Churn Modelling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ba212d37",
   "metadata": {},
   "source": [
    "## Churn Modelling\n",
    "\n",
    "## Content\n",
    "\n",
    "This project data set contains details of a bank's customers and the target variable is a binary variable which is reflecting the fact whether the customer left the bank (closed his account) or he continues to be a customer. So here I am going to build a model that will determine if a customer will continue with bank or not."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f076eb84",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import necessary libraries\n",
    "%matplotlib inline\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5028a897",
   "metadata": {},
   "source": [
    "### Loading and Reading Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "fdab918b",
   "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>RowNumber</th>\n",
       "      <th>CustomerId</th>\n",
       "      <th>Surname</th>\n",
       "      <th>CreditScore</th>\n",
       "      <th>Geography</th>\n",
       "      <th>Gender</th>\n",
       "      <th>Age</th>\n",
       "      <th>Tenure</th>\n",
       "      <th>Balance</th>\n",
       "      <th>NumOfProducts</th>\n",
       "      <th>HasCrCard</th>\n",
       "      <th>IsActiveMember</th>\n",
       "      <th>EstimatedSalary</th>\n",
       "      <th>Exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>15634602</td>\n",
       "      <td>Hargrave</td>\n",
       "      <td>619</td>\n",
       "      <td>France</td>\n",
       "      <td>Female</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>15647311</td>\n",
       "      <td>Hill</td>\n",
       "      <td>608</td>\n",
       "      <td>Spain</td>\n",
       "      <td>Female</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>15619304</td>\n",
       "      <td>Onio</td>\n",
       "      <td>502</td>\n",
       "      <td>France</td>\n",
       "      <td>Female</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>15701354</td>\n",
       "      <td>Boni</td>\n",
       "      <td>699</td>\n",
       "      <td>France</td>\n",
       "      <td>Female</td>\n",
       "      <td>39</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>93826.63</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>15737888</td>\n",
       "      <td>Mitchell</td>\n",
       "      <td>850</td>\n",
       "      <td>Spain</td>\n",
       "      <td>Female</td>\n",
       "      <td>43</td>\n",
       "      <td>2</td>\n",
       "      <td>125510.82</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>79084.10</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   RowNumber  CustomerId   Surname  CreditScore Geography  Gender  Age  \\\n",
       "0          1    15634602  Hargrave          619    France  Female   42   \n",
       "1          2    15647311      Hill          608     Spain  Female   41   \n",
       "2          3    15619304      Onio          502    France  Female   42   \n",
       "3          4    15701354      Boni          699    France  Female   39   \n",
       "4          5    15737888  Mitchell          850     Spain  Female   43   \n",
       "\n",
       "   Tenure    Balance  NumOfProducts  HasCrCard  IsActiveMember  \\\n",
       "0       2       0.00              1          1               1   \n",
       "1       1   83807.86              1          0               1   \n",
       "2       8  159660.80              3          1               0   \n",
       "3       1       0.00              2          0               0   \n",
       "4       2  125510.82              1          1               1   \n",
       "\n",
       "   EstimatedSalary  Exited  \n",
       "0        101348.88       1  \n",
       "1        112542.58       0  \n",
       "2        113931.57       1  \n",
       "3         93826.63       0  \n",
       "4         79084.10       0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 10000 entries, 0 to 9999\n",
      "Data columns (total 14 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   RowNumber        10000 non-null  int64  \n",
      " 1   CustomerId       10000 non-null  int64  \n",
      " 2   Surname          10000 non-null  object \n",
      " 3   CreditScore      10000 non-null  int64  \n",
      " 4   Geography        10000 non-null  object \n",
      " 5   Gender           10000 non-null  object \n",
      " 6   Age              10000 non-null  int64  \n",
      " 7   Tenure           10000 non-null  int64  \n",
      " 8   Balance          10000 non-null  float64\n",
      " 9   NumOfProducts    10000 non-null  int64  \n",
      " 10  HasCrCard        10000 non-null  int64  \n",
      " 11  IsActiveMember   10000 non-null  int64  \n",
      " 12  EstimatedSalary  10000 non-null  float64\n",
      " 13  Exited           10000 non-null  int64  \n",
      "dtypes: float64(2), int64(9), object(3)\n",
      "memory usage: 1.1+ MB\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "None"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = pd.read_csv('Churn_Modelling.csv')\n",
    "display(df.head())\n",
    "print()\n",
    "display(df.info())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "191b063c",
   "metadata": {},
   "source": [
    "### Data-Cleaning: Cleaning strings in column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "1bc6c349",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['rownumber', 'customerid', 'surname', 'creditscore', 'geography',\n",
       "       'gender', 'age', 'tenure', 'balance', 'numofproducts', 'hascrcard',\n",
       "       'isactivemember', 'estimatedsalary', 'exited'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# convert all columns to lower\n",
    "df.columns = df.columns.str.lower()\n",
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "152e630e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# change all \"string\" data to lowercase\n",
    "string_columns = list(df.dtypes[df.dtypes == 'object'].index)\n",
    "for col in string_columns:\n",
    "    df[col] = df[col].str.lower()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a8f55a4",
   "metadata": {},
   "source": [
    "## EDA \n",
    "\n",
    "The first thing I want to do here is look at the data. Identify numerical and categorical variables.\n",
    "See if there is a need to clean the headers and data in a way preferred for formating and processing.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "46616995",
   "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>rownumber</th>\n",
       "      <th>customerid</th>\n",
       "      <th>surname</th>\n",
       "      <th>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>15634602</td>\n",
       "      <td>hargrave</td>\n",
       "      <td>619</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>15647311</td>\n",
       "      <td>hill</td>\n",
       "      <td>608</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>15619304</td>\n",
       "      <td>onio</td>\n",
       "      <td>502</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>15701354</td>\n",
       "      <td>boni</td>\n",
       "      <td>699</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>39</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>93826.63</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>15737888</td>\n",
       "      <td>mitchell</td>\n",
       "      <td>850</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>43</td>\n",
       "      <td>2</td>\n",
       "      <td>125510.82</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>79084.10</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rownumber  customerid   surname  creditscore geography  gender  age  \\\n",
       "0          1    15634602  hargrave          619    france  female   42   \n",
       "1          2    15647311      hill          608     spain  female   41   \n",
       "2          3    15619304      onio          502    france  female   42   \n",
       "3          4    15701354      boni          699    france  female   39   \n",
       "4          5    15737888  mitchell          850     spain  female   43   \n",
       "\n",
       "   tenure    balance  numofproducts  hascrcard  isactivemember  \\\n",
       "0       2       0.00              1          1               1   \n",
       "1       1   83807.86              1          0               1   \n",
       "2       8  159660.80              3          1               0   \n",
       "3       1       0.00              2          0               0   \n",
       "4       2  125510.82              1          1               1   \n",
       "\n",
       "   estimatedsalary  exited  \n",
       "0        101348.88       1  \n",
       "1        112542.58       0  \n",
       "2        113931.57       1  \n",
       "3         93826.63       0  \n",
       "4         79084.10       0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cfd27756",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 14)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bcc77aee",
   "metadata": {},
   "source": [
    "Here, just taking a peek at the data, I can see that the headers are capitalized, the **'RowNumber' and 'CustomerId'** column is not needed since pandas gives me a good indexing system already. And just as a peek, without further exploring, I see that the data is freshly & nicely prepared as the numerical and categorical data have been separated quite nicely."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "897f0672",
   "metadata": {},
   "source": [
    "Okay now, I'll dive deeper into the data to discover the information about this data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "efa43edc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "rownumber          0\n",
       "customerid         0\n",
       "surname            0\n",
       "creditscore        0\n",
       "geography          0\n",
       "gender             0\n",
       "age                0\n",
       "tenure             0\n",
       "balance            0\n",
       "numofproducts      0\n",
       "hascrcard          0\n",
       "isactivemember     0\n",
       "estimatedsalary    0\n",
       "exited             0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "6aa3f119",
   "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>rownumber</th>\n",
       "      <th>customerid</th>\n",
       "      <th>creditscore</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>10000.00000</td>\n",
       "      <td>1.000000e+04</td>\n",
       "      <td>10000.000000</td>\n",
       "      <td>10000.000000</td>\n",
       "      <td>10000.000000</td>\n",
       "      <td>10000.000000</td>\n",
       "      <td>10000.000000</td>\n",
       "      <td>10000.00000</td>\n",
       "      <td>10000.000000</td>\n",
       "      <td>10000.000000</td>\n",
       "      <td>10000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>5000.50000</td>\n",
       "      <td>1.569094e+07</td>\n",
       "      <td>650.528800</td>\n",
       "      <td>38.921800</td>\n",
       "      <td>5.012800</td>\n",
       "      <td>76485.889288</td>\n",
       "      <td>1.530200</td>\n",
       "      <td>0.70550</td>\n",
       "      <td>0.515100</td>\n",
       "      <td>100090.239881</td>\n",
       "      <td>0.203700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2886.89568</td>\n",
       "      <td>7.193619e+04</td>\n",
       "      <td>96.653299</td>\n",
       "      <td>10.487806</td>\n",
       "      <td>2.892174</td>\n",
       "      <td>62397.405202</td>\n",
       "      <td>0.581654</td>\n",
       "      <td>0.45584</td>\n",
       "      <td>0.499797</td>\n",
       "      <td>57510.492818</td>\n",
       "      <td>0.402769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.00000</td>\n",
       "      <td>1.556570e+07</td>\n",
       "      <td>350.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>11.580000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2500.75000</td>\n",
       "      <td>1.562853e+07</td>\n",
       "      <td>584.000000</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>51002.110000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>5000.50000</td>\n",
       "      <td>1.569074e+07</td>\n",
       "      <td>652.000000</td>\n",
       "      <td>37.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>97198.540000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>100193.915000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>7500.25000</td>\n",
       "      <td>1.575323e+07</td>\n",
       "      <td>718.000000</td>\n",
       "      <td>44.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>127644.240000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>149388.247500</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>10000.00000</td>\n",
       "      <td>1.581569e+07</td>\n",
       "      <td>850.000000</td>\n",
       "      <td>92.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>250898.090000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>199992.480000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         rownumber    customerid   creditscore           age        tenure  \\\n",
       "count  10000.00000  1.000000e+04  10000.000000  10000.000000  10000.000000   \n",
       "mean    5000.50000  1.569094e+07    650.528800     38.921800      5.012800   \n",
       "std     2886.89568  7.193619e+04     96.653299     10.487806      2.892174   \n",
       "min        1.00000  1.556570e+07    350.000000     18.000000      0.000000   \n",
       "25%     2500.75000  1.562853e+07    584.000000     32.000000      3.000000   \n",
       "50%     5000.50000  1.569074e+07    652.000000     37.000000      5.000000   \n",
       "75%     7500.25000  1.575323e+07    718.000000     44.000000      7.000000   \n",
       "max    10000.00000  1.581569e+07    850.000000     92.000000     10.000000   \n",
       "\n",
       "             balance  numofproducts    hascrcard  isactivemember  \\\n",
       "count   10000.000000   10000.000000  10000.00000    10000.000000   \n",
       "mean    76485.889288       1.530200      0.70550        0.515100   \n",
       "std     62397.405202       0.581654      0.45584        0.499797   \n",
       "min         0.000000       1.000000      0.00000        0.000000   \n",
       "25%         0.000000       1.000000      0.00000        0.000000   \n",
       "50%     97198.540000       1.000000      1.00000        1.000000   \n",
       "75%    127644.240000       2.000000      1.00000        1.000000   \n",
       "max    250898.090000       4.000000      1.00000        1.000000   \n",
       "\n",
       "       estimatedsalary        exited  \n",
       "count     10000.000000  10000.000000  \n",
       "mean     100090.239881      0.203700  \n",
       "std       57510.492818      0.402769  \n",
       "min          11.580000      0.000000  \n",
       "25%       51002.110000      0.000000  \n",
       "50%      100193.915000      0.000000  \n",
       "75%      149388.247500      0.000000  \n",
       "max      199992.480000      1.000000  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "65855b6f",
   "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>rownumber</th>\n",
       "      <th>customerid</th>\n",
       "      <th>surname</th>\n",
       "      <th>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>15634602</td>\n",
       "      <td>hargrave</td>\n",
       "      <td>619</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>15647311</td>\n",
       "      <td>hill</td>\n",
       "      <td>608</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>15619304</td>\n",
       "      <td>onio</td>\n",
       "      <td>502</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>15701354</td>\n",
       "      <td>boni</td>\n",
       "      <td>699</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>39</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>93826.63</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>15737888</td>\n",
       "      <td>mitchell</td>\n",
       "      <td>850</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>43</td>\n",
       "      <td>2</td>\n",
       "      <td>125510.82</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>79084.10</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>15574012</td>\n",
       "      <td>chu</td>\n",
       "      <td>645</td>\n",
       "      <td>spain</td>\n",
       "      <td>male</td>\n",
       "      <td>44</td>\n",
       "      <td>8</td>\n",
       "      <td>113755.78</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>149756.71</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>15592531</td>\n",
       "      <td>bartlett</td>\n",
       "      <td>822</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>50</td>\n",
       "      <td>7</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10062.80</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>15656148</td>\n",
       "      <td>obinna</td>\n",
       "      <td>376</td>\n",
       "      <td>germany</td>\n",
       "      <td>female</td>\n",
       "      <td>29</td>\n",
       "      <td>4</td>\n",
       "      <td>115046.74</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>119346.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>15792365</td>\n",
       "      <td>he</td>\n",
       "      <td>501</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>44</td>\n",
       "      <td>4</td>\n",
       "      <td>142051.07</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>74940.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>15592389</td>\n",
       "      <td>h?</td>\n",
       "      <td>684</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>27</td>\n",
       "      <td>2</td>\n",
       "      <td>134603.88</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>71725.73</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rownumber  customerid   surname  creditscore geography  gender  age  \\\n",
       "0          1    15634602  hargrave          619    france  female   42   \n",
       "1          2    15647311      hill          608     spain  female   41   \n",
       "2          3    15619304      onio          502    france  female   42   \n",
       "3          4    15701354      boni          699    france  female   39   \n",
       "4          5    15737888  mitchell          850     spain  female   43   \n",
       "5          6    15574012       chu          645     spain    male   44   \n",
       "6          7    15592531  bartlett          822    france    male   50   \n",
       "7          8    15656148    obinna          376   germany  female   29   \n",
       "8          9    15792365        he          501    france    male   44   \n",
       "9         10    15592389        h?          684    france    male   27   \n",
       "\n",
       "   tenure    balance  numofproducts  hascrcard  isactivemember  \\\n",
       "0       2       0.00              1          1               1   \n",
       "1       1   83807.86              1          0               1   \n",
       "2       8  159660.80              3          1               0   \n",
       "3       1       0.00              2          0               0   \n",
       "4       2  125510.82              1          1               1   \n",
       "5       8  113755.78              2          1               0   \n",
       "6       7       0.00              2          1               1   \n",
       "7       4  115046.74              4          1               0   \n",
       "8       4  142051.07              2          0               1   \n",
       "9       2  134603.88              1          1               1   \n",
       "\n",
       "   estimatedsalary  exited  \n",
       "0        101348.88       1  \n",
       "1        112542.58       0  \n",
       "2        113931.57       1  \n",
       "3         93826.63       0  \n",
       "4         79084.10       0  \n",
       "5        149756.71       1  \n",
       "6         10062.80       0  \n",
       "7        119346.88       1  \n",
       "8         74940.50       0  \n",
       "9         71725.73       0  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# look at the data again\n",
    "df.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b7be5074",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of \u001b[1m\"being churn\"\u001b[0m: % 20.37 --> (2037 churn cases out of 10000)\n",
      "Percentage of \u001b[1m\"NOT being churn\"\u001b[0m: % 79.63 --> (7963 NOT churn cases out of 10000)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Let's take a quick look at Target variable\n",
    "y = df['exited']\n",
    "\n",
    "print(f'Percentage of \\033[1m\"being churn\"\\033[0m: % {round(y.value_counts(normalize=True)[1]*100,2)} --> \\\n",
    "({y.value_counts()[1]} churn cases out of {len(df)})\\nPercentage of \\033[1m\"NOT being churn\"\\033[0m: % {round(y.value_counts(normalize=True)[0]*100,2)} --> ({y.value_counts()[0]} NOT churn cases out of {len(df)})')\n",
    "\n",
    "explode = [0, 0.1]\n",
    "\n",
    "df[\"exited\"].value_counts().plot(kind=\"pie\", \n",
    "                                 autopct='%1.1f%%', \n",
    "                                 figsize=(7, 7), \n",
    "                                 explode=explode, \n",
    "                                 wedgeprops={'edgecolor': 'black'}, \n",
    "                                 shadow=True, \n",
    "                                 colors = ['darkorchid', 'turquoise'], \n",
    "                                 startangle=140)\n",
    "plt.title('Churn Distribution');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "56046516",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (8, 7))\n",
    "df.corr()['exited'].sort_values().drop(\"exited\").plot(kind = \"barh\");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8132351b",
   "metadata": {},
   "source": [
    "No much correlation of numerical features and target variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "9d5003da",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZgAAAEGCAYAAABYV4NmAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAAEAAElEQVR4nOydd3hUZdqH7zO9pvfeQwKE0HtvgtgFERv27tr76q67urbdtbtr74iioiIo0nuvgUAa6b1PJtPnfH+cySRDAosK6n4793VxkXlPezOZOc95n/J7BFEU8ePHjx8/fk43st96An78+PHj5/8nfgPjx48fP37OCH4D48ePHz9+zgh+A+PHjx8/fs4IfgPjx48fP37OCIrfegK/F8LCwsSkpKTfehp+/Pjx81/F7t27G0VRDO9rm9/AeEhKSmLXrl2/9TT8+PHj578KQRDKTrTN7yLz48ePHz9nBL+B8ePHjx8/ZwS/gfHjx48fP2cEfwzGjx8/fn5jHA4HlZWVWK3W33oqJ0Sj0RAXF4dSqTzlY/wGxo8fP35+YyorKzEajSQlJSEIwm89nV6IokhTUxOVlZUkJyef8nF+A+Pn90HdIajLA5kCogdBaNpvPSM/fn41rFbr79a4AAiCQGhoKA0NDT/puDMWgxEEIV4QhLWCIOQLgnBIEIQ/eMb/JAhClSAI+zz/Zvc45iFBEIoEQTgqCMLMHuNDBUE46Nn2kuD5KwiCoBYEYbFnfLsgCEk9jrlKEIRCz7+rztTv6ec0ULkL3poGX94AS66Bd2dBff5vPSs/fn5Vfq/GpYufM78zGeR3AveIopgFjAJuFQQh27Ptn6Io5nr+LQfwbJsP9AfOAl4TBEHu2f914AYg3fPvLM/4tUCLKIppwD+BZzznCgEeB0YCI4DHBUEIPoO/q5+fi8sJ214HR2f3WEc9FK787ebkx4+f08IZMzCiKNaIorjH87MJyAdiT3LIecCnoijaRFE8BhQBIwRBiAYCRFHcKkrNaz4Azu9xzPuen5cAUz2rm5nAj6IoNoui2AL8SLdR8vN7wu2EpqLe482lv/pU/Pj5b6a6upqLL74YgH379rF8+fKffI5Jkyad1oLzXyVN2eO6Ggxs9wzdJgjCAUEQ3umxsogFKnocVukZi/X8fPy4zzGiKDqBNiD0JOc6fl43CIKwSxCEXT/Vt+jnNKHUwNCFvccz/c8Dfvz8FGJiYliyZAnw8w3M6eaMGxhBEAzAF8Cdoii2I7m7UoFcoAb4e9eufRwunmT85x7TPSCKb4iiOEwUxWHh4X1K6fj5NcicDZMfBbUR9GFwzkuQMPq3npUfP78LPvroI0aMGEFubi433ngj27dvJycnB6vVitlspn///uTl5VFaWsqAAQOw2+089thjLF68mNzcXBYvXozZbOaaa65h+PDhDB48mK+//hoAi8XC/PnzycnJ4ZJLLsFisZzWuZ/RLDJBEJRIxuVjURS/BBBFsa7H9jeBZZ6XlUB8j8PjgGrPeFwf4z2PqRQEQQEEAs2e8UnHHbPudPxOfs4AxkiYcC8MXgCCQnrtx48f8vPzWbx4MZs3b0apVHLLLbdw9OhRzj33XB599FEsFguXX345AwYMoLS0FACVSsUTTzzBrl27eOWVVwB4+OGHmTJlCu+88w6tra2MGDGCadOm8e9//xudTseBAwc4cOAAQ4YMOa3zP2MGxhMLeRvIF0XxHz3Go0VRrPG8vADI8/z8DfCJIAj/AGKQgvk7RFF0CYJgEgRhFJKL7Urg5R7HXAVsBS4G1oiiKAqC8APwVA/32wzgoTP1u/o5DQgCBJwsROfHz/8eq1evZvfu3QwfPhyQVhwRERE89thjDB8+HI1Gw0svvfQfz7Ny5Uq++eYbnn/+eUBKiy4vL2fDhg3ccccdAOTk5JCTk3Na538mVzBjgSuAg4Ig7POMPQxcKghCLpLLqhS4EUAUxUOCIHwGHEbKQLtVFEWX57ibgfcALbDC8w8kA/ahIAhFSCuX+Z5zNQuC8Bdgp2e/J0RRbD4jv6UfP378nCFEUeSqq67ib3/7m894bW0tHR0dOBwOrFYrer3+P57niy++IDMzs9e2M5kefSazyDaJoiiIopjTMyVZFMUrRFEc6Bk/t8dqBlEUnxRFMVUUxUxRFFf0GN8liuIAz7bbPNlkiKJoFUVxriiKaaIojhBFsaTHMe94xtNEUXz3TP2efvz48XOmmDp1KkuWLKG+vh6A5uZmysrKuOGGG/jLX/7CZZddxgMPPNDrOKPRiMlk8r6eOXMmL7/8Mp5bJ3v37gVgwoQJfPzxxwDk5eVx4MCB0zp/fyW/Hz//zVTvg+I1YDdD+nSIHQryU9eK8vP7Jjs7m7/+9a/MmDEDt9uNUqnkvPPOQ6FQsGDBAlwuF2PGjGHNmjWkpKR4j5s8eTJPP/00ubm5PPTQQ/zxj3/kzjvvJCcnB1EUSUpKYtmyZdx8881cffXV5OTkkJuby4gRI07r/IUui/a/zrBhw0R/wzE//1VU74V3Z3cXqQoyuPJrSJ7w287Lz08mPz+frKys33oa/5G+5ikIwm5RFIf1tb9frt+Pn/9Wjn7vq4AgumHzS+B0/HZz8uOnB34D48fPfys2U+8xayt4c2P8+Plt8RsYP37+W+k3u/fYqFsldQQ/fn4H+A2MHz//rcQOh8u/hMRxEJ0LF78DqVN+61n58ePFn0Xmx89/K0o1pE2FxLGSW0x18loIP35+bfwGxo+f/3b8LjE/v1P8LjI/fvz48cP3339PZmYmaWlpPP3006flnH4D48ePHz//47hcLm699VZWrFjB4cOHWbRoEYcPH/7F5/W7yPz48ePnv4yle6t47oejVLdaiAnSct/MTM4f/PPFYnfs2EFaWppXDWD+/Pl8/fXXZGdn/4cjT45/BePHjx8//0Us3VvFQ18epKrVgghUtVp46MuDLN1b9bPPWVVVRXx8d7eUuLg4qqp+/vm68BsYP/9/sHVAe7W/kt3P/2ue++EoFodvMa3F4eK5H47+7HP2JRl2OlSW/S4yP/8VNJlsHK0zYXG4SIswkBh6XEpuxU5Y/SeoPQiZZ8P4uyEs/TeZqx8/Z5Lq1r67Tp5o/FSIi4ujoqK7y3xlZSUxMTE/+3xd+A2Mn9891a0W7v18P1uKmwAI1Cr56NoRDIwLknZoLIKPLuiWTtn/CbSVw/xFoAn4bSbtx88ZIiZIS1UfxiQmSPuzzzl8+HAKCws5duwYsbGxfPrpp3zyySe/ZJqA30Xm57+A3WXNXuMC0GZx8NLqImxdboKmwt66XKWboK0CP37+v3HfzEy0SrnPmFYp576ZvZuJnSoKhYJXXnmFmTNnkpWVxbx58+jfv/8vnap/BePn909pY2evsQNVrXTYnKiVclAZeh+k1ILCX4Do5/8fXdlipzOLDGD27NnMnt2Hvt0vwG9g/PzuGRAb2Gvs7IExBOlU0ouILEg/Cwq/795h8qMQnPwrzfB3hChCaxk4rRAY75eP+X/K+YNjf7FB+TXwGxg/v3uGJARz/8xMXlxdiM3pZnJmOFeMTkQu82S56MPgnBekBlxtlRCeCTGDQfY/5gG2mWD/p7DqT2DvgIzZMPOvEJr6W8/Mz/8ofgPzX05Nm4VNhY2sP9rAsKRgJvWLIOn4DKv/cgJ1Sm6cmMqsgdHYnC7ig3Xo1cd9dAOipX//y1TtgeX3dr8uWC69J7OeBbn/q+7n18f/qfsvxmJ38veVBSzZXQnAsoM1DNpbxdsLhxNmUP/GszsxoiiSV9VGncmGWi4QH6wjKbyPOEoP5DKB5LBTN5xmm5OjdSZqWi3EBmnJiDKiU/2Cj7vLCRXbYde74LTAsKslFWPlz8/cOe3UHeo9duhLmHg/GKN+/fn4+Z/Hb2D+iylt6vQaly72V7ZRVN/x2xoYUy0c2wgl6yB2iNSjJKQ7HrLjWDMvrC5ka3ETWqWca8YlcU5ONP2ie8dafg52p4tPtpfz5PJ879gf52Rx5agklIqf6Tar2gXvz5HaEgMcWQaXLYH06adhxqeJwD588hH9+06C8OPnV+B/zEn9GyOKULUXNr8Im16QXBp9VNCeKu4THOtyn8I5re1S5fvpxmmFDc/Dl9fBvo/gu7vhi+ugox6AJpOFT3eWs9WTdmxxuHh1bTEFdadvLiWNZp7+/ojP2N+WH6Gk8RdcI/+bbuPSxdZXf33VAKcdGgqk2h/XcdeOHQpJE7pfq/Qw7TFQ+w2Mn98G/wrm16RqN7w3G5w26bVcBQuXQ/zwn3W6pBA9M7IiWZlf5x1LDdeTFnGSG0pnMxxdAVteBKUBJj0AyRN/ek8Rl1NyFamNvuPNx2DX275jVbug4SgYImg0O1h3tKHX6Y41mn/a9U9Ci9ney8g63SItnb/EGPTxLCbI4TTIaZwybZWw8R+w5z0QZFJ75FG3gDFS2h4YBxe/LbnK7GYIy4DwjF9vfn7+q7nmmmtYtmwZERER5OXlnZZz+lcwvyb7Puk2LgAuO+z98GefTq9R8Mdzsnl4dj8Gxwdx6+Q0/n3FUCIDTmIsClfC17dIN/zq3fDJPKjc+dMuXHMAvv0DvD0dNvwdWsq6t7ld0pO+MQpL2hzEkDRp3PO0HROgIT3C2OuUsX1VIZsbvSufn0JcsI5ArdJnLFinJC74F8RLss8F2XHPY6Nv+XWD50dXSMbb7ZLez80vwLENvvsYIiB1MmTN8RsXPz+JhQsX8v333//nHX8C/hXMr0lnU++xn3ED7Ul8iI6LhsQxPj2cEL3q5MbF3gnbXu89XvgDLREjcLkhzNg7dlNUb2JjYSOVLRbGpwQwbN8/MRR8KW1c8wQ0HIFzX5IC3sFJlM96n29rQ1heJjAyEi4f3UmKJhDyv8WoDeaeaelc80EbZrtUiT8iKYQhicHdF7S2SzGOdX+TbqTj7oaBF4Mu5JTfkzevHMq9n++nvNlCYqiO5+cOIi5Yd0rH90nsUFj4nfSQYO+EIZdD/Kiff76fissFBxb3Hj+6HHLm/nrz8PP74MBnsPoJaVUbGAdTH4Oceb/olBMmTKC0tPT0zM+D38D8muReBoeX+o4NXfiLTrmvooV7PttPcYOZyAA1z108iPHpYX0rocoUUs3IcXQqArnsze00ddq5fUoaZ+fEEOwpYixtNHPF2zuoabMC8PYmeGbqdVxStUFaYQDkfQ4T7oHwfpjR8NcjsazMr/ccL+fyRDu8O9O7ihk59Bq+vPERjjTY0Knk5MYHEW7sYRiPbYClN3e/XnGfFEfIXXDK78uI5FC+vGUszWY7oXoVob806UEmh4RR0r+edDZLq1Jj1Jl1l8nlED+i92ozetCZu6af3ycHPoNv7wCHR4+srUJ6Db/YyJxu/C6yM0in3cn6gnru+GQPf1yaR6EuB3Heh9JNISoH5n0ASWN/9vnr263c8tEeihuk+EVdu43rP9hFScMJ4hkKFYy9U7pZdqE2sk02hMO1JurabTy69BCbChu9mw9Vt3mNSxfPbe+kPmuh93XDgOv4sVbHGxuKWV9Qj9CjwPGqQXqStj7qG5De/Q6ZtjzOy41lenaUr3EBKbX2ePZ8CLV5ULQKWstP+r50EWZQkxFp/OXGpS8cNjjyHbw1DV4fDWufgrZf3j/jpOReDsYetT7BKdDv7DN7TT+/P1Y/0W1cunBYpPHfGf4VzBlkS3ET172/y/t68c4Kvrh5AgMXTgHE3gHyn0h1m5Xq427+NqebiuZOUk8U6E8YDVf/AKUbQanjgHIgd37rm121aEc5c3KiEQQBh6t3Rlqn3YVTIbmbzMmzeM4xj88+PujdfsHgWIYkBLOnvIWUADeytrJe56Cjd6Dfi673KgtDJOx4A/a8D/pwuPyLkz69250uiuo7qG2zEh2oJTXCgOrnpij3Rc1e+LTHimrDsyBXSjUnZ4rIbLjmB6jPl1QKIrIl94if/y3aKn/a+G+IfwVzhui0O3l1bZHPmN3lZkNBo+Tu+YXGBSBQq+ilqgoQYlD5Drhd0tO1uUkKSscPl/qljLqJFfWhtFudPrsnhuq8LrZ+0UY0St+PyfVjE4mOTYLh11M08gk+2+drLJbuq2JChmQk1lYK2BMm0IuTyZckTwRNj5oYpRb6zYaafZK+lrlBSi7omTDRA6fLzVe7ypjz8iaueX8XZ7+8ka/3lJ1a+vapUrG999iud6QMroqdYKrrvf10EJwImWdB+gy/cflf5UR/99/h5+GMGRhBEOIFQVgrCEK+IAiHBEH4g2c8RBCEHwVBKPT8H9zjmIcEQSgSBOGoIAgze4wPFQThoGfbS4Ln7icIgloQhMWe8e2CICT1OOYqzzUKBUG46kz9nifD3ccNzfUL6l6OJylUz1/O7+/j+r9jappvmnJrOfz4J3h1BLw9TcpEctq9m88eFI2xh+yKViln/vAE7+t+UQEsun4UswdEkRlp5M/n9mfB6GSEgRfB2c9jUYX2mpcoQmakkXnD4hiTlYhzxlMQO0zaqAmCC9+UCgBPRGQWTHoIJtwHE+6FyY9iUkfyesyTPBn3L7ZMWUJFzAx+PNLAkt0V7Ctvwe7srlE5VtfCo98eoevtd4vwyNdHOFbbckrv6ynRl1KzMQrWPS29z29Pg+p9//k8pjqpNqq59PTNzc//b6Y+1ltBQqmVxn8Bl156KaNHj+bo0aPExcXx9ttv/+eD/gNn0kXmBO4RRXGPIAhGYLcgCD8CC4HVoig+LQjCg8CDwAOCIGQD84H+QAywShCEDFEUXcDrwA3ANmA5cBawArgWaBFFMU0QhPnAM8AlgiCEAI8DwwDRc+1vRFE8jXeYk6NTKbh5Uio3fbTHO6aQCd4n+9OBIAickxNDZmQAFS2dRAZoyIzsIYnidsPOt2HrS9Lr5g749FLJRZYwEoABMYEsuXkMeVVtuESRgbGBZEX7NukanBDMi/MHY3e5vRpgFruLDpuDxBAdsYEaqnq46lLD9YxIDmHWwB7xgsu/BFMVqIwQFM/x2JwuKpstyGUC8SHJyLMvgIZ8EEWaFBFM/7CaZrMUx+kYEU9BrY7d5fs97wO8ccVQpmdLciiNLW29XHt2l5vG1jbSYk4tEw2QkhgaC6XkiLB00AZ1bwuKh6CE7niQTA4D58Kqx6XXreWw/H64fMmJm55V7oLPF0pBWrUR5rwAWedKsTI/fk5EVyD/NGeRLVq06DRMzpczZmBEUawBajw/mwRByAdigfOASZ7d3gfWAQ94xj8VRdEGHBMEoQgYIQhCKRAgiuJWAEEQPgDORzIw5wF/8pxrCfCKZ3UzE/hRFMVmzzE/Ihml0/8OnoTx6eG8fdUwPtpeTrBWyYKRCeTEBp30mGONZnYca6LeZGNYUjC5cUFoT6KhpVbKGRgXyMC4bpdSh9VBXbuNSFkLhj3v+x4girhqD5Iny0Qpk5EcpiMzykhm1Ilddi1mG0dqTZisTpLD9FjsTp5fWcDhmnbOyYnhz+cN4MNtZRysamNwQhDTsyJRyI7LqNIGSv960GiyoVbIaLc6eWlNAZ/vqkQhk3HzpFSuHJNIaOpkAFZsKfEaF4CoQC2LdnQ3ExNFeOzrQwyODybMqCbaIKBXyb1p0ABGtYJow0/I8mosgi+vh2rPA0L6DDj775JRATDGwICLQKGWEhgismHNX32TGSq3S0aqLwPT0SApHHQ1RbOZpOvdmAlRA099nn7+N8mZ97vLGOuLXyXI73FdDQa2A5Ee44MoijWCIER4dotFWqF0UekZc3h+Pn6865gKz7mcgiC0AaE9x/s4pue8bkBaGZGQkHD85l+MXq1galYkU7MiT2n/siYzV7y9ncqW7gyRly8dzDmDTr03dkGdiceW5rHtWDPXDw/mIWMsMovvwq3GruX8VzcjinDJ8Hjunp5xwvqZBpOVx745xIqDtQCoFTIent2PbSXN2F1u3t1SSkG9icxII7FBWvJr2nnwy4P0izKSm9D3k3htm5Uluyv4YGsZaREGRqeG0trp5PYp6QBsP9ZEZpSR2Z4VkNXle7zT5T7+lNS1W7F4OlwmhQfx2tlh3PlDMy2dDkL0Kl6cGUxieNApv48c+KzbuIBUoFq8DoZeKb2OHiTFgPZ8BLYWiBwILcd8zxExALTB9Imppvf+olsqWvUbGD//TzjjQX5BEAzAF8Cdoii2n2zXPsbEk4z/3GO6B0TxDVEUh4miOCw8PPwkU/t1yKtq8zEuAH9bnk9TR9/B7OPpsDp4/BvJuAB8sLeNokH3+aQlu8IyWd0W65VAW7yzgu0lfRSAeufU7jUuIGWpfby9nJn9u43m5qImDBoFn+woZ29FKwCy41cwPViyu4LnVxZQb7KhUcpRyWU0dth4cXUhL64uJFCrotHU/TsPSwz2WRGpFXKOP/35ubFEyDug7jAAEzPCWXaRga/PkfPtRXrGp4f7urhOhsMKRSt7j5dt6v5ZEKRkjfB0CE6SkhKm/6V7uyYQ5vwddCcwMLoQ0PWOX2E4tYcRP///EE9jfPZM8HPmd0ZXMIIgKJGMy8eiKHYVN9QJghDtWb1EA12l7JVAT+d8HFDtGY/rY7znMZWCICiAQKDZMz7puGPWnaZf64zR9QTekzaLwyeA3WCysq+ijbImM2kRBgbFBRGsl1YK9SYbW4ubGRwfxMiUEKpbrdyxvYUP531HeGcxbrWBb+sj+PNK3xXNztIWzs3tuzteYx/Grai+g2k9VmUapQyHs/vDN2tAFCknkNZvMFl5f2t32rLN4aK23crusu45/XColjGp3TffgXFBLLlpNHlVbVicbrKjjbxxxTCeWHaI6lYr5w+O5bahWtRvjpWUEaIGwnmvEtt/jM+y1WRx0NBhI0Cj7FOxwItSI3XIrN7rO540vvvn+nx472zoWh1ueRnmfQg3rJfGgpN8FKR7ERgH570Kn10pSQYBTHxQ6s7p538OjUZDU1MToaGhfRdJ/8aIokhTUxMazU/TLDxjBsYTC3kbyBdF8R89Nn0DXAU87fn/6x7jnwiC8A+kIH86sEMURZcgCCZBEEYhudiuBF4+7lxbgYuBNaIoioIg/AA81SNDbQbw0Bn6VU8b/aICUMp9a0+uHpvkdV+ZrA6eXnGEL/Z0F/TdMCGFe2ZkoFbIMagVPDgrk91lrby96RgJIXouG5lMjT6A8OhEUGjYV1yD+7hch5w439hIT5L6MBQTM8Ipqu9gQGwARfUd3DUtg8gANecOimZsWjjjUkMpbjBT3txJhFFNVrSRAK1kBNUKOWF6FQ2eFYpGJWd7SXOva+RVtXl/brc4+DG/jn+tL8HlFhmWFMyzF+Xw1S1jsTrchDuqUf1rBLg96da1BxG/vAFh4XLQS4bqcHUbf1x6iN3lLcSHaPnbBTmMTTvJlzlnLpSs6U5HzpwDKZO6t1fu7DYuXax9Cq75HmJyT/h+IopSt0mlHtJnwo0bJLeYIULqxOlvcfw/SVxcHJWVlTQ0nKQ+7DdGo9EQF/fTUqHP5ApmLHAFcFAQhH2esYeRDMtngiBcC5QDcwFEUTwkCMJnwGGkDLRbPRlkADcD7wFapOD+Cs/428CHnoSAZqQsNERRbBYE4S9Al67GE10B/98z2dEBfHTtSP65qoDKFguXjUzg/NxYr7upqL6DL/ZUIQhgUCnosDt5a2MJFw6OpV90AEaNkp3HWlhztJ4wg5rKlk6C3M101Lfy/E4zEToZ5/VPZkNBAyUe9eJJGWEMigtk3dF63KIoxVJ6aHZlh8l5+uxE/rqqig6bk1GJBm6blML6oiZcosglwxKYmBFGQqie8wfHIYoiS3ZXct+SA95zXD8+mT9MTcfmdKNTKXhgVj+ufm8nogjF9R0MTwrhcI2v97SnNtneihZeXVvsfb2rtIV3N5fy8Ox+qBUyZPXF3cbFg9BwBNqrQR9Ks9nOXZ/t52itCYCKZgvXvr+TZbePIz2yO7nB5nRR3WpFJZcRG5oKl34KTcVSUWNouk+w3m239PYv2zt6zcOHxkLY84EUz0mbBkOuklYsZ3jVIooiDpcblaJ3zZSf3wdKpZLk5JOseP9LOZNZZJvoOxYCMPUExzwJPNnH+C5gQB/jVjwGqo9t7wDvnOp8fw/IZAIjU0J5Z+FwrA43IXrfILnF4eL6wXoujKghqLOMJl0KH1VGeF1rNW0W7C4390