Predict_Case_Status_new.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <center> Predict Case Status </center>\n",
    "                                                                                            M R Abhishek"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "import random\n",
    "import datetime\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import keras\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import confusion_matrix, f1_score\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.cross_validation import train_test_split\n",
    "le = LabelEncoder()\n",
    "\n",
    "import xgboost as xgb\n",
    "import lightgbm as lgbm\n",
    "random.seed(20)\n",
    "\n",
    "pd.set_option('display.max_columns', 500)\n",
    "pd.set_option('display.max_rows', 200)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### PATH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [],
   "source": [
    "FOLDER = '../dataset/'\n",
    "TRAIN_PATH = FOLDER + 'train.csv'\n",
    "TEST_PATH = FOLDER + 'test.csv'\n",
    "SAMPLE_SUBMISSION_PATH = FOLDER + 'sample_submission.csv'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = pd.read_csv(TRAIN_PATH)\n",
    "test_data = pd.read_csv(TEST_PATH)\n",
    "sample_submission = pd.read_csv(SAMPLE_SUBMISSION_PATH)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "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>CASE_NO</th>\n",
       "      <th>CASE_SUBMITTED_DAY</th>\n",
       "      <th>CASE_SUBMITTED_MONTH</th>\n",
       "      <th>CASE_SUBMITTED_YEAR</th>\n",
       "      <th>DECISION_DAY</th>\n",
       "      <th>DECISION_MONTH</th>\n",
       "      <th>DECISION_YEAR</th>\n",
       "      <th>VISA_CLASS</th>\n",
       "      <th>EMPLOYER_NAME</th>\n",
       "      <th>EMPLOYER_STATE</th>\n",
       "      <th>EMPLOYER_COUNTRY</th>\n",
       "      <th>SOC_NAME</th>\n",
       "      <th>NAICS_CODE</th>\n",
       "      <th>TOTAL_WORKERS</th>\n",
       "      <th>FULL_TIME_POSITION</th>\n",
       "      <th>PREVAILING_WAGE</th>\n",
       "      <th>PW_UNIT_OF_PAY</th>\n",
       "      <th>PW_SOURCE</th>\n",
       "      <th>PW_SOURCE_YEAR</th>\n",
       "      <th>PW_SOURCE_OTHER</th>\n",
       "      <th>WAGE_RATE_OF_PAY_FROM</th>\n",
       "      <th>WAGE_RATE_OF_PAY_TO</th>\n",
       "      <th>WAGE_UNIT_OF_PAY</th>\n",
       "      <th>H-1B_DEPENDENT</th>\n",
       "      <th>WILLFUL_VIOLATOR</th>\n",
       "      <th>WORKSITE_STATE</th>\n",
       "      <th>WORKSITE_POSTAL_CODE</th>\n",
       "      <th>CASE_STATUS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>9831643</td>\n",
       "      <td>23.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2011</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>2017</td>\n",
       "      <td>H1B</td>\n",
       "      <td>XTRON SOFTWARE SERVICES INC</td>\n",
       "      <td>CA</td>\n",
       "      <td>UNITED STATES OF AMERICA</td>\n",
       "      <td>ANALYSTS</td>\n",
       "      <td>541511.0</td>\n",
       "      <td>5</td>\n",
       "      <td>Y</td>\n",
       "      <td>70907.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>OES</td>\n",
       "      <td>2011.0</td>\n",
       "      <td>OFLC ONLINE DATA CENTER</td>\n",
       "      <td>71000.0</td>\n",
       "      <td>91000.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>CA</td>\n",
       "      <td>95054</td>\n",
       "      <td>CERTIFIEDWITHDRAWN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5075895</td>\n",
       "      <td>28.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2011</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "      <td>2017</td>\n",
       "      <td>H1B</td>\n",
       "      <td>GLOBAL DISTRIBUTORS INC</td>\n",
       "      <td>MD</td>\n",
       "      <td>UNITED STATES OF AMERICA</td>\n",
       "      <td>ACCOUNTANTS</td>\n",
       "      <td>423210.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Y</td>\n",
       "      <td>50877.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>OES</td>\n",
       "      <td>2010.0</td>\n",
       "      <td>OFLC ONLINE DATA CENTER</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>MD</td>\n",
       "      <td>20814</td>\n",
       "      <td>CERTIFIEDWITHDRAWN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4040912</td>\n",
       "      <td>17.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2012</td>\n",
       "      <td>18</td>\n",
       "      <td>10</td>\n",
       "      <td>2016</td>\n",
       "      <td>H1B</td>\n",
       "      <td>MARSHALL WEALTH MANAGEMENT LLC</td>\n",
       "      <td>KY</td>\n",
       "      <td>UNITED STATES OF AMERICA</td>\n",
       "      <td>ACTUARIES</td>\n",
       "      <td>523930.0</td>\n",
       "      <td>1</td>\n",
       "      <td>Y</td>\n",
       "      <td>28454.4</td>\n",
       "      <td>Year</td>\n",
       "      <td>OES</td>\n",
       "      <td>2011.0</td>\n",
       "      <td>OFLC ONLINE DATA CENTER</td>\n",
       "      <td>28455.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>KY</td>\n",
       "      <td>40202</td>\n",
       "      <td>WITHDRAWN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4491642</td>\n",
       "      <td>22.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2012</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>2017</td>\n",
       "      <td>H1B</td>\n",
       "      <td>XTRON SOFTWARE SERVICES INC</td>\n",
       "      <td>CA</td>\n",
       "      <td>UNITED STATES OF AMERICA</td>\n",
       "      <td>ANALYSTS</td>\n",
       "      <td>541511.0</td>\n",
       "      <td>5</td>\n",
       "      <td>Y</td>\n",
       "      <td>61963.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>OES</td>\n",
       "      <td>2012.0</td>\n",
       "      <td>OFLC ONLINE DATA CENTER</td>\n",
       "      <td>62000.0</td>\n",
       "      <td>82000.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>CA</td>\n",
       "      <td>95054</td>\n",
       "      <td>CERTIFIEDWITHDRAWN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3332446</td>\n",
       "      <td>22.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2012</td>\n",
       "      <td>14</td>\n",
       "      <td>4</td>\n",
       "      <td>2017</td>\n",
       "      <td>H1B</td>\n",
       "      <td>XTRON SOFTWARE SERVICES INC</td>\n",
       "      <td>CA</td>\n",
       "      <td>UNITED STATES OF AMERICA</td>\n",
       "      <td>ANALYSTS</td>\n",
       "      <td>541511.0</td>\n",
       "      <td>5</td>\n",
       "      <td>Y</td>\n",
       "      <td>71947.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>OES</td>\n",
       "      <td>2012.0</td>\n",
       "      <td>OFLC ONLINE DATA CENTER</td>\n",
       "      <td>72000.0</td>\n",
       "      <td>92000.0</td>\n",
       "      <td>Year</td>\n",
       "      <td>Y</td>\n",
       "      <td>N</td>\n",
       "      <td>CA</td>\n",
       "      <td>95054</td>\n",
       "      <td>CERTIFIEDWITHDRAWN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   CASE_NO  CASE_SUBMITTED_DAY  CASE_SUBMITTED_MONTH  CASE_SUBMITTED_YEAR  \\\n",
       "0  9831643                23.0                     3                 2011   \n",
       "1  5075895                28.0                     3                 2011   \n",
       "2  4040912                17.0                     2                 2012   \n",
       "3  4491642                22.0                     3                 2012   \n",
       "4  3332446                22.0                     3                 2012   \n",
       "\n",
       "   DECISION_DAY  DECISION_MONTH  DECISION_YEAR VISA_CLASS  \\\n",
       "0            14               4           2017        H1B   \n",
       "1            10               3           2017        H1B   \n",
       "2            18              10           2016        H1B   \n",
       "3            14               4           2017        H1B   \n",
       "4            14               4           2017        H1B   \n",
       "\n",
       "                    EMPLOYER_NAME EMPLOYER_STATE          EMPLOYER_COUNTRY  \\\n",
       "0     XTRON SOFTWARE SERVICES INC             CA  UNITED STATES OF AMERICA   \n",
       "1         GLOBAL DISTRIBUTORS INC             MD  UNITED STATES OF AMERICA   \n",
       "2  MARSHALL WEALTH MANAGEMENT LLC             KY  UNITED STATES OF AMERICA   \n",
       "3     XTRON SOFTWARE SERVICES INC             CA  UNITED STATES OF AMERICA   \n",
       "4     XTRON SOFTWARE SERVICES INC             CA  UNITED STATES OF AMERICA   \n",
       "\n",
       "      SOC_NAME  NAICS_CODE  TOTAL_WORKERS FULL_TIME_POSITION  PREVAILING_WAGE  \\\n",
       "0     ANALYSTS    541511.0              5                  Y          70907.0   \n",
       "1  ACCOUNTANTS    423210.0              1                  Y          50877.0   \n",
       "2    ACTUARIES    523930.0              1                  Y          28454.4   \n",
       "3     ANALYSTS    541511.0              5                  Y          61963.0   \n",
       "4     ANALYSTS    541511.0              5                  Y          71947.0   \n",
       "\n",
       "  PW_UNIT_OF_PAY PW_SOURCE  PW_SOURCE_YEAR          PW_SOURCE_OTHER  \\\n",
       "0           Year       OES          2011.0  OFLC ONLINE DATA CENTER   \n",
