subsample=0.8,\n",
    "                    colsample_bytree=0.8,\n",
    "                    objective='multi:softmax',\n",
    "                    nthread=4,\n",
    "                    num_class=9,\n",
    "                    seed=27)\n",
    "\n",
    "fitXgb(xgb2, training_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 5 candidates, totalling 15 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done   2 out of  15 | elapsed:  6.4min remaining: 41.8min\n",
      "[Parallel(n_jobs=-1)]: Done   4 out of  15 | elapsed:  6.4min remaining: 17.7min\n",
      "[Parallel(n_jobs=-1)]: Done   6 out of  15 | elapsed:  6.5min remaining:  9.8min\n",
      "[Parallel(n_jobs=-1)]: Done   8 out of  15 | elapsed:  6.7min remaining:  5.8min\n",
      "[Parallel(n_jobs=-1)]: Done  10 out of  15 | elapsed:  7.9min remaining:  4.0min\n",
      "[Parallel(n_jobs=-1)]: Done  12 out of  15 | elapsed: 10.9min remaining:  2.7min\n",
      "[Parallel(n_jobs=-1)]: Done  15 out of  15 | elapsed: 11.0min finished\n",
      "/usr/local/lib/python3.7/site-packages/sklearn/preprocessing/_label.py:235: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n",
      "/usr/local/lib/python3.7/site-packages/sklearn/preprocessing/_label.py:268: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score=nan,\n",
       "             estimator=XGBClassifier(base_score=0.5, booster='gbtree',\n",
       "                                     colsample_bylevel=1, colsample_bynode=1,\n",
       "                                     colsample_bytree=0.8, gamma=0, gpu_id=-1,\n",
       "                                     importance_type='gain',\n",
       "                                     interaction_constraints='',\n",
       "                                     learning_rate=0.1, max_delta_step=0,\n",
       "                                     max_depth=9, min_child_weight=1,\n",
       "                                     missing=nan, monotone_constraints='()',\n",
       "                                     n_estimators=1000, n_jobs=4, n...\n",
       "                                     num_class=9, num_parallel_tree=1,\n",
       "                                     objective='multi:softprob',\n",
       "                                     random_state=27, reg_alpha=0, reg_lambda=1,\n",
       "                                     scale_pos_weight=None, seed=27,\n",
       "                                     subsample=0.8, tree_method='exact',\n",
       "                                     validate_parameters=1, verbosity=None),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'reg_alpha': [1e-05, 0.01, 0.1, 1, 100]},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring='f1_micro', verbose=10)"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_test2 = {\n",
    " 'reg_alpha':[1e-5, 1e-2, 0.1, 1, 100]\n",
    "}\n",
    "gsearch2 = GridSearchCV(estimator = xgb2, param_grid = param_test2, scoring='f1_micro',n_jobs=-1,verbose = 10, cv=3)\n",
    "gsearch2.fit(X_train, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({0: \"mean:0.9100938241422663params{'reg_alpha': 1e-05}\",\n",
       "  1: \"mean:0.9120502334359006params{'reg_alpha': 0.01}\",\n",
       "  2: \"mean:0.9102716018947398params{'reg_alpha': 0.1}\",\n",
       "  3: \"mean:0.8956869228924701params{'reg_alpha': 1}\",\n",
       "  4: \"mean:0.450422638495132params{'reg_alpha': 100}\"},\n",
       " {'best_mean': 0.9120502334359006, 'best_param': {'reg_alpha': 0.01}})"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getTrainScores(gsearch2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have found that max_depth of 9, min_child_weight of 1 and reg_alpha of 0.01 are optimum hyperparameters for the xgb model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting model...\n",
      "Fitting done!\n",
      "[0]\tval-mlogloss:2.02665\tval-merror:0.39307\ttrain-mlogloss:2.01937\ttrain-merror:0.35963\n",
      "[100]\tval-mlogloss:0.41486\tval-merror:0.06507\ttrain-mlogloss:0.26570\ttrain-merror:0.01201\n",
      "[200]\tval-mlogloss:0.22389\tval-merror:0.03307\ttrain-mlogloss:0.09071\ttrain-merror:0.00027\n",
      "[299]\tval-mlogloss:0.16446\tval-merror:0.02453\ttrain-mlogloss:0.04494\ttrain-merror:0.00000\n",
      "-- Model Report --\n",
      "XGBoost Accuracy: 0.9678972712680578\n",
      "XGBoost F1-Score (Micro): 0.9678972712680578\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb3 = XGBClassifier(learning_rate=0.1,\n",
    "                    n_estimators=1000,\n",
    "                    max_depth=9,\n",
    "                    min_child_weight=1,\n",
    "                    reg_alpha=0.01,\n",
    "                    gamma=0,\n",
    "                    subsample=0.8,\n",
    "                    colsample_bytree=0.8,\n",
    "                    objective='multi:softmax',\n",
    "                    nthread=4,\n",
    "                    num_class=9,\n",
    "                    seed=27)\n",
    "\n",
    "fitXgb(xgb3, training_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see that a high feature importance score is assigned to 'unknown' marital status. This could be due to the fact that there are only 44 customers with 'unknown' marital status, hence to reduce bias, our xgb model assigns more weight to 'unknown' feature.\n",
    "\n",
    "I will tune gamma, subsample and colsample_bytree and learning rate if I had enough computational power. Due to the limited time I have, I only focus on max_depth and reg_alpha (applying regularization to reduce overfitting)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Deploy\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "pickl = {'model': xgb3}\n",
    "pickle.dump(pickl, open('model_file'+'.p','wb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_name = 'model_file.p'\n",
    "with open(file_name,'rb') as pickled:\n",
    "    data = pickle.load(pickled)\n",
    "    model = data['model']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([6])"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#input transformed values to make a prediction in FlaskAPI\n",
    "model.predict(X_test[4,:].reshape(1,-1))"
   ]
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}