zPpLy5k5ggDWFhWi7/8IC3JiRqj4X3Lu9Ps0OBSi5Dr5Zzw4e7KW3qBCA6UMN7Vw8nM8pzM63ay/ydtzB2zE10KgJR6+ws/Hw/pc1SWvLq/HquHpPEw7OzUCpklDV18vg3vp0V39x4jP4xATy94iipEXrum5nJkpvGcKCylQCNktQIPXvKW7ySN6NTQxmX3p3OfajK1/hEBaoZlx7K7Yv2UtbUydzB8Zyb+wei9r3YvZMmiA6rHcPap6iJPNdrXLqwOd2UN3d6DUx5k5l/rirg633VnpVgFufmxmA4QSsFU1gugTKFj0ExDb0FYx9ab4BU5PrlDd2JAw1HpIZsVywFw+mL/x2obGXxzgoqWzpZMDKRqAANi3dWcKi6jbnD4pmeFUFk4O+oC6ef/9f4pWJ+hyhlMsQ+/jL9gkQGut7HuF4KZ0UDD+XeBEFDANCr5AyICeT5lUe9x6SG67l8VCIfeOIete1WjlQ1c/7obABeW1vkNS4ANW1Wlu6r5oGzJAMjtpRBexXxW/8IwLrxi7zGpYsPt5Uxf0Q8mVEBtFsddNp9Y0nBOiX91Q0sn9GCRSHy0sbDXDMlh6vHSk9s1a0WrhufQm2bFblMIC5YQ6PJxhvri7E73UzqF0F6hIHCeknS5pFZWdz56X6vYX3qhw6aR5/LfWErkZuqwN6Be9JD5NkjeKFgEgsDNb3SlgECdUo67U4UgsDH28v5aq8U2mu3Onn4q4MkhekYk9q3wfi+OYrsqR+SXvg2Gkst5elXsKihP9eabH3Hd5qLfbPSAOryoKnI18C4XVIcSanrldb9nzhc3caCN7czJCGIIJ2KgloT9y85QJtFSp3eX9lGeXMK98/MRCH3i3j4OfP4DcwZprTJTGGdCZVcTr9oY5/pwPUmK06XSFSAhn0Vrfx7QwllTWYWjEzgrP5RRHiOCekshS6ZfA+B+9+AUVdCQH8cbpEPtpb6bC9uMHPucWnOFlf3wnKfJ+urJ7tKm2k227DYXej1vr3cRQSOT8hziyJmm3Tzjg3Skhyq45jHaKkVMj6a7ib963OkWg/gqWHXUdgRBUirh+8O1Pi0Nwa4e3oGi3ZU4HSLfLa7kifPH8AjS/MQRWjutPdKiHh3ZwOXX/kaceXfQmgasvItBOqGsO1YC7kRcv40OZT7fuhujXBFjpEARyNz/3WIecPi+HJvb6HKg5VtJzQwCrmcc78SGZv0B0L0MtZuNJMSLnDrcbI6dDaDtQ0UWinz7PhMHGsr/PAoRPSD6FzYvwj2fQxBiTD9CSmxQHZqxqCgvoO7p2fw3cEaKlraGJIY5DUuXby3uZTLRyWQEPL7ivWUN3XS3GknMkBNtH+F9f8Gv4E5g+RVtXH529tp9XRS7B8TwGuXDSExVPpym21Ovj9Uy9PLj9Bhc3LVmETarQ5+OCSlBT/29SHaLQ5unZwmBaOdlt4XEd3ecZPVgdXZu0akpzyNUi4wID6E/Jo2FDKBGf0jWXnYVzdrWlYkC9/ZSavFwb8vzkQYeR/NijDMgoHkYCXxQTIqWrvnMicnGq1Sxv6KVkINKh6ancU/fizgSK2JW0eGkL3rNq9xAZDveovUlKm8vNpNeqSepft639yP1ppICNFR0mhGFGHl4Tr+dsEAqtusKPt4+taq5CgKf4CdnnySkTehUCoRBBhkaGNSxdvEXfo4pSYZETrIqf6cytJ0DlUbWZ1fT2KIzpt40EVKuJ6aVgsymdDrwWBIYhCxQVo2HZN+L7lM4N6ZmRg0PRqdlW6CZXdD41EpQeCCN6Q2BF1utYyZsPs9KPheUtcO7wcHP5O2WVrgowvhujUQc2qS/DqlnHuW53s111o7HcxK13FZcic6t4kKInl+txuZwwJ1ZZK0zSn22DlTuNwiqw7Xcu/nBzDZnIQb1bx86WBGpfSRwn2KOFxu9pS38JUnGeaCIZL4al+fGz9nFr+BOUM4nG7e2ljiNS4Ah6rb2VbS5DUwe8ulXi5d/Gt9CdeOSybMoKKxQ0pdfWNDCRcNjZOe6kJSISBGCl53ETkAgiVXU4BKxsU5oSze1y23r1fJGRhtJDXcQGygilsnp7GtrJ1Pdx5Bq5Jx57R0rh6TxIfbynCLInOHxjMgyMa5qZtR2lqx2i9gsW4+f/+xEJvTTf9oFU9dlMWKvFoaTDYyIgwMSQjm0rckQxpmUHHL5FRy4gKZlh3J3GQzwu6CXu+PvaOZylYLKoVAZqSRQ9W+cZboIA0bC7szahwuN0t2V7GrrIWXL80lJlDjoyR9/0gdUQd7hNx2v4sm9VJunphKcqyDzZo7uOWzMuyeIs0Hxk5muFEE2tlU1Mgf52STV92G1SFtv2F8CkdrTdy9eD9KhYx7ZmRwTk60NxsuOczAR9eO5EBlGx02B9kxAQyI6eHSaiyEj+eCw+N+LFknVfWf96oU5I8dJrnHutpLp02DLS/5vklupxSrOUUDU9Vq8RH0TNE7WBj4FUEbJDWHIQoNY8/9kNDlV0k1PeFZcP5rEDsEu9P1myQBFDd0cNuivd7MyQaTjds/2cvXt40lpq8up6fA7rIWLn1zm3ex+NmuChZdP4qRv8Bo+fl5+A3MGcLicHGwR6ptF4V13dL4W/tIz117pJ5RKaEsO1ADgEYpx+ly43KLyANjYcFnsOE5KNsCadNh7B+8qbjhrjpuD9xE7OgRfFHgpF+InFuyLMQ495EZFYkMyKsx8fcfCxidEobN6eLGD/fwzlXDuGJ0Im4R5G2lJH9xluTWAXYEj+Wp77tv9IdqTBTUmegfHUC5upPMaCOf7arwGtLGDjsvry7ivQWZyCwtKBRKxJjBCD1qStwB8ezVjCIywIXVKTKzfxTrC+pp8sjBpIbpUSvkuEUI0Chotzq5dHgCf19ZQIBWQWWLhRfnD2ZLSRNVLZ3MSNUxcv+jvt1BnTYOlDfy2jobqqlpvL/V7DUuAM9sbuOludnAYdwivL6umHcXDqeq1UKAWkmb1U5li5VrxiUjCFDR3MnO0hYfVYakMH2fadyAZDwcnb5jdXnQXAJVu6WU5C7jApKrTB8GplrfY1QnaLvQBwE9V09ArK2QoLweUkFOK6Fr70NIGicZmIZ8SgsP8W2+nh+PNDAxI5wLBseSEn7q1/xZdDRI70VnI6GaBFJC1Bxt6H5YaOiwUd9u9TEwLZ12atusGDWK/9iZdNGOch9PpFuExbsq/AbmN8BvYM4QAVol5w6K4Z+rCn3Gey79YwJ7x2OiAjU+xY0LRiZwsKqduxbv56IhcVwY60YdkgoBsb1uPjYXxO15jju0IVyVNBVNewnqlZuoO/s9lh+s5c6p6RTXd3DLpDR+PFyHViXnoVn92FXawsRMSbHHfGQ91QNuIl83FKtbTps8BuiWNJk3LJ4v91T5rDgWjkkiJUxPSaMZpVzg9ilp7CyuJ1xmYo+o44LJTxLw3c3QWgYqA5umLOGaTwtwep62U8P1vL1wOK2dDmSCZFRMNhcC4HS7GZ0SyoikYPrHBGBzukkJ16NSyBmeFIzL6UDeXgE/9FQZAmvCJJZXqAAbThFaOn1jEQAOR/fYndPSeOzrPArrzYxPD2PWgGje3XzM28ogMkBN+ol67PRFXxIxSp1UVNlSCmVbIW0qFK2WtuV9CePuhFV/6t4/Kgeic075krkJQYQb1DR4Pj9GZ29tV6GlVGo/DbRlLeDBw/FsK5faShyobGPtkXrev2bEL2rSdqi6jQ0FjXTanUzMCCcnLqi7F4+5Eb67B/Kl8rdQQeD1ya9z4bpQ70OKUa0gQKv0Od+9n+8nv8ZEoFbJkxcMYEZ21An7+/TVlqEveSE/Zx6/gTmDXDgkjmONZr7eX41KLuPWyWkM7VHfMTo1lLggDZWt0tObWiHj8lEJ7C5rJTpQS3qEgQ6bg9fWFZFX1c60OBfKzTdCe4+YxdHlWC79imq7DrkiEsvgO4jY8QyBhz4AwBmYSJU6hRB9PSarneQwPU8uP+I9fGdpMy/Oz8VkdUgB9NCh3Lyrk/x6C+Dkvpm+qdKxQRo+21XhM7ZoRznXjkvmtXXFXDg4lmHqCrL2P4yyIQ8CYikKeAbnhZ+jtDVi00Xz3JdVXuMCMDI5lE93VLBkdyUKucD141Moru9geZ70NP/2pmO8umAI9y05QIfNybxhcdwyPJD4fS8gby+HgfPpmPcZyo3Pom44gDXjHFbozue7VVJ8pLnDTmyQlqoecSO5TCA9CC4bmcCwxGA2HK2lutXKWZlBpEXp2V7S5NMnp67dxrFG3xWJyy1S1mTG5nQTH6LFoO6xgojIgsFXwN4Pu8dG3yppnAGUrIVx90iB/cNLJZdZ2nSpIVxtnrSaiRnSp/L0iUgNN7DohpFsL2mmscNGZGIwbDpup9ghkgoBcCxsEtv2+nY/zatup7ih42cbmEPVbcz711Zvxt6ra4v48NqRjE3zJEvUHvQaFwBEkZTtf+T63Hd4bosDpVzg5kmp3rT6tk47Dyw5QH6N9Ldsszi4Y9Fevr1tHP1j+86yu3REgtcD0EXPFhR+fj38BuYMYlArOGtAFFnRAchlAgNiAtCpuv3cKeEGPr5+FIeq27E5XYQZ1BTUmlhzpA6nW6TT7iQj0kiepw5kkK4RWftxAfG6PPLy9jF3mYPRKcEsHHw+SZOSiKrbQEdgOsWhk2lwhjJ7gIz0iAAW7fRtNzwkIZj2Tidz/7UVl1tk4ZhEgo21UC/djA9XtzNrQBQr8o5z3fTA5nSjkEuZaXPSlAxYfyeylhJpY3sVATVbWeHux1f7BdIiOjg/N5biBjP9YwKwO91oVXLe3iStkpxukZfXFHH39Ax+OFyHyy3icIl8vruS+BAteVXtLNpRQZBGzuDoGyjXWMkW9KzcZ+KY7Q8MSBbQGyL4cHu3PurSfVU8enYWL64upK7dhl4l5+mpgcQFabh8VBRy3JRXufhuYiUJRR/RKZ/J5c29m6RVtVr4Pq+GFrODAXEBbCps5IVVUmxqfFoYfz6vf7d7SRMI0/4sKS6baiEkBZxWqSsnSPL8cUMhYxaMuUOq7Fd4DFTCKEBKAtld0MC3B6oJN6iZNSCKgXFBJ/w7AKRFGEmL8BSPOuLh3Ffg+wfAbpaSCMbdDZ9L7ZHkQt/aUvKT6Mj9J9YfbfBJB3eL8Nq6YoYnBUsxnuPVDwDMDUxN0WPRSCrci3aWe+ug6kw28o6Lz7lFKGvuPKGBGZYYzMfXjeTDraUgCFwxKtHnwc7Pr4ffwJxBVh+p497PuyvaBQHeXTjcK4MyLj2MnNggb9C/usXC6+uKGZcWTqBOSUyghoe/yvMebxf6eKoUBBqskqtga0kLYQYN07Im8oU5l4QgHUalkns+3QdIKyZ9j+ZiMgGmZ0Xy6Nfd13hk6SHunZHJztJmHC6R8uZOrhqdyOTMCGQyCNGpCNQqfdJfJ2eGMyE9HJ1SwQB9fbdxAVyhmbwvns2r30iB/l1lLVw+KoFnL85hc1EjoXoVg+KD+HpflTexAaQsskRPFhlIcjGGHnNfur+WhrQwluyWEh7+fl4qyWF6EgKVfHuomZsnpWC1uzHbnQRqVZwfXs/UEfnUEEqQYMYam8bCJeUcqDyIQa3goZlpRB36AlnNXgwtJcwdMoF9vgs10iL03PrJXlxukfvPyuTZ77vrjTYWNfLu5mM8fk7/7hoTfSh4Wg54uWE9dNRJ0jAhKdLYCUQ41xc0cMvH3bUz720pZclNY8iOOUF/GaCwzsTm4kYaTDbGpYUxeMB8NNGDpJhaSDKoA2DhcjDVkRzSn9k1bSzvIWY6Pi2U1F8Qg2m39nZFtlvsON0iKoDQNEl81d1thGwJE3hph5nlBdI8hiYEkxgixVkCNAoft18XoccVIfdErZQzNi2MMam/T12v/yX8BuYMYbY5eWujrxy7KMKaI/XYnW46bJLMy9sLR3ifrmKCtTx/8SDyqttotzhIizAQH6ylwqOwvKRMx8h+F6I50l0L0zzwet4/2r0qWn2knitGJxIbrEWrlPPM9931JQ3tNm6YmMLWkiap62SUkd3lvZ8otxQ3MiguiF1lLSwYGU+71cnao/XUtFmZnhXJcxfnsHhnBQX1JsakhjE6JYToQDU3TEjBWdki6WnZJcNQnXklb23ofgINN6joHxPIbZ90B/2DdUqevGCgz800KlDDxqLu5IJp2ZFY7S5GJEsinpUtndR6sshigjSkhWn5YEcN7++wMicrgMGRSq5ddIR6k43IADVZU3SM3v13Ii0tmFPP5qaiDA5USgkXHTYnj3xzhOT5zzGmaixY25hm+oaacZfy5rYalHIZ141LZndZKy63iF4l75XSDLAir5bbp6YTYTyJIGBI8slFMD2YLA5eXO2bfddpd7GztPmEBqa4voNL39zmNdSvri3mX3MzOGvdPKk9QPRgOP9V7wrJADwy28LEjAg2FzUwMjmUCRnhBOl+fsOzKf0i+PeGEp8g+3XjUrqb4EVkwyWfwHd3SdmQqdOwT/oz8o12ogIcTM+OYOHYZAI9c4gK1PK3Cwdy00e7vW7Vq0Yn0i/6xEa2C79x+e3xG5hfQLvFwd7yVvZWtJAQomN4UjDxngI2uUwgxOD7RR2ZHEJKuJ6PtklZLlePS2FbSaPP8j0mWEtMsJQ9U1zfwdxh8TR22KhotpAUG8CeiHsYlXMhsqYibCGZvLBfz/YK6Uapksv4y3kDuH3RXurabSjlAteMTcZkdbG3opXUCD0/5NXw0Kwsiuo7iApQ02y2czwRRjW58cGMTg0lKkDLTR/txuaprymq72DBiATumJpGfo0Jm9NNiE7FhsJGyps6WZATS+zEh5CtegwMEQiArMcX/caJqby5ocTneu1WJ4IAd01LR0RScB6dGsrW4gbsRrhhdDR1Zif/WF2IKEJGpIH7z+rHDR/sAuCNi1O4alEBTZ7f5aWNZo7UW3l8Tj/y68zIBYEnttfwRu5dxG99jMboSWxc1cHxFHeoGOP5OfLgv7jjghHMHzsLmSDw/sZC1hyRstQsDhdBWmWv4wfEBmJU9x7/ObgRfRSqu3C6jwtWO61SZprbxb7KQJ9VIMBzaysYnTSDwLz3oWav1GVzwaeSiw6IDdZyyfB4Lhl+6rGe6lYLxxrNqBUy0iMMXmMAkBsfxPtXj+C1tUW025zcMD6FSZkR3QfLFZB5FsQMlrTbjFEYVXqen+eiw+okUKvspTIwuV8Ey24fR1lTJyEGFf2ijBg1p+d99nNm8RuYn4kk6ljBE8u6VwgDYwN588phRAVq0Cjl3DopjW3FTVL6r0xgSr8I/vTNYe/+//yxgL/PG8RH20rpsDoZnhzCoLgg7xesqL6Df/xYwOzsYMbEalicV8MahYwnzh1LK8OJDtIyPNvMBwek1cClI+L59/pi6tqlp2uHS+TfG0q4Z0YGeytaOVzTzuiUMJ5ank90oAa7080NE1LQKGXe+g+VXMb83GCu/vgIVqeLJ87J8hqXLqpaO2ky23G5RVLC9LRb7BysbOdQdRsyQeD8jHMomjOLfVUmBoUFcONEJy+skjKVNAp5LymZO6am8/wPR71aZCOTQxgUH8hnQ4/irt7LLvs87l/VHWAvqOvgqz2VZEUHcKi6nbo2i9e4dLHyaAtZcaG8sqYIpVzgjqnp1CvNxAOG9kJSQtMoafItXA3S9LixaYKoVSWS4EmJHRHh5N+eTW5RktwZkxLCFk+qeYBWwV3TM9D2iLG53CIljR00m+3EBGqJDzl5em1PArUqbp2cyj09XKxapZw5yTIoWQ+CDIzRsP3fsOstkCmwjFnR6zwmmxuHssfTftkmMNV7DcxPJb+mnWvf30m1JzFlelYET5w/wFt9r1LImZARzojkEGm1pz7BLcYYCXSnfKsVctSGvutw5DKBftEBp7Rq8fP7wm9gfiaVLRaeX+nrwjhY1caR2naiPOnHI5JD+OzG0WwraSIhVMey/TW9zvP13ipMVid7K1qRywTevmqY94nPoFGw5BwN2UX/RHfoCOcPvpVVulkseGs7NqdbClZflMPTFw2gssXKyKQQn14rXRg1SpJCdUQHaTkvN5qC2nb2VUpxoGCNwF/OzeZovVQxnx6uJaqjgMwoAyUNZoJVvsZgTGoo4UY117wnrR4UMoGnLhzIjtJmiuo7CNQqMdtdvLel1HNENTdOSOGPc7LYUtxEc6eNy0ZJNS0AiaE6Kpo7vcYFYPuxZmb2j8Sw700wN3BIexXgm8G1q6yFecPiqGvvu7JfKRfososOl8gra4oYN0/qDhOa/yFPzF7ItV+Ue43n9KwIMoIEqQ2tJhB3cCo2WXcdxlBtLc9PC+T5HTbsLjcpahM3jVNQMW4gZpuTzHA1ST2CzjaHi6V7q/jj14ewu9wEapW8ftkQxnRlU50C07Ijee2ywXywtYzIAA2Pj1YQ/NU8SRlAroJJD8FOT9KAy85ARSUKmconQ+/6XA1hRz7vPmlQImh+nnHpKh7uMi4AP+bXc15uC3MG+RZFapR+5WY/fgPzs3G43H02CLP2GFPKZQxLCmFYkiTHsau0d7wjSKeiyvOFdblFXltXzKjkUDQqOUP0jTTZj7G6319otIrEhwfy5fpj3pui2e7ika8O8sxFOTy4Jo/kuTk+OmAgPfWmRxi4eWIqRq2STYWNxIfomJgZQWywBntHK49+X4FeJUcmCJhsTh6ZFM5lI6I4XGuiv66N2EAVVW3SCmF0aqjXOICU9fXMiiM8PzeH3eWtDIkP4voPd/v8jv/eUMK90zOobbNyoLKNhWMSeWR2Fl/trWJyvzDWHmnkePJr2qke9iAOq4lBoRGwuYkZmWGEGxV8eaCekfEGFDIZ07MjiQ7UkhujY1919+89d2g8q3pI4NicbsyCASbcC/oIBgjHeH1uNgWtbkI0Mvqr68is+xYqdoLdhKyzGcOlK9la3IhcJiMnKpOLDjzIpOwM3HI14XVbacz8C/vKrQiCQHtnOzqxE4suhg6rE5kAj36d561Qb7M4+MPifXx721iiTqC1ZXe6OVzTTnF9BwFaBQNiA5k9MIaZ2VHIZALCmr9KxgUgKAF3Q4FPy4CBu//IR+e/wasHZdS0WblqVDxnCVu7izflKjjnBSnB4GfQYXOy/Vjv4uCjtSbmnJrYgJ//MfwG5mcSG6TlvEExLN3XLdtiUCtIjzjx0+FFQ+L4bFeF1x2lVsgYGBfIN/u7z9FsttNmsaNRaWmxwv2F2Wwp6e6FctuUNCpbrdR4AtztVqc3rXR3aTP3nZXJg18cpN3qRCaD5y7O4c7F+7xB6QsGxyCXCby4upBB8YFEeOJEPVNLfzjmRFVexZaSZh6ON/PBCBNb7OlUmQWitb3jAk1mOx02F6IoBaL7KnRTKmTe4sznfijg73NzmJ4t3ejGp4f16geTExfEWd8bsDi0XDmig03XJRCz468IrY08Pudy2gMCeONgHYt2t2Oxu/jb2YkcrO6goN5CbnIE3x9p4Whdt/6ZXiUnXGWDDc/D+HsJ+v5BxiVMIKPflehaCwnZ9aJUp9IiJWYUTXyZW5Z1UFAvNRwblRTM6+c9R1jLQXA7Kc+5nK8Knby/tYx2i4PZA6MxBkRx95sb6bS7yI42cv9Z/fjb8nxvTcetuQrC8t6BklWQOgX6zZHSk01VoA1hfZWSGz7e6w2QD0sI4uXLhkjuJ5cLjq2XNshV4LRhDhuEkUXe31HWXs7A6i94fcGTWJwi4UYNdAZjj/gO0dKMIiwV+S/oPROgVTI1K4L3t/iukgeepGHdacNpB3M9qIwnVJmua7eikAm/qEjUz+nFb2B+JmqlnLtnZBIbrGXp3mr6RRm5fUoaqSep9h4UH8SSm8aws7QZUYTkMD33Ldnvs8/5uTEEeVIwD5j0PsYF4P3NpVwyIt6boWZQK7zpu+MzIvjrd/lcOiIBlUJGfLCWj7eX+WQ8fbW3mn/MG8TSfdWUNXUyMT2MH/N9u+jlxAbwnafupVKTQVLpQ6TWHQBNAIfD/4xM0NPThqRFGKhq6eTVtUXMHhDFmNQQthR3P+kmhOjoHxPA7VPSpFWSxcHKw/X8cKiWuGAtd05NZ0hCEHvKWwGYkR2JTiXH5nTjcIm8vbWCFHUAl5VvBWsbqspdhM78G2Ozz0YbZCMmUI3SUk+S3ERiaihyvYrJmeGsP9qAyebEqFbw+DnZCE7PnCwtHJv8Ki+XxvH1tx3EBQ3nzxM+Y5zikPSFCErky+ZkCuq7jd620hbWlAYzrqMEt0LHfpeaf67qjr99s78arUpOuFFNWVMnh2tM6PJqmZgRwdqj9VyaE8CVTX9HtnOtdEDxGji6HIZdBw2HaYoaz+Pfunyyr3aVt3KovIHogQkgl0POJZA5SxIOlauwR+ZgjR6GpkZyV7oC4jkQeR6bNxxj5eE6rhydREqYjvuW2KlukzMjy8L9Z3WSGiwHqwl0wSD/D8Fym0nq3lmyHnlgPFcOOZuDlW3sKW9FJsDC0Yk+zeFOhQ6bk8I6E62dDhJCdaSE6U+e8dVYBBv/LhVohmXCzCchcYx3c327lSW7K3lzYwl6tYKHZvVjSr8ItCr/7e23xv8X+AUkhOi4d0Ym145LRq9SoD6B37nBZEUhkxGsVzEgNpABHl99a3Uxz0w28M9dNtpsbm7IUTMqVcOGggZMVid9qVuYbE70ni+ORinjtilp3q5uBrWcmjYL//ZkaS0ck0Rli4U7pqYBIBcEjjVK6b2pYXoUchlZ0YEMig9kf4UUk8mKNqDXKL2JAl9VaNGEP8z4lBq07g604ck8NCuEV9cV09rpIDVcz4IRCZg89Q/L82q5c1o6GZEBrDlST/+YAC4YHMvdn+2n3mRDLhO4cUIKzWbp/BmRRt7efIykUD1/mJoOwI7SZlYfqWdkcgjVbVaK6jvY125k8KQ3cZibSWzZRtCutxm54FySomKJCdKibHPjth6CimXsy7qbj7aVc9moBOQyGU6Xm092lJMyO4PPZ+8nKUzHB1vL+PawlBVW1mzl2hU2vlowjhyZAnvsCDbV+N7wJmWE0epQcNbWLBQygUtH9M5CW3Oknin9IijzuCh3lbVw59R08qrbuDfHjmzJWt8DyrZA7mWw+z0sQ8KpMyUwLc3ArESROouct/ZbaTd3u/3c4VnIPr5IyhwDQtWvUTLnc+ob6pCLLvZ0RlDXGMq3ByppMNl4dGke141Pxu5yI4pSrc7dVXsRV72KULULMmfDmNsgLKPPzy0A+d9KCtAeUs9y8E5iPWWZGagFJ0nuA2jcC5G6E/1n2i12XlpT5H1A0ihlvH3V8O5K/+Oxm+H7h6BopfS6eg98eIFUTxTRD5A+c8/+ILkOWzod3PrJXj6+buSJz+nnV8NvYH4hgiAQou97Sd5gsvLV3ire3HAMnVrO/TP7MSQxiNJGMyCQ5qhn6trzGZF6Dk5VABbieGSVQEpkEDqVnKhADWqFzCeLa1x6GLnxgfxxTjbhBhVyQSBEr2TJTaMJ1imYNSCa7w5KyQSHq9u4Y2o6j39zyJu5NSwxmHMGRTMkMRiTzUlbdRFvZedxrF8kbgSSZYdZa+t+jJYJAs9uaedZ9ICeu6YH8eHWEi4aEodWJaem1cILqwpYMLJbiuOFVYW8umAwdqebQJ2CF1YVUO9ZRXXFmf50TjY7S1soqu9gZHIIn++u9FELeGR2FgV1JowaJbdPSaOmtZMfO5JxiUlgzOXc2Ok0V7fx+JoS5g9PYPaAKCzJc2mNOh+3w8mRWhN7j+t1s76whRdXF3L7lDS+y6v32eZyi+xr13N45FLcCi2TA3Qc8CRCyAQ4Z0AY93wpSezIZUKfT9xxQVrq27tXi09PDeKCwC3cMrgApStHcokdWeZzTJMymvdT3yVXr+OLi11kHnwe7eaVoA9j9oQ/Yw9LlHYURVx7PkDm7NHszWYiouJ7rjo4kcpWC7MG6IgKEH1WrA5TM4tnujG3thIQkUDipvuh3pPJuOd9aDgqCah63E52p5ujte2UNnUSphXIKt5JUNfJonKgfBtBh5d2jwFEZ0H/83u9H31xuMbkUx9mdbh58MsDfHnzWML7atTWVtltXLpwWqGxACL60W5x8GEfiS1bipv8BuZ3gN/AnEbaLHa2lTTzzb5qkkJ1DE4I5qXVRXTYnNABt36yhz+dm+1NVc6K1PHqqCdJ2XQPAAdm/UiOPIhV+XWoFTJiArW8cEkur64torjBzLSsCEYmh3DN+7u8rpQ7p6Vjc7nJr2knLljHNWOTCDeqWZFXw+CEID7dWe6TFryrrIV2qwODRkF0oIazNdsx7P4X9uyrEQU54Xnvce6AS/khYyJVbRZCjiu6O9ZgJilU75V2AbhpQorXqAGcPTCKTrtU2zIqOZTX1/nWvQAo5DJumZRKh02Sw5Ey8KSYyYT0MCpbOvl8dyWCAJP7hXOktsMb70oLN5A2ZRAWi5tpWQqqWy2sPtrAk8vyMdmcxAVruf+sTJbuqWBgmIydNQ6O1nfSZRNC9SrCjWrvKq0LlULOg2vb0SrNfD4vgj1JejaWmhkUH8S2HsrXLrdIh83p02VTrZBx+ehE7l8ipRXfOCyAuRVPIt+8ufsCw6+X2ivUScoJzvjRvHBQyYf7TYxNEnkv8G2Uxzw3U3MjSWtvxX3lt0AGLrcb+fEyQYCyo5pPrh+F1eGitMnM9R90J1hc0l/HvcKHGJd9Kg3IVTDtcVj7N6kGBaBiG7SWglaK0q88VMvtn3bHgS4dcD5/nBCLDquUgbbh2V5zoC7vlA1MfR8FqhXNFtotjr4NjEIjqQ/YfGN0XWnWaoWM2CAtxQ2+K8rIvs7l51fHb2BOI8sP1vDQl92yKwFaBdeNT+aFHorKh6tNRAVoqG23kl/XyYrUTG41REJHHVZtDFqrjcwoI/XtNvpFG1HIBcamSRXWKrmMhg6798ufHR1AXbvV5/wTMsI4f1AMYQYV2VEBfLL9OL0ToNnUyTf7qkgJN3B5upuP0l7j5c0duEWRW4e/wpRwFZO0YZhtThJCdV7JfJB0vZ6+aCBT+kXQanEQblATrFdSv0W6cQyKCyA2SMt9Sw4CMD070udG3EWIXsXj3xwiSKskfZqepy4YQEWLBbVCRk2rlT8vk4xwsE6JyyX6JFMUNXSw5mgjFw2JpaLVyujUEK59f5fXpVjZYuG9zaV8PL6RuB1P0hI/gdKJC1hR4+DRs7MIN6i4ekwyz/xwxPte5sYHEayVc2H/QGKDdWTveJjXAuM4Nn0qxkAlK5p8Zfnf2XyMv56bhYiM+g4bEXoFI8nn0xkO2pxKxoY1IF+62ecYdr8L570Gez/EGT+WJY4xfLhOMqpzkuUot33T628lM0lGRSaTUZ+5gMgyX/XKttRzvPU1giAZz66aoHlxbRjXf9q9s8su1c0MuBD2fOC5gIJGq8CmvVXEBWt59Os8nzjQxDjQlKyG8i0QOxySxsG+T3wnGXXqis8JfUjt58QFoFfLKW8yE2JQ+YqGBifCjL/At3/oHkuZApH9ASkWetuUNLaWNHoz9iIC1IxO9Uvz/x7wG5jTRIPJyj9/9JXmb7c4e3XIDdAqMNu6VXo31yq4NSYXStYjyBW8tvaQ92a+qaiRmyam8NXeau+T39i0UManh7GxsJGpWRG8ts43CWBDQSPnD4rl+ZUFvHNpJrMyDXx2oNVnn/QAN+1WJyUNHWweNIlnfux2Mbh0YSz8ppyGDinFN1Cr5KVLB/Pj4TqKGyShyp3HmvliT5XXfbdgRDxPnNefihaLFMzvIQPT2GFlwcgEXlxdSGunA0GAS4cn4HBJ2WZNZjsWh8jV7+2izeIgJUzPwNjugjq5IFDa5Kv4C7C1pIkRycG8sqaIMakh3D09k+d+6NYGK2vupN1sgZZjBLeW0hQ3jYYOFYt3V5EdbeSykQncPzOTTrsLhVxGuNrJxI4VzDS9hd0ZiTj4cowbniXn6Oeg1DL1olW820MTS6eU09/Yye3L66lotrDoLIhbuYC4rglMvL/XnHE7oa0CXHZKlSk8+GO30S1pB3dgErJWX3khtNKNUhAEzHETqBz3NHF5r4FMQXnOH3CFD4GiNeB2kBaexaIbRrG5SNIiywrYSy9ay33SlG2j7+Kab5s5UF3BbVPSfBrkjUs2MKHiDWTlW6SBqp1SFX5Ef6g/JI0NvhLiR/a+zgnoF23kmQsH8qdvD2NxuEgK0/HQrCyufX8Xh6rbGZUcwmPnZJPds3nbwLlSnKixAPSRUgO2Hr/DsMRgvrplLIeq29Ao5QyMDTzzPW38nBJ+A3OaEJF605+MMIMKo0aJqYeBGZYcAqNeBpeD+iM2H4l4gMU7KzhnUAwfePzMm4uauH1KGhsLGxGEvntfODxyIqU1DdwUX0FdZzzri1rRq+Q8MC6Yfu4iRFFGdJCW5QXd1exapRyrw0Wj2ca0jFBC9XL2VpupbLZQ3mRGJsDh6lbOz43B5nBwqMbMrAFRyAR44AtpxXLH1DTUChlXj0wkNUKPXJCx9kgNFw+NQ6uUo5AJrD5SzwCPnlZWtJG95S1e8cySRjPzRyTwzYEaRFFq9xzdR2fDQXFB1LdLT+pbiptRyGSMTA7x1mmE6FUE2aWYjinjYh7ZG8j2csmNt7WkmUM17Vw3LoWX1xQRblCxZtwRNKseBUBFPlRsgql/hB8fh6AEMo+8zmejBnFYSMcpyugnryBzz2tkhT9IRbOFdGWT7wSdNtCHg7k7Q09MmoBQuQMqthOtCmVa2i2sKpJcPx8d6OCS2X8m7cdrulsqp82QGpPVHACFmqToZPIVC6iMmoIoyAkxGsis+QYsTVLr7IYCMlInkzF2gHR8ZW/D7IwfQ13wcOTDH6Jen8bKtjgSw1QcqDZT2mgmOzrAmzI+M1GObscKaWkUlCQ1RVv3FMz/VIrZKDSSeKXqBE3X+kCjlDNveDwjUkIwWZ1olXLmv7HNu+radqyZ2xbt5bMbRxPWlW6s0ktZYz0yx3oikwk+yTMgKW2UeIp3U8L/Q5aanzOG38D8AlrMdnaWNrOxsIFBcUHcNDGVv37XnbqqV8npF2XklkmpKOQCQxOCfVxWkzLCGZIQDEbpaUwm9A5WCoLgswoyqBUkhOgYHB9EpFHDwNgADlZ1+6fDDCrCDGoUMoF39ncy9rwBzNEGMCc3AaVchsxlo0WjQS6rpt3iYFxqKD96pqxWyjCqBLZfpiVo/4soTI3snfEycz/KIzsmgDC9mqX7anDabTyneRd7gsCuuIe59qN93utXNVt46dLB7DjWzJd7qsiIMHDTxFSeWHaII7Ud6FVy7piWzsc7pLYB0YFaypt9q/QX7yzn2YtyWHm4FodLJCpAw/TsSH70FE7Gh2iZ0T/Spx5jY1Ejt0xKY/uxZpRygRemB2DvDGXnuDcRwvuxfZGvq7Dd4iQ5RM1n0ywkhSsxrn3D9413O3GgZPXZW4kINDC44n2SN95HT5lKceA87hmdyZXjXBjJ9z1+x5uIM/4KZVsQqnbh6Hcu1rRzUG5/CY02GIVM4M5J8cRHmll5qI4BsQFY41PghnVSu2VNEBij4OtboHQTyBTIxt1F/1E3Q0ymdI38ZbDhGamJF0g3fGOkJCgpk0nxngvfgOX3gbUNd/RgtmQ+yDWfd6BRDpZigzTx4FmZfHughu/zanlgVj80B2XsKW+lxqKgfeDV7A6exbpaNQkGkQm6UhLcoP4Jq5bjEQSB5DBphbG+oKGXzE9Jg5nKFku3gemD8iYzpU2dGDQK0iMMPtpklc2dfLO/mnc2S6vBa8clM2dQDPE93HMFdSbWH22gqtXC5MxwhiaF+Kh1+zk9+N/Rn4nbLbJoR7k3PfJDyrl8RAJ/nzuIJbsrSA4zMCcnmns+30+DyYZbFLloSCzTs8PJiQ9AFCFIp8TWo/I/OlDjE+8AqSHWpzu7b463TEpl0Y4ytCoFr6wt5PFzBvDt/iq2lUgqu3NyohHdIvfPzCQ1XE++xcmLqwuo9Cgyn58bS8LIaFZcakeQyRE1ar7YI7l+WjsdzI9rJnTxXO9TdFVDMw/PzqKkoQOr081Vo5MkLTJjFnFbHkUz4H4EAa8RnNE/kk93lvOjJwV4x7FmtpU08czFOWwsbMTudBOkVXhbR+8tb+GS4fE+HTKLG6QmXja7C6VcwO1y0i/KyICYAFyiiNst0txh92lJnRCiY2aqhnOignFoIqlq7eTOw5kkhukZqDL0ysYDUClkjKj5BEfgBbhVRo4XnLHLNNz9TRmZUQF8MS4NITBOymoC0IZQnHs/q4+2YHO6aApLYej5S4hbKnWLRJBh10bgGv8IOmMQLQ4NC9/bRZDqekb3u5EdtW5qvynj/WuGc/uUNPRqBWqFHAiBqIGSnP2av0LSeOnJXZBDZyNU7oKMmdI1Knd2GxeQsqvyvpS2awJBqZZqZxJGg81EWaeGq988isuNx7hI2nMZQZ7D3SL/WlfMFzePweUWUcoEvjkSzKPLigDJNRhpjObf8/uT2/sr4eVAZSurDtfRYXMys38UufFBuEWwO10+wpggyfEfj0ouw3iSm/2+ihauemend9U7d1gcD8zsR5gnsL+pqNH7vQR45vujhOhVXOJpOlbS0MFlb273ujvf21LKP+cN4oIhcfg5vfgNzM+koqWTl9b4xlw+2lHO9P6RfHzdKGSe6vr3rxnB0VoTaoWMqhYLz60sYHpWJAgC728tIzPSyMwBUg1BbLCW2yanUVjfQV27lZn9ozDbnVw+KpFOm5MArZLihg5umJBKfbuNiAA1z35/BK1KzvTsSJrNdkxWp9dd9dplQ1iyu8prXEAK0p+XaCf3h+kgUyAOvZpvL7yAze3xaBRy1FVfd7togNDwKGobRZxuaOqwkxiqJycuCL1JOmdm+SdcNfoc3vOsJlQKmde4dBGsV1HXZkWrlKNTyqlssfLKgsHUtVvRquSkhaqwdrTx0d5mSRp/fDIFde0MjA9Gp5SRou1kaZ2JLI/YYYRRw/rCbteTUi7w4MwMclacB02FFE19g1L3MOJCHOwuayFMr+SGCZI7rItJGWEMEgrB2kqLIZXOQXeR9ON13ZPWhVBj6E9mpIOjdSbMijCM/S8ElQ5EkYK4udz6dbVPFtmzF+VQNf1r5JZ6jrkj2ZOn5uGZBtCFUFHWzGFPV8auhdekJC3G2u0YO0ql1UrMYOl/kBpzBcbCykfB4fn7hWdKAW5TLQhK6Owt24KpRhLC7EmQdGMNPraX2ek6vj3a7Tq7c5SRYHcTaREGDGo5f5yTTVKY5PIqaejghXW+DerqTHaONDnJTe19aYCDla3M+/dWr1rFD4fq+PO5/flgWyktZjvnD45jTk40kQGSXl96hIGrRif6aOg9MCvTO4fj6bA5+OuyfJ9+RJ/vquTsgdFeDb/vjutmCVICTpeByatq69Vf5tkfjjI+PdxrpPycHvwG5mfS1WnxeKxOt9e4gFRImBEppVT+88cC6tptfLRd+tIKAjwyux/rjzZgd7pIjTCQEq6nvLmTyAA1mcEu5r13hAijhvhgLYdq2rl1UgoVTWa2l7YwPCmE2yanUdVmwe50M7N/FH9bkc9d09LRqOQEahXsO64WBKC81SZVcLscCDvfxBYwhAdWtBFuVDFulG9ltx0Vr60/7A3+ri9o4MYJKTQlz6RUkUC0tYhrokoYMn8ode02tEq5z4om3Khm7tA4jtZ1sLe8hdhgLRmRRtxuN3/8WgoUj0sL47bJKRgDggnQKtAqZdS121i0oxiXW6RqSCS3D3RyoK6GdoeM2IhYEkJ03DU9A6fLjUIu461NJYyMm0pIUyHtxnQ+W1XhbVb29uYyLh4ax0vzczlQ1YZRo6TZbMNRtQ2qdhNcvJT8xJuxzvyE0Lot2NUhtEePpVGdzNCkJoYlh7G0TcP8dD2K+oNSvMPZSUVLt2vP5nTz/tZSZg1I4MlVZqADmdDBVWPTaG5qRCkTiA/RUtEsGYv4EC1P9yvCWFktNeDqqIWi1dgmPkKtTYVOpiS8YKVkXDRB4LJJ7ZXdDikDTBCk1OC8JeDo4WIcfQc0FYPDDCGp3QYLCAoI4I/BSzh/5iRqbBqSNR3EalvY6RrIwNgWhiWGeG/8IMVLLPbeensna3i5rqDBa1wArhiVwP7KVgbEBGK2u3A43azKr+OykVJ9j0Gj5M7pGczoH0Vdu5X4EB3Zng6wfdFmcfqsXLvo6g0EEBfcO2YXG9TtHrP38b21OFy4xD4qm/38IvwG5mcSF6xl3rB4Fu3ofsIL1Cq9xqQv5uREs2R3hVfc8o4p6by18Zi3JWygVsnzc3OQCdJKIL7zKO9d1p+VhSZKmzr5w9R0zDYHz62UVk6r8uu5bXIaEUY1te1W2q0OHpzVj/waEwW1JnRKOcOTgqlttzIxI5zWTgfLD9YQE2IEV/cTYFDdNhaOuYmWTjvm6FEEBsTSmHoRDnUQJquT1k4HccFawgxq8mvaWbSznAkZQ9jsVBNoHM6QMCN1lSYK6zsYmhDEuYNi+NqTVnzlqAQOVbf7PKGGG9S8ND+XR8/OotPu5LNdlZQ0Wtha0kiAVsm4tHA+29XtFvxkdx3jEhJI1zRiVYKbSD7ZXu7j8hIE6MhIIgQwuZRe49LFkt2VjEwO4cOtZdhdbgwqBbMumk75uASMeh27qq28vBYiAyZjsbv4w7QInv58N0q5DLco8swkHYpv70JolrL2MvRh/Hv6W1y1vPvmXtFiQewRMAszqClvd7DsQC2p4XoenpXFcz8cIVCrYmF/OVEaF6x9S5Jjkck5dsEynl9axPJD9UQaNTwx+QEmDb0ehaMdZEpkar3U7tjm0VjThcLF78DyeyVDNPMpKN8EX17j+cMmYb3oA9a0RtBgstEvKoicYZcy9btboS4P85j7eKBwGsvyJEP/1d5qxqeH8cqlgwnUqYgO1HDdhBRe7JEGr1XKGRQfdMLPuLWHQZLLBGKDdTz5XT617d0G4OHZ/WhotxLuMWbBOhXDEoMxWZ0E6pR9qmN3EaJTMTE9nJX5dT7jCT1aIZw/OJZlB2q8yTRGtYLzcmO827Ojjb1cpjdOSCEyoG8RUj8/H7+B+ZmoFHJum5xGcpiOL/dUMTA2kKvGJJF8gqU9QHqkkcU3jqakoQOZIFDZ0snIlFCmZ0fiEkXkgsC3+2uoaO5kb0Ur56Wl8NC3hV4l401FjVw8NM6b6ZMRacBkc/LK2m7Xz+yBUVjsbtYerafV4uDyUYksO1DNO5tKCTOouHdmJgE6q8+8hKiBbNrVSIfNye6MTDaP+YbXN0minLMHdvLKpYNZX9hAXZuVWyalUtzQwZaiJm+K9O1T0rA6XHy2q5KcuCAijGoemtWPrSVNDE4I5jpPY7AuLh0Zz783lrDuaAM6lfQ+Rgao0auVJIfp2Vnq6/qZOyyOTw9ZaDEHYdQoqNhZyb0zM3myR0LFrAwjkWXfARDYWQq9IiqSkKjN6UYpF7h7RgY3LS2ktVMFOJk/3MGE9DA2FDYyKC6QvMo27p6eQZPZjl4lZ4r8e69xAcDcyKD6paRFXEBRvWTMLhgUyd6KFmb2j2RXWQs3Tkzh3iUHaLdIN7orR8bxzeQGtIc+hbAr4YcXvMbCFjOSF/Y4+S5fUtyubbdy4ze1LL3AwKDl10irl/B+3cYFoLNJSgC4cYP0wFCzH3a90729tRRhwzOEpd+GsbOevPwQamJSOf/Kb8HWSkGrjmWrtvu8RxsLGzlc3c7otDAEQWDBiATC9GoW7SgnOVzPdeOSyYw6cV+Wyf0ieH19MW4RRESazXYf4wLwyfZyzs2Jwe50oVLIOVTVxourC9lb3sqUrAiuH59M2glEY7UqOffMzKC8pZMjtSaUcoG7pmX4ZJCNTAnlg2tHcKi6DUEQyI4OYHBCt15adkwgi64fxVsbSyht6uSykQnM6B/Z1+X8/EL8BuYXEBus5YYJqVw2MhG1QtarE9/xuN0iTR029lW0IYoiufFBvL6+xKtdpVPJeWR2FqIosreilX3NCq9x6WJfeQt3TE3naF0H/aIM3LXYVyxz+cFa7pqWztqj9QyKC+L7vFq+2iutJqrbrDz29SFeuaS/10VmjxhEedAIjtRKwet6kw2zXWRqVhQWu5Mwo5rSJjPrjzZQb7KxobCRO6els7O0icxII6VNZpbtr+ZP5/VHrZTjdLtp7LDxyfZyUiMMuBxWnyy4flFGShs7WXdUiqF02l08+8NRXrtsCLnxkkROz14icplARoSBflFGthQ30dbp4OJh8XRY7EzJDGfN0QYmZoTxh6Fu1F9vB0MEqXWrmDPwSpb16DV/3qBor/TLzP5RLN5R4VPz8enOCu6ZkcGGwkYMGgUjUkJ5zCO3nxZh4PqYQ73+nkHNBxgYuYBjjZ3Mywnh8jQrn5eoWXm0lavHJNHUYfcaF4CPd1ZxRYCZ9JLVEDu4W0YfqE86h29Xm3zOL4pQ1AaDADQBUsX98bRXSSsZgLwvem1WV24hO+tC6oR2hlh+pMR2ATXW0URrg2i3WXvtD9Bm7X5fIgM0XDE6kYuGxqKSy1A4zVDwA+z9EPQRkLsAYofSJZMwKD6Ij64dyZsbS+i0O31WdCB9xq8YlchLawrJq27n/NwYKpstrPRkCC7eWcHRWhPvXz28V0JAF5lRAXxy/Ugqmi3oVHKSwvS9Vj2DE4J9jMrxDEkM5sXYwTjc7u52zn5OO/539hfgdLkpqJPcVyF6qZXryfqZ76toZdv+PEbr65HLBWyWdJ8AfKfdxY+H69B6brBmu++XMyVMz+ycaO7+bD9Ot8gfpqb7NJfqwuX5UgsC3i9uT2pMLvaP+xd2UcaGllBai+SEGVS4RQgzqvloVSETM8IJ0Cr5el8VlwyLZ1x6GF/ukarK39l8jD+d058VebVcNDQGtULBTR/uweJwYVAreHH+IBJD9RyobCNK3s71o6N5ZaN07LDEYB9ZmS4Kak0oZAIWh4v0CCNJoTpKmzoxahSEGdTc/8UBb8xrV1kLd01L57rxSWTHBrKnrIXyNjeps55HaCpCGTuc+W4DI1L6U1RvIi3CiOgWyQjXMDzBSHqkkcve2d1rDk7P+a12F+uP1nuvV9ZkpnrweDKOLPY9IGkcD7KFuyZriC55HuXRbII1V3C0zsTRlSbmDosjI9JAgSdjzuUW6QgewP6xrxIeHEpMRLZXF0zXUUFMUCaVLb43/aCumHPVbhjzB0gcCzKF9Me1myG4R+J0aFrv3yl+DPcdiGbFES1jEufxp/hOVA0HYeVt9B/9MEMTItntUbEGKeje12fYexM+sgo+X9i9Yf8n1C/cRpHFgCAIpEUYGJMWxvCkEEREth9rRiETvJ/ThWOSeHltkde4H6hs48IhsT71N/sqWilr7iTnJN+lEL36hBqAp4pSIUPZx0oXpIZxRfUd1JmsxARpSQ03nNR156dv/AbmF7DuaAM3frTbW+x45ahErhqbRG2bFYNaQWq4AUOPNEyhuYjrj92FskXyabtC0vjn1L9zx4/d8YKy5k76e4oQzXYn0QEaajwuhnNzY3hlTZH3y1rW1ElmpNGn74mU+iz5lvNr2okO1Phk3IDU1+OipTrPDbWDucOCUMplWB0uLA4Xc3Ki+XBbGWabk7MHxtBhcxLVM7tGlJpM/Xi4jn5RRv6+odDrz+6wOblz8X4enp3JmNRQ6t1qLjesIHl6Fl+VyOgXpqQw0tircZVOLedvKyTplln9o/jTuf2pbbMiCFDVaumVULF0XzVTMkJICdUzKUYkd9sfUFRuA0AODBlzP387PIZ6s4sWcwWTMsNRKiL403cFzOwfyaC4QPZX+gaLR8Rp2XW5FsEQyE3fd6dNO1wiH1TH8cDIuzHufhXcTsScSxDsHUTueqn7BKZKtLmXe19+vbeaa8clew3MtH5hvJkvY/nBYEL0Cl48733GNywGAUJ1ETwRF8F1X5R7WyGMSdAxIMizArKbISheyirr9BR1BsTBWX/zXs+si0WXcwnCAckQisFJbIm/gRXftQKwpayDv8oMvDbJBo2FhH17Nc+e9SEfx6ew/VgL/WMC6RdlPHE9iM0k9dPpQfGov3Hz4mMU1Eur8IGxgbyyIBe9SolLdNPaaefh2Vl8d6CaVouDyAC1z8oR4Nv91Vw/PsVrYOQy4TftiGlzuFi0o5w/LzuM6Gl3/s95gzhnUIy/YPMncsYMjCAI7wBzgHpRFAd4xv4EXA905Zg+LIrics+2h4BrARdwhyiKP3jGhwLvAVpgOfAHURRFQRDUwAfAUKAJuEQUxVLPMVcBj3qu8VdRFN8/3b9fbbuFh7486DUuMgGSwvRc8u+tNHZIbq0FI+O5Z0YmoZ4nreT61V7jAiBvLmJk5wYiA8Z4hRcvzI1ErVKhV8sZEG1kSmYE647WU1jfQWyQ1mfFsvJwDW9dOZx3Nh9jS3GTp+dKOtuPNZEabsDpFrllUhp3fbbPO8/MSCPRgVpumZSKyw2r8+vIjDDy+a5KBsYGoJAJPum8S/dVsXBMEi53d0D00hEJ3pugCD7BUoVM4PYp6SzbX8vm4iayoow8PPMsptV+wdjcJJw6N7LcOA5WtfkoPJc3dXpdaSsO1ZIRbeRf64qRCQI3Tkzp9f6r5AJvbCrn2wM1bJ4n9xqXLnTbX+D+qVO46hvJkK3Kr+fsnGjvz4/MzqLBZKO6TWpSdfeUJHKOvoBx31ugMnDFxO/Y2aMD6UcHOzn3snl0BM6gzWxjVowZzecLfK5pT5yEOiiK26cEU2+ysepwLQNjDWRGGjkrOwwnMl5dK8VxOu1OAjrL4cCnUm2NysCEWX/n63mJFDdaCVSJZIcriRRaJNkZpQ7R5USwtnZfsL0Sag9A9rkAuKoPILSWS62URRdHI+Zw3SfVPnPcV2vFoY6WOnu2VWKo2cawpByiA7QYtApUcjltnb5uWWwmT72NIKkTdBEQwzJTmk/fHI1SxrqjjfxrfTGddhdXjk4kPljLxcPiQMQnw7ILuUzwUcG4cUIKkUY15U1mArVKArRKjjWaqWjuJFivIj3CgFYhQEcdKLRSX5tfiNXhpKjeTEunHa1SzhMe4wLSyvPBLw+SExd0wvRpP31zJlcw7wGvIBmBnvxTFEWfxyBBELKB+UB/IAZYJQhChiiKLuB14AZgG5KBOQtYgWSMWkRRTBMEYT7wDHCJIAghwOPAMKT7325BEL4RRbF3v+JfgMni9Mmln5Aezoq8Gq9xAfhkewUzs6OY6MnPD6jf0es84U07yYyaSbPZzmU5AZwTUsk927S0dDrQqRQMV8hJjTBg1CoxahTcMikVracZ18ikEJ5akS/JbQyPp6TRzL2f7+PBs/qRExuACFgcTv59+RBKmzrRKOWE6JV8urOc5QdrUcoFrhqTREKoZHCCtEqKjhOlBFh9pI4bJ6QyNi2U4UkhlDWaGZcezm1T0kgM0aFRyLB6jMw5g2JYsrvC+9SeX2vipkV5PDjrPB778hAxgTb+dqGWO6am02l3EqxTUd9u5fX1vorLhXUmYoO0lDSakcsE9Cq5T9fNK0Yl0mZ1khSmR81ReuGyE2XwdWlU92hN/ewPR3hnfj9UDQcwGINIal6KZt9b0o72DiYce4Fn59zLa1sbUMnhriFyMlvW8pl1FK1mkSJZOBnhOagaJPVkc8JkFoXfzXNL87E53SSF6njqgoHEB8DzsyJRqrS8saXSO5ercw0M3HZbd+GmvQPFNzczcNJDDNz8pFTLMvNJSarGJX2mhOAkGHMHbPpn9y/VeBTWPQtNRegj+0P5Vukf0DBjIvYeTYWiA1R8PNlM8JKLoaMOMSwdy7jnuPuLfG9q8X0TIjlb1wRr90JoCoSmQ9lmqZtmQAyMuEGSwSnfghiUzOY631vIlH4RPP5Nd7zqvc2lPH3RQDYWNtLYYePCwXHEBWt9XMM3T0xlUHwggVrJzRxmVHHt+7vYVdZCdnQAD8zK5P4lB7wPYV9eGkduzWfI9n4gzWnmk5A8CeQ/73bWYXPyzqZj/HNVAaIIf5iaxvGe5067i2az3W9gfiJnzMCIorhBEISkU9z9POBTURRtwDFBEIqAEYIglAIBoihuBRAE4QPgfCQDcx7wJ8/xS4BXBGn9OhP4URTFZs8xPyIZpe7esqeByACNTxfG1AgDn2wv77VfVWv3F4msc6F4te8OSeN41rEde7yC6KJPsZQO45rRd3KkwcKUfhH89bvD7C5r9e5+z4wMFu+soLLFQuolg8jzyMT0fNp2ixCoU1HTZuFf60t4bE42u8taSArVUdUiJQKA5Pp5a+MxMiIN5MYHIooiWlVv10RiiI4gnZKzB0ZjsjoYkxbGP1dJ6gBDE4J55bIh3PbJHqwONwkhOr7a6ysrb7a7qDfZEEXpfXtvSxlrjkjFmAkhOkYmh/S6ZnKonrVHpIXuWxuP8dDsLCqbO2gw2ZicFcXq/Dpv8sKMyxII04X4FB5akmeQbw0BpEr3zEiDpw+PhMstElq/mcz1t+Kc+DCK/e/6XD+o6CvmRSYyI92GzNZKwLEKxNwFXLv/JmRuG23ht7Mh9xki7ZWo3BbKQsfx10+PeI8vberkrY0lvDHgEMH738Ccdg53jTiPY20u9la0MyDAguz4oL3olupdAFInw6GvvMYFgJZSqaJfruoeD8+Gjc+By4488k8gyLDFj8epDSO77lvmZJ/PssPS+/LwCAUpqz0GAhAaC0lYdyfL5v6LjvpjWPXxZLSvQbP4ye5rxg6VpGcKPW0EDn8Nc9+D7W8iBCdyljaYHR53Z1ywtpdq9nXjk3noy4NedYp1Rxt4aX4uFS0WCupMTM2KZFRKCCF6NZMyI6lvt3Lh61u8BuhwTTt3LNrH/OHx/HtDCZkROpKLP0R24E3pAg3t8PFcXNf8SLEqE4VMICFE9x8TbnpypLadf/xY0P1ngF5pzKF6FVGBmj6O9nMyfosYzG2CIFwJ7ALu8awsYpFWKF1UesYcnp+PH8fzfwWAKIpOQRDagNCe430c44MgCDcgrY5ISEjoa5cTEqBV8uQFA3nwywPsr2ijormTUSkhrD3q2344IaT7iaczYTLyAQvQHpJsnbv/xcg6aonypJY2Zl3B05aLWbJYqsRPjzD4GBeAf60r5rJRibyxoQSVXIZcJvQSvAzQKmk220kI0ZMeaWRfZQuF9R3oVHIOVrUhCJARYcTqdFHZYpFkPFwiTpebAbEBZEToKfCk3mqUMq4cnQSIJIRosbvU3PjhHm9MZHd5Cy+tLuT9a4Zhc4holHL+tb64lyxL1xc+O9rI+sIGbp2chlwGcpmMYYmSoe7q6TEyOYT0KAMWj4xOm8VBu8VBqEGD0y1gdbi9xgXgixIlrknvkVbwFobG/TQmnk1F0kVUNthIDtMzKcXAnEFxPLhUCqiH6lU8MyuGtHzpfTfJg9BEDkXbscJnzq6QdIJK35GC6sOuQfjqRm/30MBV95A68UVmr4nC4tBz2xRfkVKAnWWttKdDcGMh+sZ/oG48xD/O+wdlDe2oNTqcMcNQVPdI4ZarpJv5hPskcckDn/U6p+h2ImgCwWZCHH49QsMRr7FxF61h10XbeX1bPTV1bq7IDeSGnBDO7R9Ch83JRF2R17h0IWstJa1xNWx6DsbdDTtf871g1W5In9H92tKCu6mERclPoVHKGR5pYNYAMyvyamm3OojtIUwqCFLCSU/pI6db5C/f5fPhtSO4cWJvOYDKFovP6gakv79KIX1+5mYqCN5/XLsA0U1z6X5mfFeHSi7jlsmpXDU6iWD9iZMEelLX5ptY8fmuSu6ansHr64pp8/SpeXn+YGL6EF31c3J+bQPzOvAXpIeEvwB/B64B+oqciScZ52ce4zsoim8AbwAMGzbs5FLIfZAVHcAH14ykrt2KUa2gzeKgrMlMSWMnMkHSDRvYIz9/eZmMrxouYcHYuYBIbkoMUStvoat2Pi/qfJas6F6JmG0uFDKByf0iiA3SsqusmbyqdtSeL5tR7uCykQlepWWQBDRD1W72V7bS3GHH7nJzy6RUiuo7UMgEJmSEcdGQOPZWtKJVynl4VhZ/W5FPqSdVOiZQwz8uySW/ph2Hy83AmEAcLjfL82qpbrUwPTuKyZkR3uw0hUxg1oAoPttZxfZjzYxLC+Mf8wZx+6K9XjfDrAFRdFid3DYljVCdkqGJwTz29SFvIdyolBAWjkmkscOOTJASCAQR7puZicPlJjpQQ6hOyQHP7261+97M399WinZCCtWZj+BIMBGgdDNQY+Hmknu4IjEHQ2g88m/eZ9GQW2iIHIfRVErc2nMld0/xj+gKllI29CEy6nZDh7SyMqXMxhmYSfCwa0FlRCxY0euDlVSyiIUjn2Xl0RbSwnu7TjLCtQS0ddfqKIp+ILbfGpI3PI0pZiyNE58m/MfbkTfmgyBDnPs+Qle3SVM1Yv/zEUSH1FrZQ1vUKIJGB4NMgdNhQ7n1Ze+2vORruGxRkdf4P/K9ievHq1i6r4oGk433Z6qZePwkVYbuolu5wtuO2QfRt5rf4RZ4ZKnU9ygqQMP71wzntilpyAQQEPh0RwUNHbY+v4ggrR7VJ1hhGDUKn6wz73vn2b/aDG59JDK7r0u12SnFOe0uNy+sKqR/TCDTs0+ttiX2uB41NW1WVh+u44ubRmO2u4gIUBMd6DcuP4df1cCIoujNmRUE4U2gq39sJRDfY9c4oNozHtfHeM9jKgVBUACBQLNnfNJxx6w7Xb9DT5wuN9WtFo41mgnVq4gN1vLYOdkU1psJ0CjoHxNIoK5bemX5wRq2lpnpsgcPzlJhDv8TkzIaUYoOikkAupMAEkK0PDoniy92V7GtuIlx6WFcOiKB5Z40X8HtIi3CwL0zMmnssBGsV6FTyjBbbR7dMB23TEpF5pGk0auVxAZpuOrdnYCkvByoVXqNC0i1Mj8eriUjwohFAGTwyBd5VHieKjcUNnL9+BRv07T5w+NZtKPCq4j86c4K9le28eqCIRyqaUfn6c/xxsYSNhU1cm5ONPUmm0/Lgm0lzYxODeOlNYWIomSYX1pbRLFnFTU+NZQLhsbx3cEa2iwOHp6d5bNyE0VoaOtgWmQJKlUTmJtoapWxcfA/KGvuZHawhfpJ4/h3vop162vJigrn7lnLiRZaKB2XRaDcjlqh5/Pc90kSqrEJGpJiIon9ZqEUSDdEwoCLen8ADBH8QVjMzfEtuGrjuXrkLN7dLn08DWoFfx0NwWs+7t5fpkDVXg7tVRjbP0NXt4vWOW8RWrYcQjIQLM2w5WVJFh8QBl6Me9RtyKr2gDaY6hEPo1boJa04UcQdkSPJzLhdIFdxwBpOXLCTWQOiUMgF8mtMfLOviqn9Ivh0ZwVvH1UxfNTd6Lb9Q5qPIMDYO2C3JwemYgekTYWiHm5cTRD0SPBAHUCxMg2QXLO17VY2FTZy7XgpEaPZbOPxc7IpapAa2I1ICkGnkvt0Vb1uXDJadd9ZYklheu6dmcnTK4747q+USUYMEXPmExiXXuHVI3KEZbPJHA90u+d2ljafsoHJjDLwl/P685dl+dhdbiKMah4+O4u0k6hy+Dk1TsnACIIQCTwFxIiiOMsTlB8tiuLbP+VigiBEi6LYVQRxAdDV/vEb4BNBEP6BFORPB3aIougSBMEkCMIoYDtwJfByj2OuArYCFwNrPNllPwBPCYLQlVoyA3jop8zzVDk+Tfn83FhMNger86WnYI1Sxle3jPWKNA5NDGZdgeRCizCqKWkwk1clA+MAZIJAotH3ScrhFnl6xRFvAHZFXi2ddieJQRo2AwqNlr9+uk+SPlEr6LQ7cYvw3lWDmTssjpQwvaQlJUiy+KIo9WufPTCK5QdriQrU+MQlujhQ2Ua4Uc2ByjbsTjd3Tcvg7s+7CzoX7yrnoiFxvLu5lOQwvVdbrYv8mnacbrenGVgo20qa2FgoxUKMGiUbChs5nnaLA7VChsbTk6bLuACMSQ/j7s+6r//i6kKevSiHl9YUUt1qYfbAKG4YomV3lZ5GZyiJajNhkUHc/nkjJpuT2PmZvL61jt3l0vPN9tIWrq818edz+3PXqiZAy2UjFJw/OJEjtTHoVHISOo8gdHg+quYGhITRsPej7ta9chVC/wvQfHM7Gk/74XvGyTn33MG0d1pJDFaRmPeir0tq0KW4aw9h7jcPff1e5M2FyNvKpdTf816Dba97jQsAB5dA2gzWTvuWJisMD7AQ07QbCn8AmQK1UkvnJV/i3vsxClcnYVHxTHGYeXvTMWxON8MSg7lhQgrVnjjg2KwEnms4i+mThhHibkEfEkP8rqekYk2AknUw4T6swZloCr7FHNqf1sG3EOUoR540HjEkhYqki7jkC9/2Cj3bIO8sbeG2Rd2NzqIC1Lx86WBWHKylqrWTs3NiqGuzIDtBuq9SLuPykYkMSQiissVCVIAGnUrOPZ/vp7jBjFGtIHfeYCZdvRKx7jBoA9lsjufJpb71XukRp95wTKtUsGBkIqNTw2izOIgN0hDlX7GcFk51BfMe8C7wiOd1AbAYOKGBEQRhEdJKIkwQhEqkzK5JgiDkIrmsSoEbAURRPCQIwmfAYcAJ3OrJIAO4me405RWef3iu/aEnIaAZKQsNURSbBUH4C7DTs98TXQH/00ltu4WHvjroE/9Yuq+Ke2ZkeA2M1eHmcHW718DMGhjNkj2VlDV10m51MCQxCJkAr68rxukWGZMayovzc7n38/04XCJ1bVYf4UCA9QWNLLuuPzdlmmlw271ZQh09VgSdDpGRySHEB6upaXOw7EANyw/WEGZQcfXYZC4aGsfyg7UU1XcwaXwE6woa6B8TgEwQyKtuY0hCMK+tLabd6mT5wVr+fG42Y9PC2FzkMQwiTMuKQK9WkBbeuwUuQLBW+mhN7hfBk9/lkxZhYFRKKHIBxqeH8c1+32LL9AgDbjekhhm8Wm0gybl3ZQ91UdbUyevri7hrWjoWh5tso4UPdtXx0QEBSVZewb/OMfDVxRpMrY20q/U+xYQA7VanN8YD8PGOSqKCdPx9pRTsTQvT8PbYZ0js2C9J31dsp/3ctxGaixHcTsSwDAxHv0SIHwHFawAw7H2TwZmzYY9nRTDyJkxp5+CoO4IyYTj1QiiflqhZW+FmYtyVXDqwlGiNgeaB16JXh6Cuy+N4BGsrE/fdh1MfhTL4Alj7FPaE8eB2oVr5MKo5r3CD9XpazHZuAt7e1N0Rc1dZC4khGh6bkcgVo5OobLFQ0WxhmyONunYbCYKWiSOeYUDbAklVQB/OUXV/Hs4PY3DsbI60uNn0cQfvzIxkiuhCaCzgB5UGQbAya0AUnXYXm4saGZfY/RnYUuz78FDbbuOZ7494tPAU/GXZYZ6fm3PSIkmDRsGI5FBGJENrp50Fb26n2NM8zGRzcu1H+3n7ymHktQehtynIjg4gSNdCs6evTG58EKNSeieOnAy5TCoS9XN6OVUDEyaK4meeWpWuoHpvmdUeiKJ4aR/DJzRIoig+CTzZx/guYEAf41Zg7gnO9Q7wTl/bThcmi5MGk63XeM/gdla0EZ1KzmNf5xGoVTI9O5LFN4ziSK1UGNna6fDp9dJVy/KPuYNo6LCh76NXhkGtwCo3sLrYzdiBKlLD9BT3WIUY1QoCtCquencnr1+Wy6r8Bpbuk55Qq9usPLk8n1cWDEalkILlDoeLf8wbxPKDNbhFeObCHCpbzT6B2cU7KzlvUIzXwFw/IYUPtpbxw6E6Prs6h0kZ4d6VGUh6aFrBzq2T08iN0XPntHTKmztZdqCa+GAdN05MwelysTyvHo1SxhWjEgnQKLhufDKiKCkw/3Co1vt+hut7vw9GjZKCWhPfH6rjgtxo3JpgQHpfYwJV9BOLSFrxMOiC2TvpvT77wQRqldw+Rap+/z6v1seYFzVa2aMZQeLBl0CQcWj6x8xfXEV8SH/UCjmHatr45MLrGLbznu4TRudAU7eL012yge8GXUVLyCTC3DI+2FrBQU8xYVE9RM4aTUNhB8sL5zCwXc2tc75j4MpLpDiIp2WCqDIgq92HKmE05or9bJmwhDcPy1HK4MbJdzGieRsLR0+jqdNOZVNvleFVRxq5fnwSdrcDo0ZOWZMZha2VoaGwsayViqYg5s/+jtrKIkIiYrj123oaOtrZ3SNNRhGSBMqZoA9nmkaJW5PMkj01GDVKXpqXzSD3YUBSR87sw600IDaQLI9+2cKxyQxOCOq1z4moN9m8BZhdiCIU17fzD0+L8mlZEbx91TCazXbUChkZUUYijP8546vN4sDqcBFuUPdZn+Pnl3OqBsYsCEIonmC5x2XV+9P8P8TxacogFVt2BeCVcoGLh8Zx88d7vNvf3FjCB9eMoMPmRCOXcbCqldGpoYxKDsEtQlVrJ3q1nIYOO82dDhJD9QxNDGZ3WXfg/8rRiXQ6XDh0EXQ65Vw2KpHvD9Wy41gzWdFG5g2Lx+5y8cjsLEL0am+qb08qmy2MTQ1DLhMYGB/E7Yv2Sg2eBFhzpJ6HZ2f5SO7bXW5CDUrGp4Vy1oBo0iP1HgXcECxuOUMTgxgUH0RNm5WYIA0NJitRKhvnRDYRKHNT0mDlm/1SbKK1s417PtvPG1cMJSsmCLkgEKxT0mZxEhesQRAE4oJ1nJ8by7cHqrG73IyPcfNRgIoaT4tkuUzgnJxonlp+BKdb5O+rirhmXBL3zsjA6nQzPFxEXXuAzwe9za4GORNMLm6amMKLq7sLSCdnhrO5qIlPdpSjkAlcOTqR1HAdlw0KpMkisqrIhNVqo7r/jcgE2Fdn5w/T0tlb3ord5eaeAZl8X9nJYEGBHMAYhXvQAmRfXCtdICCWXVMX8eEP5Ryp6+CxOVkcrOmOEUzMCGd9YSObi6Sq/KpWC9sqlCy9+GuSqpeBUoc1KJ12bSwRE+8HXRhbFBO4/vPu4PbmY/DGFVdRWN2O2y2SHRNGclg9Y9PCCdAoWF/QgFGt4LmVRaw+0sDghCDenGgnaM0DKI4c5by4MexPfIxGMYI8MZnANgXXjE3C5nLjdIuoFTKO1bWRECDwXvvZxMrV1Ne38bfvC9Gr5Djdndy2uIXPFyQx1DOnsWlhPrIvIXoV145Lpn9Md7JLn7RVSWnmxkgwRHiHA+U2IoxqHzccQHJAt+dgY2EjV45Owmxz4nD13WKgJ06Xmy3FTfxtxRFq2ywsGJnAghEJvYL9fn45p2pg7kaKeaQKgrAZCEeKe/zP0pWm/MAX+zlQ2U6wTsn9MzNxuEVmD4zqU3PL6nCz7mgDi3aUo5QLPH5Of6parfzTI4d+9/QMvjtYw1HPCueuaemkRxiYmBGOxe7CqFGwtbgRu8st1YbM6sfT3x9hQno4t05Oo7TRzFPL83loVhZPrzhCUpiOO6el88raImp6pGIG65Ukh2hQKxVsKmzk/pmZnq6bEBmg5nBVG4Pigry9ZK4YlYBCJjCzfzTlLWZ+PFzns2L505x+rM6v50i9mUi9nL9MC+e7/Fbe3N3JXVMiWXbAN+PH5nSzp7yF/ZVttHU6PIH7fmwqbECjlKFRyGky27h5UipahYzkii/5ZHgw+4VMOpwyoqJjeeKHMm+mUVKojnCDmtfXSW69hBAtd065mEeXHsbmdLM4r51bJyTy8vxcjtSaiA/RUdVq8SoWON0iormBGR3ruLDhLezaCBouvpMf2g1MW5+GUi7jlomBbCtq9Kah/3i4jr+c1x9X6j3IWwpwhaSzrCGM0AmfoHObMEUM55ZPjniLQ4+XRxkUH8hLPQxe1z7F1Q0keQopVREDCcs6G9Y/iythLO+7B/X6HH6zv4YjtSaK6jvQKuX885Jc/rg0j5ZOO7MGRHHOoBhu/EjSXbs83UHY15eDJ2akaiujwSrjrqV7sbvcKGQCj83J5outpRR5XFI3jU/krT3tfLSzljk50bR12nnwrH40dNhQK2SoFDKqrQJD64+AIJAYnMQ7C4dztKYNh8tFelQgibIGyFsF1naIGkB7UH8O11to6bSTHKojvXMf8i+vBXODlJ590RvgackcKTbx7AQ1N3zv8LqDrxtiJN7ZHfe7Z0amt3MswOCEIJ66YKDXNX08h6rbufq9nV739qtrJRf1AzP7+Vcyp5lTMjCiKO4RBGEikImUBnxUFEXHfzjs/z1Z0QG8s3A41a1W1AoZt32yh+IGM8lhBmKDNL3iJwAOlxutUk6YUUW71cn3ed2KuqIoeo0LSK2D601WPt1Z4c2aumNqGp/tlEqDIgPUxAZpWV/QwHrPDV+jlNHcKcVmCuo6+NO3h7h9SprXnZAZZWRwYCcX5T2KMzST/2PvrMPkKLN//6l27+lxd89k4u5CQoyQENzdWdwXFll8cVjcLQQNCRZ3d5nJuLu3a90/atIznQksv73s3ru/5TwPD5nyrq6u877nfOWHtJu4+6vDwfKRUi7w2KLBxLTYUCkEZubFEK5T8XNRM98daOTWU7JDkgvAM6vK+PD8XOSuTiJ1