       "1           Year       OES          2010.0  OFLC ONLINE DATA CENTER   \n",
       "2           Year       OES          2011.0  OFLC ONLINE DATA CENTER   \n",
       "3           Year       OES          2012.0  OFLC ONLINE DATA CENTER   \n",
       "4           Year       OES          2012.0  OFLC ONLINE DATA CENTER   \n",
       "\n",
       "   WAGE_RATE_OF_PAY_FROM  WAGE_RATE_OF_PAY_TO WAGE_UNIT_OF_PAY H-1B_DEPENDENT  \\\n",
       "0                71000.0              91000.0             Year              Y   \n",
       "1                51000.0                  0.0             Year              N   \n",
       "2                28455.0                  0.0             Year              N   \n",
       "3                62000.0              82000.0             Year              Y   \n",
       "4                72000.0              92000.0             Year              Y   \n",
       "\n",
       "  WILLFUL_VIOLATOR WORKSITE_STATE WORKSITE_POSTAL_CODE         CASE_STATUS  \n",
       "0                N             CA                95054  CERTIFIEDWITHDRAWN  \n",
       "1                N             MD                20814  CERTIFIEDWITHDRAWN  \n",
       "2                N             KY                40202           WITHDRAWN  \n",
       "3                N             CA                95054  CERTIFIEDWITHDRAWN  \n",
       "4                N             CA                95054  CERTIFIEDWITHDRAWN  "
      ]
     },
     "execution_count": 181,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <center> Exploratory Data Analysis </center>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Errors found in the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''\n",
    "PW_SOURCE_YEAR has year information from 2007 - 2017 and 2001. One of the record has a value of ‘1’, which is \n",
    "not an year. Hence, I replaced that value with the mode of all the values in that feature.\n",
    "'''\n",
    "\n",
    "year_data = train_data['PW_SOURCE_YEAR'].value_counts()\n",
    "year_data = pd.DataFrame({'Year': year_data.index, 'Count': year_data.values})\n",
    "\n",
    "plt.figure(figsize=(10,5))\n",
    "g = sns.barplot(x='Year', y='Count', data=year_data, capsize=0.9, color='blue') # #1f77b4\n",
    "g.set_title(\"Number of applications per year\", fontsize=18)\n",
    "g.set_xlabel(\"Year\", fontsize=15)\n",
    "g.set_ylabel(\"Count\", fontsize=15)\n",
    "g.set_xticklabels(g.get_xticklabels(), rotation=90, fontsize=13)\n",
    "g.set_yticks([0, 50000, 100000, 150000, 200000, 250000, 300000, 350000, 400000])\n",
    "for p in g.patches:\n",
    "    g.annotate(np.int(p.get_height()), \n",
    "                (p.get_x()+p.get_width()/2., p.get_height()), \n",
    "                ha='center', va='center', xytext=(0, 10), \n",
    "                textcoords='offset points', fontsize=14, color='black')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "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>WAGE_RATE_OF_PAY_FROM</th>\n",
       "      <th>WAGE_RATE_OF_PAY_TO</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>71000.0</td>\n",
       "      <td>91000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>51000.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28455.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>62000.0</td>\n",
       "      <td>82000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>72000.0</td>\n",
       "      <td>92000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   WAGE_RATE_OF_PAY_FROM  WAGE_RATE_OF_PAY_TO\n",
       "0                71000.0              91000.0\n",
       "1                51000.0                  0.0\n",
       "2                28455.0                  0.0\n",
       "3                62000.0              82000.0\n",
       "4                72000.0              92000.0"
      ]
     },
     "execution_count": 183,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''\n",
    "It is obvious that WAGE_RATE_OF_PAY_FROM must be greater than WAGE_RATE_OF_PAY_TO. Some of the records in \n",
    "WAGE_RATE_OF_PAY_TO contains the value zero(0). So, I replaced those values with the values in \n",
    "WAGE_RATE_OF_PAY_FROM.\n",
    "'''\n",
    "\n",
    "train_data[['WAGE_RATE_OF_PAY_FROM', 'WAGE_RATE_OF_PAY_TO']].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "null_values = train_data.isnull().sum()\n",
    "null_values = pd.DataFrame({'Feature': null_values.index, 'Count': null_values.values})\n",
    "\n",
    "plt.figure(figsize=(15,5))\n",
    "g = sns.barplot(x='Feature', y='Count', data=null_values, capsize=0.9, color='blue') # #1f77b4\n",
    "g.set_title(\"Missing values across different columns\", fontsize=18)\n",
    "g.set_xlabel(\"Feature\", fontsize=15)\n",
    "g.set_ylabel(\"Count\", fontsize=15)\n",
    "g.set_xticklabels(g.get_xticklabels(), rotation=90, fontsize=13)\n",
    "g.set_yticks([0, 1000, 2000, 3000, 4000, 5000, 6000, 7000])\n",
    "for p in g.patches:\n",
    "    g.annotate(np.int(p.get_height()), \n",
    "                (p.get_x()+p.get_width()/2., p.get_height()), \n",
    "                ha='center', va='center', xytext=(0, 10), \n",
    "                textcoords='offset points', fontsize=11, color='black')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imbalance dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdgAAAFmCAYAAADOGYucAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xm8VVXdx/HPV64MCggqAgIF5oCahnIzLCtCQRwSfRweLRXNeSjtsdJSg8pyyKJyqDAVyQzNchYVUdJKxKuSiGYiOKCAICBeBxD4PX/sdS/nHu4Id98Bvu/Xa7/OOWuvvfbae597f2ftvfZeigjMzMyscW3S3BUwMzPbEDnAmpmZ5cAB1szMLAcOsGZmZjlwgDUzM8uBA6yZmVkOHGCtwSSdICkkLZXUtWheSZo3uhnqNTqtu6Sp190QkjaR9CtJ8yStlnRnA5cfnLZz8Dqse7SkIQ1dLi+Spkiasg7LdUnbsmcO1WqRJI2T9Gpz18PqzwHW1scWwPnNXYlW6AjgHODnwBeA7zXhukcBLSbAAmemqaG6kG3LRhNgrfVp0b/0rcV7CPimpDERsaC5K9MUJLWLiOXrWczO6fVXEbF6fevUmkXEC81dh6bSSN8da0XcgrX1cUl6vai2TBWnbqtJr3LKS1LfdOrzdEmXSpov6T1JN0vaTNL2kh6UVC5plqSRNaxyZ0mPSvognYb9saQq33VJ3ST9TtKbkpZL+o+kU4vyVJwK/5Kkv0haCjxZx7YOl/SEpA8lvSvpTkk7Fcx/FRidPq5K5Z9QS3ndJN0iaVk6JT+erPVWnG+YpPvT9n4g6XlJ50lqU5Cn4hhcmNZbeSpf0mcl3S5pbqr7S5J+JqlD0XqmSPqHpBFpHRX77qiG7ouC8qYUfK44/X2IpKslLUrTzZK6pDx9gTlpkesKtuWENH9/Sf9K6yxP2/LDmvZxWqbwWN+ZlntH0jXV7IPNJF0uaY6kFen1wsLvWMF2/I+k6yQtBGr9ESqpn6Q/pu/9ckmzJf26jmV+JOmZ9P1YJOkRSYOK8nSUdJWk11O5b0t6WFL/gjznSHoxHaslksokHVbbuq1ubsHa+pgHXA2cK+nKiHitkcr9PjAFGAnsAlwBrAb2AK4DrgTOAG6UVBYRM4uWvxO4AbgU2B+4OC0/GkBSZ+AfQIeUNifl+62yVsZVReX9Cfgz2andGv9mJA0H7gMeAf4X6Aj8GPiHpAER8SZwGPAt4ARg77ToK7Xsi78BnwF+ALycyi2uH8B2wOQ07yOgNG1bN+CClGdv4AlgHPD7lDY3vX4CmJ7mvQfsCvwwlXt00bq2B36Tyn+b7FhMkLQwIh5twL6oza+Be4GvATuRfQdWkX0n5gH/k/bNpcDdaZlXJG2XPt+e1rcC2CFtR33cDNwGXAvslfbB5mTHC2XX9x8k+17+BJgBDCL7jm0JnFdU3lXAROA4oH1NK5XUD5gGfJDW+TLZMRlWR317AWPIjuPmwLHAY5IGRsSMlGcMcAhrvkNbkV2aqPjB8nXgF2T763Gyv4vd0/bY+ogIT54aNJH9swmyf7RbAkuBG9K8kjRvdEH+0dlXba1yxgGvFnzum5Z9pCjf31L6sQVpXYGVwKji9QAXFC1/HVnQ6JI+X0wWhHaoJt8ioKRoO8fUc7+Ukf0DKylI6wd8DPyyIO2S6vZHNeUNTes/uih9YkofXMNySsfhQmAJsEnBvAAuqWO9FcsfS/bDZKuCeVNSGYMK0toA/wEeX4d9MQWYUvB5cCr/pqI6XZ2OmYq+KycX5TsipXdex+/074rSLyQL7Dumz8elfF+qJt8KYJui7bijnusfD5QD29aSZxwFfy/VzG+TjttLwK8L0p8v3OfVLHc18ExD9pen+k0+RWzrJSIWk/36Pb749N96mFj0+T/p9cGC9S4haz31qWb524o+TyBrQX06fR5Odqp3jrJezyUFLZOtyFonhe6oq8KSNifrcHNrRKwsqOcc4J/Al+sqoxp7k/1z/2tR+oRq1t9T0u8lvUb2j/5jskDeBdimHvXvnE57vgIsT8v/kSzY7lCU/Y2ImFrxISJWAX8B9lLWQ7ox9sV9RZ9nAO2A7nUsNz3VfYKkIyTVue1FqvvubELWmoXsu/Ma8K+i785DwKZkrdlCdX53kmHAvRHxVkMqK2k/ZZdD3iH7wfkxsCNZq7/CU8AJkn4gqVQFlw0K5g9Ip5H3k7RZQ+pgNXOAtcYwBlhMdoqpMSwp+ryilvTqTrsVX+uq+NwrvW4DfInsn1Hh9Jc0f6ui5efVXWW6kgWj6vLOZ91Ot/UElkTEx0XpVbYvXfu7GziYLKgOAT4L/DRlqfHUZIEbgdPJTv0OTcufVcPy1V1LXAC0JTsl3Rj7YnHR54rOQbVuS0TMIjvdvwnZD4T5kqZKqu8PnPp8dz7J2t+daWn+unx3KpabW2euAspuUbqfrOV7Ellw/yzwb6rup2+SXRL4BlkwfVvSmIJAOp7sNP/nyH5kLpb0t3St29aDr8HaeouIckmXkrVkf15Nlo8AJLWNiBUF6cX/jBpLd2B20WeAiut+75C1fs+pYfmXij7XZ0zHJSlfj2rm9WDtgFEf84CukjYtCrLFrbhPkV1zPS4ibq5IlPTV+qxEUntgBNlp/V8XpO9WwyLVtSK7k/3gWUh2Da+x90W9RXYd+FFJ7ciuNf4YuE9S34hYVMfi3YGZRZ+h6ndnDrBWp67k1eLq1LPai1gTxOvrcLJW6/8Ufj+U3Zu+tLICEeVk/Rq+L+mTZKfRLyM7XudHdp7498Dv07LDyP6WbyULuraO3IK1xnIt2T+hS6qZV9H5qeIULalH6OdzqkvxP7+jyX7lV3T6eADoD7weEWXVTO81dIUR8T7wNHCkqvbc/STZdk5Zh+14guy62uFF6cWdjipaIoX/ZDcFvl5NmSvIAmChdmk9xS3lE2qoV5/Cnqppe48EpkXE6pz2RbGKFm3xtlSKiOUR8QhZB6nNya4B16W6785q1vQef4DsskR5Dd+dugJ4TR4CDpbUswHLbEZ2CaEyiCt7iMgnalogIl6LiF+Q/S18upr5SyLiVrJT5WvNt4ZxC9YaRUQsl/RjYGw1sycC75LdUjGK7B/698iCXh5OSadNnyI7XXgyWevs3TR/DFnP1scljSFrsW5OFnS/GBEj1nG9F5NdO7xX0rVk131/RLbtv2hoYRExSdI/yFoWW7OmF3HxP74XyX7E/FTSKrJA+e0ain0BOEjSA2St7rci4i1JU4HzJM0ja019g5pbVAuAW9OxXEh2enHH9FqhUfdFDXV4Bzha0nPA+2QtyyPJTv/fD7wBbE3WenuLrLNPXQ6U9HOygLcX2cMsxkfEy2n+n4ATgcmSfkF2OrYt2VmEQ4BDI+KDddieUcCBZNd2fwbMItv/wyPi2BqWeQA4Fxgn6UayY3Axa1rbAEh6guwSwgyyv7kvk/VMvynNH0vWCfAJsjM7O5J15npoHbbDCjV3LytPrW+ioBdxUXoJ8F+KehGnefuQBbwPUp5jqbkXcXHP0NEpvaQo/VXg5mryfRp4FPiQ7JrfTyjoSZvydiULtHPIWnVvk92icG5d21nHvhlO9o/qQ7JgchewU1GeevUiTnm7kd0i9B7Zab/xZKdzq/QiBgaQ3Xr0Adm1vB+T/bAIoG9Bvi+QtS4/KjxOad9PTOt5m6xn6UHVrGdKWs8hZAFrOdkPlP9dx30xhep7Ee9Xw3eucFsOJfvB8HGadwJZx7C7yILrcrLT7H8pXm8t3+kvpeXLyU5lXwN0KMrbPn3X/pPWsZjsuz2aNT3Qq92OOurwqXSsF6Xj8wpVe1yPo6gXMdn11TlpHz8F7FfNPr0ceDYdg/fJAu23CuaPTMu8nbZnDtnfRoN6Yntae6ro8m5mVidlD4UoiYh9mrsujUnZQypuJLt1a1YzV8c2EL4Ga2ZmloNmCbCS2kh6VtK96XM/SU8qe/zdrZLapvR26fOsNL9vQRnfT+kvSdq/IH14Spsl6YKC9GrXYWZmlofmasGeQ9Yxo8LlZE/L2Z6s48VJKf0ksvsAtye7JnA5gKRdyHr37Up2nefaFLTbkF0zOYDsYQHHpLy1rcPM6ikiBm9op4cBImJcRMinh60xNXmAldSbrPPEH9Jnkd0Yf3vKchNZ5wXIOnPclN7fDuyb8o8AJkTWDX8OWY+7vdI0KyJmR3a/5QRgRB3rMDMza3TN0YL9FdktGhXDdG0FLI01j1Sby5rbA3qR9QYkzX835a9ML1qmpvTa1mFmZtbomvQ+WEkHA29HxNOSBjfluutL2ZBlpwJsvvnmA/v371/HEmZmtjF5+umnF0VEt7ryNfWDJr4AHCLpQLJ7yTqTDUvVRVJJamH2Zs2N0m+SPTVlbnqg9hZkN5dXpFcoXKa69HdqWUcVETGW9LCE0tLSKCsrW78tNjOzDUoaVKNOTXqKOCK+HxG9I6IvWSelRyLi62QPBTgiZRtJdqM3ZE8fGZneH5HyR0o/OvUy7kc22sc0shutd0g9htumddydlqlpHWZmZo2updwHez7wf5JmkV0vvT6lXw9sldL/jzRwdGQDbN9G9hSXB4CzImJVap2eTTYixIvAbbFmMO6a1mFmZtbomi3ARsSUiDg4vZ8dEXtFxPYRcWRELE/pH6XP26f5swuW/2lEfCoidoqIiQXp90fEjmneTwvSq11Hc7nmmmvYfffd6dy5M507d2bvvffmvvvWDIEpqdrprLPOqswTEYwePZptt92WDh06MHjwYGbOnLnWuh588EH23ntvNttsM7p06cKQIUOqrdOiRYvo1asXkli0aM0zyz/66CNOOOEEdt99dzbddFMGDx7ceDvCzGwD1VJasBud3r17c/nll/PMM89QVlbGkCFDOPTQQ3nuuecAmDdvXpXpnnvuAeCoo9YM9nHFFVfwi1/8gquuuoqnnnqKbbbZhqFDh/Lee2sGg7nzzjs5+uijOe6443j22Wd54oknOOmk6m8BPvHEExkwYMBa6atWraJ9+/acffbZHHTQQY25G8zMNlzN/TDkljwNHDgwmlLXrl3jd7/7XbXzTj755Nhxxx0rP69evTp69OgRl1xySWXaBx98EB07dqwsY+XKldGnT58YO3Zsnev+1a9+FUOGDInJkycHEAsXLqw231lnnRVf/vKXG7BVZmYbFqAs6hFD3IJtAVatWsWECRMoLy/n859fe4jU8vJyJkyYwCmnnFKZNmfOHObPn8+wYcMq0zp06MCXvvQl/vWvfwHw9NNP88Ybb9C2bVv23HNPevTowbBhw3j22WerlP/ss89y+eWXM378eDbZxF8JM7PG4P+mzWjGjBl07NiRdu3acfrpp3PHHXew2267rZXvlltuYcWKFYwcObIybf78+QB07969St7u3btXzps9O7tk/cMf/pAf/OAH3HffffTu3ZvBgwczb948AN5//32OPvporrrqKnr18rM3zMwaiwNsM9ppp52YPn06Tz75JGeccQYjR47k+efXHhP6uuuuY8SIEXTrVud9zVWsXp09LOvCCy/kiCOOYODAgYwdO5YtttiC8ePHA/Ctb32LffbZh8MPP3z9N8jMzCo5wDajtm3bsv322zNw4EAuvfRSBgwYwJgxY6rkmT59OmVlZVVODwP06NEDgAULFlRJX7BgQeW8nj17ArDLLrtUzi8pKWGHHXbg9ddfB2Dy5MmMGzeOkpISSkpK2HfffSvLv/DCCxtxa83MNi4OsC3I6tWrWb686t1DY8eOpV+/fuy3335V0vv160ePHj2YNGlSZdpHH33E448/Xnkdd+DAgbRr146XXnqpyjpeeeUVPvnJTwLw0EMP8e9//5vp06czffp0/vCHPwAwZcoUvvWtb+WynWZmG4OmflSiJRdccAEHHXQQffr04b333uOWW25hypQpVe6F/eCDD/jTn/7E9773PbIBgdaQxLnnnsvPfvYz+vfvz4477sgll1xCx44d+drXvgZA586dOf300xk1ahS9e/emb9++XH311SxZsoTjjjsOgB133LFKuRX3v/bv35+tt966Mv2FF15gxYoVLFq0iPLycqZPnw5Q7W09ZmbmANts5s+fz7HHHsv8+fPZYost2H333Zk4cSL77185djy33nor77//PieeeGK1ZXzve9/jww8/5KyzzmLJkiV87nOf46GHHqJTp06VeX7+85/Ttm1bRo4cyQcffMCee+7Jo48+Wnn6uL4OPPBAXnttzeM399hjDyC7zcvMzNYm/4OsmR/2b2ZmxSQ9HRGldeXzNVgzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4Fv02kE/fq92txVaJXmzOnb3FUwM8uNW7BmZmY5cIA1MzPLgQOsmZlZDhxgzczMcuAAa2ZmlgMHWDMzsxw4wJqZmeXAAdbMzCwHDrBmZmY5aNIAK6m9pGmS/i1ppqQfpfRxkuZImp6mASldkn4jaZak5yTtWVDWSEkvp2lkQfpASTPSMr+RpJS+paRJKf8kSV2bctvNzGzj0tQt2OXAkIj4DDAAGC5pUJr33YgYkKbpKe0AYIc0nQr8FrJgCYwCPgfsBYwqCJi/BU4pWG54Sr8AmBwROwCT02czM7NcNGmAjUx5+rhpmqKWRUYA49NyU4EuknoC+wOTImJxRCwBJpEF655A54iYGhEBjAcOLSjrpvT+poJ0MzOzRtfk12AltZE0HXibLEg+mWb9NJ0GHiOpXUrrBbxRsPjclFZb+txq0gG6R8S89H4+0L2xtsnMzKxYkwfYiFgVEQOA3sBekj4NfB/oD3wW2BI4P+c6BDW0nCWdKqlMUtnChQvzrIaZmW3Amq0XcUQsBR4FhkfEvHQaeDlwI9l1VYA3gT4Fi/VOabWl964mHWBBOoVMen27hnqNjYjSiCjt1q3b+myimZltxJq6F3E3SV3S+w7AUOA/BYFPZNdGn0+L3A0cn3oTDwLeTad5HwSGSeqaOjcNAx5M85ZJGpTKOh64q6Csit7GIwvSzczMGl1TD7jeE7hJUhuy4H5bRNwr6RFJ3QAB04HTU/77gQOBWcAHwIkAEbFY0k+Ap1K+H0fE4vT+TGAc0AGYmCaAy4DbJJ0EvAYcldtWmpnZRk/Z5UirTmlpaZSVldWZr1+/V/OvzAZozpy+zV0FM7MGk/R0RJTWlc9PcjIzM8uBA6yZmVkOHGDNzMxy4ABrZmaWAwdYMzOzHDjAmpmZ5cAB1szMLAcOsGZmZjlwgDUzM8uBA6yZmVkOHGDNzMxy4ABrZmaWAwdYMzOzHDjAmpmZ5cAB1szMLAcOsGZmZjlwgDUzM8uBA6yZmVkOHGDNzMxy4ABrZmaWAwdYMzOzHDjAmpmZ5cAB1szMLAcOsGZmZjlwgDUzM8tBkwZYSe0lTZP0b0kzJf0opfeT9KSkWZJuldQ2pbdLn2el+X0Lyvp+Sn9J0v4F6cNT2ixJFxSkV7sOMzOzPDR1C3Y5MCQiPgMMAIZLGgRcDoyJiO2BJcBJKf9JwJKUPiblQ9IuwNHArsBw4FpJbSS1Aa4BDgB2AY5JeallHWZmZo2uSQNsZMrTx03TFMAQ4PaUfhNwaHo/In0mzd9XklL6hIhYHhFzgFnAXmmaFRGzI2IFMAEYkZapaR1mZmaNrsmvwaaW5nTgbWAS8AqwNCJWpixzgV7pfS/gDYA0/11gq8L0omVqSt+qlnWYmZk1uiYPsBGxKiIGAL3JWpz9m7oOtZF0qqQySWULFy5s7uqYmVkr1Wy9iCNiKfAosDfQRVJJmtUbeDO9fxPoA5DmbwG8U5hetExN6e/Uso7ieo2NiNKIKO3Wrdt6baOZmW28mroXcTdJXdL7DsBQ4EWyQHtEyjYSuCu9vzt9Js1/JCIipR+dehn3A3YApgFPATukHsNtyTpC3Z2WqWkdZmZmja6k7iyNqidwU+rtuwlwW0TcK+kFYIKkS4BngetT/uuBP0qaBSwmC5hExExJtwEvACuBsyJiFYCks4EHgTbADRExM5V1fg3rMDMza3TKGndWndLS0igrK6szX79+r+ZfmQ3QnDl9m7sKZmYNJunpiCitK5+f5GRmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpaDJg2wkvpIelTSC5JmSjonpY+W9Kak6Wk6sGCZ70uaJeklSfsXpA9PabMkXVCQ3k/Skyn9VkltU3q79HlWmt+36bbczMw2Nk3dgl0JnBcRuwCDgLMk7ZLmjYmIAWm6HyDNOxrYFRgOXCupjaQ2wDXAAcAuwDEF5VyeytoeWAKclNJPApak9DEpn5mZWS6aNMBGxLyIeCa9fw94EehVyyIjgAkRsTwi5gCzgL3SNCsiZkfECmACMEKSgCHA7Wn5m4BDC8q6Kb2/Hdg35TczM2t0zXYNNp2i3QN4MiWdLek5STdI6prSegFvFCw2N6XVlL4VsDQiVhalVykrzX835TczM2t0zRJgJXUE/gqcGxHLgN8CnwIGAPOAXzRHvVLdTpVUJqls4cKFzVUNMzNr5Zo8wEralCy4/iki/gYQEQsiYlVErAauIzsFDPAm0Kdg8d4prab0d4AukkqK0quUleZvkfJXERFjI6I0Ikq7deu2vptrZmYbqabuRSzgeuDFiPhlQXrPgmyHAc+n93cDR6cewP2AHYBpwFPADqnHcFuyjlB3R0QAjwJHpOVHAncVlDUyvT8CeCTlNzMza3QldWdpVF8AjgNmSJqe0n5A1gt4ABDAq8BpABExU9JtwAtkPZDPiohVAJLOBh4E2gA3RMTMVN75wARJlwDPkgV00usfJc0CFpMFZTMzs1zIjbialZaWRllZWZ35+vV7Nf/KbIDmzOnb3FUwM2swSU9HRGld+fwkJzMzsxw4wJqZmeXAAdbMzCwHDrBmZmY5cIA1MzPLgQOsmZlZDhxgzczMcuAAa2ZmloN6B1hJx0uqdvQZSVtKOr7xqmVmZta6NaQFeyPZiDfV6Zfmm5mZGQ0LsLUNTr4VsGw962JmZrbBqPVh/5JGACMKki6WVDxIanvgi2Qj3JiZmRl1j6azDbBbwedPAT2K8qwAHgIuacR6mZmZtWq1BtiIuI5sAHQkPQqcERH/aYqKmZmZtWb1Hg82Ir6SZ0XMzMw2JA0acF3StsDBQG+ya6+FIiLOb6yKmZmZtWb1DrCSDgP+DLQB3ia79looAAdYMzMzGtaC/RlZZ6YTImJxTvUxMzPbIDQkwPYBvungamZmVreGPGjiX8BOeVXEzMxsQ9KQFuz/AX+SVA5MApYWZ4iIDxqrYmZmZq1ZQwLsc+n1RrIOTdVps37VMTMz2zA0JMB+g5oDq5mZmRVoyIMmxuVYDzMzsw2KB1w3MzPLQUMGXF8o6e3apnqU0UfSo5JekDRT0jkpfUtJkyS9nF67pnRJ+o2kWZKek7RnQVkjU/6XJY0sSB8oaUZa5jeSVNs6zMzM8tCQFuw11Uy3AAuAD9PnuqwEzouIXYBBwFmSdgEuACZHxA7A5PQZ4ABghzSdCvwWsmAJjAI+B+wFjCoImL8FTilYbnhKr2kdZmZmja4h12BHV5eeWoi3AR/Xo4x5wLz0/j1JLwK9yMacHZyy3QRMIXvs4ghgfEQEMFVSF0k9U95JFQ+9kDQJGC5pCtA5Iqam9PHAocDEWtZhZmbW6Nb7GmwKfn8Azm7IcpL6AnsATwLdU/AFmA90T+97AW8ULDY3pdWWPreadGpZh5mZWaNrrE5O2wFt65tZUkfgr8C5EbGscF4K2LneDlTbOiSdKqlMUtnChQvzrIaZmW3AGjKazpnVJLcFdga+DvylnuVsShZc/xQRf0vJCyT1jIh56RRwRYepN8megVyhd0p7kzWneyvSp6T03tXkr20dVUTEWGAsQGlpqe/7NTOzddKQB01cXU3acrLTsNcCP6qrgHS99nrgxYj4ZcGsu4GRwGXp9a6C9LMlTSDr0PRuCpAPAj8r6Ng0DPh+RCyWtEzSILJTz8cDV9WxDjMzs0bXkE5OjXE6+QvAccAMSdNT2g/Igt5tkk4CXgOOSvPuBw4EZgEfACemuiyW9BPgqZTvxwWj/JwJjAM6kHVumpjSa1qHmZlZo1N2OdKqU1paGmVlZXXm69fv1fwrswGaM6dvc1fBzKzBJD0dEaV15WtQq1TSdpJ+mx7k8GZ6vVbSduteVTMzsw1PQzo5DQQeBT4C7iV7wER34HDg65K+EhHP5FJLMzOzVqYhnZyuBJ4FDigc91XSZmTXSq8EhjRu9czMzFqnhpwi3gu4onhQ9fT5SrJevmZmZkbDAuyHwFY1zNuS7NSxmZmZ0bAAex9wmaR9ChPT50uBexqzYmZmZq1ZQ67B/h/Zwxn+noamexvYhqyj07+A8xq/emZmZq1TQx408Q6wj6ThwGeBnmQj4zwZEQ/lVD8zM7NWqdZTxJJ6SvqrpP0r0iLigYj4SUScGRE/ybLpr5K2yb22ZmZmrURd12C/QzZSTm0t1IeAfvgUsZmZWaW6AuzBwO+ilucppnm/JxvQ3MzMzKg7wH4SeKEe5bwI9F3v2piZmW0g6gqwHwKd61FOx5TXzMzMqDvAPgMcUo9yRqS8ZmZmRt0B9lrgJEkja8og6XiycVqrG5DdzMxso1TrfbAR8VdJvwZulHQ28ADwOhDAJ4D9gVJgTETckXdlzczMWos6HzQREedJmgKcS3bbTrs0aznwT2BERNybWw3NzMxaoXo9ySki7gHukVTCmgf+vxMRK3OrmZmZWSvWkGcRkwLqgpzqYmZmtsFoyGg6ZmZmVk8OsGZmZjlwgDUzM8uBA6yZmVkOHGDNzMxy4ABrZmaWgyYNsJJukPS2pOcL0kZLelPS9DQdWDDv+5JmSXqpcNB3ScNT2ixJFxSk95P0ZEq/VVLblN4ufZ6V5vdtmi02M7ONVVO3YMcBw6tJHxMRA9J0P4CkXYCjgV3TMtdKaiOpDXBKs/M+AAAey0lEQVQNcACwC3BMygtweSpre2AJcFJKPwlYktLHpHxmZma5adIAGxGPAYvrmX0EMCEilkfEHGAWsFeaZkXE7IhYAUwARkgSMAS4PS1/E3BoQVk3pfe3A/um/GZmZrloKddgz5b0XDqF3DWl9QLeKMgzN6XVlL4VsLTg8Y0V6VXKSvPfZc0jH83MzBpdSwiwvwU+BQwA5gG/aM7KSDpVUpmksoULFzZnVczMrBVr9gAbEQsiYlVErAauIzsFDPAm0Kcga++UVlP6O0CXNCBBYXqVstL8LVL+6uozNiJKI6K0W7du67t5Zma2kWr2ACupZ8HHw4CKHsZ3A0enHsD9gB2AacBTwA6px3Bbso5Qd0dEAI8CR6TlRwJ3FZRVMWj8EcAjKb+ZmVkuGjSazvqS9GdgMLC1pLnAKGCwpAFkg7i/CpwGEBEzJd0GvACsBM6KiFWpnLOBB4E2wA0RMTOt4nxggqRLgGeB61P69cAfJc0i62R1dM6bamZmGzm5IVez0tLSKCsrqzNfv36v5l+ZDdCcOX2buwpmZg0m6emIKK0rX7OfIjYzM9sQOcCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1gzM7McNGmAlXSDpLclPV+QtqWkSZJeTq9dU7ok/UbSLEnPSdqzYJmRKf/LkkYWpA+UNCMt8xtJqm0dZmZmeWnqFuw4YHhR2gXA5IjYAZicPgMcAOyQplOB30IWLIFRwOeAvYBRBQHzt8ApBcsNr2MdZmZmuWjSABsRjwGLi5JHADel9zcBhxakj4/MVKCLpJ7A/sCkiFgcEUuAScDwNK9zREyNiADGF5VV3TrMzMxy0RKuwXaPiHnp/Xyge3rfC3ijIN/clFZb+txq0mtbh5mZWS5aQoCtlFqe0ZzrkHSqpDJJZQsXLsyzKmZmtgFrCQF2QTq9S3p9O6W/CfQpyNc7pdWW3rua9NrWsZaIGBsRpRFR2q1bt3XeKDMz27i1hAB7N1DRE3gkcFdB+vGpN/Eg4N10mvdBYJikrqlz0zDgwTRvmaRBqffw8UVlVbcOMzOzXJQ05cok/RkYDGwtaS5Zb+DLgNsknQS8BhyVst8PHAjMAj4ATgSIiMWSfgI8lfL9OCIqOk6dSdZTuQMwMU3Usg4zM7NcKLskadUpLS2NsrKyOvP16/dq/pXZAM2Z07e5q2Bm1mCSno6I0rrytYRTxGZmZhscB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtW4JprrmH33Xenc+fOdO7cmb333pv77ruv2rynnXYakrjyyisr0xYvXsw3v/lN+vfvT4cOHejTpw9nnHEG77zzTpVln3nmGYYOHUqXLl3YaqutOPXUUykvL6+S55xzzqG0tJT27dvTt2/fRt9WM8tXiwmwkl6VNEPSdEllKW1LSZMkvZxeu6Z0SfqNpFmSnpO0Z0E5I1P+lyWNLEgfmMqflZZV02+ltXS9e/fm8ssv55lnnqGsrIwhQ4Zw6KGH8txzz1XJd/vttzNt2jS23XbbKulvvfUWb775JldccQUzZszg5ptv5rHHHuOYY46pkme//fZju+2248knn+SBBx5g5syZnHDCCVXKWr16NSNHjuT444/PbXvNLD+KiOauA5AFWKA0IhYVpF0BLI6IyyRdAHSNiPMlHQh8EzgQ+Bzw64j4nKQtgTKgFAjgaWBgRCyRNA34FvAkcD/wm4iYWFudSktLo6ysrM669+v3aoO312DOnL7NXYV62XLLLbn00ks57bTTAHjttdf4/Oc/z8MPP8wBBxzA2WefzXe+850al7///vs5+OCDWbp0KZ07d2bs2LH84Ac/YMGCBbRp0waAGTNmsPvuu/Pyyy+z/fbbV1n+yiuv5Oqrr+bVV1/NbRvNrP4kPR0RpXXlazEt2BqMAG5K728CDi1IHx+ZqUAXST2B/YFJEbE4IpYAk4DhaV7niJga2S+K8QVlmVVr1apVTJgwgfLycj7/+c8DsHLlSo455hguuugidt5553qVs2zZMtq1a8dmm20GwPLly9l0000rgytAhw4dAPjHP/7RyFthZs2lJQXYAB6S9LSkU1Na94iYl97PB7qn972ANwqWnZvSakufW0262VpmzJhBx44dadeuHaeffjp33HEHu+22GwCjRo1i66235owzzqhXWUuXLuXiiy/mlFNOoaSkBIAhQ4awaNEiLrvsMlasWMGSJUu44IILAJg3b15txZlZK9KSAuw+EbEncABwlqQvFc5MLc/cz2dLOlVSmaSyhQsX5r06a4F22mknpk+fzpNPPskZZ5zByJEjef7555kyZQrjxo3j+uuvr1c55eXlfPWrX6VXr15cccUVlem77rorN910E7/61a/o0KEDPXr0oF+/fnTv3p1NNmlJf5Jmtj5azF9zRLyZXt8G7gD2Ahak07uk17dT9jeBPgWL905ptaX3ria9unqMjYjSiCjt1q3b+m6WtUJt27Zl++23Z+DAgVx66aUMGDCAMWPGMGXKFObNm0fPnj0pKSmhpKSE1157jfPPP5/evXtXKaO8vJwDDzwQgHvvvZf27dtXmf+1r32N+fPn89Zbb/HOO+8wevRoFi5cyHbbbddk22lm+WoRAVbS5pI6VbwHhgHPA3cDFT2BRwJ3pfd3A8en3sSDgHfTqeQHgWGSuqYex8OAB9O8ZZIGpd7DxxeUZVar1atXs3z5cs4880yee+45pk+fXjltu+22fPvb32by5MmV+d977z2GDx/OqlWruP/+++nYsWONZXfv3p2OHTty66230r59e4YOHdoUm2RmTaCkuSuQdAfuSHfOlAC3RMQDkp4CbpN0EvAacFTKfz9ZD+JZwAfAiQARsVjST4CnUr4fR8Ti9P5MYBzQAZiYJrMqLrjgAg466CD69OnDe++9xy233MKUKVO477772Gabbdhmm22q5N90003p0aMHO+20E5AF12HDhrFs2TLuvPNO3n//fd5//30g643ctm1bAK6++mr23ntvOnXqxKRJk/jud7/LZZddRpcuXSrLnjVrFuXl5bz11lusWLGC6dOnA7DLLrtUlmNmLVeLCLARMRv4TDXp7wD7VpMewFk1lHUDcEM16WXAp9e7srZBmz9/Psceeyzz589niy22YPfdd2fixInsv//+9Vr+6aefZurUqQDsuOOOVeY9+uijDB48GIBp06YxatQoysvL6d+/P7///e857rjjquQ/+eST+fvf/175eY899gBgzpw5fvCEWSvQYu6DbYl8H2y+Wst9sGZmhTaU+2DNzMxaJQdYMzOzHDjAmpmZ5cAB1sxanMcee4xDDjmEXr16IYlx48ZVzvv44485//zz2X333dl8883p2bMnX/va13j99derlPHKK69w2GGH0a1bNzp37sxRRx3FggULquT573//y6GHHsrWW29Np06dGDRoEA888ECVPE899RT77bcfXbp0oUuXLuy7775MmzYtt223DUeL6EVstr7c0WzdtNSOZuXl5Xz605/m+OOPX2s0oQ8++IBnnnmGCy+8kAEDBvDuu+9y3nnnMXz4cJ577jlKSkp4//33GTZsGLvtthuPPPIIABdffDFf/epXmTp1auUTsw4++GC22247Jk+ezOabb87vfvc7RowYwQsvvMCnPvUpysvLGT58OAcddBBTp04lIvjpT3/K/vvvz+uvv06nTp2afN9Y6+FexLVwL+J8NeY/dx+DddNSA2yhjh07cvXVV681nF+hF154gV133ZXnnnuO3XbbjYceeojhw4fzzjvv0LVrVwDeffddunbtykMPPcR+++3HokWL6NatG4888ghf+cpXgGwwh3bt2nHrrbdyxBFHUFZWxmc/+1lmz55Nv379gOw2qe22246nnnqK0tI6O5LaBsi9iM1so7Fs2TKAymC6fPlyJFV5RGX79u3ZZJNNKkcs2mqrrdh555354x//SHl5OatWrWLs2LF06tSJL3zhC0D2XOpu3bpx/fXXs3z5cpYvX851113HJz7xCXbdddcm3kprbRxgzaxVW7FiBeeddx5f/epXK58JPWjQIDp27Mh3v/vdyqdpfec732HVqlWVIxZJYtKkSTz//PN07tyZdu3aMXr0aCZOnEjPnj0B6NSpE1OmTOG2225js802Y7PNNuPWW29l0qRJlUMMmtXEAdbMWq2VK1dy7LHHsnTpUm688cbK9G7duvGXv/yFiRMn0qlTJ7bYYguWLl3KnnvuWXn9NSI488wz2WqrrXj88ceZNm0aRxxxBIcffjhvvpmNBfLhhx/yjW98g0GDBjF16lT++c9/ssceezBixIjKR2Ca1cSdnMysVVq5ciXHHHMMM2bMYMqUKWy11VZV5g8bNoxXXnmFRYsWUVJSQpcuXejRo0fliEWPPPII99xzD4sXL658BvS1117LpEmTuPHGG7nooou45ZZbeOWVV/jnP/9JmzZtALjlllvo2rUrd9xxB8cee2zTbrS1Kg6wZtbqfPzxxxx99NGV4/T26NGjxrxbb701kAXUt99+m0MOOQTIeiMDa43Bu8kmm7B69erKPJKq5Nlkk02QVJnHrCY+RWxmLU55eXnlkICrV6/m9ddfZ/r06bz++uusXLmSI488kqlTp/LnP/8ZScyfP5/58+fz4YcfVpZx44038sQTT/DKK69w8803c+SRR/Ltb3+7cuSjvffemy233JITTzyRf//73/z3v//lu9/9LrNnz+bggw8GYOjQoSxbtowzzzyTF198kZkzZ3LiiSfSpk0bhgwZ0iz7xloPB1gza3HKysrYY4892GOPPfjwww8ZNWoUe+yxBz/84Q+ZO3cud911F2+99RYDBw6kZ8+eldOtt95aWcZLL73EYYcdxs4778yPf/xjLrzwQq688srK+VtvvTUPPPAA5eXlDBkyhNLSUh577DHuvPNO9txzTwD69+/PPffcw4wZM9h7773ZZ599mDt3LhMnTqzsULWxq+2hIMVOO+00JFU5Dq+++iqSqp1+/vOfN8EW5MeniM2sxRk8eDC13aNfn/v3L7vsMi677LJa85SWlvLggw/Wmmfo0KEMHTq0zvVtrGp7KEih22+/nWnTprHttttWSe/Tp09lz+4Kd9xxB2eddRZHHHFELnVuKg6wZma2zg488EAOPPBAgBofBvLaa69xzjnn8PDDD3PAAQdUmdemTZu1rqH/7W9/Y7/99qt8uEdr5QBrZma5qejtfdFFF7HzzjvXmX/27NlMnjyZ2267rQlqly9fgzUzs9yMGjWKrbfemjPOOKNe+f/whz/QrVs3RowYkXPN8ucWrJmZ5WLKlCmMGzeO6dOn1yv/ypUrufHGGxk5ciSbbrppzrXLn1uwZmaWiylTpjBv3jx69uxJSUkJJSUlvPbaa5x//vnV9sK+5557mD9/PieffHIz1LbxuQVrZo3Goxqtm9YwqtG6OPPMM9fqCbz//vtzzDHHcMopp6yV/7rrruPLX/4yO+64Y1NVMVcOsGZmts7Ky8uZNWsWQJWHgmy55ZZ84hOfYJtttqmSf9NNN6VHjx6VD/yo8Prrr/Pggw8yfvz4Jqt73nyK2MzM1lltDwVpiOuvv54tttiCww8/PKeaNj0PuF4LD7ieLw+43vwa+9Skj8O62VBPEW+oPOC6mZlZM9qoAqyk4ZJekjRL0gXNXR8zM9twbTQBVlIb4BrgAGAX4BhJuzRvrczMbEO10QRYYC9gVkTMjogVwASg9T8qxMzMWqSNKcD2At4o+Dw3pZmZmTU63wdbRNKpwKnpY7mkl5qzPo1ga2BRc1eiOlJz16DJ+Bi0DD4OLUOLPQ4N8Mn6ZNqYAuybQJ+Cz71TWhURMRYY21SVypuksvp0J7f8+Bi0DD4OLcPGdBw2plPETwE7SOonqS1wNHB3M9fJzMw2UBtNCzYiVko6G3gQaAPcEBEzm7laZma2gdpoAixARNwP3N/c9WhiG8zp7lbMx6Bl8HFoGTaa4+BHJZqZmeVgY7oGa2Zm1mQcYM3MzHLgANsEJJVKulPSQknLJP1X0q8k9ZQ0TtLHksqLpoPTsqMlrUxp70mandKU5j9UsMxHklYXlbO3pEskPVBQn39IWl6U73dp3smS/lOQ9+Ya6jc8zb+koH7LJL0l6T5JhzXtXq6bj0Nu+/Vbkl4pSvumpJB0QEFah7RvDpH0qqRjJX29aHtC0oeF+0NS35Teu2gdVdIlnVCw35dJelvSI5JOkrRJwXKF+colvSHpN5LaV7Ntn5C0StKjRem7p3X3KUjbLaVdXpT375J+md6PS3mOL8rzsKTRDdjtuZE0JX0v35P0bvqu/1HSwGryFH8fd0vz67WdKc8+6f3g9Lm4zD+n+RXH+/1UtyWSytLf4RZNsnMayAE2Z5KGAv8AXgIGRERn4MvAO+kV4KaI6Fg03VtQzJSI6Ah0BkYC30uvRMSwimWA04HZReU8UUPVRhXlO72Wzbi+mvo9UDD/4ZTWGdgV+BswTtIVDdlXefJxyNVkYDtJhTff7wvMBIYUpH2BrAf/lIqEiPhT4fYAq4AD6rk/qjO7YB9sB1wNjAb+UkO+jsBw4CigugFATgaWAoMl7ViQPgN4u2j71tpmSZsBg4CHC/K9A1wiqUMDt60p/SQiOkXEFsBXgNeAqUU/2H5SzfdxRsH8ddnOVdWUeUxRnp0iohOwDfAtsv1eJmmrhm9mvhxg83ctcEtEnB8RbwJExLyI+ElETGhIQZF5nOyPuEXeqB0RSyLieuBc4DxJ2zd3nRIfh/zWNROYR/aPrmJgjS8DoyrSkn2BpyJiWV51KapXeUT8Dfg68D/pR1Z1+WYCj1N0LNN2nARcCjzPmie8EVnv0EdYe/uuAHaS1DWlfREQ8PeCfHcDi4Fvr/PGNaGIeC0iLgLGA1dJ9X7uVK7bGREfR8S/yJ4pvwXwf3msZ304wOYo/eLdHrilkcrbRNJXgE+TtcRaslvJ/rEMbuZ6+Dg0zXEoDDYDgflk/2A/VdCy2JeqLbkmERGPAW9RNRhWkvQZsh8ExcfyYLJW0h+BG4CRktoVzJ/Mmh8VJcCXyO6zf4I1+3tfYGpEvF+w3GrgO8D5krqt84Y1vQlkz2/fqZ75m2Q7I2IxMIkajm9zcoDNV8WXaq1HMhY5TtLSoukTBfO/LGkp8CHZP7Ibgd+tZ91GFa2vtpbYidXUb9vaCo+ID4AlQEs4bePjkP9xeJg1p0b3BR6JiI+BfwFfSdfI9mT9AuzMwm0HnmvAsnOpug/6pXI+BKaTXT4YXbTMacB9EbGALMh2Bv6nYP7DwLaS+gOfBeamvIU/Nqr9URERD5Ptmx81YBua29z0WrEfLyz+PhYvsA7b2aaa73h9xu4uPr4tggNsvham17pG7fljRHQpml4vmP/3iOgCdAJ+QPbreH2v3/yoaH1lteS9sZr6vVVb4enaU1ey6zDNzcch/+MwGeihbIzlIWRBBuDR9Hkw8BFZ625d7Vq47cDuDVi2N1X3wZxURkey6+iDgC4VM9P15P3JWq5ExCKyFvlpFXki4lVgNtn2rbXN6TTxAGr+UfFd4BuS6tsibG4Vncwq9uNPi7+PNSzXkO1cVc13/LJ61q0l/K+pwgE2RxHxX2AWUHyRfl3LWxERl5IFjJb+y/eo9DqlOSsBPg7pdUqeK4mIN4D/AgcBe5MFGVjTmtsXeCy1apuUpC8C27ImAFaKiFURMZ7sFONvCmadTPb/8Q+S5kuaTxZwv1wUKB5mzfZVlP90Wt/XgXJgWnX1iojnyVrGLaYzYB3+l+wsUIMui+S9nemHzFCqOb7NzQE2f2cCX5f0s4rTeZK6S/q+pKPXscyLgDOLem22CJK6SjoRGAOMiYhZzV2nxMchf5PJOpq8nK6LATxLdh3zSJr4+qukzSUdCtwM3BURD9WS/UfAQZIGpeupJwGXkbWSB6RpR+A/FHR2ItvmIcDnSD9iImIVWaepC8nOeqysZb0Xp+U/0+ANbCKS+kj6EXACcE7q4NVQjb6dkkokDQLuAN4DftlYZTcWB9icRcQkYB9gF2CGpPfIrvdsw5pWxUitfe/XmbWU+TjZH3BTtZ5OqqZ+hf9k9ktpy4AXyVpNp0bEeU1Uvzr5ODSJh4EeFLQkUrB5LKU3RYDdLu2D98huLTkX+BlweG0LRcRssl6ylwJfJTutPiYi5hdOZD9YCjs7TSbrwfqfiFhSUOQj1GObU5lXko2R2pJcrOxe02Vkx2974PMR8deiPNXeN16sAdvZppoy/1WU56V0fBeR3R3wODAwncZvUfwsYjMzsxy4BWtmZpYDB1gzM7McOMCamZnlwAHWzMwsBw6wZmZmOXCAtVZLHiatIm2DGCbNbEPjAGutmYdJq2Gb1XqHSTPbYDjAWqvlYdIqbZDDpJm1dg6w1tp5mLQNf5g0s1bJAdZaOw+TtuEPk2bWKjnAWmvnYdI2/GHSzFqlkuaugNn6iIg3JBUOk/a/adYjwJ+AlbTQYdKA8ZL2JRsm7bA0q3CYtNUpbTOgk6SdIqLidHLFMGldgatSWr2HSZPUmoZJM2uV3IK1DYGHSWPDGybNrLVzgLUNgYdJW6O1D5NmtsHwcHVmZmY5cAvWzMwsBw6wZmZmOXCANTMzy4EDrJmZWQ4cYM3MzHLgAGtmZpYDB1izFkDS4WkM2aWSlkv6r6RfSto2za8Yg/bg5q6rmdWPA6xZM5P0C+A2sucLHwcMI3uwxL7ANc1YNTNbD34WsVkzkvRVssc8nhQRNxTM+ruksWTB1sxaIbdgzZrXt4FnioIrkD3uMSIm1rSgpOMl/UPSYklLJD0qqXjw9l0lPZDyvC/pRUlnFczfR9LjkpalabqkI2tZZ8Wp6qMk/V7Su5LmSvqRpE0K8vWXNEHSG5I+kDRT0rlFeQansvaVdFeq38uShklqI+nnkhZJelPS/1VTly9K+nsq/x1J10nqVOveNmtCDrBmzUTSpsDngQfWsYi+ZM8xPhL4GvAG8Lik7Qry3AOsAo4FDiEbeadTWn9n4F6yU9OHA0eQjT/bhbpdQTZizxFkgxr8ML2v0ItsIPkzgQOB68gGNji/mrJ+TzYu7mFkz3G+Hbg61fNr6fMvJH2uYgFJXyB73vL8tN5z03purEfdzZpGRHjy5KkZJrKH8gdwWj3y9k15D65h/iZkl3z+A/wwpW2dltmthmVK0/xODahzRT3GF6VPBybUsIxS3X4AzC5IH5zKGlWQtktKe6Ro2+YDlxekPQ48WrSeIWnZTzf3sfXkKSLcgjVrAdZpxA1JO0u6Q9ICslbqx8BOZMPbASwma9X+TtL/StqmqIhXyFqht0gaIak+LdcKxUPwvUA2uHxF3dqn08azgOWpbj8F+qVh+QpNLng/K70Wjoy0mqyV3SuVvRnZ2L+3SSqpmMhawR8DAxuwHWa5cYA1az7vkAWfTzR0wXSt8SGgD1knqS8CnwX+DbSHysA0jKz1dwMwP11v3SPNXwIMBTYl68W8UNJ9RaeYa7K06POKivUmlwPfAcaSnbr9LHBJmte+6qJryoqIFfUovyvQBriWLKBWTMvTtvSpR/3NcudexGbNJCI+lvRPYH/gogYuvjdZi3FoRPynIlHSFkXr+A9weLre+0WywHefpN4RsToipgLDJXUA9gN+CdwCDFrX7UqOBK6KiCsK6nbQepZZYSlZq380cH81899qpPWYrRe3YM2a16+AUkkji2dI2kTS8BqW65Belxfk/zzZNdK1RMTHEfEIWQDtSVFHpoj4MCLuIWvp7tLQjaihfoV1awMc3QjlEhHvA1OBnSKirJrJAdZaBLdgzZpRRNwj6ZfA9aln7F1k10X7A6cDr1J9L+OpKd91kq4ga82OBt6syCBpd+BK4Faya5hdyXrx/jsiFqcW5TeAO4HXya5xnkbB9c/1MAk4K12DXQycBbRrhHIrfA+YLGk1WS/j98hOtR8EXBgR/23EdZmtEwdYs2YWEedJ+hdwNtnp2Q5kgfVusgBZ3TIL0v2qV5IF5ZfJAvL3CrLNBxYAFwLbkp1afZQ1t8rMIjvV+jNgG2Dh/7dzx0QIxEAYRv8Yw9EZQAktgwBMnBFKFFAuxfYU3G33noFMqi8zmyT9bOd6wra2JLf0T1SfJPckz/RM9rCq2tdal/TTn0d6JvtKH0beZ6wBR62qvy4wAgA/mMECwACBBYABAgsAAwQWAAYILAAMEFgAGCCwADBAYAFggMACwIAvD60Ky2KPpOQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "datapoints_per_class = train_data['CASE_STATUS'].value_counts()\n",
    "datapoints_per_class = pd.DataFrame({'Class': datapoints_per_class.index, 'Count': datapoints_per_class.values})\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "g = sns.barplot(x='Class', y='Count', data=datapoints_per_class, capsize=0.9, color='blue') # #1f77b4\n",
    "g.set_title(\"Number of datapoints per class\", fontsize=16)\n",
    "g.set_xlabel(\"Class name\", fontsize=15)\n",
    "g.set_ylabel(\"Count\", fontsize=15)\n",
    "g.set_xticklabels(['CERTIFIED', \"CERTIFIED\\nWITHDRAWN\", 'WITHDRAWN', 'DENIED'], rotation=0, fontsize=13)\n",
    "g.set_yticks([0, 50000, 100000, 150000, 200000, 250000, 300000, 350000, 400000])\n",
    "for p in g.patches:\n",
    "    g.annotate(np.int(p.get_height()), \n",
    "                (p.get_x()+p.get_width()/2., p.get_height()), \n",
    "                ha='center', va='center', xytext=(0, 10), \n",
    "                textcoords='offset points', fontsize=14, color='black')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Explore EMPLOYER_NAME feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "INFOSYS LIMITED                      13188\n",
       "CAPGEMINI AMERICA INC                 7103\n",
       "IBM INDIA PRIVATE LIMITED             6870\n",
       "TATA CONSULTANCY SERVICES LIMITED     6842\n",
       "TECH MAHINDRA AMERICASINC             6099\n",
       "GOOGLE INC                            4250\n",
       "Name: EMPLOYER_NAME, dtype: int64"
      ]
     },
     "execution_count": 188,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''\n",