Ap8esfLSNqntVdrmwqxV0nnCCzbWpGWdvZUIvYqrJqdx1Yd7grBmjVLGQ/NzuevroyhkApcvjCbth5tJE2QgU/Dj+M+o6ehLmIuGJ/LMzyXBF0ZNh5MX1lVw2tB4lu2W7tUrG6v55GITxU1WTFpliBxOlEHNlZa9GNY+BICKChIaziNv2me96Cc/j/9YHFRbPn6en440ca7qbahYgxwYNf4hTts5iDabkRun+0KUB/ZUdzI9NzpIMPX5RVRyWQjTHkAj6ytNyloOQf586Q9LOurugcgrlUKGp/e7c3r9LNtTS1yYhlabm+8ONmJQK3h/gRmZo41sU1swuQBUFd7M7T93Bq9B4qgc5bqpmbywRnpeXttUzW2zsoEm3N4AswvieODbw8HvKlyv4pN5Kvj7Qkl9YMRlxBacQWz7TqnMpxsFK27tKx0KMpzzP+DirxW4fVJSe2OuienHiaNdVZKY5hVrwRQHGjOTK+5kxYzF1PgjCVe4yW5ZTrnpTm6c7iPGpGFbeXtIuXpfTRc7Ktp/McEca+rBHxCRywSUcqlUvL+6k321nTg8fpIsuj8Y+79T/Ca6qyAI1wMGURSPiKJ4GDAIgnDdv/bS/v+P0hYrdyw7yGkvb+Hy93dzyfg04sxayltttFo9zO/VvjoeaoWMCZmRnD82hVMHxaGWy1DK+0ZMJ2L/lx9oIDVCz/NnD+GuU3P5+wXD2XCsNeivUdPh4PKJaUHETLxZwx2zc1i6s6+A7vWLWHQqHpifz0ML8nlxtonsT8eiqN6ApnIV+6vbQ3oTXr/I5tI2JmVHcun4NFIjdKwqag4KaHr9A5Omze1jXbWL0z6pp9imDSYXkIQMLxmfGrJ9QYKJww3dHKzrxu7x8+H2mhBpDpc3QE2blTizGqNGgU+hR0wY1ct09zBILCHJ0ldj9/oDA8im1e2OEHfGGJMalFpm5EWTG2vk1II+COvpOWrij57QshMDpDiOoOunbHCkoSfU70f0Iwb8oJDOE3/kLS4fYWFcegQaZehPa2NpG6mROp49PZOLR0YzNDLA9ZNC+b+jEnXkdG0+4TqkJCW0lzIrPzrETVIpF8iNNQUh4iA5O0Ya+hroPx5pIrdjLRM3X0y0OxTt1yaEhQAdQPr+/SdI7B8fKPnFAF/vqw/5rjrsHnY1C6DUSXDpXW8iVqyDDU/Cmocl2Zp+fSnEAJHbHmVJvvTM+gIi925005J/ed82PQ19yDZXN7Kc2WQffp6ZW85j+K7bOZx3M2d+UMJLa8vYUdHOkYaB1fqy1oGSR8fDoFZw9eR0bpiWyYVjU3nqjMEMSjCz5LVtXPj2Tua/tJnt5e2/uP8f8dvjt5bIrhRF8ZXjf4ii2CkIwpXAq7+yz//qsLm8PPjtEbb2Poh1nU4eXH6Ee+bk8sjKIlqtLgbFm7h9Vg7rjrWgV8m5YGwK939zOMhUTjBruGF6Fs/1ylSoFXJkAiE/YIUMPtxWzZ6aLu6fl0eMScNts2Jw+wJEGtQcqu8mP97EnMGxJIbpeGFNyQC/caNGKQkF6hSoOkoluRHRByo9XbZQyKlCJjA0KYzSZhsur0T0PGN4Ild9KJVZJKVbWcjsbHiyhaJev3mrL5Qo12bzsO5YK8+fPYSjjVZiTRrabO6gWZlaIcPeDwF3PJwuN/fNyaO+20mnroNVmY+SlNmIKuCiVp7E82dksKvOTmO3K4TMejwsOiWO3uOG6ZRcNzWDy97fjcsbQBDg5hlZdDm8bK/ooNsNAU0Ysu7akGN4FPqghD9Ifbdd/VSgrxhuRtkzGuILwRDNEfVQKisU+AIewnVq5g2OZWWvLE+kQcWNiZVYVt/GYoUaKl2Mnvc6eWfnsLnGRZJFy6nao0SufK3vAmQKvOY0iif9HbXgJVPdxd1z8ihttiKXCczIi+b+b0KRZ5Ozovh4Rx/xNsWiQmetkv6o3IhYsATh8BfS5wm0YFRnhXCS1AoZhn4cFUEAnVJ6LvUqBZVtoc8LQJdHBgq1pPgMCPW7ITJbEuL0DQTCyO3NxPcTzmjqcWHVJRLsvKj0oO1lGSi1sPFZyJ4FBYvpNuXx0Jqm4KzrQF03Y9MjQrhcAMOSflkAMyVSzxM/Hgsm5ttmZfPOlqrgepvbxx1fHuCraycQZfxltNsf8Y/jtyYYmSAIgtjrHiQIghz4bToM/0ujrtMZTC7HwxcQMeuUvHL+MLKjDFS2O3hpbSkmrWQDu6OyPUQGo77bRafdTZxZQ2O3Cxkijyws4J0tlTT3uDm1IJbJ2dF8vlsaze2t6STaqA5CaQH+NCOL9CgdxY1WfAE/N07P4u6vDgXXj00LZ1NpK1/2erkkWyx8MOYhUrfdBy1HWTTIxbf9vLSumpzO25srqevlTqjkMp49awhPLB5MeZsdk0bJ384cwivryihtsTEtJ5pFwxK47hMJzPDm1npumJJCfY+PGLMGm8tHQ5eTVUdbWF3UzNyCOPbV9oEWdlZ2cMP0zBCwhCDA9HQtN313lG6nF+P8bO796RjSI6cCupmRV4fT7aO8zU5RYw9XTkrjzU0SRFcpF/jzvDyGq2o519SNL3Y4539eFkyKogjPrS7l+bOHMCo1HLVSToPuNhJ/uKTvRugjOSrPxeOXRsJRBjXTc6Jo6nbh8QW4eFQ0Y/bdBeUSar5i0rOct8pJt7NPG+6JxQUsyVbQZnMzJUHA8uWiEEfKsG/OZ8bZnzCr50VqE6/kqD2SmJFXojryBaI5CeeUB3ijIpwklQ27T8DWIqMwWUdVUwcKMUBkoI0bp6TywroqXF4/l01IxuUNBBGASrnAXcPBsOZb6YQV6xBy5sCSd6GrhqjoAu4ypvDUTyX0uHzoVXJumpHJsEiRl2aoafNpiAkzMNrSwmkzavBpnaxNHMIj3/c9f4IAYy3WEB04wjOg8aA0o1EZpI0ishC14QiN+2jPO58vjvWVTEcmm4hu/ar3gDKY9xyE9ypohyXDjAfguxsBsI97kIr2vllkTYeD+YVxTMyMYHNZOwqZwAVjUxiV+ssJpqrNHjLrOxFdCFDbIfWI/kgw/3fxWxPMT8DngiC8hoQkuwb48V92Vf8B4QsECNergtj746FVyvH6Ati9fl5YU0JOrIlEi5Z4s5aVBwcabZU1W/lgtoyOThcdFgW3fHWUWfmxTMlWsbG0jcP13czMj+b7Q01Mz43m9mUHQ/Z/ZV0Z103LCKoR3zozi7cvHklVm50Ig5rqDhseH9w0IxO5TEZps5WDigJSBQHEAKPKXuTlM//Gq5sbehFdqmByAQlB9vGOahYOTeCVdWUsGZ6IL2Dg0vGpBBAQRRFfIMC9c/LodnnRq+SkRehZfkhqcAvAn2Zm89IayTNmc3krt56Swz29SdDtC3CgtovnTs/k/d3N6BRwbaEMlVzk/DHJqBQyrN6BjdfD9d1Mzopia0UHzT1uLDolT51RSE2ng3CdgnG2VcStu0XaduLLdNgH8iJsjr5G/97MKJ4/6ytMld/3jqDDSYrP4JGFPuS9Aoq7qzpQKQRijGqGWzeiLe/TLisSU+l2hsqs/OW7o6xcrGNa1ZMgTA61Owbw2JHX7YLyNQix07hxfSYp4fM4PW8h0/IS6Ojq5IrAMgx+AfxuGjXZlLln8tiQNnB1EmirYYgpkVnDq/DLdcREttLU0sLIWYOxB5Rkx5rJ/emCIOQZkAQzi1fA4S85PPFNntvm5ZzR0n0OBAKMUNczfOO9yJoPgz4SZj5MW0MHTcRidLuZq9yCfN5E3tpaQ5hOyS1TEhmy//bg4cWILARtmNSwBzi2kgNn7+Tzg51Ud/s5Z6aR3DgTS4wSRDjCoGJ8mgWj/1oYukhKKFG5QYdMAAafAZFZ0FZCpCWT+V06vuynRffq+nK+uW48N80IoJTLyI01ov0Vl0q728+lE1IJ06nw+QPkxBhJi9RT2Q/unxVtINLwXz2G/l3ityaYu5BIkdciNfl/Bt76V13Uf0IIgsBlE9L426pjQTjvpKxIHB4fd3xxiCUjEtGrFEEXSJ1KziXjU9l0nKzYG/OzNGRZN4HMy/u2QlzeQBDSC1LJYnK25MHhO8F0C6SyTVa0keunZaJXybHolFz5wW4ComSAlhtn5LHvi4NEzImZEagjUmHQElCo0EflMsS+nbTIXBIt2gHlNYD6LheZ0XpumJbJqFQLj39fTFGTNQg8uGBMMskROl5eW8ZlE1Kp6XBw+rBEajscxJo1lDRZOWtUEs+vLmVKVjRqucDdp+ZytLGHMJ2SJIuORDO8OM6BiECtPJmvS718vqecgCgyf3AcF45N4cPtfaWfG8ZFMkRVz4WWbqqJ46NSFwXqRmKFRpKNYcQtvwtb9mI6o0cRplUFZ4nHQyZAfjgsGpZApEFNuF5Bd+33mI58BV4XeB1kzzWTIYgEAgJVXZlcmuhD1fYucqsLeeoiiB8O7WWg0nEy8JFcEPCqwyS5+4wZUq+mv9aXXBWUmI8t+ZS7p7zEvHgbZns5dqWFVIrQ4YJ9X4I2jLgRiSjlNVC3G7x2ZJY0XB4nEYVzEL1uZC0Hid/+EPExBVKJqUmDb+j5yLY8I80m9FF4J9yGcrnUPo0ONOPxmVl1pJERcUpSwlQM3XMvstbespu9Db67kaLpX3LhCgcyQccNo01cPKidCQsjUStkWDq3UT/8Djzp50uzj8gcNB3HSNFHgdfJgVFPc+4nFUGpmM3lndwzJ5d3NlfSbvegkAm8efFIsnNGDryBx0Olh5RxkDIOFXC90UaPy8+qombMWiV/np9PbpwRjbLvdVbX6QgCZnJijST2gyBnxRj4eEc1uXEmwnUqXttQzo3TM/nL8iP0uHxEG9U8teTXyaB/xG+L34oiCyAJVf79X3s5/zmRaNFS1NjN7bNycHr9qOQyylpsaJQKbpyeiVIhY1pONLuqOxFFyU+iy+HlgtGJfNpb8jpveBRTu76C/ZIrYuRp5w04z7ScKHZXSeUHi07JnIJY8uNNuH0BjGoFkQYVty87gNsXQCbANVMyyI4xUtxkxeXz8/W++hCW/+ayds4cHs/+mEUo8JFW+iEWUyKH6uPZVNrGE2cUDriG04cl8PW+Bj7eUcPIFAvXTs2grtOJIEgyVZvLWhmbIc0QpmWHs/xQS4j8+fDkMC7ubfQXJpp4dUMFZS02UiJ02Fw+2u0ebpyeyd83BJALAldOlvFpPwLqdwcbuWl6JhF6Je12L3dMjOCs1pdRH10mHVOpZc7C11Esv05CSU1/gEMzPuShvRr2HLIzMc3EkwsSufWbMtpskqHUY9OMpPmqSNObidaLjJYVk7Tj4b7RfvZs1J5u2PMuyJUUjL0etn4KtTuk9eU/4V70Dh+3ZdDuCDAqOozLJjgw69T4/AECIqSaBDLVjTD9z1JPYsaDsOrP0jlkcph0q6TUDChsTVySWIfsyysk6f1Zj6HsrpXOD5JEzfrHCD93qdTvkMuhrQR16gSExOEABDrKkJ3+KpStBWsj3tzTeK81l4TxYzCINordFsQ6JVeZEhE6K0g58CKrznkP1bHvCK9bgyfvTyh2n6AmEPBjcdUBElfrxR09jElJYsLySSCTs3bWKi77qAk4/jKu4C8LhpA79VP0SpED9nAcntCZ+9ubK1k0LJ63NlfhC4jc8+Uhlt8wgWjTPyZHAqRHGXjx3GE0dDvRKuUDVI5Lmq1c8u7OoP9PQpiG9y4dTVYvCVSCcMfxwbYqWm1uTh0US2OXky+uGR9UhI77Qzn5d4nfqkU2Acl7JaV3HwEQRVEcaDX4XxI+v8i0nGje3lxJUZMVs1bJnafm8Oq6UoqapLr9wqFxvH3xSPZUdWLQKJiYaiBn76NcfMGlIFeS0rKerg6BjZM/wR2QE6eXSlnvbq7C6vYxJi2c80Yn4/R4abd7SbBoiDFpQnowV0xMIzlcR2mLjYAIr20o508zsyluspIXZ+KHQ00Drr2u28PtqwW8fgVn5N/EDTkaLo6KxOUN4PMHuG1WNh9vr6HH5WXh0HgMKjmVTi9DEs2khOuQCQJhOiXdTi/xFg3njU4OosyUCgXf7m/glPwYcmKNNHY5WXmoEaVcEiuMMKjpcXqRCVJ/R9GLovP6peSSHqWn6ATmNkhkurtOzaWuy8m5MRWov1rWt9LrRLH+EciZA4eW0WgewmUroNUqlTw2VfRQb63i07MSsdYcwCJ2kVL5LTLVbG4qfQt04QQGLe5LLkotxI+A1Q/2nWPFzXDKw1C3M8hAle17nzWeO9lS0c30dj8xRjXvbJG+mylpeq7P3oPio0ek7eUqaf8Zf5ZKZTIlHP4C8hdKx596J8K6x2gbfBV1hgJSIxIJ23HCeG7ynQgrboGu3plcVA4kjOCjrZUcaujhr8NSkX26JDhLUlZvYc7ExzhtWzYddgVgRaOUMW/SpSRu/TOkTyZ2473QuB8AVcsh0EX0WTL3Ro88tJ/RYJe+MzFuOB8dGai+vK+qjUt8j0DjQfZOXjdgvVwmMCIlnLc2VwG9TX6Xj+iTo4pPGlqVnIyok0u7fHegIZhcQJqBH6zrotPhweML4PYGePKn4mDl4buDjagUMkxaOdmx/10Cly6vj6peN9nkCB36Xykt/jPxW4/2NnALsAfJ0vi/PtrtHu75+hBzCuKYkR/D6FQLb2ysCCYXgB8PN3PG8ERm5ccgk0GYTs3y5Dt5Z1U1IiKXjj8dU7qC4mYb/oCIvinAqCQzjtFJaJVy9CoFn++uY0UvYfONC0fw3taqkOt4f1sV10zJoLS3lxAQCfIi3N4A4zMj+OlIqBBgnKlv6v/lUSsjsyUeSVK4hklZ0SzbU8u8wXHo1QpWH21mU2kbV0xMp8flY9GIRF5bX8bmsr6X0J2zc5iSFcEN0zMRAyJ3zs5hxcFGVh1tJjVCx52zcwn0il/GmNTcPDOTLoePQ/XdWHQqksK1GNRy5hTEoZSDUavkxMiLM/Lz0WbsHj+XGQaqE9BWCnnSy7q600WrNbRmVdHqoN3pZ2zt+4gxgwhMvRtsjXDKQwTkKo76EjBNfJLk0g/wps1AUb5moNx8wz6IyJTOBYjIg/DosWnhPNZPAfjSbDe6DQ/37ev3wManoPAc2N4LvlRqIW64ZEQXlU/RsD9z3RYtVe0u5ucJvGROQDiObAtLga4ahK6+MiGtx6C9jGbvcFYeauSvCVUD5PYTD73KmXmv8/puqVcoiuCPHwXjbsQfPwz5ocv6Nj60DMbfKMGLRekZ6h58KR9X6IG+pvhx+TnB7+Jkkw6z3A0ddeC1EWNUY1QrQpBqZ45MxOHt+3tEShjRpv9ZOaqu00FVmx2DRkFmlAGFXKC204lSLqOpO9RP5sKxKSzdVcfOqg4EAR6Yn88JSGx+ONzErafk/I+u4T89mrpdPLfqGJ/vqUMUJYmne+fkkRj++yka/NYE0y2K4g//eLP/nojQq0gO1wX7Jc+eVRiChAJ4eOEgSpttbK1ox6hWMCU7itu/6EN43fnlIZ49cwg2txeVXM7KAw0UzMllUJwJm8tHQriWT3b2cRdOBuf1ntCXUcgE4ozSC9rltDEnP4pWq4e9NZ1olDIuGJtCnM7PmYNMOHzwQ4mVXTU9TMqKxOv3MyjexNubfXzaj0tz7ugk3t5SQW2HkwWFcSHJBSSgwfDkMHZVdrBgUDSf7qwJihNWtTv4aHsVj54+mHvn5hJpUKNVyLj36z7omkmj4KklheytqUKjlHHP3Dw2lbZR2mu7nGDWML8wjg+31+D3e3EZT2IOlz6NHsGAdfzDKCLSEISqkJeITAC5IYrdo19gb6OL+W1lxP94LYgBZEB68lReD7+DGsNTWJw67jNUMYDWqLWENOpbBl3Kzq+k2ZZeHjruMvtOwqNwdIC6d5iu1OGb8zd8ShOe1NkE1PE8dljGzBSR0cO8WAM+Ki23kF5/geTXEp4GrcUDj9l6jAbBSVqkHp+gHPiDlinwiSJapRyn18/lY2KJrfkCSr6D2MEnXF877P2AjiVf0NDYgE0RjtOSzc7v+pLaGcMTUBznWDcd4vyZDr4r6rNOUMllLEq0QnEvDD3g4PppmVS02WmzuRmRYsHt83O4F0mYH2fizBFJAyyKfy0O1HZx6Xu7ggCbp5YMZlNJGysONaKSy7hyUjoz8qJZU9SCWiHDrFOys7fMLIp9ck79I86sRRD+x5ZQ/6MQRZHGbhciEmdN+AVF6X9XbCxtZenuPj/H7w81MSIlnMsnpv1u5/itCWadIAhPA18hydUCEsP/d7uS/7CIMEiNwKs+2EO73UOn3cukrKigSOOQRMmo65GVkuFUQYIpREDyeBzX29pZ2cGZI5PQ+nuY5duGItBDtyeLe2flsrO2B51aQbRRjUmroMfZd5yEMG2QLKdXyXlyupHJTS8xae5IosLlNLcdZErGMCZmRSCKkKnpIc9ZzLjOJ/HJ1Fw/+0aKdCZWHu1AFAXCdUpumZnFaxsqcPn8zCmI7eXRSKNCp9fPuIwIxqVH4PFLqJ1lu2txegMERJHiFkcwuRy/pnPHpHDROzuD3JgZudHMKYjlh14Vgx6Xj0P1PTR0OfEFRMqabYxNC2fxsERApM3m4VizlR0V7WhVCjbbEhk3+UkStj8CHhvetOnsG/Usj/5cRVGLk6fPMHLmiCQ+392XJM8elYTc72L41msYFDcarS5R6oP08il0NeuZnnYBC7frgG7OW3IRmSU/9lkTq42QPVvqhfg8kD6VMLMJpdyJ2xcgxSB9J0Jv6a9FHtPn1XI8zIk4U2eyy5dPc8BMUZWJW+1PYyr/ntol33N1vo9x++5CcbQYFGqaJz5K+xlfENG8RYL7ehxQt4v+EcicibpGRmaUgWZjASlqU5+lgCBQMesdsqxRXBfhJjtaz+CejcgO76AraQYmvx9yF0Dxd30HDEumxG7gnNVST+3KiV7um5dHXacDg0aJWaOg09E3I87uXM+nF1zGpuIGZDKBqRlhDPmpT4M2Rt7DYZ+eXZUd6NRy1hY3c88paeASMWgzqWyzc/83hxmVGo75JDPXE8Pm8vLY90XB5JIQpuVAbTff9SI03b4AL68r47FFBWwqaSXaqKbD1of0FATQqRTkxhqDorNymcD5Y5IRftEi7beFPyDS0uNCrZQTfoKbZofdzbLddby4phQRyfPonNHJIaTYf3ccV5XoHysPNnLJ+FTkv5Nkzm9NMGN6/98f6iEC03+Xq/gPjREp4Sy/YSL1XQ7CtCoEQaDH6WVbRTuzC2L4Yk/f6MDp8WNQD7zderWCuiYrXr/IJztqSFCGM+LYI2BtItySgWzSUr470Eirzc2+jAheOmcYj/9QRHGTjaFJYTw41UJ03U/MnRlLhNlMeMcG9oTPYXerhlMTUklteoir3MupT5yD2ttN0u43aZ36JLSXoQj4yGq+Au1pn/NItRwE+PFIM+F6JU8tKUQQpBnRNR/1jSMkyLUm2MRXyATun5/HgboudlV1Mio1PISIuWBIPO9tqQpRKVhT3MKtp2QHEwxIkicyQSA5XEuH3c3UdAObyyQZk2nZ4eyqdfLoosE0dDkxGnSUmhdRM3ciCsGPRxvNlR8dCCKVqjtcNPe4uG1WNm5fALVCxu7KDtxiBKtSbydK6SKrZiX6cTdgb61BZa1B2bALi2ADpPLANkciiUs+Qt24C1GhQYjMRvjqKtBHgVwJRcvRz32Gbeem43HYMBm7+f7cSFQdJei8HdiMBXjnvYjyp7sk4IExluYZL9JmU3LRGgVgB+wsPPcOhiTkYzGbiN55L4r23lmKz03M+jvoXvI5bHpWSnBzniEw6AxkR74EQYZ/2MW8VJvOJztrkQkwNDGX1PHvkd+1Dq2jkfohN3L7qh4ONfTd50cXjuFYWArrypzMlpm4IDuGtJh86KyUynD2NkyeZkDOlMxwfAGRNQfKmZngo7NTzrOH/TyxKJ+60z5HUKhAbWT0FxMYrTGDNoKK7BcJGsjK5IT7O7iu+BrmjboRh9xEUtNP+I9GsCXhUl5aWxR8pkza3/Yq6nb5qGi1cd3UDOQyqWf3wurSkG0UMgGdSsFL5w1HFMGokfP57lpJc06Ehi4no1PDmTM4Dp8/gEYpx+72hoAM3F4/Za02WnrcxIVpestwvyx8Utfp4P2tVXyyo4Zok5r75uYzOTsqaPW8tUwS1zwez/xcQkKYlkXDE3/pkP/yGJViCZGqApiQGfG7JRf47Siyab/bGf+XRYJFS4JFQpw0drvYoGjlxumZZETqQ4wAy1vtnDcmmZ+ONAWJXSq5jKFJYazox4/5otjNBckzMR/9iGOFd3LtF6XB8kNTj4tD9d3kxJqYmR9LRYuNxrYOhh5+gwRrA4Sl8N6Qj/nLjzWAl/hEke7UK0h1H6PNa0SvMSEb+yAN/gii+tngGpz13D9vNh5/gBijmp+LWnj6p2JUCjl3zs4JmTWVttiCpE2QyKUvrSljwZB4QJqRXT4xnVfWST2hSKOa+q7QmjgQosGlVsgYmmjm6inp6NVyCqPVXPThoWBS+mRvMx9cUEBi12ayPHaq2lKQxRfw8VE3ZS12FgzRMa8wLqg7tra4hbw4UwgY4spJ6Tz+cwUH60VAzU1jLuHsGJHLDkGWReCKU+zkpiZwNMmGX6HF1V2HZulZoNRJZmPNR3DEjKIq5Qz8okBWYQeaYz8SXr5aOoE+irxp9yNs+FPwnN6Fr7N83FJ0vm6OOU28+bWDvy0MbSK72iph0zMY5Gqo2TjgPmmdTXjnPIvMY6PKJqcp4xbyh1wBCOzv1jNOKTDV3IRc9OE3dnPveiWdjmmYNEpuzAnjUEN9yPGe/rmUM4YnUNfZwds7nBxt0vNGeBXGxu1w9Fvwe0k9ay4bZ1bitaiosHmZ3PoU6i2bQG1ixpgHKOu2kPjzxRIoYtSVcOoTkrxLWDJvFikZM/wt8iKVIMhIsu5H3naM9PU39PtQFtTRi4Pf/ROLBxN1grx+q9VFq1VyaY0za3H7/LTZPKjlMu6Zm8tflh+lx+VjaFIYWdHGECb/jdMzeebnY0FSc5JFy9/OGsItS/cTECWJp8nZUby4poyWHhenFsRy7ujk4IvV7fXz2a4a/vLdUURRSljPnT2U+YVxJy1rBQIi72+tDhJ9K9scXPnhbr68djzDkyWAxHHLjP6xbE/d/9MEMzzZQnaMgZLeUnRKhI4pvZSI3yt+M2RAEIR5wCAg+CSIovjwL+/x3xcTMiMI0ykpbbESZVSzYEg8B+v7dJLkgiRL0djtAhGGp1h4ZMXRkGNkWWRorFLNu9oXjj/Qx3ieXxjPc6tLQ3S31pfIyZ94E6lb76Yu7zKeXtf3IH9/uInLxqez5BtHkAOyaGg816a3BJFQDcNv44nKQSw/vB+QvDUWDUvE4wtgdUkKtvfOzWNXZQeN3S40yoGCi+12D4Ze75raDiebSlt5/4JBdDTXkhITYH9mOJvL+pjegiD5hoxIsRChVzE6LZzGbhevrC8jLUJPdaolZMYjivDhznpe872OrH4XWQoN5XM+Jj0yiRUHmyhusnLOqCTSI/VUtNk5UNdNboyeHZdGI9gaEI3xPHvQxcH6PnTaD5UBhuVkMmuQDYVchhAB6p3PIRxaChozhql347roB/wtx0ChpluTyLJaIwG/xCI3GxtJLr+37ybYWxGKvoWUCRLvBVCuvo/anPd4eisctxiusfeNgtMiNKTZd0p/tBRBdD609HsetBbklhR8HdX4tZHERWWw9rCd61fYUMgEli4USF9/A8JxrS9dOB+c/hmbu+JweHz4vQMRXlaXN4SEuK3aTk32ZAYVLQXAMetZOgMG5NHZyHTRDOvZgLp2k7Sxu4ekjbejOfMb8PYOGgzRsOYhqb8kVzLltL0Ut2p57LsaXF4/a+cYOLFlLBpiyU2K5p2L40mO0A1Ag+2q6uDmz/ZT3+Uk2qDmxfOGsXRnDd8fbiIz2sCFY1NQK+Xg8rG/VpJQ2lnVQbfTS6xJQ4fdE6KYUdvppKrNzqpbJuPw+rG7fVz67m5mD4olwqBiQ0krZS1H+fDyMZi0Sipa7TzUm1xAGkTd9eVBBieYTyqC2WJ1s3RXqM6bKMKxJmswweTEmlhdFFqSyv8FMc5/V/xc1MzwZAtzB8chipKu3LLdtYxM/Z+Ztf1a/FaY8mtItYNpSATLJcDO3+0q/peEQi5jSJLkjdLtcLO5tIO7T81lW0U7iRYtVe1O3ttahUmrQEBAq5KHiCnqVHKuzXOj/ln6QYfLXdCv1ewXxQGijg6PnzbBQirgletw9ZO98AdE3t9WHUIw/Hp/A6fFqskGUBnYoJ7G8sN95M/VRS2kROhZdbSJToeXHw438dBpgzhY341RrSRcpxygl5YVY6Ag3sT10zLRKuVYHQ6GH3wIY8lXIAj8edZH3OMJY29NN2E6JVdNSuf1DWWYtCrqu5w8urKIyyemoeotQdjcA4GKbl+AgEwlqbP6XCQeeIHCkS8F13+9r55Lxqfy+sYKEi1q7ks5imnZLZIWlkLDNZOeZbM5hvpuNyatgjOGJ3LZh/uDL5FtyUZeiA8nJuADRzvC93egWfwW/HgzKLVsX7Cd0s4OVh4qRRThcLaFB8c+RPL2flDmpkOQMzeYYHB2Eqnpu1Gj08KJN2uIN2sYkxrGNamNRGuyIOM9EEUchRei++oiCUggCIhznkL++XnIezW+xJgCxo95ir86vBjVClQNO/qSC4Cjg4hDb3J6y1FoPsKRc7YNMFo7JT+Gbf1cJ+UyAVnqBBwR7yILS8Thhag1d6Nq2gPGWFonP4618FKMB98N7mOwVdM46a+oVFoiHDWQeQpYUkEQSFXbuGZ1n6vm9+3xXBCRhayfmnLruPuIiY4lOUF6tmvaHWwoaWFvTSfjMiI5Ut8dnPXmxZt4fnUJ2yukAcqRhh4eXH6Ea6ZkBBWfn1tVwkvnDsMvikQZ1DzzU9/M9XjsrenkTzOzAfh0Zw1Orz9kVqFRyrC5fZi0Slpt7gGgA4fHT4fdc9IEo1XKiDFp6HGFCmya+5X9FhTG8enOmmDvyKxVsmh4qNjpvzsaupx8u78hZNn4jAgCAfF3sy34rTOY8aIoFgqCcFAUxYcEQfgbUsP/j/jFEEi0aPh4Rw1+UXqgup3Sw3W83PT6xgrumZNDQJTkK1ItSgo1B7GOugmnoCfL5Oe6qVmolQq8/gDZ0YYBLwyzVklERBRY0oimi8VDR7Jsr1Ryy4kxhvSBjkeNGA1jr4fwNNYWDWwyHqjtIifWGPxRL91VyyOnDcLq8uEPiNw+K4c3NlXQ5fCSEaXnxmmZ3PTp/mDZ6+bJ8ai7ez1gRJGcny/klbO+55NaSW2gsdvF4YZQ2RSLToXLG6C81c7Zo5JZcagxBAV2SY4Pxfrtwb/VXeVUNfSNCDVKOZMyw4kzyJgbZ8O09Lw+oUWfi7TNt3HrqM94fr+MmbkxvLu1MuT422usHMmdRAx9SYvWYjDEgiGGsk5fSClzTUknIyOHc60hBmy9Te+U8UFOCYCv4GwqPWZiTV6m5UaRFqmXNMMmpmEUPGRHKGHrZ1C1CTEii9Zxj7J/zFISAvVExiSQsu/poIAkgNB8mFzvEb6fHYZcoye6MbT3gCCjLnIiR2Kvwu50kS438sHCMJ7Z6aaiw8vCQWEkR1v4y4q+XsDFY5Nwuz3kf6Zmx/UWwn+4DGVzrxyRtYmoH66kc/FS6JdgnOpIZn0fh0mr5O5p45lpfQnthicByEjcyEOT7uLBTdJ1P7TZTtisFxmtqsbr6KRWmc6TmzW8luEhzqyl3ermls/3sae6C4Cv9zVwakEso9PC2VnZQWGiOcRaAXoHG/2+PLvHj1YlZ1xGJABzCmPZfIJ185yCPmXzk/Fn5hfGB6Vh4sM0A35nEXpViDp3/zDrVNw3L4/L3tsVTEw5MYYQEdbcOBNfXDOOoiYriCJ5cSbSf4HH8++K04bED0gw541O/l09cX6TXD9wfL7pEAQhHvACvx+W7X9h6JUC+bIaChOMTMmOIjPawPTc6JBt/AERmSA9yCIibc4AXzqHcU7ZKZy6ZwTfOAsRBIHnV5fw8toyXlhdwl8XFQTBAmE6JXfMzmGfmMWe2V+xynImo9IiuW5qOsnhOtQKgTFpA6e7USYtraKJVr+WwpOozubEGqnqp5rr9QfYVNrGX747isPr550tlZwxPJEbpmdy0/RMbl92MKSn8tLmRipyrug7oBjA4GzkpbVlrD9az9npLkb22uaq5DKunpzOzso+WO/nu2t58ewhzMgOY1J6GO+ck82Y4idDEFnOQWfz3oG+JHXDtAwmbrqYS9aOJsxVN1DF1+tkdJSXYUkWYswaFhTGD/jcPf4TtKcMMeDqArmCIw0DyZ/fV8twx0pejt644TiHXYEoV0NEFoHhF3Mw/Qp219qZmBXJ3upOUiP07K7u5OMdNSjwwrrHoUqarQrtpaT8eDGygIclq/WU2ZQIbQNH4vLuGvIPPkbOqotQRYXynGvGPMgVe1K45rtWbltt5cx3D9GjiuPvg8v4YVIFl0cXY1Z4efr0XK4fZeKNWWquizyAzGuXvFHc3X3J5XgEfMjsfagxR/qp7HQlYXX7qO9ycuPXFeyPWhhcr6zbynT5viAizB8QWVGt5Lr9KUxdm8p5P/iJ0CpR4aep20VJizWYXI7HT0eaGJceAUioReNJwDHG3pKsSi7jnjm5IS/z9EgDC4fGI5cJyGUCC4fGkxLRV6grSDDx5BmFmHqPMS0niuumZqBSyIP7v3ze8CDwIMqg5uXzhgV7rSeLCZmRfHXteJ46o5C/nz+cty4eRVJ46GwnPcrAvMFxzCuM/3+eXADGpIXz3NlDSYnQkWjR8vjiwUzMivxdz/FbZzArBEEIA54G9iLBRP6rtchOFm6vn4N13Ryo6yJcp2CEvYrrhGIqZUNRKeR06AdzxaQ0Pt9ViyjC1VPSqWi18dEOqX5795xc7vn6cLAM5vAGeGVdefD45W0OPthSxVsXjaDD4UEllxGjFzBby9B01xOnj6HYn8jz+xoYmRrO3pouzhqZRGmLjZoOB4IA54xKorg9wJ93j2BMejjTczXkxRmDcvtZ0QYijaqg5wzA/MI43thYgd3jRyYImDRK3t4slUFunJ45wDTLHxCxBfrNjBJH0SKPAnoYFy8j//szeSflFGoLTkXjt9OVNASFTGBYsgVBkBB38So7rxZWIcpVaIpeQSw8DTqLwG0lMOQ8yhIWc6pPTpvVQ1aMgTSjH9qkkbkoihKJ0dsPXKDUsa1FyfIDDSw/0EBBvIkzRyYGgQFymUCGqp8icGQ2xORLxEWfh/w4E8sPhEqejEq1sDn8RgKxF1MdiOAsOWgHnwk99bjjx/DS5m52VzvZXd3JkhGJ/O3nEo41S/dZ4/QiNO4LfYB8LlJoBMy8e8jH9Jy5yHa9GbpN9CDYKs2yhPo9BMbfjGznayD62W+YzLGWvmv0BUSe31DLyJQx1Lf3sLPWxpQsO8/IXkRdtxUOt9GVtZjOweO5apIOhxzCtRZwdoacUm6Ko3PO3xE1ZvZ6U7nmi6qQ9TvatYzTR0raZUBM23aSw0dxqL4btULGouGJvLq+HH9AJCfGwMXj4rnsw/1Utzu4ffZAcqMoEpxNfHeggRtnZPLY932zrqnZUcwpiGV8eiRatZzUCH0I8undLRU090iOqAKwoaSV97dWBmc4OpWCs0clMSEzApfXT3yYFl2/vpRMJnBKfgwrb5xEu91NtFEzQI7mxFDKZQxNtjA0+ZdVnP9/C4NGyaJhCUzLiSIgiv8S7bXfiiJ7pPefXwqCsALQiKI40OXnvzzWH2vh6n6Q3pTwOD4Y2snoTZdB4kgu4WGOtbpYMiIRo0ZJrEnNy+vKuGBssjRKO6HHcqKMuFYpZ+HwBF5aV8bB2m5GpVq4fZyJ5K8XSmQ8wDz+DjLCJvDTEckz40BtN387awilLVZkgsDGklbic7W02z2oFTLWHG2hIN7M7EGxgAREyI0zMSs/hi6nl8XDElh5qDHozljX6eDUglg8/gB1nU5yYo0DVKXjzRqSYrQw9DwwxeMzp/HMHmn/Xc0B3InjMBUvZVDxUjxpM9iln8I3++qp7+0VDY43kx8Vy4hNj0roJJmc2rQzOTx1BU5vgFiLidu+Lcfh8WHWKfn2QD2RejWFQy4jes/fUK++F++cZ1H+eLtUYlLpqZj0HH/b1McfOtzQw1mjkvh6bz2JFi1/mWwkX9wLU+6U/HIc7Yj1+xAsKSCXE2NUUhBv4nDvTCYlXMvZkZUoHM245HomGxtQr3gJOqTSoBZ4ZNKTLGrLpNXqJiFMG1Ku7PSperktoXX7cIuFuYOjabe56cw6k3B7K0LRcknwccLNksfK8ShegW/k1fQs+YZAwEtj88BRcWO3i2iDiuUH7CzJ03NRoQz1V6vBY8MdN5rX1Zfw948luPDX+9SsPvUZDN9dEQSBeIZciLb4CwwHPsaZNot3XX8acI4YoyqklOdInsrcxFim50Yjlwl8sL2Kv82y0Frbhj7SwLbyGp4cVI/B10GrPo7UCF0ICmx8RgQpETpumJ6JViljvKWHZafpqLApiFT7KEhQEGNWDZghHA+LGn6u6+ZAXd8ralB0aPWgsdtJcaMVq8uLyxsgL840AJ6bFK4j6Xdktf//GmG6f51q9K8mGEEQFv/KOkRR/K/uw3j9AY41Walss2PSKIK2xcejusPJXtNMUrgP2sqYNlLD+rIu3tlSxYzcaMJ0Ch5ZWMAbGyuwuX1cMDaZa6ak88nOGkwa5QDG8bmjk3htQ0XQHnbtsVZqOuwsLbiSiAOS/Ih+69M8c85M3iqPxaCW0+PycpyXIAgSsMDu9qGUS+zrKTlRIf4xTy8p5N0tVczIi0ZAQK9WhPjefL2vgcXD4/lmez06tYLX1pfxwtlDePLHYxxu6GFYchj3npKK0nWUsohpdIp69rfHctb4RDITu4g0qGi13Ex880FkPbV406azu6ozmFwADjV0U9cTBSoJ0ls19q9cutFEZYeErlP1QlUfXVkUJK92OT14NFJZBWsDAZmajjO/JmBrIWBK5tylrTT3hJbNnB4/V05KZ0KSkonrzoH2E0pS0/9M0YIVgEiGTE1+vIlT8mMQgblRbWR/e2afftm0e4PJ5Xgk7n6CK4Z8yOOb3SSaQ0mEr+73MmXMAyRtujO4zJ97GqUkc6yplbQIHVUeAxHmJDjzfUm6xd4Cax/pO4jKQGXaOeyvsePxCwxKVhBjUod8zjOHRLEwXeSUhHgMMjemw+9CwWLY+wEVWRfz+qq+0l9Tj5vrdsfy/PmrkHdVgj4SdesRyv2RlE9YjEklcmtsIjtreoIKEvFmFWMSNMHZoidjFsu683hyy7HgcadlR5Fy6CWyD39K0aKfuaT5MbT7pH5aokzOi4vX82V5FPtquxiWbEEll9HWbSNTqEejMpK75XYUTfsZdfyAggyuWg9xQzhZnJXh56uDQvAaVXIZS9L7BhcNXU6u+3gv+2u7AAmG/N6lo5iY9ftCdP+IfzyDWfAr60T+yxv9a4qaue7jvcHG3sy8aGYPignR/ur2q+GKNSBXE9lsZnqOH4fXj04pIz5Mx9W9TpEAT/9Uwj1zcrl0fBo1HQ4i9Crun5fH0z8dw+0LYNGrQrzHAcpaHdQMHUpEv2U+awtvbmpHEODpMwp59ucSUiL0dDokFMyYtHASLVo0CjkWvYqbZ2bxxZ66YK3bpFXyyAppVJsVbeD+ubk81Pt3eauNlDAlN83Moq7TyexsE39bVUxSuJ4pOdGUNFt5Y2M5s3LiuGPl8VlNE29elICI5Np5xlc9XDL47xTq2gmLzePovoGM4sp2h2QXLFexW1ZAZUffi9DjD/DdgUYmZUayvjepnz0shhjr93jG3IBgSuDztiQeWNqKKMqYnOWiIN5Ec0/fACDaqKbV5uatTZUsD9Owfvy5KNc+1HcBMjm2qGHMffMQckHgsrEJnJslsry4BacPIuJ7zbTcPZIicn/PleCF2ojWyQjXq5ho6WFRvpGvj0olsoZuN8usg7np3GUo6ncgqs2sFkdz9bKq3vtsZ3N5O98snktuzXJwWxFNiTDqSoQ974JMTtG8Lznniza6ndLsVSW38eq5g3ngu2O02z2cOzKBM5M6SfxwutS/MkTTduobGOiA+j3YjekExL5ymEmrYFJePGPfPYbHrwecXD1pKocbbGwpl8qHk9Or+fLMaEranGgEP4P9R4lp2ANXbQTRT00ghtfe64Naq+QyLhyXwg9156CduIgRgVa0DX1gDQJ+ClcswLfgRyrblHx/qJGpWeGcWvMqukMfIM54EKFpf+h9FQPQUfmLCWaYooovTtWzwxoFiIwxtjFYUQOMA+BgXVcwuYBUSnx0ZRGfXWX+l47m/xvjVxOMKIqX/rsu5D8tGrud3Pf14RA44+qiFv525hCyY4wo5DLabW5izRpIlAQQNJ3NzMiLYeXBRiKN6pDG9vH4fHcdC4fG8/W+er7eV8/zZw/hxXOH0WX3EGMeiGKRCaAV+pmeKTT0aOKBZiZnRlLd7mDekHg2lrQSaVBTmGjG6wsAAjq1ArNWSWKYliXDE4gxaPD6AiHs3tIWG6Krm50X6PA4epDrLZiq3+JY2iWI7gAtDoH15VaGJskxaBQ0djs5XG9lcGpcyHX2uHy8vLaMqTlRJFp0PLm1E1By7WQ/EzIjOXqCgnJhooWqhLsJ+H00u8M5ziM5Hk3dTiZkJnK4oZvZg2JZMCSeB/ZdxeGyHh4+JZaHPigJosQ2lrZxzZR08uOM/NRbEsyLN/HMT9Iou77LxXvW0Vw2+3Hke99H1EfhGXM9Z/0oQxTBJ4q8sbWOwXodD7bdJZWPvmuD0/9OoKsGwW1FTByF7IS+T2DohcwPr2f+mKPI3YXcZdjN9JmnsL9DQWG4j9G2n1DYM2DjMzSMf4SbNob2PlzeAMccenL3fQSaMITJt0twYG0YyNWsrQ0EkwtIifeL3dU8uSiPnTVWLk7pJPLT0/oOaGvBvPF+uk59kbBhF5LhOUZqRDJV7dLs8fShCbyxoSKkp/b6phpum5UdTDAbK2xcnOVhyeYlfdc5+28QL73sf1hTys0zs/AFRLx+kXC9kk92VOPxa9lYaufIkoFgCdzdxKg8mHVK7jwlg0nty9Dt+AAAwdogcW1sJwxC9KElr7oOB6UtNtQKGVkJ0xHqKgl098r3GKIR0voIjV0OLydGTYcDp8dP2K9UxJwePx12N2atEoPmH8va/BF/EC3/6bC7JR+TE6O63R6EVRbEm5hbENtvrch9vR7qtZ12zho5ULQxQq/E4+tDSz32fTH3zs2ltstJjEnD7EGxQb0zgCsnJJHa/Ln0hzGW6knPsKxSTVqknoVDEzjWbOX1jX2lmy1lbTx31lDabC6aelxkqEWm9qyhIzwXpVrPZzWh/vZn5OlZ3PMRYZt7MR26CJh8J8Pbv2N48Qr25t7OPXNy2VnVwYZjrQxLtjC3IA5vPxmDSIOK+N7kuLOynbfOzuaKpT04PH48AZEIg4pJWZFsKm1DJkjyMh5fgNsOx1PWaufZxQMl1BcNjaUgSsas/BiGJZu56bN9wdJQoy0qhKgJ8NqGCj69aBDHmvU0dDlJsGhDelwNgTDWa08hbtwQXCipdcXi9oeWvNY2KFmAAG3HYNJtiD/cgaxb6qsIkbn4z/wA2bZXELqqCOSdDgnDUC67WNr5tJeJLf2EBZ43WGCIgWMtoNDgz34VOSBTKNGp5AP6biqfTeLFuK3w3Z9gyXuw8w3QhtMSM3bAfWmwBvjqQDNf72vg8tM7B6xXthxG4e6B9U9gSZ3Et2fexk3rPOyq7CQ9Un9Sw7njsj/Ho8PX1wwWjXEo0idCTyMIAkqFwJ+/PYIggEyQyrDDk8MYkxZOfZcLhUEt+dn0R/llzsQnKMiN1TNaXUPsjr/2rTv0BUy8Fdb8pQ9FOPIKCYDRG0caurn4nZ209WqOjUiWuGjvbKmS7qFcxvuXZTCu99E+7gvTPxYPS/hVe+Tiph6e+rGYzaXtFCaauX9+HkNPgsD8I0LjD6LlPxkxJg2jUsPZVdWHPJLLBJT9+iaHG3pCSI79Mee1nS7y401E6FXBRKWQCVwwLpV7vuxr5Lq8fira7Ly8tozrpqQTZVBx6ynZ2N0+jBolkXolh9LuQ5ZxJaLWQivhhOltDIo3kRCm5fEfQxV43b4AzT1OFsk2ISqUxPqi0Wx9mEifC8IzSB37Tcj2F6R0ELa+H2DQ1Y0oCAiODlAZiLMY+XBFdZA5/dORJspbbdwyM4tYk4bcOCOTsiKx9XJ/Ts81Mmb7dSyfdj5thly0MivrO+TIBIE/zchCRGRjSRuj08I5WN+NQiaj0+7htlnZfLS9mm6nlwVD4pHL5WyvdfHJzlriw7QURMp5dpyA0d+J36gcIBFv0ipQa7TUdDgYlmRmeqaRtIgh7K7uZFC8iaH6DtL2/AVd1SpQaokfcTt3Tl3ANV/0cTByEyKg3SnBl+VqGrMv4oh2BI6AnGlxXkzLLoaEEZA4CtnhZWCth9hCqTG/9wNY+Ar8cCd014EpHuY8hduSi/28Veh1Zu6S2bn7xz4UWGKYmkGRJ3ASOsppL7gMmbuH2TE+3t8funrx0FieXiUhD2WmWE4Mf1QemvrNEvy6+Dt03fVMTHuG9Mgk8uOMQTWE4yEIErikf4zISCCQ+DHI1cjMSSgqN8DGJ0CQc9nEOzmWm83XxXb8vVPIa6dmMDkrimunZiJUr4OZD8GBz6SeVdYpEJHFsSYrOypspAMpclWfyKizE3a/DRetkNSe9REQMwg0Urbw+gK8takimFwA9tR0MTYjMshl8fgDfL2vHrkgaZklhWl59fzh/GX5EdrtHhYPS+DKyem/qDXWYXNz4yf7KG2RABm7qzu55N1dLL9hIsn/BSCA/5v4g2j5T4ZRo+TR0wfxwLdH2FHZQbRRzZWT0kNsfUFipf9wqJFwnXKAzMor68p44ZyhHKzrxun1MyQxjJUH64OILYDFwxMl2C3w5uZK7p+bx/66Lg7UdjMqNZwhiXGsLXfSZjMyKz+cI3U9vNyrA5YYpj2pNLnB30XKpjukWrY+CtvCd9nQokWpUiEgsb1XHZX6SNFiR+jOo69EOPINpI6HxJEEAt4BJYeyFhuN3S4uGpfMgdoeHllRxNNnFDAzL5qZ2VpsVdmUiMl8dsBHmlFgymATB+qtvLCmFLlMYNGwBDLMMt6cjiRmqfLwyE+1zBscj0GjYE1RC6Io8uiMCKbO9KOLk3NO989Ebv9QEqM86OaNs3+grqGWcLGLdiyExSaj8Vv5MvptNI3HkJc14Z72GZOiZaTGxxK/+3kpuQB4ncRuf4TC0/LRqeQ4PH4yIvXEhxul0bOnh5qICVy1x01xqzSAWDO7DZPXEeS1AHDkazjjHWg+KMGL1zwMQ88HUwJYm+DbG9AtfhPdJ2fCpNuZX7uf+DnXsLVVQ6I+wHhlKcktfaWhrqzFfM+pvHK4BblM4Oa4KN65IIy//liK0xvg6omJHG5yBJ8fIeDDM+F2VFuflb5rXQS+yfei/uby4DGVjXsZWdjFQb8adcDAheNSeHtzJXWdTgxqBXfOzsEgWnltpgonakYnaojfcidC+WpJ9HPsdZI/jkN6TlQ/3cFji95HqU+l0+nlvNEpjE4LR62Uo1bKqVMlE7ftduRxgyF5DFRvpU2TjC8piW5nJWX+GLqnPIx57d1997HwHJwxQ2l1gkGtIFzT1ydxeHzsO8EmA6C5x4VF1we377B7eG51KburO1kwJJ7zRiez4saJuH0BYkzqIAfmZFHT6Qwml+B34fBS1Wb/I8H8g/itCeZEomU7fxAtyYk18caFI2jodmFQyfl4ZzU1HY6Qbdw+P9d+LEGXH19UwJf76oNQ5FPyY7j8/d3kxhoxaJQs213LAwsG0eHw0dzjYmpOFFaXj7yYMJ47I4/15d08tOIoH18xmklZkWTHGHludSlrejWOJmdFBvkpAJ/srOHWU7L5y3d9TdcwnZKh/oNgTpI4HrZmbPVF3LtjMIGAh7vmiLi8fm6emUVAFPGZTnhELKm4zZmUu410eFUktNWwdH4yV69REm/RUtZio9PhIVyv4qPtNWTFGLhnTi4pYUqWZEsvme2Zt3DdZ0WAHSE7iv1rK9CpFdw0IxNRhBRVD8N2/Al15RoA/PEjeXD8A9y6pk/v6al5yYz/aT64unCf/jZNugR+GvoJRzoFZiQKTHFtZtzue6WRutZC+aTnaJAPJ7/km+AxRqhrUfYcxc7p6CtWDPh+wzqP8NqMYYg+P6naVio8YeBsB7eVfTYLxa1VwW3DDSdheauN0HwENj4DCg1Muw+2PBd8GUsPSG9PwtqAobucyeuWMNkUD65uCfo7/X5pvULDxoQruff7vgHMrcureWdeJ1+lfoNfriV857cUTXyBZb24kXa/jtSK1fjnPIMfGUpjNOpvrumbHQDIlQymjGG776Mn6SOeOqxnYmYkkUY1Xl+APG0HgtvKEa+GJLOK2MqvpeQCEjR+ywvSNVZvAwLSDLf1GBdPmIbPHyAlQo9SLtDRUIHoc1HvNXGg8AVGOLdgai+mIfsKPmhKxdRs40AvtLhj1GhuueBHTK4mMMVRpszhqWVHWVXUTFqEjkdOH8y49AhkMgGTVsm8wXG8sr6PLwYSxLjF2lc9GJFi4emfjhEQ4au9Eqz9tlnZkqbZPwi9So5CJgwoux4ne/4Rvxz/U6LlU0iulvAH0ZKKVhuvri9n5cFGMqL03DUnF68/wHtba4g0qLhuaiaf9jMMe2V9OS+dO5R1xa29EuF+3L5ACF7/lXVljM+IQK2Q8fXeej5fZCK8+BmM7fuZmLqQyy6YybKDTXy0o4YXzxkaTC4ADd1OogwqzhiZhNcfQCWX0dDl5MVzhrKhpJVki5bZEU10OJN4NP4V9EqBGZEdJHXvIdKgQquUfkgBUeT5Xgn0lkIzD0y8B93WpyDgxx6Wy0elCp7c0k1ABIM6ltfnqHnklDA+OeJg0bB4MqMNfLKjhr01neyt6SQ1QkdTTjTvbpXKP7MHxTBvcCytNg8WnZKAKLKptI1t5e0Y1ApWTigNJhcAecNu5uUfxr/kNBwePylmGcP23gcR6WCIoVsw80B1LHlJEYRHyggz9KD48Y4+boazk4z1N6A/+/vgMR0JE9gnG8oRVyyFnUpGRQ9Gbus7J4AiLI7J1l2SHJzDRZKpC458BYKcZvXFIdsGNBaIyID2fi+60VdKnBqQkvmW56DwbNjea4WsNoG+lzl95Bs45SGqu31UCwmY5F4y3UUYIrNhyp34w9L4+CRF6e+qBKZbdwZFMnMq3uWbCx+kwxWgUW1GyLmUlLUPInd14Rt9LaRORnGsL5mKIy9Hsfcd8Hsxrb2Xv0x/g6e2d7J0t41LRkbT5pRz7XJp9H7xUBPTm1YOuAavo5uPh32Gx+tlbJKOD4/6WfbiZkDqhzw2zUjO8tPA1YUp5zRWRF7Jgv1jiTZNoazYxuRsA0f6QeHf3dXEwlHjGZo5DpvLywMf7glC5SvaHFz67i6W3ziB3FgTgiBw1sgkjjVbWV3UglIucOmENJIt2t71cPbIJJYfaAgB5KwpbubKSalEKX+dQAmQGqnnllOyeLqfxtk5o5LIjP5/z8b//z1+a4J5BrgWmARsAzYBf//VPf6Xh8Pj468ri1jTa9pzuKGHK97fzbfXT+DSCen4fAFu/Gx/0NQIoK7TycaSNrqdHo40OINyGP1DQCTNJFLe5ObTJVGkLD8D7BK8Nqp+L9rB5UTqrwIYQAxbXdTCddMyuf+bw8HR1qz8GGbkxTA+IwKTRkGzFy5ZUYYoSqO715VKPj7vImbKPCBIviwGtYJbZmbhF0V6nD7eExaSM3UsMo8NkzeGxzf3Nb9tbh+3rXXyzlnRhOt9mDRKGrtdRBjUmLVKbG7J89us60PdeHwiE3MjWXmoEa8/wLmjkthW3o4vIJIcriOqdUfI5+pJn8+nttE899NhXN4Ap+RGkDr2FgyuErxyPe26bEZkeHhtQzkOj5/Js30hxD/ppDYMzt7+hj6S77Mf4faPpReGIPSw6qw/kVm/KzijCCSPR2VtgPVPSI2IU59C9v3twcMVqkI1nNpaW4nInS+VjRydYE6Eyo0QV9jvoekAXW9CMcXDtPul5jhAwMte8rlkq71Xq07BxaPncUvbLsIAmddBwkneZwl6EVqlAUrAlMyuzD/xypY2Gmx+zhmlZrd/PIbBHxCl8rKlRc1lI8LIyF+IrKcWIjIRKjb0OWV2lJO15yGenXI3TaMj0IeFs+TjquC5ijsC2KKGYugKLQPXqjJ4bI0EKLhOE8OyA30aaXtquvjmaIA71WEIzk6Uxd8yb6iF3XkXs6q4g/PHJKOUy4IlWen7ALVSKu029rhCeFggoeUqW+3kxkpqxCmRel44Zxg1HQ5Uchmf7qqhqt3BZRNTUcoEmnvc7K4OBTykReox/0ZIslIu46KxqQxLslDT6SDOpKEg0YzxDyTZP4zfmmDeB6zAi71/nwt8AJz1SzsIgvAOMB9oEUWxoHdZOLAUSAWqgLNEUQLiC4JwD3A54AduEkXxp97lI4D3kMjR3wN/EkVRFARB3XsNI5BKdmeLoljVu8/FQG9tgUdFUXz/N37O3xwNXa5gcjkebl+AyjY7cwbH4fX5GZ5i4VA/uX6AwYlm1hxtxqJXMT1dy4fbZSGw0OsKBeYW38R5hmi6exYGk8vxMBz5mDmLL+V5wBcIcOqgWH4+2kRSuI5RKRaeX12KRilnULyJxm4XPx9tZmZeNHd+eSiolJoRZeCUvBg8/gArDjbwfZmTbw800mbzcN9cSdfp1fXluLx+5hTEoVHq+LrZRKQhikLbwJJCU4+Lsi4ZS3dJDpLzBsdx0bhkRqVasOhUtFpdQZVni07JoAQTf/62zzK5qLGHdy4cQlFZBQa1DIVuNJQuD67fk3Edj3/Xd69XFbeTEJ6CKpDF1loHN06W9NqOj1Cb/Cbo3ygGUKiRa400jr6XzqQZPPZNX5lKFGHxchcrzllOdM9hRLURjbcbvrtJ2iAiE2q2hnzmIcde4Nk5j/DXjV10O73YzRmw5v6gcgCODmkGU9TPLdKUIIEAFr4KSi1idx0eXwC1xkzHiD/xwLZAiFvp+zsbmXbWJCbvvh6fLoazRp3KyqLuINLMqFYwLd0IByVV4EMjHuH8rzuCg4tHVxZzw/RMdKYkjjm9TM6BmCOvIjv0Xt/9mXgrWNIkwzGAyo0YJ9+FbtBomjp7aLf3jdp31Ng4NP8SxtVvlhrugCdxPD/b0vH4bSSFa6loPSGxA6urA1wfNwJDl3QOVdFXXHXODcSGGUmO0FLWHLrPkuGJpEZILH2dUjHAxRUkrlb/0KsV5PXK34/PiOCy93YH1103NYOUcC3Vva6sOpWca6Zk/Grf5cQwapWMz4xk/G/e44+A355gckRR7M9qWicIwoF/sM97wMtISeB43A2sEUXxCUEQ7u79+y5BEPKBc5Bg0PHAakEQskVR9CPNlK4CtiMlmFOBH5CSUacoipmCIJwDPAmc3ZvEHkRy3xSBPYIgLD+eyH6v0ChlmDSKATbIx4UolQo5i4fFc6yph+0VHchlAjdMzcCkUWD3+BBFCJO5+fSseJYettHuhgsGaYmJtPC2/S802OFGZdfAEwtyTDoVN0zPREDgrJGJzBoUQ2mzlYxoI1dPSafVKo3YRqeFc+7opODLXaWQUZhopqnbzdubK9EoZZw/JgWjRoGyF0Hj9ol02t08vaQQETCrFZS02lhb3MzGklZGnZ4/4JKSw7XsruvrPa081MipBTE82msXfdbIRMb3ztZm5sXwzb5Q86WKNgd1tVVcXXypxCMZdgFi5ikIZatAa2GvNQwIvRc/Hm5m7uBYDtc3cajZGVL+eH5fgGETHidp810SAVKmwDfjESp9EazTLiZD0NHjDB0c9Dh97KvtZuHmGyBzRmifxNUtwbP7hbZhG3OiPmTs6UvwN5dSF4hgx6T3GNr+A8r2o1izFxPQRWLZ/0nvgxFN2ykvIG9twNh+EJsljx8co5iTY0IdkYxNl8fhNaGwaIAWl4LAyMsRlEaONNm4aUYWNpcPQZAUpCvsHkbEDgZ3D4cCafgCoerZn++qZXpuNJ/tquXDWQKmQ+9JK44n3x1/hzHXwuZnpb9nPQod5ch3vUlMximcN7KAd7f3fV9Xr/by88UrcDUdQ6PVsb4jkid+lmDzLT1ukiMGNr0nJcjQtfSpRfjCM/mx1Mbf1kqzwCsmpfH8WUNYe6yFydlRTMyMCnrWJFi0PLRgELd83ve6uXpyGt1OL/d8eZAYs4ZJmZGM6OdhMiYtgtcvHM6r68oRBIGhSWGcOSKRY81W3L4A+XGmk0KV/4jfP35rgtknCMJYURS3AwiCMAbY8ms7iKK4URCE1BMWLwSm9v77fWA9cFfv8s9EUXQDlYIglAGjBUGoAkyiKG7rPe8HwOlICWYh8JfeY30BvCxIdnOzgVWiKMGfBEFYhZSUPv2Nn/U3RaxJw40zsvhr70sUYGSKBb1GzktrSokyqhmdFs7TS4ZQ2WZHKZdhc3u58gOphSUIQHMdI9ZfyYjoPFDpqbSezdmr02ixSj/+aK2JK8MzkHX01fVdo2/g/vU21pS0c/7oJJQKOaUtVgYnmDlQ14Uoiny4Xer77KzsIEKv4rmzpbHBsSYrw5PCgl71Hn+Av28o5+klhTT1uNCr5MSYVGwqbeXtLVIZ5PFFg3ns++IgMOHlDZXcMyeHZ1eV4vYFiNCruGJieogdLEBnP2TZ57vrmNwrw6GUC8hO4gqILpKe1Nn45BoqjNNRZ11HV2IRSrmcSPnAl1ZmtJ6IXt9z1Qnw0op2F7cWZfPB4g/Q2mpBDFARPYMHfmpgd42VGblRnDEsjqV7+spcaoWMbHkjaMKgowrSp0BDrxilrUUqafUXglRo6Mg8g6j6VahaDlFrKCTJug911RrQWjDue4P9I5+maeFKGhrqSUrJpKShjSnOMhSdFSiUZubFxWD65GxwdRORfTrjUq/kYIONoYlGmqw+STXBU4Li5+vxpU1nhe129tV2oZLLEJGIjPMGRXJm1ixwtKMJhFoggDTLGZNsoMMawSBz44D1eJ2S0dmk26WyXuNB2POO9BGPfsPl015CPWk4n+9tJNGs4aYZGSz5vJi6ThVKuZ9bTzEDUoJx+wJ0ObxMzopiY6k0886I1HF2qgNZUa90jEJN18QHeOaTvmv5YGs1P986mdN73R19/gAltc00dNqIMeuYlR/DV9eOp7LNToxJTXW7g+s+7tP8+2RHDa+eP5ymHhcquYz8eBOzB8UxqfeZOy5kmfb/gYLxf1v8Iy2yQ0izACVwkSAINb1/pwBHf23fX4gYURQbAURRbBQE4TgdNwFphnI86nqXeXv/feLy4/vU9h7LJwhCNxDRf/lJ9jnx812FNDsiOXkg6fHXoq7TyY+HGrljdg4Ojx+1QkZ9l9RjOd4gT4vQ8er5kpy7gAQKOB7RRjUxnVIjlBYpSR1Ou58Wa99M4KltdtJPfZFJwj7kzYfwpJ/C+40JrCmRyhPjMiJo7HZT0mzl7c2V5MeZWDQ8IWRm1W730NDl4k9jTJjCwll2INSDG2BruWSIZnf7sLv97KjsG707PL4QAc6iRiteXx1vXzySxm4XaZF6HvjmME5vH7Q6N9bIgX5SHCC9fJ5ZUkB6lIHBiWbu+epwcJ1BrSAmIoznOq9DJgjEBDQ0HO7BE0jALwaYniMpEBzsBUOYNAoWDImnplcgcf2xFi4Zn8r726oQRSlZXDA+jS7K6PBowJxISbsfUVBw+URJhsesV3PlxFS+OdBIariW20epcHmVvJr/EWaVyPxEJ6bGQwj1uwBw1R/CecZShNZjCAEfPREFdCji+FY8nTbdfKbrw7E010CvxL4MGHLoYfYOfwInasJ1ck4tfxRF7TbpM1uS4edbpNkRoC9dznOLL8HXVkd8xRdYU7Kxzl5C2L7XAVC0FTE1I8C+2lC76bFxMqnXU7eLYZOGEmlIDOGEXDctndNkm1iU1QY6CRQR9K8BiMwCazMc+BQm3AL7+hccIHHdjZw+Zym2ggTy4o28uK4iyHny+kU2lLTy6MICPt5ZjVwQmJgZwfDpmdR2OvD6RdIj9Zh97dgNS8FtxWnOZPGXnZg0SqJNaspb7NJgqxeKL4oiP+6v5JavjuH1i8gEeGRuOmeOzWR4ioWqNju3Lwu1FGixujna2MMDvWXXjEg9b10ykrTIX04olW12Dtd34/D4SI3QU5hkRqv8AxX2e8c/uqPz/y1XIb1/TwzxV5b/s/uELhTFN4A3AEaOHHnSbX4tDtZ3c7TRSn6ckYZuF43dLm6emRVcf/boZC59bydNvQzzESkWLh6fyvtbq+hyeLEa0+jvvuAR5QxPtjAlOxJfQEQmCNQICubv9CMIhZhblcwuiGVcukBGlIEoo5rn15RR1ovRP1DXTWW7nfPGJPPahr5yS1igg3PKLsGVOoN9kVeGAA8AMqMNlLfaSAjThhBDgZNaJDu9kpLy3V8d4qYZmSwYGk9qXTcH67sZnxFBbqyJR1ce7XcMGXFhGt7eXMm8Ahl2l5enzhjM+pJWYowaBieaqWqzs72iHY8vgFGj5Pwxyby3tQqfX+SUvBjGpIUzIy8Gf69CgFwmoFGILMgzs6nazkWj4kiZn0+Xw0NWlJ4J7V9g2dTnNmmZdB8HEmbx4Y5qsmOMjM+IIF9ex1WD1qGTe9nJhZyzvAVpEg3PG9R8dd7LlFeWE0BOsyqRTdvd/HAkApkg8OCCKJ75+WiwN/DO7jb+vvgK5pQsD/bN5LXbGVGwC9mWOxCLsxBGXQ6N+yREmSYMrP1mFGHJxNauhJ2SPL+5dhvmsm9gzDVQvgKsjZxmKGJtQg4H6qXve1KagakRXTj9kwkY0zBEJvLoaXkUtzjodnopiDMw3r4G+bpbezlPkTD/BVj7KLQWSRJGhedIvaHCs8EUK213Qrhs3Xy0w8+9c3Nx9RtIAGyv6GBcejjLrh4HAhjUUm8kOsScK4EOeThub4BOu5sHxloZZN2BubuY+sIZHFUNCc5qqxpaeGJ1NZdOSEOlkBEIiLy6pZ4RySZyUxIIiCIOz0Ddt+PClgDlbXa2lbf/YoKpabez/lgLx5qstNrcDE0Kw+X1MyUn+qTb/xH/fPwjLbLqX1v/T0SzIAhxvbOXOOB4IbwOSOq3XSLQ0Ls88STL++9TJwiCAjADHb3Lp56wz/rf92NAokXLIwsLqO10sr+2k6k5UaRG6oOkw/w4EwfruoLJBWBPdWcQguz2BdjkySE5dgjyJqm+nBuhIC1Sx3O9M6B75uTy2A/FKGQCYTolZa02vP4Ak7Oj+Gh7NTmxxmByOR49Tl+wnwIQa1KT370R7G1ojizlylMWs65EHiTjJZg15McZmZYdCYJASXPo8YwaBVnRhhCi2UXjUrC7fZxaEItKLuOltWWkReoZkWLhaGMPM/NiyIszcaShh0SLlrtOzUUGrClq4coJSdjcfh767ijJEToO1Hazp6aTqyel8ObgYnyiQEnkLK76tG+Uuremi8QwDZFGDQERxIDIJEMNhrI3UDpKcEw9j+1iFLlGN+audSi1U7Hsf52q8Y/TKoskOtBKyr4XGTZ0OG96Axys66ai1c57S5KI2v8y1hHX8/SWjhCHy1abm82NMp7YrECjlHP5RC3fH5Z+DnKZQFO3e0Dj+dnNzYwffzfmVbcB4IvIxaqKwYJkKMaO12H4RVC7A7wOqezW0/s45y2A3e+EPmSuUIBImv0g744No8KqQCZAhq4LrTkNh346Po+bSpuWL4obWTIiAYVcRp73CHFrb+47gL1NUmOeeo/0d9Mh+OEumHiL1INJHAmDz4SDS4O7BAzxHHDHAVaW72/g7FFJQSFUkEqehUmWX9Tmcnl9fH+oiYdXHKXb6eXd06OZsfs6hB6pyJBV8g1x4+6kzH4NJc1WLHIXF49P5aU1pfS4fGiVcm6Ynkm3XfodpUXqpfX9XC7VChlJllAeUu0JfLT+Udlm5/nVpUEdtzVFLdw0I4vCRDOWf4Enyn9z/LvnhMuBi4Enev//bb/lnwiC8CxSkz8L2CmKol8QBKsgCGOBHcBFEPS0PX6sbUjSNWt70WU/AY8JgnBcKGgWcM+/4sNUtNp4Y5OEjNlCO7EmNacPk6pxqZE6ihsH1sQ77B4uGZ9Kl8NLRGIk7wtPMDS7AaXooTIQx5d7+0b+VrePSyekopLLaLW5SQ7Xsbe6CxGpx9FsdaOUCyGjN4CMKD1Tc6IYFGdgof4Iyev7JN6HbL6GT5Z8zdpmPYIg6YSVt9q4fdlB/KLIlZPSuHduLk/+eAx/QKTb4WVqdhTzC+Nw+wLo1QpApKLNTmmzjRSLlodPG8RjPxRzpKGHKKOaNpubS8enIiKJgt7/9SFeOncYr18wHKNGyabSOuwef9DkrL7Lic3pJnHTXRCZzRsRg7h+WqbUr0EgOULLgbpuHvtBUpWemh1BYVoxliJJTELXfC+Fw2/CG5FD3K4nsKaNZNWoN7l5tQ27x49Bbeb5GW8RTt9Lx+b20WF10Dj6XsSUSdiLQmduADank1tmZiOXCWwu67PglckIKRseD6dXxC/v5VWojZSMepiwlqMcfxDrMs9jvXYWm1WnM0EMMHnumaR8s0gig4piH2emX4j6KARLGpgTESNzCF95Jf09SgMLXiTM2QVeO+742YxJt3D314fpcXrZMnegmCqtxdBVA6v+DMnj4NTHoKu3Cl23G2IGw6y/QtFyAomj2Bt2Kn/t5cEcbujh3NHJ3D8vj9VFzRg1ShYUxpMSrmX10WYEQXJDTbT09cwO1fVwa78GfYStNJhcjodh14scCUzivg1WPr10GC+tPRQs8Tq9fl5YXcrSK3qdQ/0BLDol107NoL7TiVIuMCjeTEOXM+SYY05CATgeNR2OEJFQgE931rBoaPy/NcFUt9s5Ut+Dy+cnN9ZIXpzE6/nfFP+yBCMIwqdIM4lIQRDqkJBdTwCfC4JwOVADnAkgiuIRQRA+R+rr+IDrexFkIPFv3kOCKf/Q+x/A28CHvYCADiQUGqIodgiC8Aiwq3e7h483/H/PqOt08u7WqpBlTT1uUiP0pEbqkAsSYuqNTaHIoNFp4ajlMrRKGZvK2nlrc2fvR9Ny/bRQ8czsaAOvrCsPuiACXD4hDWUv/+XHw01cODYlKOoHkvvk8v0NtNncCG6RzLJnQuG6zk4MbQd5cW0kckHgTzOz+NvPfVDUp38q4Y0LR/DkGYNxewOkReoRBZF1xW20WJ0sHJpATYeTFQebiDaqeXdrNXMGx/DIwkEUNVnpcXp5ZV05fzktn6ZuFwa1ksfPGExpq43iJitRBjWnD4uXVJf72RB3W+2SCKLHzrjsOP6ysox2uweTVsE9c/J4t99nXF/SzoeGTO6LyEbW6+ESefANfNP+LH03mmxuXHU0KNJoc/u4abWdZxdlAX1ghE5ZOKfsGEpyiYfbT8nk8R9LaegtEcoEGBavI6v4JfwqI/6U84I2DC5vgDCdckByv2pMNKWiHNnEt6jwR4Fdw9mHpR5KT86ZPNgwhjXlEiLrx2Mwvk7By2d9TXj1D7gTxqHQWJCv6+f3oo/CFT0U7dS7wJyEsKafnUBvCOVrpRKXo5XDlnD++v1RMqL0pEcaqPELxJ24Q3QeNO6X/l2zTQItJI/rW7/nXbj0Rxh7LTJAc+QQkQYPDd0uZALEG+V4RCWXTkiV/NYF+HhnLW/3DrQSLVreu3QUmdESSqv/syvd15NUokU/Cpm0fH+jY8DM0OMPYPVI69tsHnZWdlCQYKbT4SHOpAmSipVyye/ojtk5jPgVZ0nlSTTHAgERnfq3w5b/b6Oi1cZF7+wM9rPUChkfXTGGUanh/2DP/6z4lyUYURTP/YVVM35h+78Cfz3J8t1AwUmWu+hNUCdZ9w7wzsnW/atDp5Zz+YQ0zFolGVEGipt62NirEnzZhDQmZEYSaZBGST3u0Hq2Rhn64LfbPQN+oB/vrOaaKRmApPkVZVDz0rnDKG6yEmNSk2zRsrqohQN13ezXKHFNfRDdl+cFm6j+sFR2uFMQRTs5ccYBOk6XTkjlxTWlwZe/SaPgTzOz6HJ4iA/TIQPizWr+siAfq8tHmE7JhmOtlDRbeXltGSaNgkdOL+COLw7S5fCiVsi4e04uH2+vCYoopkbouGN2Dtd/0mcZXKBskBBNfg8lTbagAKiAQHX7QG7FyjIP1+XNIOK4SZhSi7yXJFltFQYoADs8fora+l5c03OiWFvaic3t42iTj1u/OMzLZ+bz158ridLJuHNaIsNd22jOmItc9DPDFGB1ip7N1dK1bClp4t0Lh/LBtioabH4uGRbGqAgX3zXEUtlhZnSKmVmBTcE+S2XMLNYcCC0/bi3voHxcBjt9I5ikNKJuOybJyTTsk8pnxjiUXZXwzTWQOgliCqC2H50/fjhCxjTY8RrI1NRE+rl3bh6H67vpsHs4LMsmftZbJK2+WlIi1kfB8Ith1QN9x2gtlspzvREwJ+EzJqGSSS/bgqRIvp7XRk27D4slnJ02P/d+3Tcj0SrlPHNmIe/JJOXkuk4nPx5u4obpUoI5/qwfj0PeRAbpIoI8GoCOgst4v0h6PrudXgxqBbZ+QqUKmUCMWZoVGVRyMqINPNNvUGRUK3j1guGsuW0KCpnsH9obD00KQ6+Sh2j+XTc1gxjTP2b1/16xrbw9mFxAAsG8tKaUNy4aedK+539q/AGb+CfDrFOwZEQin+7sm+7HmNQoZLIgiTAzysDbl4zE5pb6IikROtT9yF2tVhd/mpHFR9urg9tcPC6FD7ZXI4pgcw1sZrp9gZDyTKJFS6fdzSu9ApdXTU5naJKZMenhmDUKHj6g5PSpn5LkLsGjNGONKOSBz5qQywQ6HR4GJ/TJ86vkEou/ptPBOaOS0KsVrClqZlt5O9FGFauLW7hgTDLFTVbu++YwDo+fhDAtd8/Jpb1X5v3WU7J5dEVRsBfl9gV4dGURN83I4rlV0kuhqt0hzSpmZKJSyEk2yig8KPUt0EdS09bXe7C7JXWAE6MgWoWhs68X0DjyDvwR+YgTnyBcKw/2uY6HWiFjZIqFm2ZkkmzRsqWsg6/39/E7vH4Rm9vPB6eq0Oj0yH1NvF6fxqu7bSgVMm4b08nT0/XU17fjF2WkW38gaudRxqRNwudxot36Np0T/sw0bTTz421UO8NZ6R/EzDO+pqW5gRbLcOBY/4/A8HgtKUIzFqEe0eqVtL02PAlRudCwF6xNKOb9DRJHSaZmKROgaLnUSwGpX/Jdr4WxIJA+SccNnx0MIvo2l7XhmT2Uy8/4EFXzXilBfXvDCQRUDR5jIip9JPbYsexNuxKz3UDh8YF0WCIx5gRibK20+DR8/GEo/c3p9VPcZCUxTBMkMh6o6wquL0w0MyIljD3V0rLHt7mZfdEXWI4thYa9eArO4f36TI42Scn3q7313Dwzi/e3VpEeZaCmw851UzMxqOXsrGzHqFGG/OZAKiU397iCsOR/FLlxJj69aiyf7qyhss3OuaOSmZQd+Y93/B2jqcdFQpiGqwqV6BQBPi+VSndur/+PBPNHQJvVg8sb4Pppmeyr6SQlQkd8mJZj/RBaZa02jjT0MHdwX6Giqt3OkV52f3yYlrc2HWXu4Dh0KjnLdtcyLj2Ch08bRIvVzeg0ywAW89yCONIjdVw+MQ2LTsXBui4MGiU3TM9ErZCRE22gx+WjtMVGkkWHX1Bwzg8B1IpsvP4AZwyXBWc8SrlEQvvhcBPdTq9EuFTIuHJSOh9vr6HH5WXh0HiMGgUJYToyoo2IYoAnfiwONsTru5w89/Mx/rp4MPML44gzawd4ivgDIl5/6IzC4wvw4hopKS4cEsfwYddh6DwKPjczcyL49qh0j3wBkcZuV8hLyqRVcPmkNDqqpqIxpdISM5kGUyGPr2mgpCWZxUObuX9eHg+vOIrXL6KSy7h5ZhatNjcvrS3j/DEpbK9o57JhJsZEOOj0qXn5gJ86q5/bvu7iydMikXl8PLmlF87rhvvWenhrUTwz998fbMyLaVMQ5UqULTuwn/I0moY9DNp1MwCpgozyGW/wddswnlztZH6hZCd9fMYYZ9bw96GVRC+7lePYJTH/dITMmVDyIwDW9LlsU0zlUyGPWJWMs90iQ0ZfiyB6JRRacz+mgExBQ3tPCFwc4KMdtZydu4nw/X/HNuUvkLUQw5GPg+u7J97Pd56JtGS8y65mkW0rbHx6ZegxEAQwRqO0u/H4BqLMAqKItd9gaEFhPB12Nz6/iCiKTMqMZHJWFH5RRCmTUSaEM2rOk+D3oRDkDDa1EHX0IAgCYVoFUxNl2AsjWV3Ww6yccNLC1Zz+yhZabR4uGpcy4PwACtnJpfaPR027nSMNPXj8AXJjjRQmhlGYGEYgICKT/fv7HnOz9JzLIeJ3Pw0+J7NyzqJ60HW/Wb7mPyX+SDD/ZJh1SnZUtNPl9JIba2LDsVYaul3cNis7ZDtrv2bikfpuLn63zxgpQq/i2qkZQca7WiFjem50sAGpkMl49syhLN1dQ2mzjZl5MYzLCOepH49R0+HE6ZVmENFGNR/tqGHxsHhMGiVf7q1jTFoEG0pacXn9XDg2GZcvQLfDS368mT99tj/IpUiyaHn5vGEcaehBFEVSInRc93Ff6erTnbXcMC2Tpi4nWyrayYtUhKCtACraHfg8Lg7XdzO3IBaLThlCtBSEUDKkXiUP8d749kAjEzILqU57C6cPUv1a/jQji3e3VuL3i6gVMq6bmsHh+h78oog/AA9+X87o1JlsrGjjdHM8b/xcFiyLfbW/Eavbz9/OHEJJs40oo4oki47rPtmLKML3hxr4ZomF+DU3oSg6Ckod08c/wHp1AvdMMJFqFHlh28DZ408VbmbGFPQlmLzTuKdiGOXdw3k90EnMrlf6NhYDZOx8kGEzl6GQCXicVh6aFsuGunDWl3Vy91gNMT+GCmYKR7/Bu+AVlPW78Ubk8VP2w9y+tG/W89VhGV+ekk7B+kslyHH+6f3OJ6J0DbSeljgm0n05KqbxpWso50w9hTBfOy2KOF4t1rFknJb7v5MGRokWLelRJ5egt+jVXDEpnbv6+RUpZAKFiWG8vqECpVzg7jm5BESRha9sweUJcOsp2Ty/pizkOIMTTHx8xVhMWiUyYGa8lynzO6GlCCFuCB8WN/DcBonQeqi+m9UlnczMj+HTnbX8cKiJs0YmhsDwTRoF+fGmk14zDOx3aJQyPrliLMNTLP9PkgtAtvco8q19MHpz0acMik4BCn95p//A+CPB/JMRbdTw1JJCrvhgN3trpB/DhWOT2V7RTrhehcPjw+MLkBPXJ0mx/EB9CAmu3e6hpsPBA/Pz6HR4yYkx8vRPJRQ1Sb2Em2dm8fzqUuYNjmPR8AR+OtzEB9uquX1WFo/9cIxhyWGcNzqZbqeXBxfkMzjBxK6qTuLMWj7YVkVKhI6rJ2fQ3OPkg201TMyKYHNpWwhRr77LSafdg14llZWO9Gu89113Aw+fNohwg5oYzUC72SijGl3AjsPjp9Xq5u45uTy4/AgubwC5TOD2WdlolXLSI/WkR+mZOziOh74L5el22F3ExiUhE8DhCfDZrhrOGJ6IXCbw4+EmXF4/q4qag66VI1IsZMcYaepxEW3UDOi5rCpq4YahAsOVezG6HBR1DeeaKen4ApAfISN++y0o2nqvwesg2naM010NqA6/BxXhPD/2fq50RnCgsa9OnhqhB1s4ZEyH9KlYTZl8faBJkqkZchIlImsDaXovz81PJFXWSs6qC4mb/SKdLhMuW6vkUnlCVFgVfJ/5HoawCD7bEiqo6fYF2OlOpkBtBEcHgYTRyA4tk+DMAR9D1Y3oVeaQ3sINk1OwmMZBXBpWzWCW/lzM0iMqCLb/HZw/UU6UQc2Y9HCun5b5q72IOQUxaJVD+XhHNeF6NRePT2VYYhg/3zIZQRCo63Rw4dtSn0guE6jtHAgXrm53YPf4JD0xezusuAVl6c/B9QuHXM3KxLnsrpP6XeWtduYXxgMSfLy0xcZTZxTy05Em0qMMnD4snuxfkX7ZUt4W0u9weQO8tqGcl84bFlKy/neGvHbHwGUHl0q8J93/HqfMPxLM/0VMyIzk2+snUNFmJ0KvItEoR6zdgap+Oz5tNLKMqdSL8NamCnQqOckRekakhDE2PRJBkBjo5S02wpR+WrushOlUweQCwb48Kw81srJXykkuExibqOLbS/MQVQZK2t10Ob2UNFnJijawrbydjaVSjb6k2cY9Xx3i+mkZNPW46HJ4BxApr5uWyfOrS6los6NWyLhqcvqAzxln1uAXA/S4vEQqnDx2aiLtAR1uXwCNQsb8eBvR9r2sG3YUn6aLDa5M/nbmENrtHsJ1KkqarRjUCq6dmoHfH+DH3pLc8RiebCHWrGXZ7jr0agWzB8WgVcpDkGMBEW6flUOPy0tlm13i2dR3BxPkiaFXyQlr3kbiVknzdLQlnZ0pL/Ly9h5emG1GUd3PGMycBDI5qh29Wq7uHqJ/uIIHZ37K4l5l+wi9irHJBmx1GQiGFCrFDH4qt/DOucm0WN3IIy0gyEKIiu7MedyxupuNlTZkAjxz2vu0N8k5WNfG0KgY/LFDkTft77sOhZoSXwwv7Ohh3mAJRn5ieFRmuHwVIPBBkUjc6PfI8xeh8LvpVkTx4IJM9td20dTjYkZuDGMyIxAiFwGQVNMY4qAKkpz+0BgZP/xpEkatArVCTl2ng/ouJ+E6FamR+hDUlUmr4rShCcwZHIdMEIKK3um9Miwf9zPc8wfEk3qmLB6WiEWrpNvhwdxWAv2SC0D4wTe5dMKp7O5FT0cZ1GTFGLhheiYquYz1x1rJijFw1qhRA2/QSaKuwzlgWXmrHY838P8swWA5iZ1WdD6o/ncZmP2RYP4voqzFxmPfF7OhpJXsGAPfTG1F922vW6AgY7v+O85fXhlsyofplNw/L4/nV5UiAqNSw5k/KJIZe65FsDWxNDlU9Nnl9RNlUIf0NOYXxmFHx76GLiZkKllXXMsPhyX5l0lZkcHkcjw8/kBQXXdnZQcXjE3haGMPaoWMMJ0Sm8sXRHe5fQECouSEWdfLK1DJZZw2NJ7lBxr5dn8D3okJeLwB3tkhNezlMoGsBXGkbroviAyamn8ef+o6m/1NHqwuHyLwyGmDuOvLgyRYtDy4YBDlrTbKW6WktmREAjcv7Wsery5q5tHTC7j3a0lOZkFhHC6vnzu+kEozmdF6xqaFMzMvhoxoA7mxRuYUxAbvA8B9E40kV3wBqROh9RiyzgqGDuoE5GgMFgL9IM6kT4HigT4nWd5jPLzgNERAo1Jy/sdHiDOPRimXCKnDkmykhMnYV91NvDme1NlvEVHzM06FEYPMyyrTGWz8WZqlBET4y081nD0qmd3Vneyr7SJp8eMMPvwEiqoNiOEZNEx6jEe+l0pz64+1cPWUDJ5d1YeW0ihlJEcYIFqaffhLKrn6BycaZTpyQcDucbJ4WAf3z8vDGxCJ0KtCSpHZnqO8viiRv+/q4VCDlalZEdxc6CW6eQt0VkFUNjXafBa+X0qnw4tSLvDn+fmcNTIJTcAJcjUolLi8Puo6XSjlAkkWXUiZKdYcSnhccbCR++fl8dqGcjodXs4bnczcwlhuXnqA4qYePp7uHKjjJAZQIg1ADGoFt5ySxe3LDgRnqVdMTCPR8tsRXxMyI3l9Yyhd4NzRSRi1JyeH/lsidYKUUHp9fFAbYdKtElT/f1H8kWD+ybC5vTzw7WG2VUgUm6lJMnQbHg6ud6dM5ZWDYgjiq8vhpb7TyYz8aFzeAOmRenqsnQiN+8BtJcdfgiAogzOX97ZW8cTiwWyv7KCkycrY9Ajy44xc/v4ulHI5eXFGfjzS91KtaLMTplMOsDA+/pLpcfnocXp456KR2L1+wrRK3t9WFbLtGxvLuXN2DiIEkW2ddjcbSyT5E51Wz2ub+156/oDIPavbGDziGuJ3Sihz09FPuGTaGVxc4UWrlHPrKVm8ubmSgAi1HU6u/3gv7186CmXVWsxxmdyzMVRdOSBCcaOVRIuWhi4ngxPNPPZ9H3+lrMXOz0dbuHakgbhwN2qFQH6ckbmDorA6vaSEKRjTtRIheTQ0HYRBi0Cpwy1oAC9OuZny0Q+TtfoyiVHvaEc0xiH0l24B2gJGHl55jIAoctusHFxeyY7heEzKiuTBnyqxe/z8XO7kyTMm8E53PNXtTk4bEofDFiBM5yIhTEt5q40ely/ocyIXBNxuN0J4BiSPQnB2oRV8zMiJ4bM9DRg0CuLMGv525hBWHGzErFWSH28iQt/3UpySHcm7YRrquqRZqVGt4JIJqZi0SnwBcaDHvCWVkW9M5dWkCVjHjCOspwRFsQj7+5r+EXlnMTbhAn4o9eL1izTWVSITf4QDH0B4BjXjHuHpzR2sONSIWiHjlpnZnDM6CbNW1XtPoojQlwdnSaXNNkamWFh50yRcXj8eX4BFr24NwpC/bzRyuTEembWvHCgmjyc8Pp1FQ7qZXxjH/cuLQ0qgb22u5NSCWKKMJ3ERPUkMT7bw5BmFPPljMXa3j8snpgVLbsejrsPBnppO6jqdFCaYGZoU9q9NQOFpcP4X0HwYfG4JORiV/Y/3+w+LPxLMPxlN3a5gcgEIU9Fnfwv41BZaOgeSyhq6Xaw/1kJzjxuFTODR0wukkUztDgYdfII3F7zAQ5vttPS4mVMQS4JFS77LwPDkMMJ0SraXtXPNlAxsLj+CIIQ03L870MB1UzOwuSU7AJVChtcXIN6s4cbpmShkAqNTw/nmQAPf7KsnTKfk8olprO7niun1i3j8ATaWtNHQ7eScUUkUN9qCTXuPfyCKqMPuoVsVQ/+fbF6Umj/PSyfaqMbm9nHmiEQcXj9KuYzGLidVbVbO2XQlPXNeRi5kDjimIIObZmTSZffQ6RhYAttR2c7iwije29bKuaPCuCC2HNPqO5F3luObfDfyirVQ18sZqdmOmDiGuqS5QBtyuYxzViq4b+ynpApNuBQmUrIhoelCSd4fCISlYY0aTlJ4D0q5QEF4gCdnx/C3rV14/AEuGZtEaas92O+4aHwq13+yLwiNfm1jBcsuzGZCbBztdi9x5gjaHF4yTG4WzmxGGZ1J0sY7kLf1Jc7wPe9x3Vk/Em7MpMvh4ZGVR7l2agYddjcddjfJEVpUchlOrx8ByIw28ulV4zja2IPPHyAvzkiHw8sNn+yjvsvJReNSmJ4XTcRxdnpEBlzwBep1j6Ou+xmGnAfLrw+5r/qiz7lk5pn8UAopEVou02xA9fNzAIhuO0t31/HdQQnh5/IGePyHYrJjjUzr1fHKiTXy+dXjOFDXhccXoCDBzKD4Pob6D4caQzguz2y3kzP7VSa0fIq8djuBnLnIRl3ByMhMRuZDSbN1QFkXJIHL3xoGjYKzRyUxNScKnz9AnFkbMutq6XFxw6f72N9PoPUvC/K5eHzqv5ZZb06Q/vtfHH8kmH8ydCoF4XoVHb0jtU+KPJyZfylReyRfDX31ai4dfil3N4fulxapZ+ku6cfhC4hsK2vjHJVkrqRsL2Jq6eOoT3uVHg90Ozz8fKSZt7dUIopSueCV84dhdfmwOn0gwrzCOFYelEbeDo8fs1bJs6tKcHkDqBUy/np6AXd9dQh/QCRcr8Lp9QeNwZp73JQ22zh7ZCLL9tQhAqcNiafV6sGsVRIfpiU9Uh9iFauSy5D3kuqOR260jrju/cG/fTGFvHEY3toroeMyovSMS4/gox2SjcCwpDDmFMTAuOvRCV4WDktgez8FZ6VcYFRKOEcbe0CExPCBdemhSWG0OvysL+vknjFKLF+fL4lIAgpB7EsuvSHU7WDUYCvXTslA20vWu3VNB2AERJIsar46eyXOuoN45VocEQVc8k0rEzOjCNMpGOw7zJRdtzF98IUEFHo2Ks08f6ivHOkPBEJ4N0/OS+H9nY18V9zXyP/zzHiGNL6C7MgymHw79EsuAAR8yLuqeWVdX/lHo5AzITMSh8fPJztqEBB4ZpWkNHztlEzGZUQwOzwWgMP1XZz3xo7gIOCOLw7ylwX5XDKhr95frMhjd/YLVLXZuVrVStSJkEAgQ+cC5FyUryLqwBvB5V2J01heNjDZ763uDCYYgIxoAxm/YCesUoTOqty+ANeucvP04vs5Km/ALxqY7YxgKNLsOMqoJjvGMEAj739SIjseMaaTz3iKGntCkgvA0z8dY0ZeDEknefb+iN8efySYfzLiw7Q8vHAQN366D1GEuk4X2y3zmTvTgnzvu4jmJIYkGLh+ahTfHmhAr1Jw4dgUPt1VE3IcEZGO+Kl4LcOJKfucssxLufSD/fgCIrfNyuatzZXBbWcPiqGy1c4zP5dgc0ss+kcWFjA00cyKg42cPzaFF9eUBssJM/NieHdrJWeOSCTapCYlXMebmypDzr9sTx03Tc/kjtk5Emmu0UpWtJr3emVwRqVYuGFaJm9trsDlDbC/ppMnFg/mse+L6HR4yY4x8NiCDGT7vGBJxZUygzWmhbyzqm82V95qZ8GQvvnNvtouup0+2PMeCr+HQQt/4v55eeyv7UKtkDE+IwKZTJL9CARE8qNVnF0YxtKDXQCkhmu4YXw0K3tf3iZnbTC5AJy0Ow4EEHhjUwUXeZNZPCwBh8fH4foejGoFZ4xIZofLQEnAAgHw1Aa4YEwy72ypJifGgCM7gfAh5xG15y3we8mdNwWVvM+NVC4I/6e98w5zo7r68DvqbXe1Rdt7d++9F8AY03tvCYQSCIQEUj8IJUBCKiSB0AKhF5tejLuNjXvdtdfr7b2rrLo03x+jlVbeNTHgssC8z7OPPaMr6WikmTP33HN+hyXFRi7NdWAIOnCZ9VHOBeDR1S3Mv/QGMp1dqFFE95cJ0RowAQHUSoGfzkpG43egdvUQcAv0egL4AkG21vRIGWXVW3jp+slMyksgIEpq2ofPMP+1poozx6STaNJS19XLAx+Wsz60Tqd0x3J3UimK/o4uLguvOhboxekH1IZw+2mD9SDDE1XUHya8lJdkHPR4D8bwtFhKUkwc6Ocwrpiaw0OfVIYyvTqJM+jYeKiTz8rbmFWUxO/PG8VPXt1JfbcLnVrBfWeNpOQYNgw7vHYIwBkK58l8M2QH8w04dXgK79wcySIrSjaxvTsXq+E0jEYDqyrtvLm9nrklFlzeAJ5AICoNeExmHLOKLJy1chQOj58fTDsfW3cQf1DKxPEc9gOfX5rMna/vCu/vcfr49bK94TqWeL06Kh0z0aRhQk4W/9lYQ22nkzGZcWQnGAbI9fd6Azz/eU1YYHBqfgL3LCrB6QsQo1fx8YYWrp2eh0opUN3Ry5+WV/C3S8bxRU0XHXY3DXaRuMk/o9l9E26thVuf3DKgVuZwXN4AZfOfRaVUYdZChlmPQaNEIUCCUcMfP60IH6t3dxl4Ypafy1J7cYtK8nq3kPDF09TnSRqmAU0sXcMup6PkUpRKFSa1QHL2WoR+bY7dufN5sUJNIOjmhU11/PaMYYzJMDOvJFlqEaBVsepAG29tb0SpELhwQiaXjTJyc2EsgkqLUPmp1P544vWgUDJy1/28fP7veK4sSLPNwym5aq7teo3YtVIHy5WzXh3wmT3+IHs8Kdzh+ClneJRcMz8T9Yc/CWeeiWMvpyBQzbKFZoy4yA9uIujToq58jqAunutP/zkbgkL4+y9OMdFodfHEs1tw+wOcNWZguEWnVoQzvQ62OMLOBeDp7XYmL3qUuW0voKxeA+njIH8OLlUsfzxXambXZr6H5FWSyoK2fj23LLqTjfWRfkOjMuKYlHv0abVpZj1PXjWRL6q6qOlwUJIWy9vbGsK/21OGp7C6op1NofDzttpuPt3XwovXT6bH5SdOpyI3yYggCMesSLLAYhogHbN4ZCoZ8Ue3xiNzZGQH8w3QqJSMzjIzOssMwPu7m7i1n77Wz08rodBi4u3tjZLSbIqJB88dydLtjYjAVVNzuP21neHxf1xRzc9OKyE/yUCHw0tKTHRGic3tH+B0rC4fLVY3y3Y2MaMoiRHpseELc7fTS32Xk9pQY65dDVZ+sySdDZUd4ZMpPU6HQaMMXzAunZzFivI2NhySMsJSY7VcNzOfhz6MyLKcOTqNd3Y18eY2KY90bUUHl0zO4k/La3jthhSunpZDkkmH2y81YqvrdNLRL2auUgjE6tUsfimAViVy42wTk3KVKFx2Ek061ta5oxzxoXYn79XH8rO634E1JBMSm07CSGkRdqs3hwnF55G38deoO8rozT0Vx9x7cVdvIaZpHWL+XF7pHsYb6yXHGgiKPPTRfp65chyB3k6MegOfVrt4a3tj+PFJsd2MWHMHysatkFREIGMy9NTC51IqswBMCNzNjoLH6fWKpLgqiC17OWxzntiASZsdtd4wMj2WT/d3srO+h7ImBXMvLqLw1IcQXB2gMSGY0ohd/SBje+qkxmAjz0P5+V8BUPS2k/vpDxDOe5/nT4EgCkhM5/rX9oSd+exiy4Akj5+eWoI5VB3uOKyPSiAocv2HDj68+qcM1xigq5agMZWMQ69RsPu/+JOG4T/lIbjibahcAeZsRhdms+yWNCrbHGhVSkrTYo4YejoSuYlGqaYIWFneGpX5ODwtlr+uOBg1vqzZTqvNE1ZI7ur1sLaig1e31FFgMXHxpCxGZ5q/kg398QdE7jqthE/2tVDd0cvsIgvFKSY8fpEjdCGQOUpkB3OM6HJ4ePiwtsGPfnKAF66bjNmgRqNU8PqWeh5ftY8p+QnE6zVsrIqWU1cpBATg2hl5tNs9ZMYb+OvFY/jZm3vwBoKY9eoB6x8GjZK+ZKEdtd1cMimLZzfUUN3RS2qMjlf2R+s2/W1FJY9cMJpWqxuFQiA5Rktlm4OfLCxCAMwGTZTWU4vNw+6GHhaUWFh9sINFI1I4f0Imjd0uRoYWb5/dUE1hsonkGC1rK9rIt5jC3QUBrp6Wg1GnQqNUkJWg5/Ip2TRZpTtWjz+IQaPizjd20273ML80eVC1289bFPgSilCHHIwtfwkejZk5RYlMSnCS8fpV4VCO8dD7+FxtWBc/Q23qQpSGeB79eHvU63n8QcRgkJKmZQQTCnlvZ0SLKj1OxyzXKoKBILUz/4BSpcGQmIFl10tRr9FRchmz8pIxxxjRuaMvinnbHuQ/pz3NA9sN7Gt2MK84kXE5iTz6ifQbuWK0ibw1dyB09vvNKFQw+2ew+vdQMG9g6vTMO8jecj85oSI9T8Fp3DPtJp7Y4UWtUPDvtVXcc3opdrefVrubhcNSGN9PVTjeoGZqfjzTCiz4AkEUApQ1WvHqEtkYs4ik/FQKyh5HVya1QVA1bUH18tlwwxpJ1j9EPpG6l2/KsLRYZvdLrz/Smnr/xfalOxrDPWk2VXWxbEcjb988g5LUrxc2q+pw8Lv3y5iWn8jsYgtba7p5Y1sDc0qSiTuZqczfAWQHc4zwBoJRekx9tNrczC6WRPhi9Cr8QZENlZ0YNUqumBqtq3TdzDw+K29lUm4CGpWClzfXSfUdt8+k2+lDFEVumlvAP1cfIhCUNLZunVeIGPBz/9w4yrvt+IMmpuQmsGR0GikxOsZmmaN6mVhdPmo6esNqtH+5eAzPbqjB6vKRZNJw6vCUAZ9hR10PD54zkvMmZpJl1vPKlvqwE1II8PvzRpEVb+DuRSVkxRu45vktUc//z8Za/nrxWG6ZV0AgCAa1ii6Xlx/PL8SkVUnCm71eJuXGk2hQMyw9jk/6pV8DnFZkQr1Puoi78k+lvfQKVh/sxOYOoLVWh51LH+qmzWjd7XxSo2dGfpBrp+fyz37yIvedNYLPKjq5deckcuI13HFKHn/+rJJmq5vhKToCpjQejH+A8kMCQRHGZ8VxwZnLyKl6mSAC3eYRvGsfzliPjwMtNk4pyo0+aK5uxlT8nWevfBFXQIm118XpT0SqtyckuFCVD1zkDyo1kgy+14VoTEHoCa3ZxaSCx47QrwJce+gTrj5lHpfNDSJ4rLRYZqIWqsnuWIbgtoLzdPAvBK1055+qCzKz0MJjyyvCraUfOGckAZ+XSz+Bvy+0UhRyLmF8TqkN9HFKoU0z6/nF4mEsru+huqOXMZlxLB6Vwod7ItkxE3PMFCZLM542m5vH+zUbAynEW9Zs/doOJtGkRRSl1uF9JMdoiZOnL98Y2cEcI1JidVw/M5c/LY/cyaqVAsPSIhpJZ4xO57Wt9bRYPfR6A8QZ1KTEaGkNhY+yE/SolUk8v0GqrZiQE09PqOAt32Jka003H+1p4cfzC/EHRZSCwMub63hsRpApmy/CmzyGRuP/UTRmBF9UdxEMOaTKNgctNmkR/Nxx6aSb9dw8twCzQc3Oemu4qr6r10tK3MDsnNnFSTzwYTmVbQ6evHICzSHNNY9fylR7+YtaLp2cwz1v7+G3S4bj9A5cNO1weOjq9WLSKlEqBZ5YWRkO0108KZPfnT2C93Y309HrI92s468Xj6W6sxdRlFSqXV4/rw1/HIMyyPu1KoQtPn42Rc0lOUpUykF0qNQGWlwCL31Rx6E2O9dNzybOUIrD4yc9TseW6i6W7pRqL3a7fPxq2T7uOrWE33+0nwabn436OaSmqjlga0enVFCQHMM+hZlfd12LxxfkNEsqBRYlU/KSmFKQguh14l70Z3Qrfw3eXoLJIxEX/B/xjauI764hxTKCX8xL5ferJMfZ5tUOushfpRtBy+IVtLuCjM5tp6D5Sklq31Ia6ePSD13tKnS2JmjZQ+ECraTG7A+FI8uWwhl/gklS8a/X3Rt2LhBRun7yvGwArD4laIwDnDWCADtfBl0cpI09pqm1VqePf64+xAd7mkkyavisrJlnTtMyK9HImmYVU5MDzMsWSdAEw6YoB1l3UQzaKf3oGJYayxVTssNZjiqFwEPnjSIlTl6D+abIDuYYIQgCF0/KwqBR8dIXdWTG67l1XiEj+onwFafE8MaN0znQYkdEqhlYNCKVvY1WXL4ABo2KXy+LhJa21XYTp1czPtvMZU9v5uenldDQ7eQvn0WcmNmgJsOxD4IBNC3byfv4amyLl/H3lU1cMimLVpubRSNTidWrSI7Rsau+hztf34VGwIK2zAAAVAZJREFUqaA0LSaqX4dU4Gjj4klZvLG1nqAIY7PimJafyIpQrUwgGMSkVYWblCkEScIlziDd7Zn1anITDdR0RjSodGoFxakxJBg1xBs0PLm2KuxcNEoFuYnGcNU+wJqKNknxIPQ5TVqpJ83dqyIXvpRYL3Gp60ja+hht5y/FW3ImmgPvhR/vnf1bfvxxD05vgJUHOqhoc/DP083Ut3WRYszkN7ujiyp9ARGzThHquqmios3Bwx9F9NJmFCbx0Id7wts7G6z844Ii6eKvUCJoDOimXIs3fzYBtw1tTCLqD+6CSkkGRQlcPv0elIvPZ2O1Fbs2jq55j5Dw0Y/Ci/z+yTfRoCngnf124vV6srPzKV/8NjHd+wjok0h3H0TdsDX6h5dUArUbQakGjyPiXPpY9xjNqfOotCnx+BiQfGF1+ej0SpeBf+32M3/SL0nf8KvIgIIFsPdt2PumtJ0/j8oFT3GgM4BOrWBYWuz/7L/yZVS223l3l+ToW+0e7p0RQ+6yc8hVabk0IR/21IJjFqJ4FkLNBiwpw3n1/Cmc+kIjfZHiviLUr0ucQc3PF5Vw9th0Onu95CYaKTqGWWrfZ2QHcwxJidXzg1n5XDghE61aOWhfh6wEw4Dc+r549vMbqgeM31DZwQ9n5dHQ7eKjPc389szhPPLxfmwuP4lGDX+bryFz/V8jT3D3kOJvBATmlVj41bK9YYFNs0HNRROzACmkV9Zk4/aFRazcHym0/GhvC3+8cDTDUmPocnoRg6BQCKGCPy8qhYL3+12cgyI8ta6KP180FoDabidXTcvljW31lDfbyTDruW5mLtXtDn77bhm/WlzK/n56a5Py4qMKPftes++5jT0uHB4/u+p7ouohFuRqiDv4NnjsJC+7iLpzlqItugidp4OAOZc710FNZySjrqHbjautisXrf0DrxJ8Rb5gSJTxamGzCZNCyqboLi0mLSasK95TJTTSwr8nK4bywrYOF4woIC6wLAprkUNFozfqwc+nD9MWfSJ4+naYegS+qungvJonXr1hOvKcR0ZjMFkcK+9qCBIIgKAQ21fTg86fzxGongaDII/Nmcl76SpRN2wAQc2chBH1Sga9CJVWnHo5Sw783NvHsditPXDIalUIISweBJFRab4sco6XBmdx41fuoOg+CyQJ1m2Dj4+HPt6PwZi5/cnt4llqaauLJKyeSk3j0qcr96T1stlsc65f61QS80LwLkorAlILw9g/DYwqSSvjsB//lsU0O8pKMLB6V+o0dQqxew6S8SJvl+m4n6w92sLOum6n5iUwrSCR1kNm9zJcjO5jjwFH3dBBFAnWbYft/EJwdJI14bMCQwmQT7aEQWml6HH9eXiFpQ6mVTE0VmPHJ4gFhloSEJN6/dRhatVRL0keP00enw8MNs/P576Za9GolOQkG7jyliBc21hIIilw4MYt4g4a73pB0v66cmsM/VlVS1ixlYP12yfABNvY4fexvlq5SzT1u3q9tojQ1lgXDUmi3e3hiZSVXhPp4uLwBZhVbeDcUngoGpZDE4RyezNBsdWMxaalodTAmPYbr0utRV4Rk451dZL9+KisXr2aTtYh0dGxuOsAtk2IpifVQbtPyn1296BKzID6XWG8Lvzk1l9vflmZhWpWCq6blcOvLO8J3+JYYLTfOyedvKyolUc9BbhZMGgGF1w2iBwyHtbr1DVQRJuBFLXopb5YcW6XXT6NuIvEFE2mzutiwv5bHVx0KD0+J1fLbJcPDbZl/vqqXlIufYHaCFQQlDrePmJdOlwYH/dIsRmcGd0/4NWpG/4Q3Vkszv1aHn1+eXswfl1fi9AZIMmm4c2EhhckxPHHZOOL0aopSTKhi9ZA/C3rq4c3rwq/lyZrD42U6nN5Imvv+Fgebq7u+toPJSzRGFSy3K5MpiEkLdwKl9EzY9ETUc4SOA+QHanji8lPD+9ps7rC2XUGy6Rstznf3evn5m7vZGFqTeW1rAxdMyOT+s0dGnU8y/xvZwZxEPPXb0L5whtTJEJgUP475pQtYuV/S/dKpFdwyr5CXN0uxYX8gSIfDy9OhYslPLXpeGn0Dlo2RTtP+sVfy8iE9D67cwP1nj+C6GXlR0virD7TzxwtHoVMpCIqgUysZkR7HvWeOAAF0KgXv7YrMUJJitGHnAlL3wMPvgoenxVAeqq15Z2cjf754LH9feZAP9jSTYdbz01OL6XL6+PH8QswGDacMM9Jh9/D5oU72NFp54KxSvjiskj8vycirWyLZbJeUqhmtqMSdDzlJIuZlt0YfTH08++q7WXXIT1aCnk/PVZG++hYU1lqWxOVy2Xl/YicZvJn8ONePiOG09Xfyxuk/Yr8rjpyMdB5eWRcVPmq3ewiKki3NVjcFFmNUl0yFAD+YmIjqpXOhtw2m3AgjzoOYUJJEYtGANRZ/1gw+a9YB0sX0hhk5FKg6YP96HKYJvLgpugi31eYJX3hBCm/ZlWbIHgaAo7Ob7tNfIH33Eyh8TppUWSRe8hb66s/A1og19zTu22TA7pFmct1OL29sbeCqaTmolAp6PX5+/9EBfrygkAc/2M/wtBh+fcZwUmL1BIIiSqUGSs+AfUsBcMbmUVE1MJFlMEn+oyUrwcB/rp3EHz+tYEd9N3XeWAwzH2f45l+i6jxAUBuLIjCwRQTByL6KFjs3vLg1HJY9dXgK9501grSvGbqrbHOEnUsfb25r4LoZed8oFPd9RHYwJ5Fg5eqwcwFI2fwwj05TsmHMZXQ6fagVAi6vP1wc5/YFo5p5VbS7+GPndG4683WE7lr8xhTW2jO4b7lUn9LY42ZbTRf3LCqlst1BnF5NrF5Ni83LVdNyUSkFVpS38tM3dkfZdfuCInRqBW5fEH8gyKyiJCbmxOMPiigE+PWSYTyx6hDtdg8j0mP50ex8TFoVI9NjMRs0bKnupMBiYn5pCsNSTVhdfv61+lA4HHL5lGym5MUzISee1Fgdkyjj5TN0vFunxaAWOT1fwyFPALNBjT8gctHETIpTFZS89iMAxJl3EEwZhaIlosBcs/i/ZFuNnKd1c0a+gsylPwhf3BXWGrI+uR7j5csZOzOLZLUbbcs2imM+xJQ0DbU+cYBAKIBBreSiCVkIAoxOM/L6GWpWt2pxB2B+UTxjV10BXaGGYB/fI63HTA85voQ8uHIprH4YGrdB6RKCOXM4r9PO1NQY0rQeRia1oX/yQgj60c19DI9/4OJ5UJQargWCIguHpTAh2xx+LC0xnnbNIirTpoAYJC4uHr1ZjzdzHL5AkMrqelZVl/d7LbC7fVHNugqTTZQ1STcHZc12bnl5O3+6cBSPr65mQlYM5w/7MSUKFZQtI962n/NHXsNfDhMnnZhz2OztKzIq08y/rpiAze3DbFCztdbAXzL/QlGpG18wjrNHX45614v9vpgESJacrD8Q5PnPa6LW/D4ta+WM0WmcPfbrJSP4g4NX8B9pv8yRkR3MScQbFDn8Hiu+5gMmjb2WNq+JJJOGVQda+fH8QjodXiwxGv5y8Vj+uuIgO+t7JHXl3BT+0+jEqMulSG/ivnd2MiYzjiWj0zBqlAgKqQlUcYqJQFBkRXkbl03JJilUxPlZ2cAuiJuqOhmdaWZzdRcFFiOVbQ7+HFpwj9WpuP/skVw9LRu3X6S6vZc7Xt/Fs9dM4qFQHdAPZuVRYDHxWXkrOYl6ngllxfXx0hd1/PuqCVS196JWKqhW5lLQ+yEPez4Fj4Ja+4V0xc/hnLEZ4YZjHm88oy5/G4XPgV1toTL5XMy91RiVftqNxTy83sH6Q9LMbpTSS/ZhYUNc3cQ1rSNh7cP4Usdz4OI1/PzDRnZt7+WcUTauGW/moVWR1GiFAKMTA7y1vRNvIMjSHY28dEUJt28+V5JUTzw34lzCB+4fMPpiae0CpMr4C56X1kjszWiemsOMvrEL74X3/hoW10zb/U+un/QkT3weSc81aJRMTlWw6/Qa0MQgZBfQq1KytaYLhUKgwGLEEqPFEiNpkQWDIl9UdfKP1VK69UOLc/i/hen8flUL3kCQleWt/POysfz+kwrKm+3MKExgbkkKj34cSZfudvqwOXrZVtvNttpu3t2t5c1Zc8mMywRHG+cbttM0YTxv7WjCoFZy56nFjAsVGn8T9BplOPw0oyCJ3AQjbXY3iSYtasXdkFIKu1+F9AlSVlyC1Leo1+Nnw6GOAa+3r8n2tR1MgcVEgcXEofaInM2MgkTyvmYY8PuM7GBOIr3pM4hTaaMyf9rH3oopNo70kPz56SPTKWuyYXP7yDDrQRQZlhrD1PxErC4f6WY9z26oprbTxQ2z8/nThaNptXtZdaCdvCQjt8wtZNnORhweqcPmuGwzCQYNH+1uwmxUk2YemIpZkGRg8ag0uqZkoVUpo/qs2Nx+nl5fRUa8gY9D+2O0KgwaJbfMK0SnVrDmQBv3njWCs8em0+vxU9XeO+A9GrtdbDzUidXl467TinkucDqxSVPxBgV0WNi8vZPVoRYBAKPjTdBVBa4OfOlz2WmL4UBrPs09LhYMi8XYr9OmXRELCqU0o+hDoULZ2wKOVryiggdWNLGrSbIrS+/h3NhDKBeO4LltnSQalNw1O4Ui+xbabBbsoWr8jytdjL3mQwh6pUX8w9EngPKw9TdHC7i6CIgKlBoTeEMXrdiMqPCZoquSqzLfIuX0K3ltlzQDvH5iAsPenBfus1M993Hu3NfOjnop4WBOsYUHzhkZThrZ22Tl8qe/CIcvL3huH69cOYwPrjdjd3vIjDeSnJLM6OwEbG4/Tls3S/69C38gEhsUBEjQRMJgLTYPFYbxZNYuA8sw2jNPo6q8lh/NKcDjD/BZWSvzSpOPubR9RryejLCgpUGaGU68TnLuisg6SIxOzYLSFJ49LEFm7GFOz+cPUtvlxBcIkp1gwKg98qUvOVbHv64cz+tb6vn8UCenDk/h9FGp7GroobbLSVqcntGZcVEZmDKDIzuYk0h7TCm2Ra+SWrMMjaebprzzcCRNYqw+cpFKMmnDhZq1nb386MVtnDIihaAI8UYNf/vsILOLknmxs5ZgIMjOeisvhLoKfn6ok6x4PY3dLl7f2oBBo+SqaTl4AwFqupwkeDRMykvgre2N4VoYk1bFrOJkbnppB05fgLtOHVhgt6/JxqwiS3j7oXNH4unt4bR0N81+AyPmFnL/++V8Ud3FldOyGZtlHqBWq9MoWV3RzkUTM3l2Qw0xWiVzJhYTCIi02z3kWYyMzTYjijAiUWD+tltQNEpFhomavzNxwQvcv9WDKMLagx1cNyM3nHX2xG6BsZN/RdqmSH8ept4MZe8A0JE2l3WfRe5OW706ktx1XL/z/zi36Cy07k6M732M45znouRVWmweiAuJdgZ8YEoGtVGqD2ndCwt+A/q4yOPl78F7t4HHjjI2ncDpf0C56XHp7lupluTzO/st6u99kquuXMyF06ejDrhR/fecsHMhJo33rblh5wKwpqKdNRXt4YLd3Q3WqLUxgNuXVfHhbbMo6ic7ZDZoMBs09HoauXNqHI9uiHw3N02MJVEZ3QFSoVTBFW/i8QV4/KVtbK3tZmttxDluqe4KS78cVwbp9qhQCFw+JZtttd3saugB4JJJWUzMiSgYdDo8PL2+mn+vrcIfFFk4LJnfLBn+pYkJRckx/HLxMNz+ICpB4J9rDkU1fzt/fCb3njWcGLkY80uRHcxJZHi6mV2BsaxUFCMikmTUMjrNfMTxLVappuXxVZX4AiI6tYKfLCgGpIvKrGILP3xhW3j8yIxYdjVYw7piTm+Af62p4s8Xj6HN5qHD4aXQYuSR80fRbHUjilCUbOKWV7aH79oPu14BMLMwidGZcfxycSmFySYKfRWkrr8XTct2UmfezxW7R4XTiV/+op5fnzGMbqeX2k4nWpWC62bkhe+aLSYt0/ISMGjVrK/sRKtUMDk/gb+tPEhtqNVtcoyGF2b9kNKQg8HbS1HZ48zLv5OVh6T1g9e21HPplGyeXlfN3lYXz+TO5spz38HRXkdBTja6NQ9CuxQKMtkOkp9YSFUojXlRroCw6SVwtJGw++nw5/S3HUCnKg2r7c4pSoBtL0LzDhhzMbvPW8OyXS002X2cf3YyUzNSCSfLtlfA2z+IzKJsTShXP0hg9t0oN/6dQGcNirm/Qlj1gDQz08fDrJ9CQj56tRJ8HnBEwmX+pGGsbBiYhrzxUGfYwejUAx836VSolIMXIfrddq5Kb2TCIgNNPiOpaifDTc1UWiNrKnkJWkos0izX6QtwsM0x4HW+ySL/saAg2cTz102ipqMXjUpBXpIRgyZyadta08U/V0cc+WflbQxLjeXOU4u/tN+LIAjo1UoOttr522H6aG9tb+DyqdlRUjwyA5EdzElErVQwMTeBURlx+IPioNP2DoeHfY02Ons9pMbq+PvKyvBdqtsX5F9rD/H4JeO4dV4hRo0Sk1ZJl1NajJyQk8DbIUHK/rRa3by2tQ6nN8j54zJIM+v597oqRBEunZLFNVMyKFC2oRX8lLsc3DA7n+c31OANBCmwmLhpbgG9Hj/pZh1F2h6MZe/wefHPacwxkBRnITPeTma8geHpsTR2u/jz8goevWA05c129BolCQY1XaEmYmsOtDE6s5hbXt6OP9Szxub2hZ0LQJvdyxstafwmIV+6GAP67gMUZQus7Pe50uJ0JBo1TC9IJM1ipjU2n4OuPDwqA5mpc0gO9YhJLH+R3516Mde/48HjD9LiBFFtGFgLrjGRlSCFac4bn8GI+AAcLIeuQ5RX13PJZ93hepCPyzr488Vqzh2XKT3XWhsdogOwNqC01kJHBcqV9xKY9TOE036PwloPRgukjIpUyRsSYPIP4dNfA6DqKGN+YZAd0dJyTC+M1G6MzTKTaNSEu0kC/Kyf2OXhKII+TOsfYsqYi0HpB9Sw6U0yZj/CGaVxjEqCU0oSSEuR1njiDRounJAVdScPMOkbLvIfC+INGuKzB/+cW2q7B+x7f08zP5ydT+xRhPac3sCAmSFI6z8yX47sYIYAWrWSwaK5Xb0efvvOXj7cI6113LOoNOqHrlUpuH1BEQ99VE5Zs513djby2zOH88AH5XQ4vDR0OylINg0ITyWatDi9kszL6Cwz//duRD3A3dPO3ZbVGDY+DkE/87Jns3vc/aQvLqWz10u8QQrLbQylFa++Kom/eM7ihU/sgA210s6fLhzLc59X8/jKSvKTjNxxSjH1Xa6wSq5aKfCPyyegEODyqTm8vLk2/LlSYrXUdg28I97RIeDLnYPSmIqiaStdBefwcUVk3eWqaTkkm7T8+oxhuHwBlu5oZFp+IpePS6CsM8AzngUsmTsMi2M/NkM2Bzu9/POysexssOHWqhFn/TSqmA99POqMMfyfOQFBEMlT27C4JUkW/B7SDSKT0zWsrok4wr+tqGReSTJmgwavPpkBlzt9PAGVgb4VBOW6P8B1n0LJokG+fWDUhVKV/6Z/gD6RM0cms7ZZyZYa6YK5oDSZOcWRUGVhcgwv/3Aqnx/qoM3uYVZREuP7ZZ0dji6lGF9cDurVD4f3+VPHk5BRyBOpbtDowZwd9ZzzxmfQanPz2pZ6DFold59WyrgveY+hwGAaZROyzUdd05KVoKc0NSaqzUW8QS0v+h8FsoMZwuxvsYedC0iNkfo3uTp3XAbPrK8O99Ko73bxi7f38vdLx/LCxlrykgxcNTWHm16KVF5PL0ikNaRLNik3nrX9FtIFAW4q7MTw6V/C+7R1axmT8iq/OHAalW1O7jilOOxcAFpI5oVdkTqbs8ak85cVB8MZOFUdvfzhkwPcc3ppeIwvINJmc3PT3ELSzbqoFOGq9l5OHZ4alqbp4/QRydzXeik1AT+XzdeTY4lhstdJhtnF4lGptFo9/PjVneHxN83JJ66nDJb+hOLCRcwf+UPWNlnwakehFRRkpeq4Z+m+cOvdrSVp/PmiV1DWrCaoS6QrZTp6t4NxFS8SFFQoiheievOKsNZKXP0X/Gr+U2yojwkXQgZFMVxLUxHMIGnSPaRuCV28lRoOTX8UQfST3/+DDVZ930dMKsy4HcZcCiotubo4/p3mpbqjF4Ug1QodfgdekhoTfUHt7YD6vVJCQUKhlN6rlE57tSkez+LH8JS/h/bQx3hyF8CIc9AmRouw9icz3sC9Z47ghtn5qJWKbyQTc6KYmp/IpNz4sGO2mLRcNzNvUNXuwUgwavnrJeP40/IDrK3oYGx2HPcsGjZop1WZaGQHM4TpW3jv49N9zdx71nB+934Zbl+Q1DhdVIMxkJyQ0u/k9+eNQqNSYInR8dZN0znYapeUi5UCVz8rqR0rFALZiZGTJNGowWI7TOsK0Bz8gCUFZ/OXNieBw2oBKrqjQwfpZn24r0ofTm8grEbQR4xOBVbJAV01LSdci+PxB6nu6OX6mXm8sLGGoAgXjM+grtvNf7dL6ajrq3r4yUITexut6DUqgiL8fVW0wu5zn9dwTqoJ2vejsjZAwmKeXW+js9eLViUVsC4amUqcXo3bF+S9XU083lDAm7u09Hr92Fw27jm9lNGjf4NWrWTsngcHCHnlVb7AtNxfsPaQtPD+4/mFxBuleUuHW8l91VO4ac4bxAW6aRaSeXh9gH9P6acSXbhg0IXrAZgi7YjNBg3jjhAKAqS1m65DkpKAwQIr7oV9IYVkhRIueQWKTwsP1yYXQvIdMOt2tIqju+CqVYqvXbl/MsiKN/DPKyZQ0WLHGwhSmGwiM/6rOYeS1Bj+eslYunp9xOnUGHXypfNokI/SECYvyYhaKYTvkGcXJ/Ofz2u4bkYeCoXA1GxjuCCyPwlakYzQCeTy+ul2eqlodZAcq2VCdjyPnDeKp9dX4/YFmFNs4YPdzbTZPSgFgd7YAowaE90lF+PTxJFc+Tq9yROIT0jEpLWRadZj1Cgx69XkJBlIMGlJjtGGZwKBoDioTf3vFn+1uJQV+9t4b1cTcXo1ty8s4k8XjebxlYfQqhXMKbGgVsBbP5pGQISaVit3vLUv6vXe3dnEyIw43t3VxIx+6xB9uH1BehRmAFpH3cjtyx3htQmPP8hfPqvgN0uGc997ZcRoVVw/K49kk9SV1Ob2S2FFEaYXhkJQFQPDLAqtkZkZySiUGi6enMWMgogdhSkmanqCXPeJDzABTuYXmck2a6TQV3wu9LZHssSOBc4u+Pxx2PBnKbR26gMR5wLSmtD7P4EbVksNzaI+zNE5l5OBzeVjZ30PO+t7yE4wMCE3nqyv6CCSTFqSCr9ZWrFOrSLdLF8yvwry0RrCFCfH8Nw1k/jVsr3UdjpJjtVyoNUR7meunG7h16fk8OsPIzUAP5yeTopJyfu7m0gyaulweLj1lUiXTYtJy2s3TmXxqDQUCoHNNZ2cOy4DvUaJQhBoSchm/+IV7Lep8fiC5M68lBxTEJNTz0UTM3lzWwMfXFuArnUnKsde9pHIjXPy+WRfK3sarDR2O7ljYTG/79d87YLx6aTF6bh1fmFIT8zOOyEtsm6nj3vfLeMXp5eyYFgybl+Ae9/Zx2MXjeEXS/eiVylYUjpQnqO/VplKIWDQKKPaBIzNMpPqkY5Le+yIcLuCPoIiBPzSIq3d42dvo5XZxRZsbj9mg5r/WzKcERn93nfYEvjiH5GaJUFAMf1Wbsgr5oa5A7+7zHgD/7luEs9uqGFbbTdnjEzmAu0XGD/4OcSkQfm7Ur/7mXf+r5/BQBxtUmjNmBS9v3EbrO+nZ9dPkyyMrUlSXT7cwQxRRFHkja313P9BRJFgbJaZp66cQPJX7KQpc+I5KQ5GEIQawA4EAL8oihMFQUgAXgNygRrgIlEUu0PjfwFcHxp/myiKn4T2TwCeB/TAh8DtoiiKgiBogReACUAncLEoijUn6OMdMxQKgZlFFt6+aTq9Hn9UdhBAeY8COwHuOrUEtz+ATqWkpsPBi3tcPLF6H+eNz2DbYRk07Q6pQ+U5oWyn93Y183a/kNaIqyfy+p6GcMOvAouRny8qpdVup6bTyYOnppKy6ifo69YAkDI7ketXJDAtP5HLpmRT1myjsayF564aS09jBXFJaby6x87P3pRCYJdPyY4q3OzDHwzS6fCgUCj4w4Wjuf/9fUxO12DUCIxODKJXK8PpwgA/nmImPcHIkiItBSYnwy8ewxOrqtjfamNusYVrpqST8+41ACSY40g0iuHwmDcQRBSh2Cw5KLNBTUlqbFQXzg2VHbx+47SIgRkT4NqPYP9H4HdKIowZE770+xueHsfvzxuF0+snVqdGqLfB3mwphGXOhXP+AfFHXu8YgLMT9i2DdX+U1JPn/lLSCtOFHGFHdCotKr20sNY/tJc/X1rb+ZZQ3+UMN8frY2d9D+XNdtnBfAs4mTOYeaIo9td4uAdYIYriw4Ig3BPavlsQhOHAJcAIIB34TBCEYlEUA8A/gRuATUgOZhHwEZIz6hZFsVAQhEuAR4CLT9QHO9YkmrQkmrQkmbT8cFYe/w6JXS4ZnsBtb5YPGP+380sASUvLMUiXTVe/8JWxXyZNklFDfbczqpvkofZe3t/VxH1nj+DmuUV0710edi4ABWX/4Lez/sD/relg3cEOFAL8eH4RVpcfT1CFX6ljU1VteHyLzU1WvD5KxBEgTq8hI16PTq2ioaWT56e1kVv1Kkp3EJf/Ul47O5v3a0RanQLn5PqYZG7FZIgFoRNH/Aje3NpNcqyWiXk57Kjtoa03CKf/AWrWka4P8vgZKeywGenq9RKrU1MYG8TsrEarUrCgNJl3d0WvG/kCIrsarIzrq3MQBMmh/A+ncjhqpYK4UOFsT9J4dp/yHg3dTjLNBkYlJfGVqigOrYIP+s14lv0ILn0tkoWWkBc9fversOgRWPsHKRyXOxsW/V5qKvYtwRcQcfsHNrBz+wbukxl6DKUQ2dnA3ND//wOsBu4O7X9VFEUPUC0IQiUwOTQLihVFcSOAIAgvAOcgOZizgXtDr/Um8LggCIIoHt5u6duFQavi1nlFzCtJptnqIidm8Dx8nSAlB6zY38ZZY9N5bkNN+DG1UohqgnbOuAxe3VKPLyCiVSuo7RyYIry9rgefXzp0Sn/04+qOfVys+CXFV73A+no3aqWCdquL7XXdrDtoIyW2kpvnFvDwx5Ju15oD7dx31gj2t9jDysTD02KI06v4y2f16NRK7l6QQ4duPK/bRuALiCxQqhmnquaemI0Iaiu+uDko8xaCwQzA7soOnloXvUZT2e5g9C2nkDviHAK93excc5BH10ZqguYXxnHdzFFcPd3KsNQYdvWrkI9w7H4ubp+ff64+xJNrI0KT183I467TiqOKAo9IMABbnx24f9/bEQeTPl7SQtv9WuhNbZAzA248UwqLxaaBNgb8PvDaJWn/Ibz2ApJkzOKRqXzQL5vSpFVRlGI6ru/rCwSpau+lze4mLU5PfpIRxSBtJWS+nJPlYETgU0EQROBJURSfAlJEUWwGEEWxWRCEvtSZDKQZSh8NoX2+0P8P39/3nPrQa/kFQbACicBAVbxvGXEGNdMLpdi7q7Oec4fHsrQs0sBrZIqOOJXkeJqtbnp6ffz6jGG8srmeDLOOH88vZJiuC9e+zxGUasamjuD1G6exvKyVgChSnDxwMXtaQSIOj4/GWhdFSUXQX1MLEIyJaNVKOh0eJuUmMDbLzHXPS5lqrTYP8UYNz10ziS21XWTFGzDrVdwyrxBfIIhGqSAzQc9tr+4MR3KanQI/XVoVTsd+cbPAi9dMwBqbhksXoNBlJbWjhaRQ/UWzNXp9BaQMvFabm9wkI7VONX/eEL2YvrLSyphcB0+trcKgUXLHKcU82C/OH6tTfeWF5C+jqt3JU+uqovY9u6Gac8dnMCoj7n+/gKAAcw7UbojeH5cV+X9MCiz+o1Sg6e2VpGj6Pw5SHc+Gv0P9Jhh2pqTvlVjw1T+Q2woK9dFlwX0DdGolP19USlaCgXd2NjEsLYbbFhSFm/QdD7z+IEu3N/CrZXvxByU1679eMo5FI789ocWhwslyMDNEUWwKOZHlgiDs/5Kxg902iF+y/8ueE/3CgnADUoiN7OzsAU8Y6ugTMvnZlBampgRZ0ahkSnKQedkqXqqULrgWk5ZFI1OZPyyZiyZmoVUpCLbsQf3cBah7pfoXf/Iois59Du3oNERRSh8+d1w6S3dIi/AFFiOnj0zltL+swx8UGZsdx0sXvYl23YOo2vbiLTmLwNRbGJeaybhCaV3HFwjywY9n0tDtxBKjoyjFhEGjYl6pdM9Q3+Vk3cFOXt9az+JRqXxa1hp2LlkJeg602MLOBeCHs/L55TvlVHVIsyetSsHzFwUxJHsx6DSkxGoH9KjJSzKSYJJCU25fIOr1+giE3tTpDfDh7mb+cdEwPj/QRLI+yMKEDuKV9cCxWQzv9foHtCu+fKSBgvYVsPNzSBkB+XMHhrlEUZKMUWokx1G2FHyh1HRdHAw/O3q8LhYyJw1uRE8d/Pf8iATNxseheTdc8lJkHed/fpAOSWNt0xOSuOecuyF3FqiOssne1yAn0cjPTyvlh7PyMWpVgzZ/O5ZUtTv45bK94SQSjz/Iz97YxbC0mG9VevZQ4KQ4GFEUm0L/tgmCsBSYDLQKgpAWmr2kAX2Vdg1A/9uwTKAptD9zkP39n9MgCIIKiAO6OIzQzOkpgIkTJ377wmeCQHr+aC6OPcDFRd1S+COpiKvSFcwekYvZoGJ4WhxqpQK1XgEBP+7N/5bi8SFUbXsQq1Zz6fJsrG4/0/ITeei8kVw7Iw+3L4hKARc9uSl88d5ZZ+WmNUk8dcnLqAK9aIxJknBjP9RKBaVpsZSmSRctt8/P5uouKlrtJJo0jMk0c99ZI7hpbgFatYJ7393HnkYpRKVRKqNSnDVKBQlGTdi5QCjN+PMensywYtBZKEox8aszhvH3lZV09XopsBi569SSsABjVryBKXkJUU3NYvWqKBXhHfU9+Io9PNBxF3js4GjDt+iPwMQjH/+mHbDrNbA1wthLIWfmES/UOQkGMuP14bqlEakGfhq7AsOyv0cGpY2Dy16NLMJbG2Hbc7DlaUmrbP5v4QcrpGwxQQkZ48N9UY6KjooofTMAatZCdw2kjZa2RVFKFuiplbLUkkqiZynl70mpzn28dD5c+wlkTzl6O74GCoVA4glSL26ze6K6qYKUadjp8MoO5itywh2MIAhGQCGKoj30/1OB3wHvAlcDD4f+fSf0lHeBlwVB+BPSIn8RsFkUxYAgCHZBEKYCXwBXAX/v95yrgY3ABcDKb/v6y5HoDSrZ586gvjuB1DgtmiYfN7+8nXa7B5VC4O5FpVw2JRujVoXL6UDfum3AayjbyhiTNZq1BzvYWNXJJ/ta+dEcKWzyxtb6ATpM6ys76PCqyIxPi+x0dkl/xiTQm6PGf7S3hTteizQHG59t5h+XTwjLzP9gZj4rytvw+IMcandw2ZQsPtgjddWM06vodEQnBICkWuBS6DEDKbF6Fg5LJifBQGevlwyzjuKU2HDtTYxezUPnjeKpNYf4pKyVkemxXDAhi98s2xt+PY1SQZG2O0rd+EsrvVv2wPNnSKEokNKOz/s3jL5o0OHJsTqeunICf/jkABurOvnFVC0Jn/4zelDzDmgriziYnS9JC/QgVeK/eQ1c8TaMv+rIdn0ZqkGq7hXK6BYDh1bAq5eDPxR2nPtLmHazdPPi6hnQvhhRlFoXHGcHcyJJi9NF1Z+BlGloiTkxDu67xMmYwaQAS0MqpirgZVEUPxYEYQvwuiAI1wN1wIUAoijuEwThdaAM8AO3hDLIAG4ikqb8UegP4BngxVBCQBdSFtp3jkBQ5NXNdeEagbPHSr1j+qrm/UGRBz8sZ3yO1D0yqDLQU3A25rbozDNX1gy66yOqAS6vn/X7m/D4/FLvDI0yqmHY2Cwzcf1lyuu/gHd/Au1lkDYGlvw5nG3VbHVx//vR77e9rofyZiupcVKa6ficeN6+eTr7Gm2olAKZZj2PXjCaN7c1oFEKFA+yoLtweAoGbeTCmKVxkKXaC3obqFJAMwb6KbwVWEw8cM4o7jy1BKPfxvJqD6eNTGVFeSsZ8XouGJ+JPrgl8gZqA2RNPvLBb9gScS59rHkYCk8BQyQ3zOMP0OsJYNarGZ4exxOXj6fH6cPiqg43G4uir8OprRl2vDjw8bqNkgLA18FSIqUpV/WTCJ16a7h5F7ZmeOeWiHMBWP0QFMyTjoVSLYXFDudow2vfEvItJv588Vh+/uZunN4AsXoVf7tkXPiGSOboOeEORhTFKmDMIPs7gUHPHFEUHwQeHGT/VmDkIPvdhBzUt51Wq4s9jTbaHR7yLUZGZcSFs45qOnp55ONIV8W+hdDDaepxMSEnHlEQaM5agrrrAMaKZaBQ0T3uJqr1o3H6pBnDVdNy2F7bTavNg1GrYnN1Lf+8Yjw76q34A0GMWiUzC5MiDaa6a+HliyPNs5p30bv6L5RN+QN1Vj+Z8Xp6nANnIA5PQJIzEdSgUjMiPY4R6dJid4vVzfrKds4ek4ZGpSDeoOGOU4p5cWMNNpefJWPSmJybwH831bKzoYezR6cy3foBCavujrzB4sekzof95NjVKgUpsToaOrzcs3QbmfF6zhiVRqvdwwMflJN0zmjyhp0FsZkw+kJIHXXkL2awCbEo0n+pb2+jlX+srmRXvZVFI1O5Ymo2eUnSehSGXBhxbrjfPSApKltCmm2CIG331EW/h+4oEgKOhDEJzvob1G2C9nLImCg5jr71E1cX2AfWKIX3aYzSmstL50c+vz4ecmd+fZuGIEqFwBmj0hiRFktnr5eUWJ3sXL4mQylNWeYwOuwe7n57D6sPRNZMHj5vFJdMlhISej3+qMXrph4XeUlGqjui76z7Zgo6tZIXykXG5PyCsSNvJoiKz1r0KNsUtFhdqJUCpakxWGK0vPxFHTaXj4smZVHd3hvWIKvrdNKRFnEYga4alP06Mwbj83kz4Uf839NSKG5CTjynjUiNKq7UKBUUarvh2Uulzo4zbqPWMJKyZjsqpcCwtFhunldITacTURTZ02AjwaDm2um5IIBZp2bdwQ5e2ypp1y8va+OXs0ZzQ1wmWEOJhct/Q33iNJbVaChJi2FclhlLjHQcFCo1GqWCQ+29HOrXbVOlj4WLB5k1DEbmJFDrIwvuALPvkmT2kRIZrn52c7g49pn11RxstfOPK8Zj0oayrxbeB8kjYN9bkDlZWsTvK7yMSZWapC29MTLTic2QUpG/CeYs6W8wTCmQWAid/XTdBCFaUTl3FlzzkRQW08VB3ixpnaa7VhoblxXl1L+tCIJAnsVEnuV/j5U5MrKDGcLsb7FFOReABz8oZ0ZhElmhReMCizF8kfxgdzO/XFzKnz87SI/Th0KAOxYWMyxVCmGolAqumJbDVc9ELnzDUmP46yXjmJqfiEKA6nY7j/WrnD7QYkerUvD0umr8QZHsBKnPS6fDQ6JJS6/CRGy/avHa4Tfx0JpI2vS22m5unlvAZVOyeH9XM7mJRn4510Lp+2eDoxmad0LlchrmvsJNH0ihvewEA89dO4nSkN2BoMglT21iUm4iMToVE3PNYefSx9822zlz+nWkbQx1sfQ52V3VxGMrpM956aQsfn3mcIwaFelmPXeeWsx970VUoNPidAxP/wqzg7TRcPX7sOO/YK2H8VdD/pzww5VtjgHKC2sPdlDX5WR4Wuh94nNgzs9g6o+kkJzisOyo7Klw7pNSozSVVupHn/ENHcyXYUyS3u+NqyVHrTFKac+WfokEKg3kTJP+AOytUmhww1+lVOpZd8GEqwfK2Mh8L5EdzEmgJVS30TezOBL2QRoa2T3+cBVzgknL45eN58EPyllf2UFhsonRWWbeu3Um9d1OzHoNBRYj2n5pnSPS41h683Qq2hxolApKUmJI6WfH8rLmqPebXpDEHz+NhOHqupy8s7OJ2cUWEtFSK2SQOu52LNv/AoBDMIYLKPv4x+pDvPLDqdw2vxiD6CT2+Xng68VZfA4qVweahg3kuPejURbgDQSp63Ky+kA7BaFahxHpcfz3+im8uqWeFqsbtXJgmmpQFPHEl4S3vWmTWN2qA6SL/Ctb6rl8ag4jQzUn543LJDNez4ryNgosJuaVJpP9VTOEMidKf4MwWCqtWimgHcR2tANrjwCIy4QR54G9WVqIN52A2+nMiVKmmrVBCn8l5H/5jOTQZ7Dmkcj2yt9BQi6MPP+4myoz9JEdzAmkx+nl7e2N4cZbty0o5NxxGSQYB89OKbSYSDBqOG1ECokmLfsabYiIUT04hqXF8uSVE+js9RKnUxEX6l74ZTHj7ETjES+mGbHR9QyDyXTE6tRUt/eyu8FKolHDU4ElLJw7kQR/O37LcIqTbVT0a61r1ChJjdVKDtXpoqPoPD5TzeOF/QJpRrhxoZfUQAvBfusalf2eLwgCY7PjGRuSbfmsrJU4vTqqncFFE7NQCzWgUBEoOp216Tfw1seRmRREy4vEGdScMjyVU4Yfn+K54hQTMwoSw+2qAW6aWxjVHuGoUCgiXS5PFDGpR6dXJoqw85WB+/e+LTsYGUB2MCeUDZUd/O79SFjm/vfLSY7RceaY9EHH51tMPHDOSO5/v4xmq5vJuQn8YnHpgNbKRq1q0HbLR4MoiuxttLK+shOlAPMtVjLNWhp6pHCVThWdqju3xIJBq+Tq56SMK0GAB88ZySPbvGyvU1CSaufu00t5Zn01nx/qpDQ1hvvPHkleX+W1IZ73TRdy70eSnloZsLZa4MmLp+EPlke9z5HwBYPcNKeAijY7TT0uJuUm4PEHUeROhdt2YhNiefDf2+mfXV1gMZ3QGoZEk5ZHLxjD1touKtscjM0yMz47/qibXH0rEAQpVFizLnp/yoC8G5nvKbKDOYG8fVgjLpDqTI7kYCpa7dz2yo5wHcrmmi4e/KCc566ZFMni+obsarDyt88qGJ1lRqkQ0NtreWlcE3tVI3CLSkri2mifkMB/tklFipPzEni0X+aaKMKjnxzg7tNKmF6YRH2Xkzte38mym2egUiiI1auiesJ3Ojz8a0N0GM4XECnvDGDUKBGB2xcUMSXvyH3eS+IVlNf0sLPORqJJyzs7G3ns7AKS4uNBlUg88K8rJvLk2kNsqOxgVmESN8zJP+F1DBnxejLiT/Ds40Qz5lLY9YpUAwVSc7TD1QVkvrfIDuYEUphiYsX+6FbAXybaV9PZO6DIcWttN8029zFzMBUtNoLAXz6TwnbbitO5P72OM9adEx6Te9EbnDd1Bna3ny6nZ8Br9Dh9NFndPL6yErVS4LELx5CbOLg4oFIhYNQOXIewxOj48PZZKBUCGWY9wpfE/fM8+7ml4XecPv467EEtWZ6DpG74A0L2q6CS1lhKUmN4+LxR2Nx+4nRq1Krv0MxhKJE6Cq5bDq17pUX+lJGQmP+/nyfzvUB2MCeQs8dk8Orm+vDaQaxOxXnjM4843jyIE4k3qDF9zXDYYLTaPVGZaqsrOvkg5xRuLN2K0F0FU2/ClDOBMUYzAPubbQN0v0amx7JkTBoTc+NJjdVRYDEdUXnWbNDw01NLuPml7VGfaXy2+ahDWIKtEW3LVoa1HNbe2dUN+kgmmEalJMl0HHWrAj7oqpKakJlzot77e0VSofQnI3MYsoM5gQxPj+Wtm6ZR1mQHRIalxVKUcoQMIqA0NZZzxqazLFQ8KQhw/9kjoxb5vym76nsG7PuswsoN1z+DIPpAGz3DKkqJ4amrJvCLt/fQavMwPtvMA+eMoiQllpKUo6vonldi4aUfTGHVgTZSYnTMLk6icBAV5/602dyUN9tx+wIUmaeRrzNHd2zMmgrGga2TjxvObtjyb1j7qORosqfDmX8FS/GJs0FGZogjfEclur4yEydOFLdu3fq/B55guno9lDfZ6ej1kJdkpDQ1Bo3qG9yV+9zQsBn2LgWThWeE87j/40NRQ+5YWMTtC7/8Qtlqc2Nz+UiJ1RF7jMJ1R6Kh28ltr+xge10PIPUD+e+FGYx9e45UhBiXCec/I9WNnCgqP5OUifsz4VpY/IcB4p8yX47bF6CyzUGb3U2G2UCBxYjqBCdDtNpc7Gmw0dXrJS/JyKjMWHRq+f77aBAEYZsoioPm68tHcIiTYNQyo+gYLk7XrJekPkIsmBHP+1kj2VEvpfQOT4s5YtJBf1JidaScoJa1W2u6w84FwOHx87eN3fxj8d/R2aqk0NjuN6RCRNUJuri37hu4b//7MPcXUl8WmaPC7Qvw0he1PPBBOaIIKoXAXy8Zyxmj//dv8FjRYfdw91vRihl/vHAMF0w4cvha5uiQHcz3CZ8b1v85alfuhnt4+txXORQ7jaAoUmAxDTnV2IbugV02y9q99NbvRrcrpEisUMGM275aj3uAYPDrdXWMzxu4L2MiaL9bwo/Hm0NtjrBzAUmg9e639jAyI+6EpZUPppjxu/f3MS0/kYz4YxeO/j4iO5jvFSIEBmaBJXZsIXHM6SfBnqNjdObAxfOzijXE1y+P7DAmSdpgRyLgh44DknikMVmaZVSthp0vS5lP466I9EQ5GjImQtEiOPixtG1IgLl3g0a+IH0V2h2eAbqhDo+f7t4T13vF5h6omGFz+XH5Bm9JLnP0yA7m+4RaD9Nvg9evjOwTFFB02smz6SgYmxXPb5YM47FPK3D5AiwemcrlqZUo9vYTZVz0iFSDcSQqPoI3rokIR06/TWobXL9ZakO853W4fjkkFR2dUXHpcO4/oa1cUoVOLJIkUmS+EhlmPRqlIkq01WLSRskXHW8KLCa0KkWUxNG8EgtpxzCZ5vuKvMgfYqgu8h9zPHaoWgub/wXaOEloMWsqKIf2vYYoitR3OfEGRDLj9eh6G6WOks4uqatj2lhQH+Gi1F0HT82OtBToY96vYFW/LhAXPi9J6MucMAJBkc/KW/nZm7uwufwkx2h5/LJxTM47cRmBoijyRVUn971fxqG2Xk4flcptC4rCWngyX468yC8TQRsDw86A4tOkvOfDFXyHKIIgROunmbOjZeS/DFf3QOcCg4QL5WLME41SIXDaiFSGp8XS1eslJVZLatyJnTkIgsDUgiRe/eFUHB4/STFatN8kU1MmjOxgvq8M8RnLMSUmDcy50FMT2adQgqJfxpnRAqmyhtbJIivBcNKbesUZNGGxWJljw/foKiPzXae+y4nV5SM1VkdS/0y4mGS44Bl483rJyejMsOQvUkht+DmQPByGnQmJBSfHcBmZ7yiyg5H51uMLBFle1so9b+/G5vKTk2jgb5eMZUxWfGRQ5kRpEd/eJGV89YXXSoZu9pyMzLcdOegsc1ywurxUtjnosA9Miz7WVLY6+PErO7C5pAyx2k4nP3ltJ52Ow947JhnSxx792o2MjMw3Qp7ByBxz9jRa+dXS3exusJEZr+fh80YzozDxSxWSvwn13U4Ch6lOV3c4abW5STQNraJRGZnvE/IMRuaY0unwcNsr29ndIEnPNHS7+MELW6hq7z1u7zmY8kC8QU2c/iss2PpcUL0Wlv8WPv/74FIwMjIyXwl5BiNzTGnqcVHdES3t4vYFqe3spSD5+NQVFKXE8OP5hfx9pVR4qVIIPHz+6K8m83FoBbx6eWRbHw/XfiTV2MjIyHwtZAcjc0yJ0akxaJQ4vYGo/fHG45f+adKq+NGcAuaXJtPp8JCVYKTwqzgztxVWPBC9z9UtVfnLDkZG5msjh8hkjik5iQb+78wRUft+MDOPov/R7+WbYtSqGJcdz8LhqZSkxqA8QsOzQQn4wTdICM/nOnYGysh8D5FnMDLHFEEQOGdsOqWpMdR1OUmO0TIsLQaTbgj/1IyJkjbZh3dF9ilUkDX55NkkI/MdYAif9TLfVrRqJWOyzIzJMp9sU46eEeeASgtfPAmmVJh1B6SNOdlWych8q5EdjMy3h2AAxAAoj8N6jtEC46+CEedJHSlVcnqzjMw3RXYwMkOfYFBq8/zFv8DeApN+AAULwBD/v5/7VdHKCroyMscK2cHIDH2ad8J/lkDAJ23XbYSzn5CahMnIyAxZ5CyyY0BnVycOVyQLqanbSVdvRKak2eqk3RHJSGqzuWizucPbXXYnrVZHZLu3l6aeyLbd5aKxyx7edri81Hf14vFE3qOhy0GvK7Ld2O2gO8omB129ke1Wq5N2R79tm4N2e2S70+6k2dpv+zCbbE4nTV228LbX66W+yxZlU32Xjd5+NnR1d2FzRT5Ha7edHmfk8W6rlR6HdaBNfhUEU+n8STPNtzfCFR/A+j/RZXXQ2B2xyep0Rh0nj8czwKb2rk7s/d6ztbsHez8b27qtWJ2R1+zu6aHHGrHJau3G2hOR/rc67bR1Rx53uRzYuiPtd11OO46uliibOro6omzq6WrD2c8me3crjn7Hobu7G7u9J7Jtt9Ftixz7LoeDDmvkc9ucNrq7+9no6qW5O/K4y9VLR1fXgOPkcEV+o/VdDmzOyHZXTxe2fr+fNquDbnuk3qnd7qbTHhnf4eilvd/vp6vXQ1O/1tdul4uenogNLpeXhq7IeK9X2na5vOF9dV0O7P1+4009Trp6I+dRh82JzRkZ32Zz02mPPG51eujot+0PBHF6I10r7S4Pjd0RG4JBkV6Pn/49s7p6Pfj7NSZrs7lweXz9PrcHhyuy7fYF8PqjU/a/jMNtOhp6PX6CwaHZ1+s7PYMRBGER8FdACTwtiuLDx/L1W+oqMBx8j8SKt3AljaJ+0l18WKvgnZ1NpMbpuHF2HpXtvby2pR6TVsU103NBhKfWVREU4ZoZuZh1Kp5cW4XN7eeqaTmkxGr5z+e1NFvdnDkmjYm5Cfzn8xoq2xwsHJ7M7KJk3trewM66HqYVJLBkdBof7W1l/cEORmfGcfGkLNZWtPNpWSsFFiNXT89lV72VZTsbSY7RcuPsfKo7nby6uQ69Rsk103NRCQJPrq0iIIpcOTWHJJOWJ9ceosfl40ezC4jRq/jvxjoae1ycMTqVKbmJvLiplv0tduaXJjN/mIWl2xvZWtvN1PxEzhydzif7Wlh3sIORGbHcNdWI8dCHJOx/DVd8KdZJN/JxZwovbKonyaTh+uk5jPNsJXbznwmqjHRNv4dNnlye/7wGX0DkiqnZpFy+iiff3E23y8dFE7IoOWM5z7y7n/ouJ4tHpjGtIJH/bqqlvNnOvFILC4elsHRHI1tqupicm8Bt49XoD7yNpfI9nKmT6B53Lcsa4nh9ewPZCQZunJXH7gYrb2xvJNGo4Uez8+i22XhmUzNalcAN0zMZp23AtOERBL8L+9Q72a4eyzMb6uns9XLB+HSmZcfwjzWHONjl48KRTZxWqCd+298xtWzFUXAGDSVXs3RXK6uqepmTa+S8sSlkHnwFc+VSnCkT8E64Em/DDmL2vIgnvojqafeytiHIG9sbideruW5GLm5fgKfW1aBSCNy+MB+7O8h/Ntbi8ga4bHIWY9N0PPbZIRrtAW6YYsGSmMiLm+qo7ujl7DHpTMqJ5ZWtTexttDG7KJFTR6by3q5mNlV1MiE7nnPHZ7Cmop0V5W2UpMRw8aQsRnZ9TML2f+GOy6N6+u9Y2yjw5rYG4vQqrpqWS168hruXlaMUBB46u4TKTi//2ViDyxvgqmk5WGK0PLehhja7h7PGpDMpN56X1lewr93LJaPjGZabxpvbGtndYGVmURKnj0zl/d1NbDzUxdhsMxdNyGTlgTZWlLdRmGzi6mk5bKvt4d1dTaTF6bhqeg6dDg/PbaglRifZFKNV8djyCgBunJ2PIMDzn9fg8Pi5aGIWI9NjeXJNFbVdTm6ak0+CScsrm+uobHOwYFgyC0pTeG9XE58f6uTU4SnMG2bhg93NrD/YyZisOC6amMWainaWl7VSaDFx+dRsypttvLW9kZRYHVdOzUYliDz2WSUxOjU/mlPA5LwE1Moj39PvqOvm3+uqqOlwctnkbE4dkUJy7JE7e9Z29rJ0RyMf7WlhSl4Cl0/NpiQ19lhe4r4x31kHIwiCEngCOAVoALYIgvCuKIplx+L1rVYHsVv+hmHPiwDo3VY+MFzNw+ukO8eazl7GZpn5U+hHDrCpqpO7F5WytVYas72um1+cXsrGqk6CIjRb3fzmnb24fdIdUlmzjWtn5LKzvoeGbhcTcuK5+81dVHdKd4IHWu3sarCSHqfnQKsdlVLgpS9qWbqjCYBup5d0s55/r6sG4JBawficeB77NGLTF9Vd3LOolC1hm3q4Z1EpG6u6CARF6eR4cRsuXyBs01XTctjZ0ENtp5MDrXaqOhx0O31UtDqoaHWws66HG+fk8/T6ah44s4S4XX/CsO1f0nFq3Ye+ejljlixjX5N0F77xUBdvneplVONWlMD+wtu4+YMdYRt31Pdw96ISNlZ14Q+K7G3cx63zCvmiqgury0d5s53LrFnsbrBS1dHLgVY7B1sdOH0BKlod/HJBNrHrfoWu8gMADG1l6KqXc+r5y7jvAzsef5BN1d08+skBADRKBVMLEnn04wNhGzZWW3n1DA1Ta9cCUN3t5Qef7MQXkO4c9zXZ+MWiEqw+BeXNdn7XbKe1J56fN2+BjnIccx7h7ncPsatJukOuaHWwocbOM4tmEfP5gxjayqD6EzSjL4HWvWi9DlbUeHngo4MRG6q6uOvUEnbU94S+3yC3vboj3NN+V4OVP5w/krJWFw09bpT6Am58cTt2j3RH/NvTDdz22h6arNJd/AUTMrn/vXJ2N1rDNm2p6WZibjz7W+zsb7GzvrKDp684hXEtt6Bz9bCmLsC9HxyIsumZqyawvU6yqc7q4+aXt4dtarF5+OXSveF2xPuabNy+oIgDnT72t9jJWVTKT9/YTUO3K/yb3ttoZVpBIgda7Swelcp/Ntby7i7pN213+8MOq+/3uL6yg1vnF7In9Dk2VXXy5JUT2V4XORcf+Xg/fTf5uxus3HvWcLbUdtHh8JIcq+PG/26jxynNPPa32GnsdpGTaOBAq50DrXY2VnViidFyoNWOVq3gxU21vLOzKTx+XWUHV07LYV+TjX1NNjZUdvDklRPYWS/ZtKGyg9dvnMbE3AQGY3+zjUv/vSl87v/6nb10Ob38eH7hoBp+Do+Pe9/dx6oD7eHjtvJAG6/fOI30IdTq+bscIpsMVIqiWCWKohd4FTj7WL24r6cWw76Xw9utRZfx5LZIaGVGYRKf7GuJek5QlJR++2tnbazqZGRGHAD+YDD8A+vjrW0NnDI8BYAkkzbsXPrYUdcTrlqfU2zhvV3N4ccWlKawLORsAKbmJ7K8rDXq+aIIle0O0vr1QF9f2cGYTDMgnZx9zqWPN/vZBPBpWStT8iInzu5GK8rQSZGr7sGw85mo5+OxkeSsDG96A0F2uSxSdlhSMe/WDbxrW1vRwficyKL+e7ubmFtiCW+/tb0xyqYV+9uYlCuNH67vCjuXPhT2JnQ9kg1XTs3h5c114ccm5Maz+kA7h/NBgx4S8kGtZ0dvYti59PHiF3X8eE5OePv5HVaaSqR1oloxJexc+ihrdVLj65eo4GgDtXRxaJn4c/67uTFqvD8o0u7wYNKqmJgTz+eHOji84/nLmxu478xSALp6vWHnAtDe6w87F4AkkybsXPqobHeQ0u+uubPXS2WnFJZqm/hT/rsl2qZAUGRLTTc/mpXHny4cxdqDEZsUArh8gahe9wAvfVHLFVOl42R1ecPOpY8vqrvIS5K6l47NMvP+7shv+JThKby1vSFqvMcfxBeIvEdQhM3VnUzMNpNk0lDX5eTwCNJrW+q5appkQ3Vnb9i59PHe7mYm9Pu9ba3tpjhFKhaeVWTh/d3NUeOtLl/4N99n06F2BxplxKZVB9o4EuUttgHn/lNrq2jp9331p7bTGXYufTR0uzjU7hh0/Mniu+xgMoD6ftsNoX1hBEG4QRCErYIgbG1vH3hB+TJEQRmVLqvy96JXR9qsenyBQduuatWKqJNBr1biCf2wFIPcqej6PX4kMeK+/f6gGDUF9/gD6DT9t4NRNoZtUinx9rsI6DVKPKG4sVo58E372wygVigGqBmr+uxQKEA10GEE+3eTBLRKEcQg+D2Y1APjydJxiDg6vVoZdeHSqZV4+x1XlUIIX1REhWrQ1tBiyAa3L3DYdzf4cTJpRAh4IRhAqxjcxv7HQatSoApIFwj1EZQFBoRMBGlb6e9Frxlog1opEAiKOLz+QW2UZHqk46A67D0P/y6PpHZw+O4+G5W+3kHfU6dW0mJz0WJ1Y+j3uAhRF93+4/u+O5Vi4CVIECLngiAw4Dc9mA2Hnzt6tRKH148vIKJRDXwPvVoZvqCrB7FBrRQQiLymIETOs0AwiGaQUNdgx62/YpJRc+SA0WA26NVKlIOcfyB9t4N9fV8WgjsZDC1rji2DfTNRVwVRFJ8SRXGiKIoTLRbLIMOPTIwlD8eUO8LbiQde5q5pkRTXzw91smR0WtRzDBolKTHa8N2SSiEwKTeBA63SAqwvIGI5TF7+qmk5fLRXuluqaLUzvSB6ir1kdBpfVEuLpR/tbebqaZE76OVlrVw2OdL7ZHN1F4tGpkY5Kp1aQaZZR2evtDiqVAhMz08Mh68y4/UkH6ZWfPX0XD7cE7mDu3RyFp+VR2ZGp41IoTG0oLuuTYd92s+jnu8351Kjzg9vmw1qxqpqIOiHnlqWZPRGncAKAWYXJ0XdbV8wIZPV/e4Ir5meG3VXefGkLFaEbPqkSU/vuBujbPCmjqfbKHWw/Pfaan4wKy/82I76buaWWKJOYK1KwaIUO1gbIOBlvKaeOH20k7xpTh6/eGd/ePvOGfGklT0LQG7vTpYMj06rPq04jlx3eXhbTB0lpWEDlo0PcuOM6L41sToVJq0Kly/A/mY7k3IT0Kkjx0kQ4Kpp2fzsrd2AdKOQ2U/wU6uEMZlx4e09DVYWjYzM+gBmFiZS1hxJCChKNlEUJ502iV88wg0zMqPGm7QqJuSYWbarhUc/PcjMwqSwAxBFEBFJOEyH7kezC3hq7aGwzeOzzVGPnzs2g88rOwD4YHeTtHYZ4qO9LVw5NSdqvMWkjZpNGjVKxufEs7/FgdXlIzlGi+EwZ33ltByeXS+FjjPi9eEZUx/XTs9l2c7ITOnM0elsPNTZz4bo7yYvyUBPv8X9RKOG3MTIa+rVSmYXH/kaMyIjjuTY6PPsZ6eVkBwz+BpMbpIxPAvsY1JuPEXHSVD26yKIh8+xvyMIgjANuFcUxdNC278AEEXx94ONnzhxorh169av9B6NjXUYW7ehqV2DL7EEZ96pbO3Ss+5gO6mxOuaWWOhweFlT0Y5Jq2J2cRJiEFYcaAMR5pUmo1MKrKxox+72M68kiVidhvWVHTRb3cwqSiIn0cD6yk6q2h1MyUugKNnE9roe9jXbGJtlZnRGLHubbGyv62FYagyT8xI42OZg46FO8pKMzChIpMnqZm1FO8mxWuaVWOjq9bH6QDsGrYrZRUkoFfBZeRv+oMi8kmSMaiUrD7RjdXlZOMyCUavh80OdNPW4mFGYSF6Skc8rO6lsdzApL4HilBh21nWzt8nGmMw4RmfEUd5iZ2ttN6UpMSwq1BLbvh119SoC8QV4c+ex3W5mZUUHFpOWOUUJFPkqUB78CFGtJ1B0OnsC2aw+0I4vEGRuSTIxWiWrDrTT1etlbomFZJOGdZWd1Hc7mVmYRH6Skc8PdVLR5mBSbjzDUmPYUW9lT6OV0RlxLMwWMLVtQVO7Hl/ySDxZM9nYaWT9oS6y4/XMK0mkscfD6ooOEo0a5pUk4ve4WHWgA41KYG6xhUJlG4qKD8HvJli8mAPKQtZUdtPh8DC3OIniBBVrDrRQ3elmToGZMRYBbe1qlG17CGTPpN0yje0NDrY32hmfEcO4zBiSOr5AXbsWX/JIVDlTCXbVoKxeRSAuF2fhEnZb9aw80E6CUcOcogS8AVhe3oZaqWDxyGTc3iCrD3bi8gWYX5xEbhx8uq+FFruPhcUJ6GLi+fxQF7VdvUwvSKI4Uc2mGhvlrb1MzI6jJC2OvY1WdjZYGZEey7iseA622dlc3UWhxcSkvHgKbV+grvqMQFwOvUVnsrtHz5qKduL0GmYWJZJo0PDK1noUgsCVU7JptLpYd7ATp8fPvNJkYnUq1h7soM3mZlaxhdx4Pdsq66hsdzOrwExSYhLb6qyUt9gYnx3PiPRY9jZa2VHfw4i0WCblmdnf0ssXVZ3kW0xML0ikodvF2op20uJ0zChMwuHxsXJ/OzE6FTMLk9AqFXy4twUEWFiajCDA2oMd2N1+5hRbyE3Qs+ZgJ409LuaVWjCoVWyu7qKqw8HU/ESGpcVQ1mxnd0MPk3Kl825nfQ/barsZnhbLpLwEKlqktZn8JBPTChJo7nGz5mA7KbE6ZhYmoVbBR3taidWpmVeaHA6FH4mDrXbWVrRT1+1kXkkyE3PiMenURxzfZnezuaqLjVWdjM40M70ggawE4xHHHy8EQdgmiuLEQR/7DjsYFVABLAAagS3AZaIoDtro4+s4GBkZGZnvO1/mYL6zWWSiKPoFQbgV+AQpTfnZIzkXGRkZGZljz3fWwQCIovgh8OHJtkNGRkbm+8h3eZFfRkZGRuYkIjsYGRkZGZnjguxgZGRkZGSOC7KDkZGRkZE5Lnxn05S/KoIgtAO1x/EtkoCO4/j6xwLZxmODbOOx49tg5/fdxhxRFAetIpUdzAlCEIStR8oVHyrINh4bZBuPHd8GO2Ubj4wcIpORkZGROS7IDkZGRkZG5rggO5gTx1Mn24CjQLbx2CDbeOz4Ntgp23gE5DUYGRkZGZnjgjyDkZGRkZE5LsgORkZGRkbmuCA7mGOMIAhZgiCsEgShXBCEfYIg3B7anyAIwnJBEA6G/o3/X691HG3UCYKwWRCEXSEb7xtqNvazVSkIwg5BEN4fwjbWCIKwRxCEnYIgbB2KdgqCYBYE4U1BEPaHfpvThpKNgiCUhI5f359NEISfDCUbQ3beETpn9gqC8EroXBpqNt4esm+fIAg/Ce07KTbKDubY4wd+KoriMGAqcIsgCMOBe4AVoigWAStC2ycLDzBfFMUxwFhgkSAIUxlaNvZxO1Deb3so2ggwTxTFsf1qDYaanX8FPhZFsRQYg3RMh4yNoigeCB2/scAEwAksHUo2CoKQAdwGTBRFcSRSG5BLhpiNI4EfApORvuclgiAUnTQbRVGU/47jH/AOcApwAEgL7UsDDpxs20K2GIDtwJShZiOQGToZ5gPvh/YNKRtDdtQASYftGzJ2ArFANaGknqFo42F2nQpsGGo2AhlAPZCA1Ork/ZCtQ8nGC4Gn+23/Bvj5ybJRnsEcRwRByAXGAV8AKaIoNgOE/k0+iab1hZ52Am3AclEUh5yNwF+QTo5gv31DzUYAEfhUEIRtgiDcENo3lOzMB9qB50LhxqcFQTAOMRv7cwnwSuj/Q8ZGURQbgT8CdUAzYBVF8dOhZCOwF5gtCEKiIAgGYDGQdbJslB3McUIQBBPwFvATURRtJ9uewxFFMSBK4YhMYHJoaj1kEARhCdAmiuK2k23LUTBDFMXxwOlIIdHZJ9ugw1AB44F/iqI4Dujl5IfsBkUQBA1wFvDGybblcELrFmcDeUA6YBQE4YqTa1U0oiiWA48Ay4GPgV1IYfuTguxgjgOCIKiRnMtLoii+HdrdKghCWujxNKSZw0lHFMUeYDWwiKFl4wzgLEEQaoBXgfmCIPyXoWUjAKIoNoX+bUNaN5jM0LKzAWgIzVIB3kRyOEPJxj5OB7aLotga2h5KNi4EqkVRbBdF0Qe8DUwfYjYiiuIzoiiOF0VxNtAFHDxZNsoO5hgjCIIAPAOUi6L4p34PvQtcHfr/1UhrMycFQRAsgiCYQ//XI504+xlCNoqi+AtRFDNFUcxFCpmsFEXxCoaQjQCCIBgFQYjp+z9STH4vQ8hOURRbgHpBEEpCuxYAZQwhG/txKZHwGAwtG+uAqYIgGELn+QKkZImhZCOCICSH/s0GzkM6nifFRrmS/xgjCMJMYB2wh8jawS+R1mFeB7KRfqgXiqLYdZJsHA38BykLRgG8Lori7wRBSBwqNvZHEIS5wF2iKC4ZajYKgpCPNGsBKRT1siiKDw5BO8cCTwMaoAq4ltB3P4RsNCAtoueLomgN7Rtqx/E+4GKksNMO4AeAaYjZuA5IBHzAnaIorjhZx1F2MDIyMjIyxwU5RCYjIyMjc1yQHYyMjIyMzHFBdjAyMjIyMscF2cHIyMjIyBwXZAcjIyMjI3NckB2MjIyMjMxxQXYwMjIyMjLHBdnByMgMAQRBWBYSy9zXJ5gpCML1giBUCIKwWhCEfwuC8Hhov0UQhLcEQdgS+ptxcq2XkRkcudBSRmYIIAhCgiiKXSHpni3AacAGJM0wO7AS2CWK4q2CILwM/EMUxfUhOZBPRKn/kIzMkEJ1sg2QkZEB4DZBEM4N/T8LuBJY0yfnIQjCG0Bx6PGFwHBJDguAWEEQYkRRtJ9Ig2Vk/heyg5GROcmEtNYWAtNEUXQKgrAaqUHUkWYlitBY1wkxUEbmayKvwcjInHzigO6QcylFarVtAOYIghAvCIIKOL/f+E+BW/s2QkKWMjJDDtnByMicfD4GVIIg7AbuBzYBjcBDSCrcnyHJ61tD428DJgqCsFsQhDLgRyfeZBmZ/428yC8jM0QRBMEkiqIjNINZCjwriuLS//U8GZmhgjyDkZEZutwrCMJOpAZm1cCyk2qNjMxXRJ7ByMjIyMgcF+QZjIyMjIzMcUF2MDIyMjIyxwXZwcjIyMjIHBdkByMjIyMjc1yQHYyMjIyMzHHh/wGV+aKgHPQucgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.scatterplot(data=df, x=\"age\", y=\"balance\", hue=\"exited\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "5d97223a",
   "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>rownumber</th>\n",
       "      <th>customerid</th>\n",
       "      <th>surname</th>\n",
       "      <th>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>15634602</td>\n",
       "      <td>hargrave</td>\n",
       "      <td>619</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>15647311</td>\n",
       "      <td>hill</td>\n",
       "      <td>608</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>15619304</td>\n",
       "      <td>onio</td>\n",
       "      <td>502</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rownumber  customerid   surname  creditscore geography  gender  age  \\\n",
       "0          1    15634602  hargrave          619    france  female   42   \n",
       "1          2    15647311      hill          608     spain  female   41   \n",
       "2          3    15619304      onio          502    france  female   42   \n",
       "\n",
       "   tenure    balance  numofproducts  hascrcard  isactivemember  \\\n",
       "0       2       0.00              1          1               1   \n",
       "1       1   83807.86              1          0               1   \n",
       "2       8  159660.80              3          1               0   \n",
       "\n",
       "   estimatedsalary  exited  \n",
       "0        101348.88       1  \n",
       "1        112542.58       0  \n",
       "2        113931.57       1  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "2f7770f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "del df['rownumber']\n",
    "del df['customerid']\n",
    "del df['surname']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "be4cd56a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# has credit cards encoding\n",
    "hascrcard_values = {\n",
    "    1: 'yes',\n",
    "    0: 'no'\n",
    "}\n",
    "\n",
    "df.hascrcard = df.hascrcard.map(hascrcard_values)\n",
    "\n",
    "# is an active member\n",
    "isactivemember_values = {\n",
    "    1: 'yes',\n",
    "    0: 'no'\n",
    "}\n",
    "df.isactivemember = df.isactivemember.map(isactivemember_values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "01e328e0",
   "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>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>619</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>608</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>502</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   creditscore geography  gender  age  tenure    balance  numofproducts  \\\n",
       "0          619    france  female   42       2       0.00              1   \n",
       "1          608     spain  female   41       1   83807.86              1   \n",
       "2          502    france  female   42       8  159660.80              3   \n",
       "\n",
       "  hascrcard isactivemember  estimatedsalary  exited  \n",
       "0       yes            yes        101348.88       1  \n",
       "1        no            yes        112542.58       0  \n",
       "2       yes             no        113931.57       1  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "dfd52bd1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# split the categorical and numerical columns\n",
    "# categorical = []\n",
    "categorical =list(df.select_dtypes('object').columns)\n",
    "numerical = list(df.drop(['exited'], axis=1).select_dtypes('number').columns)\n",
    "# numerical = ['creditscore', 'age', 'tenure', 'balance', 'numofproducts', 'estimatedsalary']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "e4d17a8b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Numerical:  ['creditscore', 'age', 'tenure', 'balance', 'numofproducts', 'estimatedsalary']\n",
      "Categorical:  ['geography', 'gender', 'hascrcard', 'isactivemember']\n"
     ]
    }
   ],
   "source": [
    "print('Numerical: ', numerical)\n",
    "print('Categorical: ', categorical)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "0d0ddb94",
   "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>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>creditscore</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>650.528800</td>\n",
       "      <td>96.653299</td>\n",
       "      <td>350.00</td>\n",
       "      <td>584.00</td>\n",
       "      <td>652.000</td>\n",
       "      <td>718.0000</td>\n",
       "      <td>850.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>38.921800</td>\n",
       "      <td>10.487806</td>\n",
       "      <td>18.00</td>\n",
       "      <td>32.00</td>\n",
       "      <td>37.000</td>\n",
       "      <td>44.0000</td>\n",
       "      <td>92.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tenure</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>5.012800</td>\n",
       "      <td>2.892174</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>5.000</td>\n",
       "      <td>7.0000</td>\n",
       "      <td>10.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>balance</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>76485.889288</td>\n",
       "      <td>62397.405202</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>97198.540</td>\n",
       "      <td>127644.2400</td>\n",
       "      <td>250898.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>numofproducts</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>1.530200</td>\n",
       "      <td>0.581654</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.000</td>\n",
       "      <td>2.0000</td>\n",
       "      <td>4.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>estimatedsalary</th>\n",
       "      <td>10000.0</td>\n",
       "      <td>100090.239881</td>\n",
       "      <td>57510.492818</td>\n",
       "      <td>11.58</td>\n",
       "      <td>51002.11</td>\n",
       "      <td>100193.915</td>\n",
       "      <td>149388.2475</td>\n",
       "      <td>199992.48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   count           mean           std     min       25%  \\\n",
       "creditscore      10000.0     650.528800     96.653299  350.00    584.00   \n",
       "age              10000.0      38.921800     10.487806   18.00     32.00   \n",
       "tenure           10000.0       5.012800      2.892174    0.00      3.00   \n",
       "balance          10000.0   76485.889288  62397.405202    0.00      0.00   \n",
       "numofproducts    10000.0       1.530200      0.581654    1.00      1.00   \n",
       "estimatedsalary  10000.0  100090.239881  57510.492818   11.58  51002.11   \n",
       "\n",
       "                        50%          75%        max  \n",
       "creditscore         652.000     718.0000     850.00  \n",
       "age                  37.000      44.0000      92.00  \n",
       "tenure                5.000       7.0000      10.00  \n",
       "balance           97198.540  127644.2400  250898.09  \n",
       "numofproducts         1.000       2.0000       4.00  \n",
       "estimatedsalary  100193.915  149388.2475  199992.48  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[numerical + categorical].describe().T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "d47bac91",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\USER\\anaconda3\\envs\\keras_env\\lib\\site-packages\\seaborn\\distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "C:\\Users\\USER\\anaconda3\\envs\\keras_env\\lib\\site-packages\\seaborn\\distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplots(figsize=(20,5))\n",
    "plt.subplot(121)\n",
    "sns.distplot(df['balance'])\n",
    "\n",
    "plt.subplot(122)\n",
    "sns.distplot(df['age'])\n",
    "\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "e99e651e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "creditscore\n",
      "[619 608 502 699 850 645 822]\n",
      "460\n",
      "\n",
      "geography\n",
      "['france' 'spain' 'germany']\n",
      "3\n",
      "\n",
      "gender\n",
      "['female' 'male']\n",
      "2\n",
      "\n",
      "age\n",
      "[42 41 39 43 44 50 29]\n",
      "70\n",
      "\n",
      "tenure\n",
      "[2 1 8 7 4 6 3]\n",
      "11\n",
      "\n",
      "balance\n",
      "[     0.    83807.86 159660.8  125510.82 113755.78 115046.74 142051.07]\n",
      "6382\n",
      "\n",
      "numofproducts\n",
      "[1 3 2 4]\n",
      "4\n",
      "\n",
      "hascrcard\n",
      "['yes' 'no']\n",
      "2\n",
      "\n",
      "isactivemember\n",
      "['yes' 'no']\n",
      "2\n",
      "\n",
      "estimatedsalary\n",
      "[101348.88 112542.58 113931.57  93826.63  79084.1  149756.71  10062.8 ]\n",
      "9999\n",
      "\n",
      "exited\n",
      "[1 0]\n",
      "2\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for col in df.columns:\n",
    "    print(col)\n",
    "    print(df[col].unique()[:7]) # looks for all unique values\n",
    "    print(df[col].nunique())\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e83f79c6",
   "metadata": {},
   "source": [
    "From the above, it seems this there is no null value and all categorical variables are properly encoded. \n",
    "\n",
    "Next, look at the **relations** in our numerical data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "6e17c2d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['geography', 'gender', 'hascrcard', 'isactivemember']"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "48e42fcf",
   "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>creditscore</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>estimatedsalary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>creditscore</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.003965</td>\n",
       "      <td>0.000842</td>\n",
       "      <td>0.006268</td>\n",
       "      <td>0.012238</td>\n",
       "      <td>-0.001384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>age</th>\n",
       "      <td>-0.003965</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.009997</td>\n",
       "      <td>0.028308</td>\n",
       "      <td>-0.030680</td>\n",
       "      <td>-0.007201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tenure</th>\n",
       "      <td>0.000842</td>\n",
       "      <td>-0.009997</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.012254</td>\n",
       "      <td>0.013444</td>\n",
       "      <td>0.007784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>balance</th>\n",
       "      <td>0.006268</td>\n",
       "      <td>0.028308</td>\n",
       "      <td>-0.012254</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.304180</td>\n",
       "      <td>0.012797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>numofproducts</th>\n",
       "      <td>0.012238</td>\n",
       "      <td>-0.030680</td>\n",
       "      <td>0.013444</td>\n",
       "      <td>-0.304180</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.014204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>estimatedsalary</th>\n",
       "      <td>-0.001384</td>\n",
       "      <td>-0.007201</td>\n",
       "      <td>0.007784</td>\n",
       "      <td>0.012797</td>\n",
       "      <td>0.014204</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 creditscore       age    tenure   balance  numofproducts  \\\n",
       "creditscore         1.000000 -0.003965  0.000842  0.006268       0.012238   \n",
       "age                -0.003965  1.000000 -0.009997  0.028308      -0.030680   \n",
       "tenure              0.000842 -0.009997  1.000000 -0.012254       0.013444   \n",
       "balance             0.006268  0.028308 -0.012254  1.000000      -0.304180   \n",
       "numofproducts       0.012238 -0.030680  0.013444 -0.304180       1.000000   \n",
       "estimatedsalary    -0.001384 -0.007201  0.007784  0.012797       0.014204   \n",
       "\n",
       "                 estimatedsalary  \n",
       "creditscore            -0.001384  \n",
       "age                    -0.007201  \n",
       "tenure                  0.007784  \n",
       "balance                 0.012797  \n",
       "numofproducts           0.014204  \n",
       "estimatedsalary         1.000000  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[numerical].corr() # the correlation matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8f6bb8e",
   "metadata": {},
   "source": [
    "Does not say much"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "39548cf2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,8))  \n",
    "sns.heatmap(df[numerical].corr(),annot=True,linewidths=.5, cmap=\"Reds\")\n",
    "plt.title('Heatmap showing correlations between numerical data')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7cf52eff",
   "metadata": {},
   "source": [
    "minimal relation of numerical features with target variable "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "b5436338",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "creditscore      creditscore        1.000000\n",
       "age              age                1.000000\n",
       "numofproducts    numofproducts      1.000000\n",
       "balance          balance            1.000000\n",
       "tenure           tenure             1.000000\n",
       "estimatedsalary  estimatedsalary    1.000000\n",
       "balance          age                0.028308\n",
       "age              balance            0.028308\n",
       "estimatedsalary  numofproducts      0.014204\n",
       "numofproducts    estimatedsalary    0.014204\n",
       "                 tenure             0.013444\n",
       "tenure           numofproducts      0.013444\n",
       "estimatedsalary  balance            0.012797\n",
       "balance          estimatedsalary    0.012797\n",
       "creditscore      numofproducts      0.012238\n",
       "numofproducts    creditscore        0.012238\n",
       "estimatedsalary  tenure             0.007784\n",
       "tenure           estimatedsalary    0.007784\n",
       "creditscore      balance            0.006268\n",
       "balance          creditscore        0.006268\n",
       "creditscore      tenure             0.000842\n",
       "tenure           creditscore        0.000842\n",
       "creditscore      estimatedsalary   -0.001384\n",
       "estimatedsalary  creditscore       -0.001384\n",
       "creditscore      age               -0.003965\n",
       "age              creditscore       -0.003965\n",
       "estimatedsalary  age               -0.007201\n",
       "age              estimatedsalary   -0.007201\n",
       "tenure           age               -0.009997\n",
       "age              tenure            -0.009997\n",
       "balance          tenure            -0.012254\n",
       "tenure           balance           -0.012254\n",
       "numofproducts    age               -0.030680\n",
       "age              numofproducts     -0.030680\n",
       "balance          numofproducts     -0.304180\n",
       "numofproducts    balance           -0.304180\n",
       "dtype: float64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[numerical].corr().unstack().sort_values(ascending = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69b7207e",
   "metadata": {},
   "source": [
    "From here I can see there is low corelation amongst the numeric data. So the best way I can determine if a feature is important is excuse it from the training dataset and see how it affects it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "9ee1b9fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "# sns.pairplot(df, hue=\"exited\", palette=\"inferno\", corner=True);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ae7f4ee",
   "metadata": {},
   "source": [
    "### Data separation  - Setting up the validation framework\n",
    "\n",
    "I am gonna split it into three datasets (train, val, and test) - training, validation and test datasets\n",
    "I will use scikit-learn for this step"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "06e7abaf",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.metrics import roc_curve\n",
    "from sklearn.metrics import mutual_info_score\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "5d696501",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_full_train, df_test = train_test_split(df, test_size=0.2, random_state=43)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "adf484ba",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train, df_val = train_test_split(df_full_train, test_size=0.25, random_state=43)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "de65a403",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.6, 0.2, 0.2)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# sanity check\n",
    "(len(df_train) / len(df)), (len(df_val) / len(df)), (len(df_val) / len(df))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "0247b108",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = df_train.reset_index(drop=True)\n",
    "df_val = df_val.reset_index(drop=True)\n",
    "df_test = df_test.reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "0943a575",
   "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>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>619</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>2</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>101348.88</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>608</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>83807.86</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>112542.58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>502</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>42</td>\n",
       "      <td>8</td>\n",
       "      <td>159660.80</td>\n",
       "      <td>3</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>113931.57</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>699</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>39</td>\n",
       "      <td>1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>93826.63</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>850</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>43</td>\n",
       "      <td>2</td>\n",
       "      <td>125510.82</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>79084.10</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   creditscore geography  gender  age  tenure    balance  numofproducts  \\\n",
       "0          619    france  female   42       2       0.00              1   \n",
       "1          608     spain  female   41       1   83807.86              1   \n",
       "2          502    france  female   42       8  159660.80              3   \n",
       "3          699    france  female   39       1       0.00              2   \n",
       "4          850     spain  female   43       2  125510.82              1   \n",
       "\n",
       "  hascrcard isactivemember  estimatedsalary  exited  \n",
       "0       yes            yes        101348.88       1  \n",
       "1        no            yes        112542.58       0  \n",
       "2       yes             no        113931.57       1  \n",
       "3        no             no         93826.63       0  \n",
       "4       yes            yes         79084.10       0  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head() # index reset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "a1450f59",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = df_train.exited.values\n",
    "y_val = df_val.exited.values\n",
    "y_test = df_test.exited.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "c415cec6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# drop target column\n",
    "df_train = df_train.drop('exited', axis=1)\n",
    "df_val = df_val.drop('exited', axis=1)\n",
    "df_test = df_test.drop('exited', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "f816bb2b",
   "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>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>604</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>36</td>\n",
       "      <td>6</td>\n",
       "      <td>116229.85</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>79633.38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>542</td>\n",
       "      <td>spain</td>\n",
       "      <td>male</td>\n",
       "      <td>34</td>\n",
       "      <td>8</td>\n",
       "      <td>108653.93</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>144725.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>701</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>44</td>\n",
       "      <td>7</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>23425.78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>521</td>\n",
       "      <td>germany</td>\n",
       "      <td>female</td>\n",
       "      <td>29</td>\n",
       "      <td>2</td>\n",
       "      <td>87212.80</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>994.86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>728</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>30</td>\n",
       "      <td>10</td>\n",
       "      <td>114835.43</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>37662.49</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   creditscore geography  gender  age  tenure    balance  numofproducts  \\\n",
       "0          604    france    male   36       6  116229.85              2   \n",
       "1          542     spain    male   34       8  108653.93              1   \n",
       "2          701    france    male   44       7       0.00              2   \n",
       "3          521   germany  female   29       2   87212.80              1   \n",
       "4          728    france    male   30      10  114835.43              1   \n",
       "\n",
       "  hascrcard isactivemember  estimatedsalary  \n",
       "0       yes            yes         79633.38  \n",
       "1        no            yes        144725.14  \n",
       "2       yes             no         23425.78  \n",
       "3       yes            yes           994.86  \n",
       "4        no            yes         37662.49  "
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "27a9006f",
   "metadata": {},
   "source": [
    "I still want to know what variables are important.\n",
    "**ROC AUC** could also be used to evaluate feature importance of numerical variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "4994c701",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['creditscore', 'age', 'tenure', 'balance', 'numofproducts', 'estimatedsalary']"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "numerical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "4c41475f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    creditscore, 0.520\n",
      "            age, 0.733\n",
      "         tenure, 0.509\n",
      "        balance, 0.570\n",
      "  numofproducts, 0.580\n",
      "estimatedsalary, 0.506\n"
     ]
    }
   ],
   "source": [
    "for c in numerical:\n",
    "    auc = roc_auc_score(y_train, df_train[c])\n",
    "    if auc < 0.5:\n",
    "        auc = roc_auc_score(y_train, -df_train[c])\n",
    "    print('%15s, %.3f' % (c, auc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "f65dc6d2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5, 5))\n",
    "\n",
    "fpr, tpr, _ = roc_curve(y_train, df_train.age)\n",
    "plt.plot(fpr, tpr, label='+balance')\n",
    "\n",
    "fpr, tpr, _ = roc_curve(y_train, -df_train.age)\n",
    "plt.plot(fpr, tpr, label='-balance')\n",
    "\n",
    "plt.plot([0, 1], [0, 1], color='grey', linestyle='--')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ae4317b4",
   "metadata": {},
   "source": [
    "There's separation in distribution of age between positive and negative classes. Does it mean age is important here? "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "08cf07a4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.histplot(df_train.age[y_train == 1],\n",
    "             stat='density', bins=50,\n",
    "             color='orange', alpha=0.5,\n",
    "             label='positive')\n",
    "sns.histplot(df_train.age[y_train == 0],\n",
    "             stat='density', bins=50,\n",
    "             color='blue', alpha=0.5,\n",
    "             label='negative')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "2464ea14",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.histplot(df_train.balance[y_train == 1],\n",
    "             stat='density', bins=50,\n",
    "             color='orange', alpha=0.5,\n",
    "             label='positive')\n",
    "sns.histplot(df_train.balance[y_train == 0],\n",
    "             stat='density', bins=50,\n",
    "             color='blue', alpha=0.5,\n",
    "             label='negative')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "8da41da1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.histplot(df_train.tenure[y_train == 1],\n",
    "             stat='density', bins=50,\n",
    "             color='orange', alpha=0.5,\n",
    "             label='positive')\n",
    "sns.histplot(df_train.tenure[y_train == 0],\n",
    "             stat='density', bins=50,\n",
    "             color='blue', alpha=0.5,\n",
    "             label='negative')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "f4b902ce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.histplot(df_train.creditscore[y_train == 1],\n",
    "             stat='density', bins=50,\n",
    "             color='orange', alpha=0.5,\n",
    "             label='positive')\n",
    "sns.histplot(df_train.creditscore[y_train == 0],\n",
    "             stat='density', bins=50,\n",
    "             color='blue', alpha=0.5,\n",
    "             label='negative')\n",
    "\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7b3f280",
   "metadata": {},
   "source": [
    "After some help collected by good friends of the datatalkclub, I was told to get the correlation between features and the target for feature importance in relation to target variable. \" Because correlation between features and themselves tells us if there is reduntant information in the set (which helps us to keep/eliminate features, but it doesn't clearly tells us the relationship feature and target)\" - Kattson Bastos Advice"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "10036bc5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age                0.285646\n",
       "balance            0.113100\n",
       "numofproducts      0.044804\n",
       "creditscore        0.025264\n",
       "tenure             0.019579\n",
       "estimatedsalary    0.005647\n",
       "dtype: float64"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_full_train[numerical].corrwith(df_full_train.exited).abs().sort_values(ascending = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b870258",
   "metadata": {},
   "source": [
    "Hence from above, it is seen what numerical variables are really affecting the target variables"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7b07479",
   "metadata": {},
   "source": [
    "\n",
    "Now, I will look at the feature importance for the categorical variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "0528d1f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# determine the feature importance for the categorical variables\n",
    "# use the mutual_info_score to determine the relation of \n",
    "# one categorical variable to the target variable\n",
    "\n",
    "def mutual_info_exited_score(series):\n",
    "    return mutual_info_score(series, df_full_train.exited)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "8eca2370",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['geography', 'gender', 'hascrcard', 'isactivemember']"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "86098665",
   "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>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6100</th>\n",
       "      <td>796</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>30</td>\n",
       "      <td>2</td>\n",
       "      <td>137262.71</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>62905.29</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4860</th>\n",
       "      <td>748</td>\n",
       "      <td>spain</td>\n",
       "      <td>male</td>\n",
       "      <td>36</td>\n",
       "      <td>4</td>\n",
       "      <td>141573.55</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>82158.14</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4401</th>\n",
       "      <td>845</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>38</td>\n",
       "      <td>2</td>\n",
       "      <td>112803.92</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>179631.85</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1416</th>\n",
       "      <td>585</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>33</td>\n",
       "      <td>8</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>114182.07</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8620</th>\n",
       "      <td>768</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>40</td>\n",
       "      <td>8</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>69080.46</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8499</th>\n",
       "      <td>498</td>\n",
       "      <td>germany</td>\n",
       "      <td>male</td>\n",
       "      <td>35</td>\n",
       "      <td>2</td>\n",
       "      <td>121968.11</td>\n",
       "      <td>2</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>188343.05</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2064</th>\n",
       "      <td>765</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>44</td>\n",
       "      <td>6</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>159899.97</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7985</th>\n",
       "      <td>696</td>\n",
       "      <td>germany</td>\n",
       "      <td>female</td>\n",
       "      <td>27</td>\n",
       "      <td>2</td>\n",
       "      <td>96129.32</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>5983.70</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2303</th>\n",
       "      <td>714</td>\n",
       "      <td>spain</td>\n",
       "      <td>male</td>\n",
       "      <td>45</td>\n",
       "      <td>0</td>\n",
       "      <td>124693.48</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>187194.15</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3392</th>\n",
       "      <td>623</td>\n",
       "      <td>germany</td>\n",
       "      <td>male</td>\n",
       "      <td>33</td>\n",
       "      <td>8</td>\n",
       "      <td>96759.42</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>174777.98</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8000 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      creditscore geography  gender  age  tenure    balance  numofproducts  \\\n",
       "6100          796    france    male   30       2  137262.71              2   \n",
       "4860          748     spain    male   36       4  141573.55              1   \n",
       "4401          845     spain  female   38       2  112803.92              1   \n",
       "1416          585     spain  female   33       8       0.00              2   \n",
       "8620          768    france  female   40       8       0.00              2   \n",
       "...           ...       ...     ...  ...     ...        ...            ...   \n",
       "8499          498   germany    male   35       2  121968.11              2   \n",
       "2064          765    france    male   44       6       0.00              2   \n",
       "7985          696   germany  female   27       2   96129.32              2   \n",
       "2303          714     spain    male   45       0  124693.48              1   \n",
       "3392          623   germany    male   33       8   96759.42              1   \n",
       "\n",
       "     hascrcard isactivemember  estimatedsalary  exited  \n",
       "6100       yes             no         62905.29       0  \n",
       "4860       yes             no         82158.14       0  \n",
       "4401       yes             no        179631.85       0  \n",
       "1416       yes             no        114182.07       0  \n",
       "8620        no            yes         69080.46       0  \n",
       "...        ...            ...              ...     ...  \n",
       "8499        no            yes        188343.05       0  \n",
       "2064       yes            yes        159899.97       0  \n",
       "7985       yes            yes          5983.70       0  \n",
       "2303        no            yes        187194.15       0  \n",
       "3392       yes            yes        174777.98       0  \n",
       "\n",
       "[8000 rows x 11 columns]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_full_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "8dd73a23",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "isactivemember    0.014091\n",
       "geography         0.012912\n",
       "gender            0.006405\n",
       "hascrcard         0.000032\n",
       "dtype: float64"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mutual_info = df_full_train[categorical].apply(mutual_info_exited_score)\n",
    "mutual_info.sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99f15028",
   "metadata": {},
   "source": [
    "The categorical feature importance"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19384c7d",
   "metadata": {},
   "source": [
    "Now, we train our model with the information we know.\n",
    "We build our first model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "930915d7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction import DictVectorizer\n",
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "cb112bda",
   "metadata": {},
   "outputs": [],
   "source": [
    "dv = DictVectorizer(sparse=False)\n",
    "\n",
    "train_dict = df_train[categorical + numerical].to_dict(orient='records')\n",
    "X_train = dv.fit_transform(train_dict)\n",
    "\n",
    "val_dict = df_val[categorical + numerical].to_dict(orient='records')\n",
    "X_val = dv.transform(val_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "b332ccab",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(solver='liblinear')"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = LogisticRegression(solver='liblinear', penalty='l2')\n",
    "model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "338f6a3d",
   "metadata": {},
   "source": [
    "Now, to test model **one**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "31f53da3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'age': 0.04477071578233233,\n",
       " 'balance': 3.154761594349631e-06,\n",
       " 'creditscore': -0.005014792377389181,\n",
       " 'estimatedsalary': -1.5986656350423042e-06,\n",
       " 'gender=female': 0.0013453554677235787,\n",
       " 'gender=male': -0.001710808876885251,\n",
       " 'geography=france': -0.0010627591209487386,\n",
       " 'geography=germany': 0.0013176869257597054,\n",
       " 'geography=spain': -0.0006203812139726306,\n",
       " 'hascrcard=no': -7.367462855608813e-05,\n",
       " 'hascrcard=yes': -0.00029177878060557445,\n",
       " 'isactivemember=no': 0.0019172458242749735,\n",
       " 'isactivemember=yes': -0.0022826992334366446,\n",
       " 'numofproducts': -0.0008480016190156052,\n",
       " 'tenure': -0.0025240875067798446}"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dict(zip(dv.get_feature_names_out(), model.coef_[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "de927fec",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_dict = df_test[categorical + numerical].to_dict(orient='records')\n",
    "X_test = dv.transform(test_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "f07752f7",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = model.predict_proba(X_test)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "a0ca4f16",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False, False,  True, ..., False, False, False])"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exited = (y_pred >= 0.5)\n",
    "exited"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "97290f36",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.785"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_val, y_pred >= 0.5).round(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c63e349",
   "metadata": {},
   "source": [
    "On the first model, we have a 77% accuracy but now let's calculate the precision & recall or the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "46789e20",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7995"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# how many people actually did not exit the bank\n",
    "1- y_val.mean() "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5af58f85",
   "metadata": {},
   "source": [
    "79% of people did not exit the bank"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92f4873d",
   "metadata": {},
   "source": [
    "Let's find a good threshold "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "54120a5e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.00, 0.201\n",
      "0.05, 0.205\n",
      "0.10, 0.294\n",
      "0.15, 0.443\n",
      "0.20, 0.562\n",
      "0.25, 0.646\n",
      "0.30, 0.705\n",
      "0.35, 0.740\n",
      "0.40, 0.756\n",
      "0.45, 0.771\n",
      "0.50, 0.785\n",
      "0.55, 0.791\n",
      "0.60, 0.794\n",
      "0.65, 0.796\n",
      "0.70, 0.798\n",
      "0.75, 0.799\n",
      "0.80, 0.799\n",
      "0.85, 0.799\n",
      "0.90, 0.799\n",
      "0.95, 0.799\n",
      "1.00, 0.799\n"
     ]
    }
   ],
   "source": [
    "thresholds = np.linspace(0, 1, 21)\n",
    "\n",
    "scores = []\n",
    "\n",
    "for t in thresholds:\n",
    "    default = (y_pred >= t)\n",
    "    score = (default == y_val).mean()\n",
    "    scores.append(score)\n",
    "    print('%.2f, %.3f' % (t, score))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "247825d0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x20c487b3108>]"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(thresholds, scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "44943df4",
   "metadata": {},
   "source": [
    "We creat a confusion matrix that will help us find precision(p) and recall(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "11ebf64a",
   "metadata": {},
   "outputs": [],
   "source": [
    "actual_positive = (y_val == 1)\n",
    "actual_negative = (y_val == 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "9b49aefc",
   "metadata": {},
   "outputs": [],
   "source": [
    "t = 0.5\n",
    "predict_positive = (y_pred >= t)\n",
    "predict_negative = (y_pred < t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "47e12fef",
   "metadata": {},
   "outputs": [],
   "source": [
    "# tp - true positive, tn - true negative, fp - false positive, fn - false negative.\n",
    "tp = (predict_positive & actual_positive).sum()\n",
    "tn = (predict_negative & actual_negative).sum()\n",
    "\n",
    "fp = (predict_positive & actual_negative).sum()\n",
    "fn = (predict_negative & actual_positive).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "8af84eea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1561,   38],\n",
       "       [ 392,    9]])"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix = np.array([\n",
    "    [tn, fp],\n",
    "    [fn, tp]\n",
    "])\n",
    "confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "b8ba5aa8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.78, 0.02],\n",
       "       [0.2 , 0.  ]])"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(confusion_matrix / confusion_matrix.sum()).round(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "b2f552e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.19148936170212766"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "p = tp / (tp + fp)\n",
    "p"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "aa6d1d68",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.022443890274314215"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r = tp / (tp + fn)\n",
    "r\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dfd56521",
   "metadata": {},
   "source": [
    "Now let's plot the roc curve to see the performace of this model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "e55d2ce3",
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "d3a07b23",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c4877c1c8>"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5,5))\n",
    "\n",
    "plt.plot(fpr, tpr, label='model')\n",
    "plt.plot([0,1], [0,1], label='random', linestyle='--')\n",
    "\n",
    "plt.xlabel('FPR')\n",
    "plt.ylabel('TPR')\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4d7a74d",
   "metadata": {},
   "source": [
    "So far, it is shown that our model is bad. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "66649716",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5136049806690279"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import auc\n",
    "\n",
    "auc(fpr, tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "890a2a5e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5136049806690279"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ade83464",
   "metadata": {},
   "source": [
    "I have tried removing features from the dataset, and what I noticed is that this dataset needs more features. As I removed features, the auc reduced but with more features the model increased"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a20a3fc",
   "metadata": {},
   "source": [
    "Buidling a **decision tree** to see if the model can improve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "483705df",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.tree import export_text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "4a4a5df0",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dicts = df_train.to_dict(orient='records')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "2ca740e8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'creditscore': 673,\n",
       " 'geography': 'germany',\n",
       " 'gender': 'male',\n",
       " 'age': 36,\n",
       " 'tenure': 5,\n",
       " 'balance': 73088.06,\n",
       " 'numofproducts': 2,\n",
       " 'hascrcard': 'no',\n",
       " 'isactivemember': 'no',\n",
       " 'estimatedsalary': 196142.26}"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dicts[900]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "9ac6dbbd",
   "metadata": {},
   "outputs": [],
   "source": [
    "dv = DictVectorizer(sparse=False)\n",
    "X_train = dv.fit_transform(train_dicts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "46eaad9c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier()"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt = DecisionTreeClassifier()\n",
    "dt.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "f9d6bae7",
   "metadata": {},
   "outputs": [],
   "source": [
    "val_dicts = df_val.to_dict(orient='records')\n",
    "X_val = dv.transform(val_dicts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "511d63e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6743374521794326"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = dt.predict_proba(X_val)[:, 1]\n",
    "roc_auc_score(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a56550d",
   "metadata": {},
   "source": [
    "A higher auc score here from the basic **DecisionTreeClassifier** compared to the **LogisticRegression**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "7edfef55",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = dt.predict_proba(X_train)[:, 1]\n",
    "roc_auc_score(y_train, y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "01f86232",
   "metadata": {},
   "source": [
    "We can see that the model is **over-fitting** here. It has an auc score of **1.0** on the **training set** but a significantly lower score on the **validation-set**\n",
    "\n",
    "One reason here, is that the model learning to deep so it memorizies all outcomes in the training-set and poorly in the validation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "c9afbf20",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(max_depth=3)"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# restrict depth search\n",
    "dt = DecisionTreeClassifier(max_depth=3)\n",
    "dt.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "4750e7f4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train: 0.7968065750867629\n",
      "val: 0.8023265788000293\n"
     ]
    }
   ],
   "source": [
    "y_pred = dt.predict_proba(X_train)[:, 1]\n",
    "auc = roc_auc_score(y_train, y_pred)\n",
    "print('train:', auc)\n",
    "\n",
    "y_pred = dt.predict_proba(X_val)[:, 1]\n",
    "auc = roc_auc_score(y_val, y_pred)\n",
    "print('val:', auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "02e22187",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "|--- age <= 42.50\n",
      "|   |--- numofproducts <= 2.50\n",
      "|   |   |--- numofproducts <= 1.50\n",
      "|   |   |   |--- class: 0\n",
      "|   |   |--- numofproducts >  1.50\n",
      "|   |   |   |--- class: 0\n",
      "|   |--- numofproducts >  2.50\n",
      "|   |   |--- balance <= 55948.91\n",
      "|   |   |   |--- class: 0\n",
      "|   |   |--- balance >  55948.91\n",
      "|   |   |   |--- class: 1\n",
      "|--- age >  42.50\n",
      "|   |--- isactivemember=yes <= 0.50\n",
      "|   |   |--- age <= 50.50\n",
      "|   |   |   |--- class: 0\n",
      "|   |   |--- age >  50.50\n",
      "|   |   |   |--- class: 1\n",
      "|   |--- isactivemember=yes >  0.50\n",
      "|   |   |--- numofproducts <= 2.50\n",
      "|   |   |   |--- class: 0\n",
      "|   |   |--- numofproducts >  2.50\n",
      "|   |   |   |--- class: 1\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\USER\\anaconda3\\envs\\keras_env\\lib\\site-packages\\sklearn\\utils\\deprecation.py:87: FutureWarning: Function get_feature_names is deprecated; get_feature_names is deprecated in 1.0 and will be removed in 1.2. Please use get_feature_names_out instead.\n",
      "  warnings.warn(msg, category=FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "print(export_text(dt, feature_names=dv.get_feature_names()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c9230184",
   "metadata": {},
   "source": [
    "now let's try for different depth parameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "3330fc73",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   1 -> 0.685\n",
      "   2 -> 0.743\n",
      "   3 -> 0.802\n",
      "   4 -> 0.819\n",
      "   5 -> 0.845\n",
      "   6 -> 0.845\n",
      "  10 -> 0.774\n",
      "  15 -> 0.689\n",
      "  20 -> 0.673\n",
      "None -> 0.679\n"
     ]
    }
   ],
   "source": [
    "depths = [1, 2, 3, 4, 5, 6, 10, 15, 20, None]\n",
    "\n",
    "for depth in depths: \n",
    "    dt = DecisionTreeClassifier(max_depth=depth)\n",
    "    dt.fit(X_train, y_train)\n",
    "    \n",
    "    y_pred = dt.predict_proba(X_val)[:, 1]\n",
    "    auc = roc_auc_score(y_val, y_pred)\n",
    "    \n",
    "    print('%4s -> %.3f' % (depth, auc))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "853bf940",
   "metadata": {},
   "source": [
    "Best **max_depth = 5**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "f20c72e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = []\n",
    "\n",
    "for depth in [4, 5, 6]:\n",
    "    for sample in [1, 5, 10, 15, 20, 500, 100, 200]:\n",
    "        dt = DecisionTreeClassifier(max_depth=depth, min_samples_leaf=sample)\n",
    "        dt.fit(X_train, y_train)\n",
    "\n",
    "        y_pred = dt.predict_proba(X_val)[:, 1]\n",
    "        auc = roc_auc_score(y_val, y_pred)\n",
    "        \n",
    "        scores.append((depth, sample, auc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "fed1f172",
   "metadata": {},
   "outputs": [],
   "source": [
    "columns = ['max_depth', 'min_samples_leaf', 'auc']\n",
    "df_scores = pd.DataFrame(scores, columns=columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "20ff722b",
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">auc</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max_depth</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min_samples_leaf</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.819</td>\n",
       "      <td>0.843</td>\n",
       "      <td>0.846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.819</td>\n",
       "      <td>0.845</td>\n",
       "      <td>0.850</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.825</td>\n",
       "      <td>0.847</td>\n",
       "      <td>0.851</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.824</td>\n",
       "      <td>0.847</td>\n",
       "      <td>0.847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.825</td>\n",
       "      <td>0.849</td>\n",
       "      <td>0.848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0.791</td>\n",
       "      <td>0.790</td>\n",
       "      <td>0.796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>200</th>\n",
       "      <td>0.775</td>\n",
       "      <td>0.775</td>\n",
       "      <td>0.775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500</th>\n",
       "      <td>0.757</td>\n",
       "      <td>0.757</td>\n",
       "      <td>0.757</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    auc              \n",
       "max_depth             4      5      6\n",
       "min_samples_leaf                     \n",
       "1                 0.819  0.843  0.846\n",
       "5                 0.819  0.845  0.850\n",
       "10                0.825  0.847  0.851\n",
       "15                0.824  0.847  0.847\n",
       "20                0.825  0.849  0.848\n",
       "100               0.791  0.790  0.796\n",
       "200               0.775  0.775  0.775\n",
       "500               0.757  0.757  0.757"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_scores_pivot = df_scores.pivot(index='min_samples_leaf', columns=['max_depth'], values=['auc'])\n",
    "df_scores_pivot.round(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "38fcd017",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='None-max_depth', ylabel='min_samples_leaf'>"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(df_scores_pivot, annot=True, fmt=\".3f\", color='red')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aed66687",
   "metadata": {},
   "source": [
    "We can see that a **max_depth=6** and **min_samples_leaf=10** are the best parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "e5d05ff0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(max_depth=6, min_samples_leaf=10)"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt = DecisionTreeClassifier(max_depth=6, min_samples_leaf=10)\n",
    "dt.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "7614d853",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "|--- age <= 42.50\n",
      "|   |--- numofproducts <= 2.50\n",
      "|   |   |--- numofproducts <= 1.50\n",
      "|   |   |   |--- geography=germany <= 0.50\n",
      "|   |   |   |   |--- balance <= 80289.48\n",
      "|   |   |   |   |   |--- age <= 37.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- age >  37.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- balance >  80289.48\n",
      "|   |   |   |   |   |--- age <= 38.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- age >  38.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |--- geography=germany >  0.50\n",
      "|   |   |   |   |--- balance <= 97709.64\n",
      "|   |   |   |   |   |--- balance <= 77121.73\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- balance >  77121.73\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- balance >  97709.64\n",
      "|   |   |   |   |   |--- age <= 38.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- age >  38.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |--- numofproducts >  1.50\n",
      "|   |   |   |--- balance <= 1884.34\n",
      "|   |   |   |   |--- gender=male <= 0.50\n",
      "|   |   |   |   |   |--- estimatedsalary <= 12428.89\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- estimatedsalary >  12428.89\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- gender=male >  0.50\n",
      "|   |   |   |   |   |--- age <= 36.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- age >  36.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |--- balance >  1884.34\n",
      "|   |   |   |   |--- balance <= 188051.12\n",
      "|   |   |   |   |   |--- age <= 35.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- age >  35.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- balance >  188051.12\n",
      "|   |   |   |   |   |--- class: 0\n",
      "|   |--- numofproducts >  2.50\n",
      "|   |   |--- balance <= 55948.91\n",
      "|   |   |   |--- estimatedsalary <= 124467.49\n",
      "|   |   |   |   |--- age <= 37.50\n",
      "|   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- age >  37.50\n",
      "|   |   |   |   |   |--- class: 1\n",
      "|   |   |   |--- estimatedsalary >  124467.49\n",
      "|   |   |   |   |--- class: 1\n",
      "|   |   |--- balance >  55948.91\n",
      "|   |   |   |--- age <= 35.50\n",
      "|   |   |   |   |--- isactivemember=no <= 0.50\n",
      "|   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- isactivemember=no >  0.50\n",
      "|   |   |   |   |   |--- class: 1\n",
      "|   |   |   |--- age >  35.50\n",
      "|   |   |   |   |--- estimatedsalary <= 56731.37\n",
      "|   |   |   |   |   |--- class: 1\n",
      "|   |   |   |   |--- estimatedsalary >  56731.37\n",
      "|   |   |   |   |   |--- class: 1\n",
      "|--- age >  42.50\n",
      "|   |--- isactivemember=yes <= 0.50\n",
      "|   |   |--- age <= 50.50\n",
      "|   |   |   |--- numofproducts <= 1.50\n",
      "|   |   |   |   |--- geography=germany <= 0.50\n",
      "|   |   |   |   |   |--- balance <= 81819.96\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |   |   |--- balance >  81819.96\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- geography=germany >  0.50\n",
      "|   |   |   |   |   |--- balance <= 98151.25\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |   |   |--- balance >  98151.25\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |--- numofproducts >  1.50\n",
      "|   |   |   |   |--- numofproducts <= 2.50\n",
      "|   |   |   |   |   |--- balance <= 80966.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- balance >  80966.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- numofproducts >  2.50\n",
      "|   |   |   |   |   |--- class: 1\n",
      "|   |   |--- age >  50.50\n",
      "|   |   |   |--- age <= 68.50\n",
      "|   |   |   |   |--- age <= 53.50\n",
      "|   |   |   |   |   |--- estimatedsalary <= 74582.09\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |   |   |--- estimatedsalary >  74582.09\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |   |--- age >  53.50\n",
      "|   |   |   |   |   |--- creditscore <= 639.00\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |   |   |--- creditscore >  639.00\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |--- age >  68.50\n",