    "Infosys, Capgemini, IBM, TCS, Tech Mahindra, Google are the top companies submitting the applications for their \n",
    "employees.\n",
    "'''\n",
    "\n",
    "train_data['EMPLOYER_NAME'].value_counts().head(6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CERTIFIED             10680\n",
       "CERTIFIEDWITHDRAWN     3228\n",
       "WITHDRAWN               459\n",
       "DENIED                    2\n",
       "Name: CASE_STATUS, dtype: int64"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''\n",
    "However, the application is most likely to be accepted if it is from an University. The below table shows the \n",
    "number of records per label that are submitted by a University(14369 records).\n",
    "'''\n",
    "\n",
    "train_data['EMPLOYER_NAME'] = train_data['EMPLOYER_NAME'].fillna('OTHER')\n",
    "train_data[train_data['EMPLOYER_NAME'].str.contains('UNIVERSITY')]['CASE_STATUS'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <center> Preprocessing </center>\n",
    "- Drop unnecessary features/columns.\n",
    "- Handle null values.\n",
    "- Create new features."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fill null values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "CASE_SUBMITTED_DAY has 3999 null values. Since, this column has day information, I imputed missing values \n",
    "by ramdomly sampling numbers between 1 to 28 (if it is between 1 to 31, some of the months may not have that \n",
    "many days).\n",
    "'''\n",
    "\n",
    "null_case_submitted_day = train_data[train_data['CASE_SUBMITTED_DAY'].isnull()]\n",
    "null_case_submitted_day.CASE_SUBMITTED_DAY = null_case_submitted_day['CASE_SUBMITTED_DAY'].apply(lambda v: random.randint(1,28))\n",
    "train_data[train_data['CASE_SUBMITTED_DAY'].isnull()] = null_case_submitted_day"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create date-time features \"CASE_SUBMITTED_DATE\", \"DECISION_DATE\" and \"DECISION_PERIOD\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_decision_period_column(train_data):\n",
    "    '''\n",
    "    Combined (CASE_SUBMITTED_DAY, CASE_SUBMITTED_MONTH, CASE_SUBMITTED_YEAR) and (DECISION_DAY, DECISION_MONTH, DECISION_YEAR) \n",
    "    to get CASE_SUBMITTED_DATE and DECISION_DATE date-time features respectively.\n",
    "    '''\n",
    "    \n",
    "    train_data['CASE_SUBMITTED_DAY'] = train_data['CASE_SUBMITTED_DAY'].astype('int')\n",
    "    train_data['CASE_SUBMITTED_DATE'] = train_data['CASE_SUBMITTED_DAY'].astype('str') + '-' + train_data['CASE_SUBMITTED_MONTH'].astype('str') + '-' + train_data['CASE_SUBMITTED_YEAR'].astype('str')\n",
    "    train_data['CASE_SUBMITTED_DATE'] = pd.to_datetime(train_data['CASE_SUBMITTED_DATE'], dayfirst=True, format='%d-%m-%Y', errors='coerce')\n",
    "\n",
    "    train_data['DECISION_DATE'] = train_data['DECISION_DAY'].astype('str') + '-' + train_data['DECISION_MONTH'].astype('str') + '-' + train_data['DECISION_YEAR'].astype('str')\n",
    "    train_data['DECISION_DATE'] = pd.to_datetime(train_data['DECISION_DATE'], dayfirst=True, format='%d-%m-%Y', errors='coerce')\n",
    "    \n",
    "    '''\n",
    "    Difference between CASE_SUBMITTED_DATE and DECISION_DATE date-time columns which carries information about \n",
    "    the gap (in number of days) is labelled as DECISION_PERIOD.\n",
    "    '''\n",
    "    \n",
    "    train_data['DECISION_PERIOD'] = (train_data[\"DECISION_DATE\"] - train_data[\"CASE_SUBMITTED_DATE\"])\n",
    "    train_data['DECISION_PERIOD'] = train_data.DECISION_PERIOD.apply(lambda x: float(x.days))\n",
    "    train_data['DECISION_PERIOD'] = train_data.DECISION_PERIOD.astype('int')\n",
    "    \n",
    "    return train_data\n",
    "\n",
    "train_data = create_decision_period_column(train_data)\n",
    "test_data = create_decision_period_column(test_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A function to impute NULL values and convert categories into numerical values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [],
   "source": [
    "def fill_fit_transform(column_name, train_data, test_data, mean_or_mode):\n",
    "    '''\n",
    "    Use MODE, if the feature is categorical.\n",
    "    Use MEAN, if the feature is numerical.\n",
    "    '''\n",
    "    \n",
    "    if mean_or_mode == 'MODE':\n",
    "        train_data[column_name] = train_data[column_name].fillna(train_data[column_name].mode()[0])\n",
    "        test_data[column_name] = test_data[column_name].fillna(test_data[column_name].mode()[0])\n",
    "    elif mean_or_mode == 'MEAN':\n",
    "        train_data[column_name] = train_data[column_name].fillna(train_data[column_name].mean())\n",
    "        test_data[column_name] = test_data[column_name].fillna(test_data[column_name].mean())\n",
    "    \n",
    "    '''\n",
    "    Label Encoder is used to transform categorical values into numerical values.\n",
    "    '''\n",
    "    \n",
    "    label_encoder = le.fit(list(train_data[column_name]) + list(test_data[column_name]))\n",
    "    train_data[column_name] = label_encoder.transform(train_data[column_name]).astype('int')\n",
    "    test_data[column_name] = label_encoder.transform(test_data[column_name]).astype('int')\n",
    "    return train_data, test_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Drop unnecessary columns\n",
    "1. CASE_NO - This is a unique number for each CASE. Since, it does not capture any information, I dropped \n",
    "this column.\n",
    "2. EMPLOYER_NAME - This column has been removed because there are many catgories(53,463). Most of the categories \n",
    "occur only once. So, it's better to drop this feature.\n",
    "3. EMPLOYER_COUNTRY - There are 4 categories(USA, CANADA, AUSTRALIA, CHINA) in this column. Out of all Visa cases \n",
    "in training set, only 9 belong to CANADA, AUSTRALIA and CHINA. In test data, all the cases belong to USA. As this \n",
    "feature is constant in test data, this feature has been dropped.\n",
    "4. WORKSITE_POSTAL_CODE - There are 17427 categories. I have compared the performance of the models with and without \n",
    "this feature. Since, it didnot effect the score much, I dropped this feature.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = train_data.drop(columns = ['CASE_NO', 'EMPLOYER_COUNTRY', 'WORKSITE_POSTAL_CODE', 'CASE_SUBMITTED_DATE', 'DECISION_DATE'])\n",
    "test_data = test_data.drop(columns = ['CASE_NO', 'EMPLOYER_COUNTRY', 'WORKSITE_POSTAL_CODE', 'CASE_SUBMITTED_DATE', 'DECISION_DATE'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preprocess each column in both train and test at the same time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "MEAN(Numerical feature), MODE(Categorical feature), NONE(Only LabelEncoding)\n",
    "'''\n",
    "\n",
    "train_data, test_data = fill_fit_transform('VISA_CLASS', train_data, test_data, 'NONE')\n",
    "train_data, test_data = fill_fit_transform('EMPLOYER_STATE', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('WAGE_UNIT_OF_PAY', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('SOC_NAME', train_data, test_data, 'NONE')\n",
    "train_data, test_data = fill_fit_transform('NAICS_CODE', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('FULL_TIME_POSITION', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('PREVAILING_WAGE', train_data, test_data, 'MEAN')\n",
    "train_data, test_data = fill_fit_transform('PW_UNIT_OF_PAY', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('PW_SOURCE', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('H-1B_DEPENDENT', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('WILLFUL_VIOLATOR', train_data, test_data, 'MODE')\n",
    "train_data, test_data = fill_fit_transform('WORKSITE_STATE', train_data, test_data, 'NONE')\n",
    "\n",
    "train_data['WAGE_RATE_OF_PAY_FROM'] = train_data['WAGE_RATE_OF_PAY_FROM'].astype('float')\n",
    "train_data['WAGE_RATE_OF_PAY_TO'] = train_data['WAGE_RATE_OF_PAY_TO'].astype('float')\n",
    "train_data['PREVAILING_WAGE'] = train_data['PREVAILING_WAGE'].astype('float')\n",
    "test_data['WAGE_RATE_OF_PAY_FROM'] = test_data['WAGE_RATE_OF_PAY_FROM'].astype('float')\n",
    "test_data['WAGE_RATE_OF_PAY_TO'] = test_data['WAGE_RATE_OF_PAY_TO'].astype('float')\n",
    "test_data['PREVAILING_WAGE'] = test_data['PREVAILING_WAGE'].astype('float')\n",
    "\n",
    "'''\n",
    "Map categories to numerical values.\n",
    "'''\n",
    "train_data['CASE_STATUS'] = train_data['CASE_STATUS'].map({'CERTIFIED': 0, 'CERTIFIEDWITHDRAWN': 1, 'WITHDRAWN': 2, 'DENIED': 3})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "PW_SOURCE_YEAR has year information from 2007 - 2017 and 2001. One of the record has a value of ‘1’, which is \n",
    "not an year. Hence, I replaced that value with the mode of all the values in that feature.\n",
    "'''\n",
    "\n",
    "train_data.loc[train_data['PW_SOURCE_YEAR'] == 1]['PW_SOURCE_YEAR'] = 2016\n",
    "train_data, test_data = fill_fit_transform('PW_SOURCE_YEAR', train_data, test_data, 'MODE')\n",
    "\n",
    "'''\n",
    "PW_SOURCE_OTHER is a categorical feature that has several categories in which most of the categories have \n",
    "occurred less than 10 times. Such categories are clubbed into one category. Around 116 categories are combined \n",
    "into one category('OTHER').\n",
    "'''\n",