      "|   |   |   |   |--- class: 0\n",
      "|   |--- isactivemember=yes >  0.50\n",
      "|   |   |--- numofproducts <= 2.50\n",
      "|   |   |   |--- numofproducts <= 1.50\n",
      "|   |   |   |   |--- geography=germany <= 0.50\n",
      "|   |   |   |   |   |--- balance <= 67335.82\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- balance >  67335.82\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- geography=germany >  0.50\n",
      "|   |   |   |   |   |--- age <= 67.50\n",
      "|   |   |   |   |   |   |--- class: 1\n",
      "|   |   |   |   |   |--- age >  67.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |--- numofproducts >  1.50\n",
      "|   |   |   |   |--- balance <= 60071.48\n",
      "|   |   |   |   |   |--- estimatedsalary <= 161088.09\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- estimatedsalary >  161088.09\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |--- balance >  60071.48\n",
      "|   |   |   |   |   |--- tenure <= 1.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |   |   |   |--- tenure >  1.50\n",
      "|   |   |   |   |   |   |--- class: 0\n",
      "|   |   |--- numofproducts >  2.50\n",
      "|   |   |   |--- balance <= 22291.04\n",
      "|   |   |   |   |--- class: 1\n",
      "|   |   |   |--- balance >  22291.04\n",
      "|   |   |   |   |--- class: 1\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\USER\\anaconda3\\envs\\keras_env\\lib\\site-packages\\sklearn\\utils\\deprecation.py:87: FutureWarning: Function get_feature_names is deprecated; get_feature_names is deprecated in 1.0 and will be removed in 1.2. Please use get_feature_names_out instead.\n",
      "  warnings.warn(msg, category=FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "print(export_text(dt, feature_names=dv.get_feature_names()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49b2211e",
   "metadata": {},
   "source": [
    "From above we can see that the the decision tree has an auc of 85.1% which is better than the 51% by the logistic regression model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "ebceff1d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val: 0.8508637724013918\n"
     ]
    }
   ],
   "source": [
    "y_pred = dt.predict_proba(X_val)[:, 1]\n",
    "auc = roc_auc_score(y_val, y_pred)\n",
    "print('val:', auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "c4a02058",
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "e810b2ee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c481ee488>"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5,5))\n",
    "\n",
    "plt.plot(fpr, tpr, label='model')\n",
    "plt.plot([0,1], [0,1], label='random', linestyle='--')\n",
    "\n",
    "plt.xlabel('FPR')\n",
    "plt.ylabel('TPR')\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1eaa8ce0",
   "metadata": {},
   "source": [
    "Significantly better."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b22f9e5",
   "metadata": {},
   "source": [
    "### Random Forests & Ensembles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "c30fdff2",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "2f985624",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = []\n",
    "\n",
    "for n in range(10, 201, 10):\n",
    "    rf = RandomForestClassifier(n_estimators=n, random_state=43)\n",
    "    rf.fit(X_train, y_train)\n",
    "\n",
    "    y_pred = rf.predict_proba(X_val)[:, 1]\n",
    "    auc = roc_auc_score(y_val, y_pred)\n",
    "    \n",
    "    scores.append((n, auc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "73abf99f",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_scores = pd.DataFrame(scores, columns=['n_estimators', 'auc'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "95f53a3e",
   "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>n_estimators</th>\n",
       "      <th>auc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>0.802486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>20</td>\n",
       "      <td>0.822447</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>30</td>\n",
       "      <td>0.826615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40</td>\n",
       "      <td>0.831803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>50</td>\n",
       "      <td>0.834809</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>60</td>\n",
       "      <td>0.839074</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>70</td>\n",
       "      <td>0.840231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>80</td>\n",
       "      <td>0.840760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>90</td>\n",
       "      <td>0.841888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>100</td>\n",
       "      <td>0.843120</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>110</td>\n",
       "      <td>0.844037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>120</td>\n",
       "      <td>0.843074</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>130</td>\n",
       "      <td>0.842257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>140</td>\n",
       "      <td>0.842726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>150</td>\n",
       "      <td>0.843140</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>160</td>\n",
       "      <td>0.843403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>170</td>\n",
       "      <td>0.843838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>180</td>\n",
       "      <td>0.844555</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>190</td>\n",
       "      <td>0.844775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>200</td>\n",
       "      <td>0.844177</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    n_estimators       auc\n",
       "0             10  0.802486\n",
       "1             20  0.822447\n",
       "2             30  0.826615\n",
       "3             40  0.831803\n",
       "4             50  0.834809\n",
       "5             60  0.839074\n",
       "6             70  0.840231\n",
       "7             80  0.840760\n",
       "8             90  0.841888\n",
       "9            100  0.843120\n",
       "10           110  0.844037\n",
       "11           120  0.843074\n",
       "12           130  0.842257\n",
       "13           140  0.842726\n",
       "14           150  0.843140\n",
       "15           160  0.843403\n",
       "16           170  0.843838\n",
       "17           180  0.844555\n",
       "18           190  0.844775\n",
       "19           200  0.844177"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_scores"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3bab1230",
   "metadata": {},
   "source": [
    "from here we can see that the best **no. of estimators is 190**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "9df0fb8e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x20c4a9e9308>]"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(df_scores.n_estimators, df_scores.auc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "4294294d",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = []\n",
    "\n",
    "# max-depths\n",
    "for d in [5, 10, 15]:\n",
    "    # no of estimators - no of trees\n",
    "    for n in range(10, 201, 10):\n",
    "        rf = RandomForestClassifier(n_estimators=n,\n",
    "                                    max_depth=d,\n",
    "                                    random_state=43)\n",
    "        rf.fit(X_train, y_train)\n",
    "\n",
    "        y_pred = rf.predict_proba(X_val)[:, 1]\n",
    "        auc = roc_auc_score(y_val, y_pred)\n",
    "\n",
    "        scores.append((d, n, auc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "566362a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "columns = ['max_depth', 'n_estimators', 'auc']\n",
    "df_scores = pd.DataFrame(scores, columns=columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "9d911a04",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c4a9babc8>"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for d in [5, 10, 15]:\n",
    "    df_subset = df_scores[df_scores.max_depth == d]\n",
    "    \n",
    "    plt.plot(df_subset.n_estimators, df_subset.auc,\n",
    "             label='max_depth=%d' % d)\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a3cdefe1",
   "metadata": {},
   "source": [
    "Hence the best value for **max_depth** is 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "c11e7d8d",
   "metadata": {},
   "outputs": [],
   "source": [
    "max_depth = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "0fc3cffe",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = []\n",
    "\n",
    "for s in [1, 3, 5, 10, 50]:\n",
    "    for n in range(10, 201, 10):\n",
    "        rf = RandomForestClassifier(n_estimators=n,\n",
    "                                    max_depth=max_depth,\n",
    "                                    min_samples_leaf=s,\n",
    "                                    random_state=43)\n",
    "        rf.fit(X_train, y_train)\n",
    "\n",
    "        y_pred = rf.predict_proba(X_val)[:, 1]\n",
    "        auc = roc_auc_score(y_val, y_pred)\n",
    "\n",
    "        scores.append((s, n, auc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "3acbfd50",
   "metadata": {},
   "outputs": [],
   "source": [
    "columns = ['min_samples_leaf', 'n_estimators', 'auc']\n",
    "df_scores = pd.DataFrame(scores, columns=columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "d8a17db7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c4ca7c548>"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# colors = ['black', 'blue', 'orange', 'red', 'grey']\n",
    "colors = ['red', 'blue', 'orange']\n",
    "\n",
    "# values = [1, 3, 5, 10, 50]\n",
    "values = [1, 3, 5]\n",
    "\n",
    "for s, col in zip(values, colors):\n",
    "    df_subset = df_scores[df_scores.min_samples_leaf == s]\n",
    "    \n",
    "    plt.plot(df_subset.n_estimators, df_subset.auc,\n",
    "             color=col,\n",
    "             label='min_samples_leaf=%d' % s)\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "fd4284c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "min_samples_leaf = 1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "42b1b4d5",
   "metadata": {},
   "source": [
    "Now, we train the final random forest model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "4c8c5b38",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=10, n_estimators=200, random_state=43)"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestClassifier(n_estimators=200,\n",
    "                            max_depth=max_depth,\n",
    "                            min_samples_leaf=min_samples_leaf,\n",
    "                            random_state=43)\n",
    "rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "943e03bf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val: 0.858858170396398\n"
     ]
    }
   ],
   "source": [
    "y_pred = rf.predict_proba(X_val)[:, 1]\n",
    "auc = roc_auc_score(y_val, y_pred)\n",
    "print('val:', auc)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee9f8581",
   "metadata": {},
   "source": [
    "The auc is just slightly better by **0.7%** compared with DecisionTrees"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9e00467",
   "metadata": {},
   "source": [
    "### Gradient boosting and XGBoost"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06747628",
   "metadata": {},
   "source": [
    "Here we gonna use another model called **XGBoost** to experiment with before we select the best model to go with"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "2d094d82",
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "e523d0cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "# we have to convert it to a form that can be processed by xgBoost\n",
    "features = dv.get_feature_names_out()\n",
    "dtrain = xgb.DMatrix(X_train, label=y_train, feature_names=features)\n",
    "dval = xgb.DMatrix(X_val, label=y_val, feature_names=features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "aa8a49cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['age', 'balance', 'creditscore', 'estimatedsalary',\n",
       "       'gender=female', 'gender=male', 'geography=france',\n",
       "       'geography=germany', 'geography=spain', 'hascrcard=no',\n",
       "       'hascrcard=yes', 'isactivemember=no', 'isactivemember=yes',\n",
       "       'numofproducts', 'tenure'], dtype=object)"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "8af1827f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# basic parameters for an xgBoost model\n",
    "xgb_params = {\n",
    "    'eta': 0.3, \n",
    "    'max_depth': 6,\n",
    "    'min_child_weight': 1,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "b2385186",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = model.predict(dval)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "a0b4dff4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8585758867371909"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "2c5a01c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "watchlist = [(dtrain, 'train'), (dval, 'val')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "f6ce40f4",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture output\n",
    "\n",
    "xgb_params = {\n",
    "    'eta': 0.3, \n",
    "    'max_depth': 6,\n",
    "    'min_child_weight': 1,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'eval_metric': 'auc',\n",
    "\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=200,\n",
    "                  verbose_eval=5,\n",
    "                  evals=watchlist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "952f5c75",
   "metadata": {},
   "outputs": [],
   "source": [
    "s = output.stdout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "03d405bf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\ttrain-auc:0.86270\tval-auc:0.84610\n",
      "[5]\ttrain-auc:0.89757\tval-auc:0.85690\n",
      "[10]\ttrain-auc:0.91897\tval-auc:0.86103\n",
      "[15]\ttrain-auc:0.93145\tval-auc:0.86358\n",
      "[20]\ttrain-auc:0.93788\tval-auc:0.86473\n",
      "[25]\ttr\n"
     ]
    }
   ],
   "source": [
    "print(s[:200])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "479efa14",
   "metadata": {},
   "source": [
    "It is doing quite well. a little bit of overfitting as the list grows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "e221faae",
   "metadata": {},
   "outputs": [],
   "source": [
    "def parse_xgb_output(output):\n",
    "    results = []\n",
    "\n",
    "    for line in output.stdout.strip().split('\\n'):\n",
    "        it_line, train_line, val_line = line.split('\\t')\n",
    "\n",
    "        it = int(it_line.strip('[]'))\n",
    "        train = float(train_line.split(':')[1])\n",
    "        val = float(val_line.split(':')[1])\n",
    "\n",
    "        results.append((it, train, val))\n",
    "    \n",
    "    columns = ['num_iter', 'train_auc', 'val_auc']\n",
    "    df_results = pd.DataFrame(results, columns=columns)\n",
    "    return df_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "d75b01cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_score = parse_xgb_output(output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "33ab6372",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c4c9e2b08>"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(df_score.num_iter, df_score.train_auc, label='train')\n",
    "plt.plot(df_score.num_iter, df_score.val_auc, label='val')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8c40ad12",
   "metadata": {},
   "source": [
    "It is over-fitting as we can see. As it grows it learns the training set well but it is quite bad at predicting a new dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5f4ea766",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.plot(df_score.num_iter, df_score.val_auc, label='val')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f671b814",
   "metadata": {},
   "source": [
    "#### So we tune the parameters of the xgBoost model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8eca45c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = {}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "74231b1b",
   "metadata": {},
   "source": [
    "We tune **eta(learning rate)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "25a34fca",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture output\n",
    "\n",
    "xgb_params = {\n",
    "    'eta': 0.01, \n",
    "    'max_depth': 6,\n",
    "    'min_child_weight': 1,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'eval_metric': 'auc',\n",
    "\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=200,\n",
    "                  verbose_eval=5,\n",
    "                  evals=watchlist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "907214bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "key = 'eta={}'.format(xgb_params['eta'])\n",
    "scores[key] = parse_xgb_output(output)\n",
    "key"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a520ce1b",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "da87b2a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "etas = ['eta=1.0', 'eta=0.3', 'eta=0.1', 'eta=0.05', 'eta=0.01'] # the etas I used. future ref.\n",
    "for eta in etas:\n",
    "    df_score = scores[eta]\n",
    "    plt.plot(df_score.num_iter, df_score.val_auc, label=eta)\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "86a38f3e",
   "metadata": {},
   "source": [
    "From the above we can see that the best **learning rate is 0.1** and it peaks at **65th** iteration before it starts to degrade."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "986bd85f",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores['eta=0.1']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ea6bd70",
   "metadata": {},
   "source": [
    "We tune the next parameter, **max_depth**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "442d9820",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "663a8905",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture output\n",
    "\n",
    "# eta is now 0.1\n",
    "# max_depth_used while training = [3,4,6]\n",
    "xgb_params = {\n",
    "    'eta': 0.1, \n",
    "    'max_depth': 6,\n",
    "    'min_child_weight': 1,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'eval_metric': 'auc',\n",
    "\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=200,\n",
    "                  verbose_eval=5,\n",
    "                  evals=watchlist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eefcc4aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "key = 'max_depth=%s' % (xgb_params['max_depth'])\n",
    "scores[key] = parse_xgb_output(output)\n",
    "key"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "55969bb4",
   "metadata": {},
   "outputs": [],
   "source": [
    "for max_depth, df_score in scores.items():\n",
    "    plt.plot(df_score.num_iter, df_score.val_auc, label=max_depth)\n",
    "\n",
    "plt.ylim(0.84, 0.876)\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "61aab295",
   "metadata": {},
   "source": [
    "we can see that the best **max_depth** for this model is **4** at the **100th** iteration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "66b8f829",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores['max_depth=4']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3a2972fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ad6738a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture output\n",
    "\n",
    "# eta is now 0.1\n",
    "# min_child_weight = [1,10,20,30]\n",
    "xgb_params = {\n",
    "    'eta': 0.1, \n",
    "    'max_depth': 4,\n",
    "    'min_child_weight': 10,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'eval_metric': 'auc',\n",
    "\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=200,\n",
    "                  verbose_eval=5,\n",
    "                  evals=watchlist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c633bd3",
   "metadata": {},
   "outputs": [],
   "source": [
    "key = 'min_child_weight=%s' % (xgb_params['min_child_weight'])\n",
    "scores[key] = parse_xgb_output(output)\n",
    "key"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e31bb877",
   "metadata": {},
   "outputs": [],
   "source": [
    "for min_child_weight, df_score in scores.items():\n",
    "    plt.plot(df_score.num_iter, df_score.val_auc, label=min_child_weight)\n",
    "\n",
    "plt.ylim(0.860, 0.8750)\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7feefd9a",
   "metadata": {},
   "source": [
    "The best **min_child_weight is 1** which is the default"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9c90a897",
   "metadata": {},
   "outputs": [],
   "source": [
    "scores['min_child_weight=1']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "60d22322",
   "metadata": {},
   "outputs": [],
   "source": [
    "# the best model after tuning is\n",
    "xgb_params = {\n",
    "    'eta': 0.1, \n",
    "    'max_depth': 4,\n",
    "    'min_child_weight': 1,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'eval_metric': 'auc',\n",
    "\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=100)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abb6c353",
   "metadata": {},
   "source": [
    "### selecting the best model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "f9c1b2a6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(max_depth=6, min_samples_leaf=10)"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt = DecisionTreeClassifier(max_depth=6, min_samples_leaf=10)\n",
    "dt.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "33bbb9ce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8508637724013918"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = dt.predict_proba(X_val)[:, 1]\n",
    "roc_auc_score(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "a915eddc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=10, n_estimators=200, random_state=43)"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestClassifier(n_estimators=200,\n",
    "                            max_depth=10,\n",
    "                            min_samples_leaf=1,\n",
    "                            random_state=43)\n",
    "rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "ab0da997",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.858858170396398"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = rf.predict_proba(X_val)[:, 1]\n",
    "roc_auc_score(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "02977e3a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# the best model after tuning is\n",
    "xgb_params = {\n",
    "    'eta': 0.1, \n",
    "    'max_depth': 4,\n",
    "    'min_child_weight': 1,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'eval_metric': 'auc',\n",
    "\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "5ac11258",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8715032306662985"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = model.predict(dval)\n",
    "roc_auc_score(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "f4b8f719",
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_val, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "9c568c4d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c4ce17a88>"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5,5))\n",
    "\n",
    "plt.plot(fpr, tpr, label='model')\n",
    "plt.plot([0,1], [0,1], label='random', linestyle='--')\n",
    "\n",
    "plt.xlabel('FPR')\n",
    "plt.ylabel('TPR')\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d672570",
   "metadata": {},
   "source": [
    "From the above, we can that the xgBoost model is a little bit better than the other models "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94f3351c",
   "metadata": {},
   "source": [
    "so now, let's test it on the test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "3e715e55",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_full_train = df_full_train.reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "ac34a727",
   "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>creditscore</th>\n",
       "      <th>geography</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>tenure</th>\n",
       "      <th>balance</th>\n",
       "      <th>numofproducts</th>\n",
       "      <th>hascrcard</th>\n",
       "      <th>isactivemember</th>\n",
       "      <th>estimatedsalary</th>\n",
       "      <th>exited</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>796</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>30</td>\n",
       "      <td>2</td>\n",
       "      <td>137262.71</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>62905.29</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>748</td>\n",
       "      <td>spain</td>\n",
       "      <td>male</td>\n",
       "      <td>36</td>\n",
       "      <td>4</td>\n",
       "      <td>141573.55</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>82158.14</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>845</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>38</td>\n",
       "      <td>2</td>\n",
       "      <td>112803.92</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>179631.85</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>585</td>\n",
       "      <td>spain</td>\n",
       "      <td>female</td>\n",
       "      <td>33</td>\n",
       "      <td>8</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>114182.07</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>768</td>\n",
       "      <td>france</td>\n",
       "      <td>female</td>\n",
       "      <td>40</td>\n",
       "      <td>8</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>69080.46</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7995</th>\n",
       "      <td>498</td>\n",
       "      <td>germany</td>\n",
       "      <td>male</td>\n",
       "      <td>35</td>\n",
       "      <td>2</td>\n",
       "      <td>121968.11</td>\n",
       "      <td>2</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>188343.05</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7996</th>\n",
       "      <td>765</td>\n",
       "      <td>france</td>\n",
       "      <td>male</td>\n",
       "      <td>44</td>\n",
       "      <td>6</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>159899.97</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7997</th>\n",
       "      <td>696</td>\n",
       "      <td>germany</td>\n",
       "      <td>female</td>\n",
       "      <td>27</td>\n",
       "      <td>2</td>\n",
       "      <td>96129.32</td>\n",
       "      <td>2</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>5983.70</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7998</th>\n",
       "      <td>714</td>\n",
       "      <td>spain</td>\n",
       "      <td>male</td>\n",
       "      <td>45</td>\n",
       "      <td>0</td>\n",
       "      <td>124693.48</td>\n",
       "      <td>1</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>187194.15</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7999</th>\n",
       "      <td>623</td>\n",
       "      <td>germany</td>\n",
       "      <td>male</td>\n",
       "      <td>33</td>\n",
       "      <td>8</td>\n",
       "      <td>96759.42</td>\n",
       "      <td>1</td>\n",
       "      <td>yes</td>\n",
       "      <td>yes</td>\n",
       "      <td>174777.98</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8000 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      creditscore geography  gender  age  tenure    balance  numofproducts  \\\n",
       "0             796    france    male   30       2  137262.71              2   \n",
       "1             748     spain    male   36       4  141573.55              1   \n",
       "2             845     spain  female   38       2  112803.92              1   \n",
       "3             585     spain  female   33       8       0.00              2   \n",
       "4             768    france  female   40       8       0.00              2   \n",
       "...           ...       ...     ...  ...     ...        ...            ...   \n",
       "7995          498   germany    male   35       2  121968.11              2   \n",
       "7996          765    france    male   44       6       0.00              2   \n",
       "7997          696   germany  female   27       2   96129.32              2   \n",
       "7998          714     spain    male   45       0  124693.48              1   \n",
       "7999          623   germany    male   33       8   96759.42              1   \n",
       "\n",
       "     hascrcard isactivemember  estimatedsalary  exited  \n",
       "0          yes             no         62905.29       0  \n",
       "1          yes             no         82158.14       0  \n",
       "2          yes             no        179631.85       0  \n",
       "3          yes             no        114182.07       0  \n",
       "4           no            yes         69080.46       0  \n",
       "...        ...            ...              ...     ...  \n",
       "7995        no            yes        188343.05       0  \n",
       "7996       yes            yes        159899.97       0  \n",
       "7997       yes            yes          5983.70       0  \n",
       "7998        no            yes        187194.15       0  \n",
       "7999       yes            yes        174777.98       0  \n",
       "\n",
       "[8000 rows x 11 columns]"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_full_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "15a38a67",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_full_train = df_full_train.exited"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "cbc45a38",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       0\n",
       "1       0\n",
       "2       0\n",
       "3       0\n",
       "4       0\n",
       "       ..\n",
       "7995    0\n",
       "7996    0\n",
       "7997    0\n",
       "7998    0\n",
       "7999    0\n",
       "Name: exited, Length: 8000, dtype: int64"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_full_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "b9d94f48",
   "metadata": {},
   "outputs": [],
   "source": [
    "del df_full_train['exited']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "id": "d535b633",
   "metadata": {},
   "outputs": [],
   "source": [
    "dicts_full_train = df_full_train.to_dict(orient='records')\n",
    "\n",
    "dv = DictVectorizer(sparse=False)\n",
    "X_full_train = dv.fit_transform(dicts_full_train)\n",
    "\n",
    "dicts_test = df_test.to_dict(orient='records')\n",
    "X_test = dv.transform(dicts_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "405b9f4e",
   "metadata": {},
   "outputs": [],
   "source": [
    "dfulltrain = xgb.DMatrix(X_full_train, label=y_full_train,\n",
    "                    feature_names=dv.get_feature_names_out())\n",
    "\n",
    "dtest = xgb.DMatrix(X_test, feature_names=dv.get_feature_names_out())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "246cd132",
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_params = {\n",
    "    'eta': 0.1, \n",
    "    'max_depth': 4,\n",
    "    'min_child_weight': 1,\n",
    "    \n",
    "    'objective': 'binary:logistic',\n",
    "    'eval_metric': 'auc',\n",
    "\n",
    "    'nthread': 8,\n",
    "    'seed': 1,\n",
    "    'verbosity': 1,\n",
    "}\n",
    "\n",
    "model = xgb.train(xgb_params, dtrain, num_boost_round=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "d2a68dbe",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = model.predict(dtest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "7b2bcb67",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.857551936856521"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "0b847c6f",
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "84c4c30d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x20c4cddf188>"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(5,5))\n",
    "\n",
    "plt.plot(fpr, tpr, label='model')\n",
    "plt.plot([0,1], [0,1], label='random', linestyle='--')\n",
    "\n",
    "plt.xlabel('FPR')\n",
    "plt.ylabel('TPR')\n",
    "\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "5b105899",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'Feature importance'}, xlabel='F score', ylabel='Features'>"
      ]
     },
     "execution_count": 140,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdwAAAEWCAYAAADM0CYnAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8/fFQqAAAACXBIWXMAAAsTAAALEwEAmpwYAABGcUlEQVR4nO3dd5xU5dn/8c+XKoqCivpQRaQKLKsCSiS6RkFFLFgCqJGiJiQaKyI/S0SfRLEFC5agErBhR0hUhCBrQaTpUkQJeWQNASIWQCkq5fr9ce7dndmd2Z0FdrZd79drX3vmnlPuc7PstafM+crMcM4551zZqlHeHXDOOeeqAy+4zjnnXBp4wXXOOefSwAuuc845lwZecJ1zzrk08ILrnHPOpYEXXOdchSLpRklPlHc/nNvT5J/Dda7qkJQLHALsiGlua2ZrdnOdl5rZP3avd5WPpFFAazO7qLz74io/P8J1ruo5w8zqx3ztcrHdEyTVKs/t76rK2m9XcXnBda4akNRA0pOS1kpaLemPkmqG9w6X9LakbyR9LelZSQ3De08DLYC/SdokaYSkLEn/KbT+XEknh+lRkl6W9Iyk74DBxW0/QV9HSXomTLeUZJKGSFolab2kYZK6SVosaYOksTHLDpY0W9JDkjZK+kzSSTHvN5E0VdK3kv4l6bJC243t9zDgRqB/2PdFYb4hkj6V9L2kzyX9JmYdWZL+I+k6SevC/g6Jeb+epPskfRH6976keuG9YyV9EPZpkaSsXfindhWYF1znqoeJwHagNXAk0Bu4NLwn4E6gCdABaA6MAjCzXwH/puCo+e4Ut3cW8DLQEHi2hO2n4higDdAfuB+4CTgZ6Aj8UtIJheb9HGgE3Aq8KumA8N4k4D9hX88D7ogtyIX6/SRwB/BC2PcuYZ51QF9gP2AIMEbSUTHr+B+gAdAUuAR4WNL+4b17gaOBnwEHACOAnZKaAq8Dfwztw4FXJB1UijFyFZwXXOeqntfCUdIGSa9JOgQ4DbjazDab2TpgDDAAwMz+ZWYzzOxHM/sK+DNwQvLVp2SOmb1mZjuJClPS7afof83sBzObDmwGJpnZOjNbDbxHVMTzrAPuN7NtZvYCsBw4XVJzoCdwQ1hXDvAE8KtE/TazrYk6Ymavm9n/WeQdYDrw85hZtgG3h+2/AWwC2kmqAQwFrjKz1Wa2w8w+MLMfgYuAN8zsjbDtGcACoE8pxshVcH6Nwrmq5+zYG5wkdQdqA2sl5TXXAFaF9w8GHiQqGvuG99bvZh9WxUwfWtz2U/RlzPTWBK/rx7xebfF3g35BdETbBPjWzL4v9F7XJP1OSNJpREfObYn2Y29gScws35jZ9pjXW0L/GgF7Af+XYLWHAudLOiOmrTYwq6T+uMrDC65zVd8q4EegUaFCkOdOwIAMM/tG0tnA2Jj3C3+UYTNRkQEgXIstfOozdpmStr+nNZWkmKLbApgKrAEOkLRvTNFtAayOWbbwvsa9llQXeAW4GJhiZtskvUZ0Wr4kXwM/AIcDiwq9twp42swuK7KUqzL8lLJzVZyZrSU67XmfpP0k1Qg3SuWdNt6X6LTnhnAt8fpCq/gSaBXz+p/AXpJOl1QbuBmouxvb39MOBq6UVFvS+UTXpd8ws1XAB8CdkvaSlEF0jfXZYtb1JdAynA4GqEO0r18B28PRbu9UOhVOr48H/hxu3qopqUco4s8AZ0g6JbTvFW7Aalb63XcVlRdc56qHi4mKxTKi08UvA43De7cBRwEbiW7cebXQsncCN4drwsPNbCPwO6Lrn6uJjnj/Q/GK2/6eNpfoBquvgT8B55nZN+G9gUBLoqPdycCt4XppMi+F799I+igcGV8JvEi0HxcQHT2najjR6ef5wLfAXUCN8MfAWUR3RX9FdMR7Pf47ukrxB18456oMSYOJHtLRs7z74lxh/teTc845lwZecJ1zzrk08FPKzjnnXBr4Ea5zzjmXBv45XJdUw4YNrXXr1uXdjQpj8+bN7LPPPuXdjQrDxyOej0eB6j4WCxcu/NrMijyW0wuuS+qQQw5hwYIF5d2NCiM7O5usrKzy7kaF4eMRz8ejQHUfC0lfJGr3U8rOOedcGnjBdc4559LAC65zzjmXBl5wnXPOuTTwguucc86lgRdc55xzLg284DrnnHNp4AXXOeecSwMvuM4551waeMF1zjnn0sALrnPOOZcGXnCdc85VWatWreLEE0+kQ4cOdOzYkQceeACA66+/nvbt25ORkUG/fv3YsGFD3HL//ve/qV+/Pvfee2/C9X777bf06tWLNm3a0KtXL9avX19iX7zgpoGklpKWlmL+CZLOK8s+OedcdVCrVi3uu+8+Pv30Uz788EMefvhhli1bRq9evVi6dCmLFy+mbdu23HnnnXHLXXPNNZx22mlJ1zt69GhOOukkVqxYwUknncTo0aNL7stu742rsrZu20HLka+XdzcqjOs6b2ewj0c+H494Ph4FKsJY5I4+HYDGjRvTuHFjAPbdd186dOjA6tWr6d27d/68xx57LC+//HL+69dee41WrVoVGzE4ZcoUsrOzARg0aBBZWVncddddxfbJj3DTp5akiZIWS3pZ0t6S/iBpvqSlksZJUuGFks0jKVvSXZLmSfqnpJ+H9pqS7pW0JGzr96H9aEnvSFoo6S1JjdO7+845V75yc3P5+OOPOeaYY+Lax48fn380u3nzZu666y5uvfXWYtf15Zdf5hfyxo0bs27duhK370e46dMOuMTMZksaD/wOGGtmtwNIehroC/yt0HLFzVPLzLpL6gPcCpwM/Bo4DDjSzLZLOkBSbeAh4Cwz+0pSf+BPwNDCnZT067AOGjU6iD903r4Hh6ByO6Re9Je7i/h4xPPxKFARxiLv6DPP1q1bueqqq7j00kv56KOP8tufeeYZNmzYQNOmTcnOzubRRx+ld+/eLFiwgNzcXOrVq1dkXQDbt2+Pay/8OhEvuOmzysxmh+lngCuBlZJGAHsDBwCfULTgnljMPK+G7wuBlmH6ZOAxM9sOYGbfSuoEdAJmhAPkmsDaRJ00s3HAOIAWrVrbfUv8RyTPdZ234+NRwMcjno9HgYowFrkXZuVPb9u2jb59+zJs2DCuvfba/PaJEyfyySefMHPmTPbee28AbrnlFubOncvEiRPZsGEDNWrUoGPHjlxxxRVx62/atCnt2rWjcePGrF27liZNmpCVlUVx/KcjfSzB60eArma2StIoYK/YGSTtVcI8P4bvOyj4t1SCbQn4xMx6lKbD9WrXZHm4DuKiv5hj/xNXdz4e8Xw8ClSksTAzLrnkEjp06BBXbKdNm8Zdd93FO++8k19sAd5777386VGjRlG/fv0ixRbgzDPPZOLEiYwcOZKJEydy1llnldgXv4abPi0k5RW8gcD7YfprSfWBRHcl75XCPIVNB4ZJqgUg6QBgOXBQ3vYl1ZbUcRf3wznnKo3Zs2fz9NNP8/bbb5OZmUlmZiZvvPEGV1xxBd9//z29evUiMzOTYcOGlbiuSy+9lAULFgAwcuRIZsyYQZs2bZgxYwYjR44scXk/wk2fT4FBkv4CrAAeBfYHlgC5wPzCC5jZBkmPFzdPAk8AbYHFkrYBj5vZ2PAxowclNSD6d7+f6PS0c85VWT179sSs8Ek/6NOnT4nLjho1Ku71E088kT994IEHMnPmzFL1xQtuGphZLnBEgrduDl+F5x8cM51snqyY6a8J13DDtdtrw1fs/DnA8aXuvHPOuT3CTyk755xzaeAF1znnnEsDL7jOOedcGnjBdc4559LAC65zzjmXBl5wnXPOuTTwguucc86lgRdc55xzpTZ06FAOPvhgOnXqlN+2aNEievTowdChQznjjDP47rvvAJg3b17+U566dOnC5MmTE65zV0LdKxMvuM4550pt8ODBTJs2La7t0ksvZfTo0YwfP55+/fpxzz33ANCpUycWLFhATk4O06ZN4ze/+Q3btxdNE9qVUPfKRIkeeeUcRGlBNX75QHl3o8KoCAkoFYmPR7zqMh65MYEmubm59O3bl6VLlwKw3377sXHjRt555x0OP/xwTjnlFJYtWxa3/MqVKzn22GNZvXo1tWrFj1e7du3Izs7OT+DJyspi+fLlZb9Te5ikhWbWtXC7H+FWYpJeC4Hyn4QcWyRdEgLpsyU9LmlsaD9I0ishzH6+pOPKt/fOuaqmU6dOTJ06FYCXXnqJVatW5b83d+5cOnbsSOfOnXnssceKFFvYtVD3yqTq/zlWtQ0Nebf1gPmSXgduAY4CvgfeBhaFeR8AxpjZ+5JaAG8BHQqv0APok6sIodoViY9HvOoyHrEh6//973/ZvHlzftuwYcP44x//yPr16/n5z39OjRo14uZ/+OGH+eKLL7jxxhvZZ599qFOnTty6dyXUvVIxM/+qpF/AKKKCugjYCIwEJsa8fyUwNkyvA3JivlYD+xa3/rZt25orMGvWrPLuQoXi4xGvOo7HypUrrWPHjkXaZ82aZcuXL7du3bolXC4rK8vmz59fpL1t27a2Zs0aMzNbs2aNVdbfQcACS/A71U8pV1KSsoCTgR5m1gX4mCj3NpkaYd7M8NXUzL4v+54656qLvFPAO3fu5I9//GN+xuzKlSvzb5L64osvWL58OS1btiyyfF6oO5ByqHtl4gW38moArDezLZLaA8cCewMnSNo/BNCfGzP/dOCKvBeSMtPZWedc1TJw4EB69OjB8uXLadasGU8++SSTJk2ibdu2DBo0iCZNmjBkyBAA3n//fbp06UJmZib9+vXjkUceoVGjRsDuh7pXJn6XciUlqS7wGtCU6Mj2IKJTzG2B4cAaotD7b83sJkmNgIeJrtvWAt41s2HFbaNdu3ZWGe8QLCvZ2dlkZWWVdzcqDB+PeD4eBar7WCS7S9lvmqqkzOxH4LTC7ZIWmNm4cIQ7mejINi+kvn96e+mccy6Pn1KuekZJygGWAiuJjoKdc86VMz/CrWLMbHh598E551xRfoTrnHPOpYEXXOeccy4NvOA655xzaeAF1znnnEsDL7jOOVcNJMqv7d+/f35ObcuWLcnMzARg27ZtDBo0iM6dO9OhQwfuvPPOhOus6vm1e1q1KriSBktqEvP6CUlH7IH1tpR0wS4sN0HSebuw3OC8FCDnnEtFovzaF154gZycHHJycjj33HM555xzgCjp58cff2TJkiUsXLiQv/zlL+Tm5hZZZ1XPr93TqlXBBQYD+QXXzC41s2XJZ09ZS6DUBTddwkMwnHPV2PHHH88BBxyQ8D0z48UXX2TgwIEASGLz5s1s376drVu3UqdOHfbbb78iy02ZMoVBgwYBMGjQIF577bUy639VUCV+EUu6iCgZpw4wF/gd8CTQFTBgPLAqvH5W0lagB/AmMNzMFkjaRPTow5OB9cCNwN1AC+BqM5sqqSXwNLBP2PQVZvYBMBroEB44MRF4MLRlAXWBh83sL5IEPAT8guihFIrZh9HAmcB2YLqZDZd0BnBz2K9vgAvN7MtC+55wHkmjiP64aAl8Lak58HszywnLzQZ+a2aLk43r1m07aDny9eKGvlq5rvN2Bvt45PPxiFdRxyM2MD6Z9957j0MOOYQ2bdoAcN555zFlyhQaN27Mli1bGDNmTMJiXdXza/e0Sl9wJXUgemThcWa2TdIjRAWoqZl1CvM0NLMNkq4gFNjQHruqfYBsM7tB0mTgj0Av4AiiIjqVKOKul5n9IKkNMImoiI8M6+0b1vtrYKOZdQvPPJ4taTpwJNAO6AwcAiwDxks6AOgHtDczk9Qw9Ol94NjQdikwAriu0BAUN8/RQE8z2yppENER/tWS2gJ1iyu2zrnqY9KkSflHtwDz5s2jZs2arFmzJj/b9uSTT6ZVq1bl2MvKr9IXXOAkosIyPxTQesA0oJWkh4DXCc8TLsFPYTmAJcCPoYAvITpKBKgNjA1JOzuIggIS6Q1kxFyfbQC0AY4HJpnZDmCNpLfD+98BPwBPhBD5v4f2ZsALkhoTHcGuTLCt4uaZamZbw/RLwC2SrgeGAhMSddwD6JOrLgHjqfLxiFdRx6O4wHiAHTt28MILL/CXv/wlv/3+++/niCOOYPbs2QC0atWKiRMncuKJJ8ate7/99uOVV17hwAMP5JtvvmHfffclOzubTZs2Va3g+D2kKhRcEYWu/7+4Rukm4BTgcuCXREWmONusIDppJ/AjgJntjLkGeg3wJdCF6Pr3D8X06fdm9lahPvUhOsUdx8y2S+pO9MfDAKIYvV8QnX7+czidnUWUBlRYcfNsjtnGFkkzgLOIxqNIkkWYbxwwDqK0oN9fWLXyKHdHdnY2v6zGCSiF+XjEqwzjkZubyz777BOX5DNt2jQ6d+7M+eefn982d+5cPvvsM0444QS2bNnCF198wV133UVGRkbc+vr378+KFSs499xzGT16NAMGDCArK6vapwUlUxVumpoJnCfpYABJB0g6FKhhZq8AtwBHhXm/B/bdjW01ANaa2U7gV0DNJOt9C/itpNqhT20l7QO8CwyQVDMckZ4Y3q8PNDCzN4CrgcyY7a0O04OK6VNJ8+R5guj68nwz+7aEeZ1zVUii/FqA559/Pu50MsDll1/Opk2b6NSpE926dWPIkCH5xbY65dfuaZX+CNfMlkm6GZguqQawDbgWmBxeA+Qd/U4AHou5aaq0HgFekXQ+MIuCI8jFwHZJi8I2HiA6Df1RuFHqK+Bsori8XxCdsv4n8E5Yfl9giqS9iI6Orwnto4CXJK0GPgQOS9CnVOYBwMwWSvoO+Gvpdts5V9lNmjQpYfuECROKtNWvX5+XXnop4fxPPPFE/vSBBx7IzJkz90j/qgMPoK9GwmeQs4luztpZ0vweQB/PT5PF8/GI5+NRoLqPRbIA+qpwStmlQNLFRB+ZuimVYuucc27PqvSnlF1qzOwp4Kny7odzzlVXfoTrnHPOpYEXXOeccy4NvOA655xzaeAF1znnnEsDL7jOOedcGnjBdc65NEoUBD9q1CiaNm2aHwb/xhtv5L+3ePFievToQceOHencuTM//FD0ibIeBF85eMHdAyS1l5Qj6WNJh+/hde9y2HxYtknJczrn0iVREDzANddckx8G36dPHwC2b9/ORRddxGOPPcYnn3xCdnY2tWvXLrKsB8FXDl5w94yzgSlmdqSZ/V8qC0iqWfJcu20wUSauc66CKC4IvrDp06eTkZFBly5dgOhRijVrFv3V4UHwlUOVfPBFCIp/kygr9mdED/c/i/jA+UbAAjNrKWkwUdGsCXQC7iOKuvsVUWpQHzP7NsTyPQbsDfwfUQJRD6LAgR2SjgeGEMX8zSXKv/0ncHFI68kFxhPF940Nz1m+kej5ya+b2Q2h/0OInv+8Niz/Y2ifAPzdzF4OrzeZWf0wPSL0d2fYzwVEiUDPxjw7+lYKhdwXN44eQB+vogaMlxcfj3gljUdJQfBjx47lqaeeomvXrtx3333sv//+/POf/0QSp5xyCl999RUDBgxgxIgRRZb1IPjKoUoW3KANMNDMLpP0InBuCfN3IiqQewH/Am4wsyMljQEuBu4nelLT783sHUm3A7ea2dWSHgM2mdm9odi3Ay4xs9mSxgO/A+4N2/nBzHqGU70fEmX5ricKXzibqFDfFto3EoUkfFxcxyWdRvQHwzGhsB8Q/kC4goI/MJKF3Bdel+fhJlFR807Li49HvJLGo7hc2oyMDJ588kkkMX78eC644AJuuOEGli9fzj/+8Q8ee+wx6taty3XXXUfNmjU5+uij49a9ffv2uPUXfp1unoebWFUuuCvNLCdML6QgRD6ZWWb2PfC9pI3A30L7EqIw+QZAQzPLS/iZSBTqnsgqM5sdpp8BrqSg4L4QvncDss3sKwBJzxIF1FOo/QWSB93nORn4q5ltAUgSvZcs5D5ObB5ui1at7b4lVflHpHSu67wdH48CPh7xShqP3AuzCqYT5NLmadWqFX379iUrK4v//ve/bN26lbPOinKp58+fz86dO4ss17RpU9q1a0fjxo1Zu3YtTZo0KdfwgOoeXpBMVf7f8mPM9A6gHtGp1Lzr1nsVM//OmNc7Kf04FY5gin2dF+mnUiyfJ7//4XR0nZh1FRv7VEzIfVL1atdkeQmnwaqT7OzsuF+a1Z2PR7zdGY+1a9fmnxKePHly/h3Mp5xyCnfffTdbtmyhTp06vPPOO1xzzTVFlj/zzDOZOHEiI0eOZOLEifkF2lUs1e2mqVyiU7UA55VmQTPbCKyX9PPQ9CsK8mwLayEpL293ING15MLmAidIahRuoBoY1jcXyJJ0YAiwPz9J/88C8m5XnA4MlbQ3QDh9DPA9UdZucSH3zrk0ShQEP2LECDp37kxGRgazZs1izJgxAOy///5ce+21dOvWjczMTI466ihOPz36I9iD4CufqnyEm8i9wIuSfgW8vQvLDyIKsN8b+JzoBqlEPgUGSfoLsAJ4tPAMZrZW0v8jukYr4A0zmwIgaRQwh+imqY+IbuYCeJwoqH4eMJNwtGxm08INXQsk/QS8QXQz1oTQ363AaSQOuXfOpVGiIPhLLrkk6fwXXXQRF110UZF2D4KvfDyAfg8LN0393cw6lTRvRecB9PH8ulQ8H494Ph4FqvtYeAC9c845V46q2ynlMmdmuUQfMXLOOefy+RGuc845lwZecJ1zzrk08ILrnHPOpYEXXOeccy4NvOA655xzaeAF1znnCkkUEn/LLbeQkZFBZmYmvXv3Zs2aNQD89NNPDBkyhKFDh9KlS5ekD+33kHjnBTfNJG0K35tIyovZy5TUp3x75pzLkygk/vrrr2fx4sXk5OTQt29fbr/9dgAef/xxAMaPH8+MGTO47rrr2LlzZ5F1eki888/h7gGSaplZqXLKzGwNBc9zziTKrn1jD3cN2LX+gefhFub5r/Gq4njkZdYef/zx5Obmxr2333775U9v3ryZKD8Eli1bxkknnQTAwQcfTMOGDVmwYAHdu3ePW37KlCn5R7+DBg0iKyuLu+66q4z2xFVEfoSbIkkXS1osaZGkpyVNkPRnSbOAuyQdLmmapIWS3pPUPix3mKQ5kuZL+t+Y9bWUtFRSHeB2oL+kHEn9JZ0QpnMkfSwpL4BghKQloQ+jQ1umpA9D3yZL2j+0Z0u6Q9I7wFWSjpb0TujfW5Iap3sMnavsbrrpJpo3b86zzz6bf4TbpUsXpkyZwo4dO1i5ciULFy5k1apVRZb1kHjnR7gpkNQRuAk4zsy+Dmk8fybKqT3ZzHZImgkMM7MVko4BHiGKv3sAeNTMnpJ0eeF1m9lPkv4AdDWzK8L2/gZcHgLs6wM/JAqZD6t4Cvi9mb0j6XbgVqI0IIjye08IqUPvAGeZ2VeS+gN/AoYm2FcPoE/CA9fjVcXxKC4kHqBXr1706tWLZ599luHDhzNkyBAOP/xwZsyYwWWXXUbjxo1p3749n376aZFruRUtJL4seQB9Yl5wU/ML4GUz+xqigPdwOumlUGzrAz8DXso7zQTUDd+PA84N008DqZxDmg38OYTSv2pm/5FUJGReUgOiopoXEzgReClmPXlh9+2IHjc5I/SvJlESUREeQJ+cB67Hq4rjkWpI/GGHHcbpp5/OxIkTATjppJPyH9j/s5/9jHPOOYcjjjgibpmKFhJflqp7eEEyVet/S9lJFvCeFyZfA9hgZplJli9VJJOZjZb0OtAH+DAU2xJD5ovpn4BPzKxHcTMX5gH08TxwPV51G48VK1bQpk0bAKZOnUr79u0B2LJlC3mpazNmzKBWrVpFii14SLzza7ipmgn8UtKBEBfwDoCZfQeslHR+eF+SuoS3ZwMDwvSFSdafHxQflj/czJaY2V3AAqA9CULmzWwjsF7Sz8OivyI6dVzYcuAgST3CsrXDaXLnXAKJQuJHjhxJp06dyMjIYPr06TzwwAMArFu3jqOOOopBgwZx11138fTTT+evx0PiXSw/wk2BmX0i6U/AO5J2AB8nmO1C4FFJNwO1geeBRcBVwHOSrgJeSbKJWcBISTnAnUBPSScCO4BlwJtm9mOSkPlBRCHzewOfA0MS9P8nSecBD4bT0LWA+4FPSj0YzlUDpQmJb9myJcuXL094GtVD4l0sL7gpMrOJRNdIk72/Ejg1SXvsqdzRoT2XEONnZt8C3WLmeYEEzGx03vIxbTnAsQnmzUow3/HJ+u+cc65spXRKOXzkpW6YzpJ0paSGZdoz55xzrgpJ9RruK8AOSa2BJ4HDgOfKrFfOOedcFZNqwd0ZnlTUD7jfzK4B/MEJzjnnXIpSLbjbJA0kukHn76Gtdtl0yTnnnKt6Ui24Q4hu/PmTma2UdBjwTNl1yznnnKtaUrpL2cyWSboBaBFer6TQ3bLOOeecSy7Vu5TPAHKAaeF1pqSpZdgv55xzrkpJ9ZTyKKA7sAHyP9N5WJn0yDlXrOXLl5OZmZn/td9++3H//ffnv3/vvfciia+//jrh8tOmTaNdu3a0bt3aM1mdS6NUH3yx3cw2xjyYH0r/XN8KR1I2MNzMFpR3X5xLVbt27cjJyQFgx44dNG3alH79+gGwatUqZsyYQYsWLRIuu2PHDi6//HJmzJhBs2bN6NatG2eeeWbCZ/865/asVAvuUkkXADUltQGuBD4ou26VjqSaZrajvPuxu3Y1KL6seAB9vPIOXM9NECQxc+ZMDj/8cA499FAArrnmGu6+++6kD8afN28erVu3plWrVgAMGDCAKVOmeMF1Lg1SPaX8e6Aj8CPRAy82UpC5mpSkWyR9JmmGpEmShhcT1H6opJkhSH2mpBah/fAQsD5f0u2SNoX2LEmzJD0HLAltr4X1fhJyXfP6sUnSfZI+Cus+KKab50uaJ+mfeSEAoV+ZMcvPlpSRykAl2ueY/Ui034WD7CdIejTs2+eKwujHS/pU0oSY7TwqaUHY19ti2nMl3Rb2dYmk9pJqSFqRt9/h9b8kNUpln1zF9fzzzzNw4EAgSrBp2rQpXbp0STr/6tWrad68ef7rZs2asXr16jLvp3MuhSNcSTWBqWZ2MlEIe0okdSXKgT0ybOcjYCFR1mqioPaxwFNmNlHSUOBBosD1B4AHzGySpGGFNtMd6BTumgYYGnJi6wHzJb1iZt8A+wAfmdl1isLebwWuyBsDM+suqU9oPxl4AhgMXC2pLVDXzBaHQIExCXZ3i5n9rJh9ppj9hvgg+wnA/uG9M4G/EWXqXhr2KTNcQ78p7GtNYKakDDNbHNb3tZkdJel3RKfML5X0DFHAwv1hHxfl5fsW+nfzAPokyjtwvXCg97Zt23jllVfo27cv06ZN44YbbuCee+4hOzubH374gdmzZ9OgQYO4ZZYuXcratWvz1/Xpp5+yZs2aXQoL95DxeD4eBXwskjCzEr+AqUCDVOaNWeZq4LaY138Gbga2Et3xnPf1aXj/a6B2mK5NVDQAviEqigD7AZvCdBYwq9A2RxEl9CwiOgo/NrTviFlHKyAnTGcDx4XpQ4B/hem9gX+FfowGrtiNfR4O1C9mvycAg2KWmQBcGNPXFTHvPQWcHaaHERX0xcBXwIDQngs0DdPHAP8I082J/uiAKMmob0n707ZtW3MFZs2aVd5diPPaa69Zr169zMxs8eLFdtBBB9mhhx5qhx56qNWsWdOaN29ua9eujVvmgw8+sN69e+e/vuOOO+yOO+7Ype1XtPEobz4eBar7WAALLMHv1FSv4f4ALJE0g4JQc8zsymKWUYK2koLaY6VyU1Z+XyRlER259TCzLeGGqL1SWPeP4fsOwhF/WH4GcBbwS6Br2EaxR7gk3mcoeb83F3qd16edMdN5r2uFB48MB7qZ2fpwVLxXguVj92mVpC8l/YKoECfL5nWVxKRJk/JPJ3fu3Jl169blv9eyZUsWLFhAo0bxVw26devGihUrWLlyJU2bNuX555/nuef8sejOpUOq13BfB24B3iU6RZr3VZz3gTMk7SWpPnA6sIXkQe0fEB/U/n6Y/pDoNC0x7yfSAFgfimV74iPragDnhekLYtZdnCeITmvPtyg+DzObZWaZCb5+Vsw+lxRQvyv2IyrSGyUdApyW4nJPED0h7EWrAjeZVWdbtmxhxowZnHPOOSXOu2bNGvr06QNArVq1GDt2LKeccgodOnTgl7/8JR07dizr7jrnSP1JU0lzYItZZr6ih2MsAr4AFhCd5k0W1H4lMF7S9USnSPOC1K8GnpF0HVHh35hkk9OAYZIWA8uJCnWezUBHSQvD8v1T6P9CSd8Bf90D+wzJ97vUzGyRpI+JAuQ/B2anuOhUov1JeZ9cxbT33nvzzTffJH0/Nzc3f7pJkya88cYb+a/79OmTX4Cdc+mTUsGVtJIEp3jNrFUJi95rZqMk7U10dHyfJQ9qz6XgJqJYq4muxZqkAURFDDPLJroGm7f8jxRzpGdmtxAdpce2ZcVMfw20zHstqQnRkfH04nexiCL7HNafbL8HJ3ttMSH1Cd6LWy6mvWXM9AKia915uhDdLPVZqjvjnHNuz0j1Gm7XmOm9gPOBA1JYbpykI8IyE83so1L2D+BoYKwkET3paugurKNUJF0M/Am41sx2lnLxPbHPe5ykkcBv8Wu3zjlXLlI9pVz43NX9kt4H/lDCchfsasdi1vEe0ZHZ7qyjfinnf4rojuBd2dZu73NZMLPReOCEc86Vm1RPKR8V87IG0RHvvmXSI+ecc64KSvWU8n0x09uBlUQfl3HOOedcClItuJeY2eexDeGzoM4555xLQaqfw305xTbnnHPOJVBswQ0Pvj8XaCDpnJivwSR/ipNzrhgbNmzgvPPOo3379nTo0IE5c+bw0ksv0bFjR2rUqMGCBcnTIj3L1rnKq6RTyu2AvkBD4IyY9u+By8qoT9WGpIbABWb2SHn3xaXPVVddxamnnsrLL7/MTz/9xJYtW2jYsCGvvvoqv/nNb5Iu51m2zlVuxRZcM5sCTJHUw8zmpKlP1UlD4HdEyUFlQhUsY7e6++6773j33XeZMGECAHXq1KFOnTo0bNiwxGU9y9a5yi3Vm6Y+lnQ5USZu/qlkMyvzh1BUcaOBwyXlADOAdUR3f9cFJpvZrZJaAm8SPaf5Z0RP3jrLzLaGgIbhZrYgZNsuMLOW4ZT/6UT/VvtIOgN4COhM9G8+KvwxVSwPoI+3OwH0eeHxn3/+OQcddBBDhgxh0aJFHH300TzwwAPss88+Ja4jUZbt3Llzd6k/zrn0S7XgPg18BpwC3E70tKJPy6pT1chIojzfTEm9iQIWuhOlDk2VdDzwb6ANMNDMLpP0IlGYwzMlrLsHkGFRZu4dwNtmNjScxp4n6R9mVjilyPNwi7E7ebh52aDLly9n4cKFDB48mMGDB/PQQw/x29/+lqFDo79dN2zYwMKFC9m0aVORdezJLNs9wTNP4/l4FPCxSCzVgtvazM6XdJZFAfHPAW+VZceqod7h6+Pwuj5Rof03sNKi0HmIUppaprC+GXkpR2G9Z0oaHl7vBbQgwR9NZjYOGAfQolVru29Jqj8iVd91nbezq+ORe2EWAO3bt+fOO+/kd7/7HQA1a9Zk9OjRZGVF7zds2JCjjz6arl27FllH3bp1mTNnTv68c+bMoVu3bvmv0y07O7vctl0R+XgU8LFILNXfHtvC9w2SOgH/JbVf+i51Au40s7/ENUanlGMzcXcA9cL0dgruNC9813js0auAc81seWk6VK92TZaHU6Eu+iWSVzh31f/8z//QvHlzli9fTrt27Zg5c2bK12A9y9a5yi3Vz+GOk7Q/UdrOVGAZcHeZ9ar6+J6CR2S+BQwNObpIairp4BKWzyUKd4CCvN9E3gJ+HwIgkHTkLvfY7baHHnqICy+8kIyMDHJycrjxxhuZPHkyzZo1Y86cOZx++umccsopgGfZOleVpBpe8ESYfAcoKZLPpcjMvpE0W9JSohujngPmhLq4CbiI6Ig2mXuBFyX9Cni7mPn+F7gfWByKbi7Rx71cOcjMzCzyWdt+/frRr1+/IvN6lq1zVUeq4QWHAHcATczstBA/18PMnizT3lUDCdKFHkgwW2wm7r0x058BGTHz3RzaJwATYubbCiT/gKdzzrkyl+op5QlEpyWbhNf/BK4ug/4455xzVVKqBbeRmb0I7AQID1Io7lSnc84552KkWnA3SzoQMABJxwIby6xXzjnnXBWT6seCriW6O/lwSbOBgyj+rljnnHPOxSi24EpqYWb/NrOPJJ1AFGYgYLmZbStuWeecc84VKOmU8msx0y+Y2SdmttSLrXPOOVc6JRVcxUz752+dc865XVRSwbUk0865JFq2bEnnzp3JzMzMfybyqFGjaNq0KZmZmWRmZsY9zCKWB8w7V3WVdNNUF0nfER3p1gvThNdmZvuVdoOSPjCzn5V2uSTryiR6GMcb4fWZwBFmViF/U0nKIorT86c8VXGzZs2iUaNGcW3XXHMNw4cPT7KEB8w7V9WVFEBfc09vcE8V2yAT6Aq8EdY9lehu6ipJUk0zS9vnnz0PN16yPNzcPRTw4AHzzlVtqX4Od4+RtCl8byzpXUk5kpZK+nlof1TSAkmfSLotZrlukj6QtEjSPEkNiLJ5+4d19Jc0WNJYSQ0k5UqqEZbdW9IqSbUlHS5pmqSFkt6T1D7MMyFse5akzyWdIGm8pE8lTYjpR29JcyR9JOmlmLCBXEl3hPcWSDpK0luS/k/SsJgh2E/SZEnLJD0W08fi1vsHSe8D56cwvoMlvRr2cYWku2PeGyhpSRjvu3bxn9CVQBK9e/fm6KOPZty4cfntY8eOJSMjg6FDh7J+/foiyyUKmF+9enVa+uycK3vlGXZ6AfCWmf1JUk1g79B+UwhNrwnMlJQBfAa8APQ3s/mS9gO2AH8AuprZFRAVGwAz2yhpEXACMAs4I2xrm6RxwDAzWyHpGOAR4Bdh2/uH6TOBvwHHAZcC88Pp6/8QPa/4ZDPbLOkGos8o3x6WX2VmPSSNIXoc5nFEsXmfAI+FeboDRwBfANOAcyRll7DeH8ysZ9jH64ELE4znu2Z2ZZjOBI4kivVbLukhoieD3UWULrQemC7pbDN7LXYlHkCfXLIA+sJB2/fccw+NGjVi/fr1DB8+nK1bt5KRkcGTTz6JJMaPH88FF1zADTfcELdcRQuYL4mHjMfz8SjgY5FYeRbc+cB4SbWB12IC1n8ZfunXAhoTFScD1prZfAAz+w6iI4livAD0Jyq4A4BHwlHjz4CXYpatG7PM38zMJC0BvjSzJWE7nxDl/zYL/Zkdlq8DzIlZPu909hKgvpl9D3wv6QdJDcN788zs87DeSUBP4IcS1vtC3oSZ3QPcU9yOAzPNbGPYxjLgUOBAINvMvgrtzwLHE//RLw+gL0ayAPriMnIXLVrEtm3bOOecc/LbWrVqRd++fYsEdFe0gPmSeMh4PB+PAj4WiZXbb1Mze1fS8cDpwNOS7gHeA4YD3cxsfTiVuxfhJq1SbmIqcKekA4iO6t4G9gE2mFlmkmXygt53Eh/6vpNorHYAM8xs4C4uD0X3w4j2r7j15ofJp3iEWziwvhbxH/FKiQfQx0slgH7z5s3s3LmTfffdl82bNzN9+nT+8Ic/sHbtWho3bgzA5MmT6dSpU5FlPWDeuaot7ddw80g6FFhnZo8DTwJHAfsRFZeNIRLwtDD7Z0ATSd3CsvtKqkV8gHscM9sEzCOKu/u7me0IR8YrJZ0f1iNJXUrR7Q+B4yS1DsvvLaltqXYcuks6LFy77Q+8X5r1mtk9ZpaZ4OvKRPPHmAucIKlROF0/kCjf2O1BX375JT179qRLly50796d008/nVNPPZURI0bQuXNnMjIymDVrFmPGjAE8YN656qQ8zxdmAddL2kYUtn6xma2U9DHRNc/PgdkAZvaTpP7AQ5LqAVuBk4lOF4+UlAPcmWAbLwAvhW3luRB4VNLNQG3geWBRKh02s6/CdeJJkvJORd9MFFeYqjnAaKAz8C4w2cx27oH1FsvM1kr6f0RjJuANM5uyp9bvIq1atWLRoqI/Tk8//XTC+T1g3rnqQ2b+PAuXWLt27Wz58uXl3Y0Kw69LxfPxiOfjUaC6j4WkhWbWtXB7uZ1Sds4556oTL7jOOedcGnjBdc4559LAC65zzjmXBl5wnXPOuTTwguucc86lgRdc55xzLg284Lpq44cffqB79+506dKFjh07cuuttwLQv3///GD4li1bkpmZmXD5efPmeTi8c26X+ZPpSxCe5/x3M3t5D6/3HqAP0ROfrt+T647ZxgTKoO+VVd26dXn77bepX78+27Zto2fPnpx22mm88EJ+NgTXXXcdDRo0KLLsjh07eOCBB3jvvfc8HN45t0u84O5hkmqZWSqZdr8BDjKzH0ucs5xUlQD6vIB4SdSvXx+Abdu2sW3btrjEKTPjxRdf5O233y6yjnnz5tGkSRMPh3fO7bIqdUpZ0i2SPpM0Q9IkScNVfOD8g4pC7T+XdF5ol6IQ+2WSXgcOjln/0ZLeCet6S1Lj0J6tKHz+HeCqFPo5lSi5aK6k/pIOkvSKpPnh67gw3yhJEyVNVxREf46ku0OI/LQQbYiigPr5ioLlxylBbmGyvlc3O3bsIDMzk4MPPphevXpxzDHH5L/33nvvccghh9CmTZsiy61evZqDD87/UfBweOdcqVWZI1xJXYFziYLXawEfAQuJsl2TBc43JsqjbU8U5/cy0A9oRxQucAiwjILc3oeAs0KIQX/gT8DQsK6GZnZC6MuFQKLTxP8ys/PM7ExJm/JiAiU9B4wxs/cltQDeAjqEZQ4HTiTKy50DnGtmIyRNJoo2fA0Ya2a3h3U9DfQF/hYzNiX1PXYcq1wAfeEg7Pvvv59NmzZxyy230L59ew477DAAxowZQ/fu3RMGZy9dupRt27ZVmnD4dPCQ8Xg+HgV8LBKrMgWXqHBOMbOtAJL+RpSlW1zg/GtmthNYFuIAIQpln2RmO4A1kvLOL7YDOgEzwrpqAmtj1hUbEv8s8Gwp+n4ycERMH/eTlBc7+KaZbZO0JGxzWmhfArQM0ydKGgHsDRxAlLaUX3BT6Hu+qhhAnyzDduHChXzzzTcMGTKE7du3079/fxYuXEizZs2KzFu3bl1ef/31ShMOnw7V/QH1hfl4FPCxSKzy/zYtkChgvQapBc4XXj5RhJKAT8ysR5J1xYbEF3uEm6SfPfL+WIhZT34fQ4TfNiuId9oJ1JK0F9FRe1czWyVpFNEfGqXpe0JVLYD+q6++onbt2jRs2JCtW7fyj3/8gxtuuAGAf/zjH7Rv3z5hsYUoHH716tUeDu+c22VV6Rru+8AZkvaSVJ/odOsWSh84/y4wQFLNcJ3zxNC+HDhIUo+wrtqSEqaDm9mzSULiExVbgOnAFXkvJGWmtstAQXH9Oux3om2k3PeqbO3atZx44olkZGTQrVs3evXqRd++fQF4/vnnGThwYNz8hcPhr7zySg+Hd87tsipzhGtm88PNSIuAL4AFwEZKHzg/mega7xKiAPh3wvp/CjdWPSipAdHY3U90+nZ3XQk8LGlxWO+7wLBUFjSzDZIeD/3NBeYnmKcs+15pZGRk8PHHHyd8b8KECUXaCofDH3vssYwcObKsuuecq+KqTMEN7jWzUZL2Jipa95nZSuDUwjOa2eBCr+uH70bM0WaheXKIrvEWbs8qbUfzthemvwb6J5hnVDHLjIqZvhm4OcHyg2Omc0jQd+ecc+lR1QruOElHEJ1mnWhmH5V3h5xzzjmoYgXXzC4o7z4455xziVSlm6acc865CssLrnPOOZcGXnCdc865NPCC65xzzqWBF1xXIQwdOpSDDz6YTp065bfl5ORw7LHHkpmZSdeuXZk3b17CZadNm+Y5tc65Cs8LbimEVKCue3idV0r6VFJpnr1c5QwePJhp06bFtY0YMYJbb72VnJwcbr/9dkaMGFFkuR07dnD55Zfz5ptvsmzZMiZNmsSyZcvS1W3nnEtZtSy4kmqWdx9i/A7oY2YXxjZKqlIf2SrJ8ccfzwEHHBDXJonvvvsOgI0bN9KkSZMiy82bN4/WrVvTqlUr6tSpk59T65xzFU2F+qUu6RaiRzGuAr4mitebDDwMHET0bOTLzOwzSYcC40P7V8AQM/u3pMOJknpqAm8C15pZfUlZwK1EKTmZROk8rwHNiR6U8UBIykHSJuAvRM9RXg8MMLOvQjfPl/QI0BC4xMzek/Qe8PvwNCckzQZ+a2aLS9jfx4BWwFRJ44EGQBOiFKCvJd0IPE2UnQtwhZl9EPZlVBijTmGcLjIzk9QNeCAs8yNwUhi30UAWUVrSw2b2l+L6BmUfQJ9bQjDC/fffzymnnMLw4cPZuXMnH3zwQZF5Vq9eTfPmzfNfN2vWjLlz5+7xvjrn3O6qMEe4hfJszwHyTt2OIypmRwPDiZJxAMYCT5lZBlGBfTC0P0BUPLsBawptpjtwk5kdEV4PDevtClwp6cDQvg/wkZkdRfQs5Vtj1lHLzLoDV8e0PwEMDvvRFqhrZoslnSgpJ8HXBwBmNiz08UQzGxPWdTRRbu0FwDqgV+hH/5h9JIzT1UQ5ua2A4yTVIYoJvMrMuhDF/m0FLgE2hjHpBlwm6bDC/wYVzaOPPsqYMWNYtWoVY8aM4ZJLLikyT0F4UoGYmEPnnKswKtIRbmnzbHsQFWaIjgLvjmk/O0w/B9wbs4154dnKea6U1C9MNwfaAN8QRd/l5ds+A7was0ze9EIK8mhfAm6RdD1RqPsEADObRXQ0XRpTY2L6agNjQ3rQDqBtoX35D4CknNCXjcBaM5sftv9deL83kBECDCA6km4DxI4FYd60BdAXDqj+73//y+bNm/Pbx48fT79+/cjOzuaggw5izpw5RZZZt24dixYtym9/9913E657T/BQ7Xg+HvF8PAr4WCRWkQruruTZxkqUYVtYbGZtFtERYA8z2yIpm6I5sonWnZehu4MwfmH5GcBZwC8JR+eSTgTGUNQWM/tZSX0ErgG+BLoQjcUPCfoR2xeRPMv392b2VpJt5osNoG/Xrp39/sKzSlpkj8nNzWWfffbJD65u3rw5ksjKymLmzJm0b9++SKh1z549ue+++zj00ENp2rQpV111Fc8991yZROd5qHY8H494Ph4FfCwSqzCnlCl9nu0HwIAwfWFYHuBDolPTxLyfSANgfSiW7YFjY96rQUGu7AUx6y7OE0SnfOeb2bcQHeEmycVNVmwT9XGtme0EfkV0Xbo4nwFNwnVcJO0bbr56C/itpNqhva2kfYpZT9oNHDiQHj16sHz5cpo1a8aTTz7J448/znXXXUeXLl248cYbGTduHFA0p3bs2LGeU+ucq/AqzBHuLuTZXgmMD6dxvwKGhFVdDTwj6Trg9bCORKYBw0IG7XKiQp1nM9BR0sKwfJHovAT9XyjpO+CvKe90yR4BXgl/cMwi/ug3UR9+ktQfeEhSPaLrtycT/THQEvhI0bn5ryg47V4hTJo0KWH7woULi7QVzqnt06dPfgF2zrmKqsIU3KA0eba5REHxha0Gjg137A4gKtyYWTaQHbP8j8BpyTpiZrcAtxRqy4qZ/pqCa7hIakJ0ZDy9+F0ssp2WMdOjCr23AsiIafp/oT2b+H25ImZ6PvFH63luDF/OOefKQUUruHsiz/ZoohuNBGwguompTEm6GPgT0UeQdpb19pxzzlU+Farg7ok8WzN7j+gmo91ZR/1Szv8U8NTubNM551zVVpFumnLOOeeqLC+4zjnnXBp4wXXOOefSwAuuc845lwZecJ1zzrk08ILr9rgxY8bQsWNHOnXqxMCBA/nhhx/i3jczrrzySlq3bk1GRgYffbQrn/5yzrnKxQvubiqLUPo9QdKZkkame7urV6/mwQcfZMGCBSxdupQdO3bw/PPPx83z5ptvsmLFClasWMG4ceP47W9/m+5uOudc2nnBDSpYKP1uM7OpZja6PLa9fft2tm7dyvbt29myZUuR4PgpU6Zw8cUXI4ljjz2WDRs2sHbt2vLoqnPOpU2FevBFItUwlH4f4EWgWejv/5rZC5JyiSIDTwyzXmBm/5J0BnAzUIcoWvBCM/tS0mCgq5ldIWkC8B1RitH/ACPM7OWSxr40AfR5YfJNmzZl+PDhtGjRgnr16tG7d2969+4dN2+i0PjVq1fTuHHjlLblnHOVUYUuuIVC6WsBHxEV3HHAMDNbIekYoof8/4KCUPqJkoYSpfecTUEo/SRJwwptpjvQKSYnd6iZfRse/j9f0itm9g0FofTXSfoDUaHOe4ZxLTPrLqlPaM8LDBgMXF04lJ7iI/tOBdaY2elhDBrEzPNd2M7FwP1AX6Iko7xnR18KjACuS7D+xkSZw+2BqUDCgrurebh52Zfff/89EydO5JlnnqF+/fqMGjWKm266iV69euXP+/XXX/Pxxx+zfXu07vXr17Nw4UI2bdqU0rbKi2d8xvPxiOfjUcDHIrEKXXCpnqH0S4B7Jd0F/D08qjLPpJjveUW7GfCCpMZER7lFQuWD18JznpdJOiTZxmPzcFu0am33LUntRyT3wiwAXnrpJY488kjOPvtsIIrS+/DDD+OyMbt06UKjRo3y2zZv3syZZ55Z4Y9wPeMzno9HPB+PAj4WiVX0glvtQunN7J+Sjgb6AHdKmm5mtyfYZt70Q8CfzWxq6P+oJP2NDaxPNK5F1Ktdk+XhVHGqWrRowYcffsiWLVuoV68eM2fOpGvX+HvKzjzzTMaOHcuAAQOYO3cuDRo0qPDF1jnndldFv2mq2oXSh5i/LWb2DNGR+FEx6+sf831OTJ9Xh+lBKfSpTB1zzDGcd955HHXUUXTu3JmdO3fy61//mscee4zHHnsMiPJrW7VqRevWrbnssst45JFHyrnXzjlX9ir0EW41DaXvDNwjaSewDYj9zExdSXOJiv/A0DaK6PT66tDfw0qxrTJx2223cdttt8W1DRtWcOlcEg8//HC6u+Wcc+WqQhfcoFqF0pvZW8BbSd5+2MxuKzT/FGBKgvVMoOC68eBC75UqftA559zuqwwF10PpnXPOVXoVvuB6KH3++lruqXU555xLv4p+05RzzjlXJXjBdc4559LAC65zzjmXBl5wnXPOuTTwguucc86lgRdcVyqrVq3ixBNPpEOHDnTs2JEHHnigyDweMO+cc0VVq4IrqaWkpeXdj9KqSCH3tWrV4r777uPTTz/lww8/5OGHH2bZsmVx83jAvHPOFVWtCm667E6YvaQK89noRHm4jRs35qijosc777vvvnTo0IHVq1fHzeMB8845V1R1LLg1JT0u6RNJ0yXVk3SZpPmSFkl6JTxGEknnS1oa2t8NbTUl3StpiaTFkn4f2nMl/UHS+0SB9KdK+igsOzPM013SB5I+Dt/bhfbBkl4K8YN5fXo+rP8FoF6qOydpgqQHw/o/l3ReaJeke8L+LJFU4rOgS5Kbm8vHH3/MMcccE9eeLGDeOeeqswpzNJVGbYCBZnaZpBeJUoReNbPHAST9EbiEKPbuD8ApZrZaUsOw/K+JAgKONLPtkg6IWfcPZtZT0kHAR8DxZrYyZp7PQtt2SScDd1CQYtQDyDCzbyVdS5QYlCEpI6yL0L8XgHYJ9uvP4elWkDhs/hyiHN4uQCNgvqR3zSzu0DNRAH2iIOmtW7dy1VVXcemllxa5RltZA+ZL4qHa8Xw84vl4FPCxSKw6FtyVZpYTpvMC4zuFQtsQqE9BeMBsYEIozHkh8ycDj5nZdoC82L0gL6D+WODdvGD7mHkaABMltSHKs60ds+yMmPmOJ4r1w8wWh/QiwutUjkwThc33BCaZ2Q7gS0nvAN2ICnK+RAH0eeHyebZt20bfvn0ZNmwY1157bZGNV9aA+ZJ4qHY8H494Ph4FfCwSq46nlGOD2PMC4ycAV5hZZ+A2Qui8mQ0DbgaaAzmSDiQKb08WbJ8XZp9snv8FZplZJ+AM4sPtNxeaN+E2JL0gKSfB18VJ9lGFvqesXu2a5BYKoDczLrnkEjp06JCw2EIUMP/UU09hZnz44YceMO+cc1TPI9xE9gXWSqpNlLW7GkDS4WY2F5gr6QyiwjudKDM3O++UcqGjXIjC4R+WdFjeKeUwT2xY/OBi+vNu6McsSZ2AjLw3UjzCTbbO30iaCBxAdBR9fWlXMnv2bJ5++mk6d+5MZmYmAHfccQf//ve/gSj3tk+fPrzxxhu0bt2avffem7/+tTRxwM45VzV5wY3cAswlCrlfQlSAIQqCb0N0dDiTKOR+KdAWWCxpG/A4MDZ2ZWb2VbgW+qqkGsA6oBdwN9Ep5WuBt4vpz6PAX8Op5Bxg3h7Yx8lE14kXER09jzCz/5Z2JT179sQs2QF+xAPmnXOuqGpVcENAfaeY1/fGvP1ogvnPSbCa7cC14St23paFXr8JvFmobQ5Rsc5zS2ifQAiLD6+3AgOS70lyycLmLaqS17MLR7XOOed2X3W8huucc86lnRdc55xzLg284DrnnHNp4AXXOeecSwMvuM4551waeMF1zjnn0sALrnPOOZcGXnCdc865NPCC65xzzqWBF1znnHMuDbzgOuecc2mgkh5E76ovSd8Dy8u7HxVII+Dr8u5EBeLjEc/Ho0B1H4tDzeygwo3VKrzAldpyM+ta3p2oKCQt8PEo4OMRz8ejgI9FYn5K2TnnnEsDL7jOOedcGnjBdcUZV94dqGB8POL5eMTz8SjgY5GA3zTlnHPOpYEf4TrnnHNp4AXXOeecSwMvuK4ISadKWi7pX5JGlnd/yoOkXElLJOVIWhDaDpA0Q9KK8H3/8u5nWZE0XtI6SUtj2pLuv6T/F35elks6pXx6XXaSjMcoSavDz0iOpD4x71X18WguaZakTyV9Iumq0F5tf0ZS4QXXxZFUE3gYOA04Ahgo6Yjy7VW5OdHMMmM+TzgSmGlmbYCZ4XVVNQE4tVBbwv0PPx8DgI5hmUfCz1FVMoGi4wEwJvyMZJrZG1BtxmM7cJ2ZdQCOBS4P+12df0ZK5AXXFdYd+JeZfW5mPwHPA2eVc58qirOAiWF6InB2+XWlbJnZu8C3hZqT7f9ZwPNm9qOZrQT+RfRzVGUkGY9kqsN4rDWzj8L098CnQFOq8c9IKrzgusKaAqtiXv8ntFU3BkyXtFDSr0PbIWa2FqJfOMDB5da78pFs/6vzz8wVkhaHU855p0+r1XhIagkcCczFf0aK5QXXFaYEbdXxs2PHmdlRRKfWL5d0fHl3qAKrrj8zjwKHA5nAWuC+0F5txkNSfeAV4Goz+664WRO0VckxKY4XXFfYf4DmMa+bAWvKqS/lxszWhO/rgMlEp7++lNQYIHxfV349LBfJ9r9a/syY2ZdmtsPMdgKPU3CKtFqMh6TaRMX2WTN7NTT7z0gxvOC6wuYDbSQdJqkO0Y0OU8u5T2klaR9J++ZNA72BpUTjMCjMNgiYUj49LDfJ9n8qMEBSXUmHAW2AeeXQv7TKKyxBP6KfEagG4yFJwJPAp2b255i3/GekGJ4W5OKY2XZJVwBvATWB8Wb2STl3K90OASZHv1OoBTxnZtMkzQdelHQJ8G/g/HLsY5mSNAnIAhpJ+g9wKzCaBPtvZp9IehFYRnT36uVmtqNcOl5GkoxHlqRMolOjucBvoHqMB3Ac8CtgiaSc0HYj1fhnJBX+aEfnnHMuDfyUsnPOOZcGXnCdc865NPCC65xzzqWBF1znnHMuDbzgOuecc2ngHwtyzqWVpB3Akpims80st5y641za+MeCnHNpJWmTmdVP4/Zqmdn2dG3PuWT8lLJzrkKR1FjSuyFjdqmkn4f2UyV9JGmRpJmh7QBJr4UAgQ8lZYT2UZLGSZoOPCXpIEmvSJofvo4rx1101ZSfUnbOpVu9mKcTrTSzfoXevwB4y8z+FDJT95Z0ENHzio83s5WSDgjz3gZ8bGZnS/oF8BRRmADA0UBPM9sq6Tmi7Nr3JbUgepJahzLbQ+cS8ILrnEu3rWaWWcz784Hx4eH4r5lZjqQs4N2QpYqZ5WXT9gTODW1vSzpQUoPw3lQz2xqmTwaOCI/rBNhP0r4hy9W5tPCC65yrUMzs3RCHeDrwtKR7gA0kjnMrLvZtc0xbDaBHTAF2Lu38Gq5zrkKRdCiwzsweJ0qkOQqYA5wQkmaIOaX8LnBhaMsCvk6SyzoduCJmG5ll1H3nkvIjXOdcRZMFXC9pG7AJuNjMvpL0a+BVSTWIclZ7AaOAv0paDGyhIBqusCuBh8N8tYgK9bAy3QvnCvGPBTnnnHNp4KeUnXPOuTTwguucc86lgRdc55xzLg284DrnnHNp4AXXOeecSwMvuM4551waeMF1zjnn0uD/AxM/01y545KAAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb.plot_importance(model)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "028d0541",
   "metadata": {},
   "source": [
    "We can see what features turned out to be important after building the final model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d107e5e3",
   "metadata": {},
   "source": [
    "### Using the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "id": "e01cd538",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'geography': 'germany',\n",
       " 'gender': 'female',\n",
       " 'hascrcard': 'no',\n",
       " 'isactivemember': 'yes',\n",
       " 'creditscore': 554,\n",
       " 'age': 55,\n",
       " 'tenure': 0,\n",
       " 'balance': 108477.27,\n",
       " 'numofproducts': 1,\n",
       " 'estimatedsalary': 140003.0}"
      ]
     },
     "execution_count": 176,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "selected_user = 50\n",
    "customer = test_dict[selected_user]\n",
    "customer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "id": "941db55e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 177,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test[selected_user]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "id": "1a940b27",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_one_person = dv.transform([customer])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "id": "ca2b4f88",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtest = xgb.DMatrix(X_one_person, feature_names=dv.get_feature_names_out())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "id": "16055f9b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7474809"
      ]
     },
     "execution_count": 180,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exited_prob = model.predict(dtest)[0]\n",
    "exited_prob"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "id": "f454ef40",
   "metadata": {},
   "outputs": [],
   "source": [
    "customer_will_exit_bank = exited_prob >= 0.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "id": "0499d265",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 182,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customer_will_exit_bank.astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "id": "4388635a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Customer WILL exit the bank\n"
     ]
    }
   ],
   "source": [
    "if customer_will_exit_bank == 0:\n",
    "    print('Customer will NOT exit bank!')\n",
    "else:\n",
    "    print('Customer WILL exit the bank')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "id": "de2a665b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# test for a range for users\n",
    "def get_predictions(from_i, to_i):\n",
    "    if(from_i > to_i):\n",
    "        print('first argument must be greater then the first')\n",
    "        return\n",
    "    for i in range(from_i, to_i + 1):\n",
    "        print(\"================================\")\n",
    "        \n",
    "        customer = test_dict[i]\n",
    "        X_one_person = dv.transform([customer])\n",
    "        \n",
    "        dtest = xgb.DMatrix(X_one_person, feature_names=dv.get_feature_names_out())\n",
    "        payment_prob = model.predict(dtest)[0]\n",
    "        customer_will_default = payment_prob >= 0.5\n",
    "        \n",
    "        if customer_will_default == 0:\n",
    "            print('Predicted: Customer {} will NOT exit bank!'.format(i))\n",
    "        else:\n",
    "            print('Predicted: Customer {} will exit bank!'.format(i))\n",
    "        \n",
    "        if y_test[i] == 0:\n",
    "            print('The truth is: Customer {} will NOT exit bank!'.format(i))\n",
    "        else:\n",
    "            print('The truth is: Customer {} will exit bank!!'.format(i))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "id": "bdf41100",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'geography': 'germany',\n",
       " 'gender': 'female',\n",
       " 'hascrcard': 'no',\n",
       " 'isactivemember': 'yes',\n",
       " 'creditscore': 554,\n",
       " 'age': 55,\n",
       " 'tenure': 0,\n",
       " 'balance': 108477.27,\n",
       " 'numofproducts': 1,\n",
       " 'estimatedsalary': 140003.0}"
      ]
     },
     "execution_count": 187,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_dict[selected_user]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "id": "3b767a81",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "Predicted: Customer 40 will NOT exit bank!\n",
      "The truth is: Customer 40 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 41 will NOT exit bank!\n",
      "The truth is: Customer 41 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 42 will NOT exit bank!\n",
      "The truth is: Customer 42 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 43 will NOT exit bank!\n",
      "The truth is: Customer 43 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 44 will exit bank!\n",
      "The truth is: Customer 44 will exit bank!!\n",
      "================================\n",
      "Predicted: Customer 45 will NOT exit bank!\n",
      "The truth is: Customer 45 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 46 will NOT exit bank!\n",
      "The truth is: Customer 46 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 47 will NOT exit bank!\n",
      "The truth is: Customer 47 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 48 will NOT exit bank!\n",
      "The truth is: Customer 48 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 49 will NOT exit bank!\n",
      "The truth is: Customer 49 will NOT exit bank!\n",
      "================================\n",
      "Predicted: Customer 50 will exit bank!\n",
      "The truth is: Customer 50 will exit bank!!\n"
     ]
    }
   ],
   "source": [
    "get_predictions(40, 50)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9d3f3b7",
   "metadata": {},
   "source": [
    "### save the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a44bffc5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "727848eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "output_file = 'model_v0.bin'\n",
    "output_file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "92d9129d",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(output_file, 'wb') as f_out:\n",
    "    pickle.dump((dv, model), f_out)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9ade26a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_file = 'model_v0.bin'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "946c0138",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(model_file, 'rb') as f_in:\n",
    "    dv, model = pickle.load(f_in)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "321f54cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "dv, model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f5e951f5",
   "metadata": {},
   "outputs": [],
   "source": [
    "selected_user = 150\n",
    "customer = test_dict[selected_user]\n",
    "customer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "78f05618",
   "metadata": {},
   "outputs": [],
   "source": [
    "X= dv.transform([customer])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d8480578",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtest = xgb.DMatrix(X, feature_names=dv.get_feature_names_out())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "37d777da",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.predict(dtest)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6dec1d7f",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}