    "\n",
    "other = list(train_data['PW_SOURCE_OTHER'].value_counts()[-116:].index)\n",
    "train_data['PW_SOURCE_OTHER'] = train_data['PW_SOURCE_OTHER'].replace(other, 'OTHER')\n",
    "test_data['PW_SOURCE_OTHER'] = test_data['PW_SOURCE_OTHER'].replace(other, 'OTHER')\n",
    "train_data, test_data = fill_fit_transform('PW_SOURCE_OTHER', train_data, test_data, 'NONE')\n",
    "\n",
    "'''\n",
    "It is obvious that WAGE_RATE_OF_PAY_FROM must be greater than WAGE_RATE_OF_PAY_TO. Some of the records in \n",
    "WAGE_RATE_OF_PAY_TO contains the value zero(0). So, I replaced those values with the values in \n",
    "WAGE_RATE_OF_PAY_FROM.\n",
    "'''\n",
    "\n",
    "train_data.ix[train_data['WAGE_RATE_OF_PAY_TO'] == 0, 'WAGE_RATE_OF_PAY_TO'] = train_data[train_data['WAGE_RATE_OF_PAY_TO'] == 0]['WAGE_RATE_OF_PAY_FROM']\n",
    "test_data.ix[test_data['WAGE_RATE_OF_PAY_TO'] == 0, 'WAGE_RATE_OF_PAY_TO'] = test_data[test_data['WAGE_RATE_OF_PAY_TO'] == 0]['WAGE_RATE_OF_PAY_FROM']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "A new division feature(WAGE_RATE_OF_PAY_FROM/WAGE_RATE_OF_PAY_TO) has been created from WAGE_RATE_OF_PAY_FROM and \n",
    "WAGE_RATE_OF_PAY_TO. Most of the methods capture information from the addition and subtraction of features. But, \n",
    "it is difficult to extract information from division of two features. So, I have created this feature to extract \n",
    "hidden information. This feature has increased the F1 score.\n",
    "'''\n",
    "\n",
    "train_data['RATIO_OF_PAY_FROM_TO'] = train_data['WAGE_RATE_OF_PAY_FROM']/train_data['WAGE_RATE_OF_PAY_TO']\n",
    "test_data['RATIO_OF_PAY_FROM_TO'] = test_data['WAGE_RATE_OF_PAY_FROM']/test_data['WAGE_RATE_OF_PAY_TO']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "A new boolean feature IS_ES_SAMEAS_WS is added which tells whether the EMPLOYER_STATE and WORKSITE_STATE \n",
    "are the same or not.\n",
    "'''\n",
    "\n",
    "train_data['IS_ES_SAMEAS_WS'] = train_data['EMPLOYER_STATE'] != train_data['WORKSITE_STATE']\n",
    "train_data['IS_ES_SAMEAS_WS'] = train_data['IS_ES_SAMEAS_WS'].map({True: 1, False: 0})\n",
    "\n",
    "test_data['IS_ES_SAMEAS_WS'] = test_data['EMPLOYER_STATE'] != test_data['WORKSITE_STATE']\n",
    "test_data['IS_ES_SAMEAS_WS'] = test_data['IS_ES_SAMEAS_WS'].map({True: 1, False: 0})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "A boolean feature ‘IS_UNIVERSITY’ is created which tells whether the application is from a University or not.\n",
    "'''\n",
    "\n",
    "column_name = 'EMPLOYER_NAME'\n",
    "train_data[column_name] = train_data[column_name].fillna(train_data[column_name].mode()[0])\n",
    "test_data[column_name] = test_data[column_name].fillna(test_data[column_name].mode()[0])\n",
    "\n",
    "train_data['IS_UNIVERSITY'] = np.nan\n",
    "train_data['EMPLOYER_NAME'] = train_data['EMPLOYER_NAME'].str.lower()\n",
    "train_data.IS_UNIVERSITY[train_data['EMPLOYER_NAME'].str.contains('university')] = 1\n",
    "train_data['IS_UNIVERSITY']= train_data.IS_UNIVERSITY.replace(np.nan, 0, regex=True)\n",
    "train_data['IS_UNIVERSITY'] = train_data.IS_UNIVERSITY.astype('int')\n",
    "\n",
    "test_data['IS_UNIVERSITY'] = np.nan\n",
    "test_data['EMPLOYER_NAME'] = test_data['EMPLOYER_NAME'].str.lower()\n",
    "test_data.IS_UNIVERSITY[test_data['EMPLOYER_NAME'].str.contains('university')] = 1\n",
    "test_data['IS_UNIVERSITY']= test_data.IS_UNIVERSITY.replace(np.nan, 0, regex=True)\n",
    "test_data['IS_UNIVERSITY'] = test_data.IS_UNIVERSITY.astype('int')\n",
    "\n",
    "train_data = train_data.drop(columns=['EMPLOYER_NAME'])\n",
    "test_data = test_data.drop(columns=['EMPLOYER_NAME'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Upsample lower class data-points"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Since the dataset provided is imbalanced, I manually up-sampled the data-points with lower class by 20 times.\n",
    "'''\n",
    "\n",
    "class_two_data = train_data[train_data['CASE_STATUS'] == 2]\n",
    "\n",
    "class_three_data = train_data[train_data['CASE_STATUS'] == 3]\n",
    "copy_of_class_three_data = class_three_data.copy()\n",
    "for i in range(1, 20):\n",
    "    class_three_data = class_three_data.append(copy_of_class_three_data, ignore_index=False)\n",
    "\n",
    "train_data = train_data.append(class_three_data, ignore_index=False)\n",
    "train_data = train_data.append(class_two_data, ignore_index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "datapoints_per_class = train_data['CASE_STATUS'].value_counts()\n",
    "datapoints_per_class = pd.DataFrame({'Class': datapoints_per_class.index, 'Count': datapoints_per_class.values})\n",
    "\n",
    "plt.figure(figsize=(7,5))\n",
    "g = sns.barplot(x='Class', y='Count', data=datapoints_per_class, capsize=0.9, color='blue') # #1f77b4\n",
    "g.set_title(\"Number of datapoints per class\", fontsize=16)\n",
    "g.set_xlabel(\"Class name\", fontsize=15)\n",
    "g.set_ylabel(\"Count\", fontsize=15)\n",
    "g.set_xticklabels(['CERTIFIED', \"CERTIFIED\\nWITHDRAWN\", 'WITHDRAWN', 'DENIED'], rotation=0, fontsize=13)\n",
    "g.set_yticks([0, 50000, 100000, 150000, 200000, 250000, 300000, 350000, 400000])\n",
    "for p in g.patches:\n",
    "    g.annotate(np.int(p.get_height()), \n",
    "                (p.get_x()+p.get_width()/2., p.get_height()), \n",
    "                ha='center', va='center', xytext=(0, 10), \n",
    "                textcoords='offset points', fontsize=14, color='black')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Final data - X_train and y_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = train_data.drop(columns=['CASE_STATUS'])\n",
    "y_train = train_data['CASE_STATUS']\n",
    "\n",
    "'''\n",
    "Split the dataset into train and validation sets.\n",
    "'''\n",
    "\n",
    "train_X, val_X, train_y, val_y = train_test_split(X_train, y_train, test_size=0.3, random_state=0)\n",
    "dummy_test_data = pd.read_csv(TEST_PATH)\n",
    "\n",
    "test_data = test_data[train_X.columns]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <center> Modelling - XGBoost </center>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
       "       colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,\n",
       "       max_depth=3, min_child_weight=1, missing=None, n_estimators=100,\n",
       "       n_jobs=1, nthread=None, objective='multi:softprob', random_state=12,\n",
       "       reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
       "       silent=True, subsample=1)"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb_model_1 = xgb.XGBClassifier(objective = 'multi:softmax', \n",
    "                              n_estimators = 100, \n",
    "                              max_depth = 3,\n",
    "                              colsample_bylevel = 0.9,\n",
    "                              learning_rate = 0.1,\n",
    "                              random_state=12)\n",
    "\n",
    "xgb_model_1.fit(X_train, y_train, eval_set=[(val_X, val_y)], verbose=False, early_stopping_rounds=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score - 0.9761484488623773\n"
     ]
    }
   ],
   "source": [
    "y_valid_pred_1 = xgb_model_1.predict(val_X)\n",
    "print ('F1 score - ' + str(f1_score(y_valid_pred_1, val_y, average='weighted')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_1 = xgb_model_1.predict(test_data)\n",
    "y_pred_1 = pd.DataFrame({'CASE_NO': dummy_test_data['CASE_NO'], 'CASE_STATUS': y_pred_1})\n",
    "y_pred_1['CASE_STATUS'] = y_pred_1['CASE_STATUS'].map({0: 'CERTIFIED', 1: 'CERTIFIEDWITHDRAWN', 2: 'WITHDRAWN', 3: 'DENIED'})\n",
    "y_pred_1.to_csv('1_prediction_xgboost.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Significant variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [],
   "source": [
    "def feature_importance(model, X_train):\n",
    "    importances = model.feature_importances_\n",
    "    indices = np.argsort(importances)[::-1]\n",
    "    features = X_train.columns\n",
    "    important_features = []\n",
    "    important_features_scores = []\n",
    "    for index in indices:\n",
    "        important_features.append(features[index])\n",
    "        important_features_scores.append(importances[index])\n",
    "    feature_importance_df = pd.Series(data=important_features_scores, index=important_features)\n",
    "    return feature_importance_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Feature Importance')"
      ]
     },
     "execution_count": 164,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,5))\n",
    "feature_importance_df_1 = feature_importance(xgb_model_1, X_train)\n",
    "feature_importance_df_1.iloc[:15].plot('bar')\n",
    "plt.xlabel('Feature')\n",
    "plt.ylabel('Importance')\n",
    "plt.title('Feature Importance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=0.9,\n",
       "       colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,\n",
       "       max_depth=5, min_child_weight=1, missing=None, n_estimators=80,\n",
       "       n_jobs=1, nthread=None, objective='multi:softprob', random_state=12,\n",
       "       reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
       "       silent=True, subsample=1)"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb_model_2 = xgb.XGBClassifier(objective = 'multi:softmax', \n",
    "                              n_estimators = 80, \n",
    "                              max_depth = 5,\n",
    "                              colsample_bylevel = 0.9,\n",
    "                              learning_rate = 0.1,\n",
    "                              random_state=12)\n",
    "\n",
    "xgb_model_2.fit(X_train, y_train, eval_set=[(val_X, val_y)], verbose=False, early_stopping_rounds=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score - 0.9794831488102878\n"
     ]
    }
   ],
   "source": [
    "y_valid_pred_2 = xgb_model_2.predict(val_X)\n",
    "print ('F1 score - ' + str(f1_score(y_valid_pred_2, val_y, average='weighted')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_2 = xgb_model_2.predict(test_data)\n",
    "y_pred_2 = pd.DataFrame({'CASE_NO': dummy_test_data['CASE_NO'], 'CASE_STATUS': y_pred_2})\n",
    "y_pred_2['CASE_STATUS'] = y_pred_2['CASE_STATUS'].map({0: 'CERTIFIED', 1: 'CERTIFIEDWITHDRAWN', 2: 'WITHDRAWN', 3: 'DENIED'})\n",
    "y_pred_2.to_csv('2_prediction_xgboost.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Feature Importance')"
      ]
     },
     "execution_count": 168,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4IAAAHNCAYAAABLmMXxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xu8dvWc//HXu5JyZsr4dU5CCHE7DWaIyAxyivRj4hdh9CNmkMMUGaYYvzHkTI5DckpGxilyphCpxF1KNShyiE6qz++PtXb3de/2ve/rvu/r2muva72ej8d+3Nf6rmvt67Puvfe1rs/6fr+fb6oKSZIkSdJwbNR1AJIkSZKkpWUiKEmSJEkDYyIoSZIkSQNjIihJkiRJA2MiKEmSJEkDYyIoSZIkSQNjIihJkiRJA2MiKEmaqiTnJLksyR9HvrbawO/5gCTnTyrGMV/zPUn+ZSlfc02SvDzJB7qOQ5LUXyaCkqSl8IiqutHI1/90GUySTbp8/Q3R59glScuHiaAkqTNJ7p3kG0l+l+QHSR4wsu+pSc5IckmSs5M8o22/IfAZYKvRHsb5PXbzew3bnskXJfkh8Kckm7THfSzJRUl+luQ5Y8a9Q5JqYzwvyW+TPDPJPZL8sD2fI0ee/5QkX09yZJLfJ/lxkgeN7N8qyXFJLk6yMsnTR/a9PMlHk3wgyR+AZwIvAZ7QnvsPFvv/Gv2/SPKPSS5M8oskTx3Zv3mS1yU5t43va0k2X9vPSJLUX95VlCR1IsnWwKeBJwP/DTwI+FiS21fVRcCFwMOBs4G/Bj6T5KSq+l6ShwEfqKptRr7fOC/7RODvgF8D1wCfAj7Ztm8DfCHJmVX12TFP417Azm18x7Xn8WDgesD3k3ykqk4cee5HgS2AxwAfT7JjVV0MHA38CNgKuD3w+SRnVdUJ7bF7AXsDfw9cv/0et6mqJ43Essb/r3b/rYCbAlsDewAfTXJsVf0W+DfgjsBfAb9sY71mjJ+RJKmn7BGUJC2FY9sepd8lObZtexJwfFUdX1XXVNXngZOBvwWoqk9X1VnVOBH4HHD/DYzjDVV1XlVdBtwD2LKqDquqK6vqbOAdwD7r8P1eWVWXV9XngD8BH6qqC6vqAuCrwG4jz70QeH1V/bmqPgycCfxdkm2B+wIvar/XKcA7aZK+Od+sqmPb/6fLFgpkjP+vPwOHta9/PPBH4HZJNgL+D/Dcqrqgqq6uqm9U1RWs5WckSeovewQlSUvhUVX1hXlt2wN7J3nESNv1gC8BtL1+hwK3pblxeQPg1A2M47x5r79Vkt+NtG1Mk8CN61cjjy9bYPtGI9sXVFWNbJ9L0wO4FXBxVV0yb9+KNcS9oDH+v35TVVeNbF/axrcFsBlw1gLfdtGfkSSpv0wEJUldOQ94f1U9ff6OJNcHPkbTK/bJqvpz25M4N/6z5h9D0yN3g5HtWy3wnNHjzgN+VlU7r0/w62HrJBlJBrejGU76P8Atktx4JBncDrhg5Nj557va9hj/X4v5NXA5sBPwg3n71vgzkiT1m0NDJUld+QDwiCQPTbJxks3aoibbAJvSzIW7CLiq7e16yMixvwL+IslNR9pOAf42yS2S3Ao4aC2v/x3gkraAzOZtDHdKco+JneHqbgk8J8n1kuwN7EIz7PI84BvAv7b/B3cG9qf5/1mTXwE7tMM6Ye3/X2tUVdcARwH/ry1as3GS+7TJ5WI/I0lSj5kISpI60SZAe9FUwLyIpvfpBcBGbc/Yc4BjgN8C+9L0ns0d+2PgQ8DZ7bzDrYD30/RonUMzP+7Da3n9q2mKq9wV+BlNz9g7aQqqTMO3aQrL/Bp4FfC4qvpNu++JwA40vYOfAA5dYCjtqI+0//4myffW9v81hn+iGUZ6EnAxcATNz2GNP6N1+N6SpGUoq09XkCRJk5bkKcDTqup+XcciSRJ4R0+SJEmSBsdEUJIkSZIGxqGhkiRJkjQw9ghKkiRJ0sCYCEqSJEnSwMzMgvJbbLFF7bDDDl2HIUmSJEmd+O53v/vrqtpynOfOTCK4ww47cPLJJ3cdhiRJkiR1Ism54z7XoaGSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDBTTQST7JnkzCQrkxy8wP7nJzk9yQ+TfDHJ9iP7rk5ySvt13DTjlCRJkqQhmdqC8kk2Bt4E7AGcD5yU5LiqOn3kad8HVlTVpUmeBbwGeEK777Kquuu04pMkSZKkoZpmj+A9gZVVdXZVXQkcDew1+oSq+lJVXdpufgvYZorxSJIkSZKYbiK4NXDeyPb5bdua7A98ZmR7syQnJ/lWkkdNI0BJkiRJGqKpDQ1dF0meBKwA/makefuquiDJrYETkpxaVWfNO+4A4ACA7bbbbr1ee4eDP71+Qa+ncw7/uyV9PUmSJEmab5o9ghcA245sb9O2rSbJg4GXAo+sqivm2qvqgvbfs4EvA7vNP7aq3l5VK6pqxZZbbjnZ6CVJkiRpRk0zETwJ2DnJjkk2BfYBVqv+mWQ34G00SeCFI+03T3L99vEWwH2B0SIzkiRJkqT1NLWhoVV1VZIDgc8CGwNHVdVpSQ4DTq6q44DXAjcCPpIE4OdV9UhgF+BtSa6hSVYPn1dtVJIkSZK0nqY6R7CqjgeOn9d2yMjjB6/huG8Au04zNkmSJEkaqqkuKC9JkiRJWn5MBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgTAQlSZIkaWCmmggm2TPJmUlWJjl4gf3PT3J6kh8m+WKS7Uf27Zfkp+3XftOMU5IkSZKGZGqJYJKNgTcBDwPuADwxyR3mPe37wIqqujPwUeA17bG3AA4F7gXcEzg0yc2nFaskSZIkDck0ewTvCaysqrOr6krgaGCv0SdU1Zeq6tJ281vANu3jhwKfr6qLq+q3wOeBPacYqyRJkiQNxjQTwa2B80a2z2/b1mR/4DPreawkSZIkaUybdB0AQJInASuAv1nH4w4ADgDYbrvtphCZJEmSJM2eafYIXgBsO7K9Tdu2miQPBl4KPLKqrliXY6vq7VW1oqpWbLnllhMLXJIkSZJm2TQTwZOAnZPsmGRTYB/guNEnJNkNeBtNEnjhyK7PAg9JcvO2SMxD2jZJkiRJ0gaa2tDQqroqyYE0CdzGwFFVdVqSw4CTq+o44LXAjYCPJAH4eVU9sqouTvJKmmQS4LCqunhasUqSJEnSkEx1jmBVHQ8cP6/tkJHHD17k2KOAo6YXnSRJkiQN01QXlJckSZIkLT8mgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDBjJ4JJtk/y4Pbx5kluPL2wJEmSJEnTMlYimOTpwEeBt7VN2wDHTisoSZIkSdL0jNsj+GzgvsAfAKrqp8AtpxWUJEmSJGl6xk0Er6iqK+c2kmwC1NoOSrJnkjOTrExy8AL7/zrJ95JcleRx8/ZdneSU9uu4MeOUJEmSJK3FJmM+78QkLwE2T7IH8A/ApxY7IMnGwJuAPYDzgZOSHFdVp4887efAU4B/WuBbXFZVdx0zPkmSJEnSmMbtETwYuAg4FXgGcDzwsrUcc09gZVWd3fYmHg3sNfqEqjqnqn4IXLNOUUuSJEmS1tu4PYKbA0dV1Tvg2t6+zYFLFzlma+C8ke3zgXutQ2ybJTkZuAo4vKquU5wmyQHAAQDbbbfdOnxrSZIkSRqucXsEv0iT+M3ZHPjC5MNZzfZVtQLYF3h9kp3mP6Gq3l5VK6pqxZZbbjnlcCRJkiRpNoybCG5WVX+c22gf32Atx1wAbDuyvU3bNpaquqD992zgy8Bu4x4rSZIkSVqzcRPBPyW529xGkrsDl63lmJOAnZPsmGRTYB9grOqfSW6e5Prt4y1olq44ffGjJEmSJEnjGHeO4EHAR5L8DxDgVsATFjugqq5KciDwWWBjmjmGpyU5DDi5qo5Lcg/gE8DNgUckeUVV3RHYBXhbkmtoktXD51UblSRJkiStp7ESwao6Kcntgdu1TWdW1Z/HOO54mgqjo22HjH5fmiGj84/7BrDrOLFJkiRJktbNuD2CAPcAdmiPuVsSqup9U4lKkiRJkjQ1YyWCSd4P7AScAlzdNhdgIihJkiRJPTNuj+AK4A5VVdMMRpIkSZI0feNWDf0RTYEYSZIkSVLPjdsjuAVwepLvAFfMNVbVI6cSlSRJkiRpasZNBF8+zSAkSZIkSUtn3OUjTpx2IJIkSZKkpTHWHMEk905yUpI/JrkyydVJ/jDt4CRJkiRJkzdusZgjgScCPwU2B54GvGlaQUmSJEmSpmfcRJCqWglsXFVXV9W7gT2nF5YkSZIkaVrGLRZzaZJNgVOSvAb4BeuQREqSJEmSlo9xk7knt889EPgTsC3wmGkFJUmSJEmannETwUdV1eVV9YeqekVVPR94+DQDkyRJkiRNx7iJ4H4LtD1lgnFIkiRJkpbIonMEkzwR2Be4dZLjRnbdGLh4moFJkiRJkqZjbcVivkFTGGYL4HUj7ZcAP5xWUJIkSZKk6Vk0Eayqc5OcD1xeVScuUUySJEmSpCla6xzBqroauCbJTZcgHkmSJEnSlI27juAfgVOTfJ5m+QgAquo5U4lKkiRJkjQ14yaCH2+/JEmSJEk9N1YiWFXvTbIpcNu26cyq+vP0wpIkSZIkTctYiWCSBwDvBc4BAmybZL+q+sr0QpMkSZIkTcO4Q0NfBzykqs4ESHJb4EPA3acVmCRJkiRpOtZaNbR1vbkkEKCqfgJcbzohSZIkSZKmadwewZOTvBP4QLv9v4GTpxOSJEmSJGmaxk0EnwU8G5hbLuKrwJunEpEkSZIkaarGrRp6RZIjgS8C19BUDb1yqpFJkiRJkqZi3Kqhfwe8FTiLpmrojkmeUVWfmWZwkiRJkqTJW5eqoQ+sqpUASXYCPg2YCEqSJElSz4xbNfSSuSSwdTZwyRTikSRJkiRN2bpUDT0eOAYoYG/gpCSPAaiqj08pPkmSJEnShI2bCG4G/Ar4m3b7ImBz4BE0iaGJoCRJkiT1xLhVQ5867UA0JS+/6RK/3u+X9vUkSZIkrbNxq4buCPxfYIfRY6rqkdMJS5IkSZI0LeMODT0WeBfwKZp1BCVJkiRJPTVuInh5Vb1hqpFIkiRJkpbEuIngfyQ5FPgccMVcY1V9bypRSZIkSZKmZtxEcFfgycDurBoaWu22JEmSJKlHxk0E9wZuXVVXTjMYSZIkSdL0jZsI/gi4GXDhFGOR1tmu7911SV/v1P1OXdLXkyRJkqZh3ETwZsCPk5zE6nMEXT5CkiRJknpm3ETw0KlGIUmSJElaMhuN86SqOnGhr7Udl2TPJGcmWZnk4AX2/3WS7yW5Ksnj5u3bL8lP26/9xj8lSZIkSdJiFu0RTHIJTXXQ6+wCqqpussixGwNvAvYAzgdOSnJcVZ0+8rSfA08B/mnesbeg6YVc0b7+d9tjf7vWM5IkSZIkLWrRRLCqbrwB3/uewMqqOhsgydHAXsC1iWBVndPuu2besQ8FPl9VF7f7Pw/sCXxoA+KRJEmSJDHm0ND1tDVw3sj2+W3btI+VJEmSJC1imong1CU5IMnJSU6+6KKLug5HkiRJknph3Kqh6+MCYNuR7W3atnGPfcC8Y788/0lV9Xbg7QArVqxYaC6j1Gtn3H6XJX29XX58xpK+3pueecKSvt6z37r7kr6eJEnScjXNHsGTgJ2T7JhkU2Af4Lgxj/0s8JAkN09yc+AhbZskSZIkaQNNLRGsqquAA2kSuDOAY6rqtCSHJXkkQJJ7JDkf2Bt4W5LT2mMvBl5Jk0yeBBw2VzhGkiRJkrRhpjk0lKo6Hjh+XtshI49Pohn2udCxRwFHTTM+SZIkSRqiXheLkSRJkiStOxNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRoYE0FJkiRJGhgTQUmSJEkaGBNBSZIkSRqYTboOQJJm0eue8PAlfb1//PB/LenrSZKkfrNHUJIkSZIGxkRQkiRJkgbGRFCSJEmSBsZEUJIkSZIGxkRQkiRJkgbGRFCSJEmSBsZEUJIkSZIGxkRQkiRJkgbGRFCSJEmSBsZEUJIkSZIGxkRQkiRJkgbGRFCSJEmSBsZEUJIkSZIGxkRQkiRJkgbGRFCSJEmSBsZEUJIkSZIGZqqJYJI9k5yZZGWSgxfYf/0kH273fzvJDm37DkkuS3JK+/XWacYpSZIkSUOyybS+cZKNgTcBewDnAyclOa6qTh952v7Ab6vqNkn2AY4AntDuO6uq7jqt+CRJkiRpqKaWCAL3BFZW1dkASY4G9gJGE8G9gJe3jz8KHJkkU4xJkjQB5x/81SV9vW0Ov/+Svp4kSbNumkNDtwbOG9k+v21b8DlVdRXwe+Av2n07Jvl+khOT+AlAkiRJkiZkmj2CG+IXwHZV9ZskdweOTXLHqvrD6JOSHAAcALDddtt1EKYkSZIk9c80ewQvALYd2d6mbVvwOUk2AW4K/Kaqrqiq3wBU1XeBs4Dbzn+Bqnp7Va2oqhVbbrnlFE5BkiRJkmbPNBPBk4Cdk+yYZFNgH+C4ec85Dtivffw44ISqqiRbtsVmSHJrYGfg7CnGKkmSJEmDMbWhoVV1VZIDgc8CGwNHVdVpSQ4DTq6q44B3Ae9PshK4mCZZBPhr4LAkfwauAZ5ZVRdPK1ZJkiRJGpKpzhGsquOB4+e1HTLy+HJg7wWO+xjwsWnGJkmSJElDNdUF5SVJkiRJy4+JoCRJkiQNjImgJEmSJA2MiaAkSZIkDYyJoCRJkiQNjImgJEmSJA2MiaAkSZIkDYyJoCRJkiQNjImgJEmSJA2MiaAkSZIkDYyJoCRJkiQNjImgJEmSJA3MJl0HIEnScvPyl798pl9PkiR7BCVJkiRpYEwEJUmSJGlgTAQlSZIkaWCcIyhJ0sB88YSdlvT1HrT7WUv6erf60ilL+nq/fOBdl/T1JGkS7BGUJEmSpIExEZQkSZKkgTERlCRJkqSBcY6gJElST+xw8KeX9PXOOfzvlvT1JC0dewQlSZIkaWBMBCVJkiRpYEwEJUmSJGlgnCMoSZKkZcE5kNLSsUdQkiRJkgbGRFCSJEmSBsZEUJIkSZIGxjmCkiRJ0lJ4+U2X+PV+v6Qvt+t7d13S1zt1v1OX9PVmjT2CkiRJkjQwJoKSJEmSNDAmgpIkSZI0MCaCkiRJkjQwJoKSJEmSNDAmgpIkSZI0MC4fIUmSJElrccbtd1nS19vlx2dM9fvbIyhJkiRJA2MiKEmSJEkDYyIoSZIkSQNjIihJkiRJA2MiKEmSJEkDYyIoSZIkSQNjIihJkiRJAzPVRDDJnknOTLIyycEL7L9+kg+3+7+dZIeRfS9u289M8tBpxilJkiRJQzK1RDDJxsCbgIcBdwCemOQO8562P/DbqroN8O/AEe2xdwD2Ae4I7Am8uf1+kiRJkqQNNM0ewXsCK6vq7Kq6Ejga2Gvec/YC3ts+/ijwoCRp24+uqiuq6mfAyvb7SZIkSZI20DQTwa2B80a2z2/bFnxOVV0F/B74izGPlSRJkiSth1TVdL5x8jhgz6p6Wrv9ZOBeVXXgyHN+1D7n/Hb7LOBewMuBb1XVB9r2dwGfqaqPznuNA4AD2s3bAWdO5WQWtgXw6yV8vaXm+fWb59dfs3xu4Pn1nefXX7N8buD59Z3nNznbV9WW4zxxkykGcQGw7cj2Nm3bQs85P8kmwE2B34x5LFX1duDtE4x5bElOrqoVXbz2UvD8+s3z669ZPjfw/PrO8+uvWT438Pz6zvPrxjSHhp4E7JxkxySb0hR/OW7ec44D9msfPw44oZouyuOAfdqqojsCOwPfmWKskiRJkjQYU+sRrKqrkhwIfBbYGDiqqk5LchhwclUdB7wLeH+SlcDFNMki7fOOAU4HrgKeXVVXTytWSZIkSRqSaQ4NpaqOB46f13bIyOPLgb3XcOyrgFdNM74N1MmQ1CXk+fWb59dfs3xu4Pn1nefXX7N8buD59Z3n14GpFYuRJEmSJC1P05wjKEmSJElahkwEJUmSJGlgTAQlSZIkaWBMBEWStyS5SddxTEuSNyZ5w5q+uo5vQyXZfeTxjvP2PWbpI5qsJE8aeXzfefsOXPqIJivJi5Js3HUcSynJHkk+33UcG2rWfzfntMtAPbz9unXX8WjDJdmu6xgmIcnDknwlya/brxOT/G3XcU1Tu+52ryW522JfXcc3Lcvx2mexmHWQZD/gucDt2qYzgDdU1fu6i2rDJXkBcABwaFV9sOt4Jq39uc15BXDo6P6qeu/SRjRZSb5XVXeb/3ih7T4awPkdCdyPZpmcr3cdzyS1NyneCmwFHAscAbwbCPCqqvp4h+FtsAH8bt4EeCewAjilbb4r8F1g/6r6Q1exTUKSTwFr/BBUVY9cwnCmIsl9gK2Br1TVhUnuDBwM3L+qtu02ug2T5OnAM4AXAie3zSuAw4F3VtWyrNI4jiRfq6r7tY/fX1VPHtk3C+8t1wA/An491zSyu6pq9+se1R99uvb1/q7CUmmTiYOA5wPfo/lh3g14bZKqqvd3Gd+GqKrXJvkg8P+S7A+8BbhmZP+y+YVdH6OJXpKD+p74LSBreLzQdh/N9PlV1YHtHdAjk5zBdf/+vtdZcBvudTQ3mb4JPKz99+CqOrLTqCZnpn83gTfQrOe7T1VdA5AkwD8DRwJ/32Fsk/Bv7b8B3gE8rcNYJi7Ja4GH0yTxL0ryWZpz/Ffg/3QZ24Q8D7hfVV080nZCkocBX2OZlusf0w1HHt9x3r5ZeG95PvA44DLgaOATVfXHbkOaqN5c+0wEx/cs4NFVdc5I2wlJHkvzS9zbRBCgqi5I8mmatRsfwaoPogX0OhGcZxa7wGsNjxfa7qNZPz+q6ntJXgJ8DNiJVedVQJ/vjFZVfbl9fGySC5bjhXADzPrv5n2r6imjDdUMIzosyU+7CWlyqurEucdJ/ji6PSP+Dtitqi5PcnPgPOBO8z7H9FnmJYEAVNVvmvsVvbbY+0fv31uq6vXA69uh5vsAX0xyLvDqqjpl8aN7oTfXPhPB8d1koTfPqjqn7/PrktyRphfif4B7VtUvOg5J6+bWSY6juUs495h2e8c1H9Ybt0/yQ5rz2al9TLvd+/lKSW5Jc/fw1sDuVfWDjkOapJvNm6e6yeh230cbMOO/m2vR+0/a8/T+w/UCLq+qywGq6rdJfjpDSSDAH5LcZf57ZpK7AJd0FNOk3CzJo2lqeYy+jwa4aXdhTVZVnZ3kk8DmwJOB27JqGHqf9eba5xzBMSX5blXdfV339UE7HO25VfW5rmOZhiSXsOoifwPg0rldNHdt+p7I/81i+/t+lzvJ9ovtr6pzlyqWaUjyM5qhWu+oGXtDTvLuRXZXVfV6eNoAfjffC5wFvHL0dzPJPwO3HZ231EdJbjGy+SXgAYwkuAv1NvVJkt8BXxlp+uvR7b7PgUxyP+A/aeZefbdtXgHsBzypqr7WVWwbai3vnVTVU5cqlmkY6Qnci6an+mjg01V1WaeBTUifrn0mgmNKcimwcqFdwK2r6oYL7OuFJNevqisWaL8f8MSqenYHYU1MkutV1Z+7jmMpJNkSoKou6joWjSfJlgv9vJJsSzM367UdhDV1Sf6yqn7VdRxas3a0y7to5sOPFov5Pk2xmN93FdsktDdhioV7N6uqet2rO+s3CaF5HwGezap5dKcDb6qqX3YX1YZL8pjl1Gs0aW2xmB8CnwT+wLwe+ar6f13EtRSW27XPoaHj26XrAKZlNAlMshuwL7A38DNmY37gt2k+yMyktnjDIcD/pRlGkiRXAW+sqsM6DW4C5vXoQtuTO/dv33t0R5PANpHfG3giTbWxT3QV1zQkuRnwWJr3mF1ozrG32uJat5hL1pNcANyY5nfzBVX11i7jm4CbVdXeSXYC7tC2nV5VZ3UZ1AQ9oO+9tmvx/TVVdp2F5SOSvKedw3pI17FMwcuYjc9fa3IYq67rN+oykKWwnK99JoJjmrtYpFmn7do7T1V1dndRTUaS29J88HwiTSnfD9P0Fj+w08AmZ9bmssz3PJrlB+5RVT+Da4ddvCXJ86rq3zuNbsN9EbgVzUXx6Kr6ecfxTFSSGwOPoblA3JbmPHesqm06DWxCkmxOM/xnX2A3mkTpUay/2T+RAAAgAElEQVQ+ZK2vngnsObJ9YVVtnWQz4LM05cP77Fjgbm3iNyvJ36hPMMM3CYEv055fki9W1YNG9h1L/8/9zl0HoPVTVS/vOoZp68u1z0RwTGtaTynJLKyn9GPgq8DDq2olQJLndRvSRG2Z5Plr2jkDQxCeDOxRVXPr8cxNwH4S8Dmg14lgVT0qyU1pkqV3tB+yP0yTFPZ6Dk/rQuA7NHeAv1ZV1RYJ6L12WZr70/wevhE4AVg5Uk2t71JVvxnZ/ghAW6Vx845imqRZv4k2pPO7xSL7+uoG7SimBc+l50vvzBWimm9uJEzvk+B2mY8Xs2q0wWnAEVV1fHdRTUafrn0mguOb5fWUHkMzafdLSf6bZtLuLFwk5mxMM/Rgls5p1PVGk8A5VXVRkut1EdCktXOR3t0Wr9iH5u9xM6DvSTw0F8J9gDcDH0ry4Y7jmaQ7AL8FzgDOqKqrk8zSxPSbjW5U1asBkmwEbNFJRJO1dZI3rGlnVT1nKYOZglk/v1lf3mRrmorLC87xpN9L7/yMZimvmZTk6cAzgBcCJ7fNK4DDk2xTVX1eAxJ6dO0zERzfzK6nVFXH0qxzckOabuyDgFsmeQvNIp99ryb6i1mYK7eIK9dzX28k+Suaocv3p1ko+NFV9dVuo5qMBdZTOhbYKsmLaP7+ftJpgBugqu6a5PY0P7svJPk1cOPlNll+A3wuyb9U1cvmtR9Gcye47y5jVTXGWTTr53fLdjRMRh7Tbm/ZXVgTs7Kq+pzsLebKGZ+/+jzgfvNG9ZzQ9hJ+Deh1Itina59VQ8fUrr+z8xr2rayq2yx1TNPULj67N/CEefMKeifJ96tqt67jmJYkVwN/YvW7onPFVDarql73CiY5B/gdTU/1CcBVo/t7PvxnQUnuRDOv4PGz9N6S5O40F8bHA+dX1V91HNIGaW+evRO4BzC3ltldaO5wP72qer2WWZLvVVXf55Gt0QDO79DF9lfVK5YqlmmY5Wt7kiOr6sA17Ft2ycS6SnJGVS1YhHGxfX21nK99JoJjmvX1lACS7Arcvt08o6p+1GU8k5Lk5lX1267j0PpJ8mVWDWOaX+q9ZuWOcFtVbO5m00/6Xpp/Me2w+vtX1bKaNL++2t7c0SJiM1FYJcm3qureYzzvjlV12lLENEnjnp+WpyQPGWfEUpKPVdVjlyKmaZlfdbKqlk3VyfWR5NvAAVX1g3ntd6FZU/ee3UQ2Xcvx2mciOKa1rKf0tKr6XVexbai2EMcngW1p1nUJsCtwLvConhfCmb/8wFwSUTRDozetqpkYIp3kgaz6MPqj5TgpedJmYY3IJNcH3kZTTexnNL+j29NUNHxmVfV2eG+SY6rq8e3jI6rqRSP7PldVD+kuuulol1rYl2Y++R3X9vxZ0NeetbUtodD3CsWLzX+EmZgDOZa+9hwuVnVyrlZFX6VZp/o/gXezanj2CmA/4ElV9bWuYpuEPl37ZuID8FJok6FZXU/plTRDmXYfKYSzEXA48Cqa9el6q6puPLqd5EY0C9A+gxlYpy3J1jRLDlzOqjfUvduLyKOr6oLOgpuC9o7a7jQXx4cDf9ltRBvspcD1gG3nhhK2S0q8iaYY1T93GNuGGh1OvwfwopHtWZijBECSrYAn0PxO7gr8K818z6HoayGuT7PAKAOa381b0hQa67PR+Y+vABYdKjrDetfj0aeqk+ujqr6W5F7APwBPaZtPB+5dVb/sLLDJ6c21z0RwTElG73bOfbC+6Vx7z+cpPRi48+gdpqq6JslLgFO7C2uy2qEVB9FUeP0gzbp7v1n8qF44EnhLVb1ntDHJ39NUotyri6AmLcm9aT5oP4qmFPqzgX/qNKjJeAxwz6q6dK6hqi5J8g/At+h3IrjYB7DefTibL8kBNPM+tgaOAfYHPtn3uVfroZc/y6radXQ7yQ40H9geDLy6g5AmqqreO/c4yUGj21r2elN1cn21Cd8h89uT3Leqvt5BSJPUm2ufieD4XrfIvr6XKb6yqq6a31hVVyW5oouAJinJFsA/0tyxPwrYbcbmX92hqq6z7lxVvS/JS7sIaJKSvJqmcNHPgQ/R3Nk+eYY+1FwzmgTOqao/zsCFf26dr42AzUfW/AowC+vsHQl8E9i3qk4GmIGf2eAk2ZmmZ/5eNNf65/R9yPkChvx72bse6z5VnVwfSTamKZyyNfCZqjotycOBl9BcG3o3lHee3lz7TATHVFUP7DqGKdpsDYuyBrh+B/FM2rnARTRj0S8F9m9GFzaq/wvKb7RQYzu8t+9DmwCeBvwEeAvwqaq6YsY+bFdbpXehDyu9ngcC/JJVaz2OPp7b7rv/RXOT4nVJbkXTK9jrKr2jkmxcVVeP8dRezmNtq/O+lGZu9WuA/cc8X/XLi9b+lOWnqn5MM5z30JGqkyclWVZVJ9fTu2jqUnwHeGOS/6GZI3hwu6RZ3/Xm2mexmA2UZA/ghVW1R9exrK8kX1psf9+T4CQvZ5G7oX0fxpXk34EbAQdV1Z/athsC/w5c3veCAO2dwz1oLoIPAr5EM3Rr24V6svumXR7jGtawKHJV3XppI5qcJDfpe7GpcSXZhmbUwROBG9KsAfmSbqPaMElOAZ5VVd/sOpZpaJfeOY9mruB1EsAZeO8cLZR2A5obodC811RV3aSTwCYkyQ/XtIvm/O68lPFMUpIDq+rIBdqXXdXJ9ZHkR7RTkpJsRpMc7TQj03V6de0zERxTkt2BtwJb0Sz4fARND1OAV1XVxzsMb0kk2aOqPt91HFpdkuvRFKd4Ck3vJ8B2wHuBl/S56uR8bYXNh7NqcfkvVtW+3Ua1NPpYoj/JWcBLq+rormNZSkluS1M19LCuY9kQbTGHN9KskfjCWVuGJ8l+i+3v+/DzWaiqvJj2RkXRzPn/FHDZ6P7q8YLsfa3EO6755zdr59una5+J4JiSfB94Hs18kIcBH6Dpwr7OHZtZ1ec/1CQPA17MqoqvpwFHVNXx3UU1WW2V0LnFx89aaN7ZLGkraz66qt7XdSxLoY9/f0m2B15P02P9rKpa2XFIE5XkB8DX269vVNXPOg5p4toeiGfSFGb6DCPDlfveYzaqrSZNVf2x61gmpY/vGetqZB7dI2iqTn4Q+FzfR4vM+s8uyaXA3PUgwE7tdu97c6Ff1z4TwTEtcPfizKq6XZcxLbUer8XzdJqlIl5Is0wGNGPRDwfeWVVv7yq2SUjyJJq/5ffPa38ycHVVfbCbyCajrX66RgNKBHv59wfX3oh5D3ASqycSj+wqpklo55j91cjXDWluFn4d+HpVfafD8CYiyV8Ar6W5ifY2Vv/59brHDCDJs2huEt6wbfojzU3CN3cX1WT0+T1jfSR5As2yO0dU1Wu7jmdDJLmKVUN5V9vFbAzr3X6x/X3uzR3Vh2ufieCYkpzN6qXqXwu8YG5jIENDe3mHKsnpwP2q6uJ57X8BfK2qdukmsslI8m3gQfPvZLfzBL9SVXfvJrLJSPLGNex6JLB1VQ2i6FWP//5uR7OMyW9pPqSNXgxP7CquaWgrFO9Ds0zNjlXV62JNSZ5Jc517LfC2mrEPDEleRpPAH1hVZ7dttwb+A/h2Vf1Ll/FtqCTns3qRitXMQKG0uXV09wEeTfMecwzN/Nxe9+zOehKf5N5V9a2u45imvlz7BvEBakK+QjP0YKHtolnQW8tT5ieBAFX1m9HqoT12vYUuelX1p3b+YK9V1f+de9wOU/vfNFXgvgW8qqu4tHZJDqdZx/J5VfXfXcczaW0ho91okon70gxvugB4J03PYN/dD7hPVV3YdSBT8mTgLlV1+VxDVZ2d5PE08yJ7nQjSVI2+ET1cPmEcSU4EbkyT/D0VmCs0smmSWyx03dey8WbgbgBJvllV9+k4nonq07XPRHBMVfWUrmNYBs7pOoD19Ickd6mqH4w2JrkLcElHMU3S5kluOFcxdE47h27TjmKaqCSb0BTD+SeaBPBxVXVmp0EtvT4W/bmKZt3Oyxd7Uo8LUV1CMy/pTTRzxmdqjmBVPWl+W5KdgH1piuHccemjmqha6Hezqi5L0velWwB+0feCRWuxPc2N+Ge0X8WqpLeA3lZcBj7SdQBTNnpzYrPOopie3lz7TATHlOT1VXVQ+/i5VfUfI/ve0/dEsR0muS9w+7bpDOBDo6V8q+oxXcQ2Af8IHJfk3cB327YVwH7AdT7o9NC7gI8meebcuPokO9B8OH1Xh3FNRJJnA88FvgjsWVXndBvRZCXZbrH9VfXz9t97L01Ek1NVLxvzqUcAfUwE9wfuQ7PW5VOTnETTE/jNqrqg08gmKMlWNEtj7AvsSlOleJ9Og5qMC5I8qKq+ONrYVgn/RUcxTdJM9gTOqaoduo5hii5KsnNV/bQdCXMU8FiaG/JPqarvdRrdhtsozfq5G408vvb3te+9uX269jlHcEyj83Nmrextkl2AE4DPAt+n+WPcjWbttt2rWdS019Is9vwPNAsHQ3sXv6qW1cKe66udy/NimmFA0BQ8OLyq3tJdVJPR3pm/ELiI1deDnJXqYqey+p1s2u0tgVv2fZ7ZOGZhPkySGwD3pBkm+lRg06patCDCcpfkAJqKjFvTDL87BvhkVe3YaWATkuSOwCeBr7H6TcL7Anv1bbmW+eaGRyZ5IKuufadV1aJrB/dJkk1ppgtce37AB6vqiu6i2nDtOnu7VdWfk+xLc0P7ITSfzQ6tqvt3GuAGygyvn7sulsO1zx7B8WUNj2fBK4HnVtUxo41JHkszB+uxnUQ1QW3Cd0jXcUxLVb0VeGs7HJSqus6Q1yT79bTK30x86FyTqtp1dLvtzX0R8GDg1R2E1IXe3pFsizLdi1XzBO9Bs0j517uMa0KOpOnh3LeqTgZI0tuf1XxVdVpb+XVfViUSXwGesbYhXT2xeVtM7HJWJbp7JzmCZumdXvdaJ7kDcBzN39rc+T0AeGmSvifyV9WqNSAfDryvHaH1hSSv6TCuiRi3Nzc9XD93HXX+fmoiOL7FurH7fsd+16p63PzGqvpYkt5/EE3ywzXtYgZ6lEYtlACOeC7NIvO9Mm4Z6b5POE+yM/BSmqTidcBzaoYXg54F7fqy29IsS/MNmp/bt/pesXDE/wL2Bl7Xjqo4Buh9AapRbcJ3VNdxTMmRwFuq6j2jje2SPG+mKWbRZ2+kWaNttaF1SR5Mc+4P7CSqybgmyf+iqTj5IFYvjLZ5NyF14v20RWU0HSaC47spzR2nueRvdHx25xn9BvrTeu7ri2tofkYfBD4FXNZtOJ2ZtZ7s+Xo54bztkXgpTY/Ea4D9q+rqbqNacud0HcB62g84dW3LKvS1N77tgZgbbbAt8HjgV0nOoCnR/5JOA9xASS5h4ev3TKzVBtyhqh49v7Gq3pfkpV0ENGFbL1Roo6q+sMiyQ31xCM0Npo2B4+Z6xZL8DXB2l4EtsVn/3HJO1wE4R1CLrTUU4KCq2naJQ5q4JLenmevyCJr5gR8EPldVV3Ua2BLq+1zWtenr+SW5mmYo4aeB6ySAVfWcJQ9qgsYpRDXr+vq7uSZJbktTNbTXFSmTHAvcimb5pw+PO/qgL5L8tKp2XqB9I+AnVXWbDsKamCQ/oRnRdMW89s1obtBc59z7pK2WfeOq+u1I2w1pPrv/sd3uvOrkNPX1vTPJC6vqNe3jvavqIyP7Xr2cbqJt1HUAfZHklklen+S/krw6Sd/vFI56B81aPPO/bkSzHlbvVdWPq+rQ9g3lU8D7gOd1HNZSm/U7a331f4BDge/QjDqY/9VbbSGqHwF3B34C/JRmDt2p7c2Zoejl316SnZN8MsmPknyoXbybqvpJ35NAgKp6FPBQmkJUb09yYpJ/SHKLjkOblP9K8o42eQCuTSTeChzfXVgT8z7gY0muLcrUzrE+hmZIYa9V1VWjSWDb9qd5Q8+PWOKwNJ7Rqsovnrdvz6UMZG0cGjq+99F8KHsjzcTdN9Csa9Z7VfWKrmOYtvYDzD7Ao2nG3D8P+ESnQS29WShesZheftgeHTKY5EZt26zMMZv5QlRj6uvQm6Norn1fAR5Jc/3r6zJCC6qq3wPvTvJemmvEG2iGmS80SqZvXkiz1Me5SeZ6O7ejmSu+bHok1ldV/UuSA4GvtlV7oZnO8m9V1fehoePq5XUvycZjToHo4/q5sHiByWX1M3No6JiS/KCq7jKy3cvu6oUkOaaqHt8+PqKqXjSy73NV9ZDuottwSU6k6eE8BvgYsNqQtL6vV5PkEcAPR9YQPITmA/a5NB/Ce73IdZLbAH9ZVV+f135f4JdVdVa7faeq+lEXMW6oJM+iuWs4d+f+j8ARVfXm7qLacEnOrKrbreu+WbMcSoSvjySnVNVdR7Zn5ro3J8lf0UwbuD/NMhIfrqqvdhvVZCXZHJgbBnpWVV06b3/vhxcuVjF7lvX1bzLJKTSFfr7ZdSzT0Kcl5+wRXAfzK4WObvc8mRgdR78HTen6OVsucSzTsD3NHflnAAeMtKdt7/t6Na8C7g2Q5OHAk2g+2OxGMwTood2FNhGv57pDKwD+0O57BECPk8CX0Sw98ICqOrttuzXwH+06YP/SaYAbZtYLUY2rr73xmyXZjVXXvc1Ht6vni1q3a5n9Djia5tpwVdt+N+j/+c2pqsuAUxd5SueLWq+vJBsDN6+qX7fbm9KM1npeVe3SZWxa1DOANyb5AfDC+UNgZ8BdkvyB5r1y8/Yx7fayKmxnIji++VVDYVXl0L4nE4t1C/e+y3gA69XUyB3exwDvqqrvAt9N8g8dxjUpf1lV1/kQU1WntvNB+u7JwF1G1y2rqrOTPB74AdDnRPCWSZ6/QHuYgZtMSbYBdqiqr7Xbz6eZWw3NotYrAarqwI5C3FC/YPUhkr8c2S5g9yWPaLLOoTmPh9Is1j16fZ+F8xvXshqqNq4k+wBvA/6U5Kc0N0WPAk6iWWR+CM7pOoD1UVXfTnIv4JnAyUk+Q1PhfW5/r4ukVVVvlpUzERzTjCcTN2jv8m7E6nd8g+vV9EHauWWX0qw3NDqccFndeVpPN1tk3yz8ftZCi1dX1WVJrlnogB6ZK0S1kFkoRPVa4D9Htp8BvB24AfAKev5htKrGWoetr0MLq+oB4zyvr+e3Dvp6w/dlwN2ramXbi/tN4HFV9amO45qaJHvQ9KDtAVBVfZ6zewua4mEX0XS09P16d60k9wC2qKrPzGt/GHBhe7N+WTARnLw+JhOjd31H7/jObQ9FL++K0gyPPIVmqOQZVXUyQJvQ/6LLwCbk5CRPr6p3jDYmeRo9r6rZuiDJg6rqi6ONSXan5z+/ARSiul1V/dfI9qVV9TqAJDM1z2wteju0cEyzfn59deVIr/v32uUyZiIJbN//3wpsBRxL8zv4bprPKa9a5NBeSPJM4AU0N9P2r9krWHIE8NQF2k+n+Tkum9EGJoKT18dk4sVV9a2ug1gGevlGVFVHJfkscEuaoYRzfsnCb0R9cxDwiST/m1WJ3wpgU5oqsH33HOCTSb7G6ud3X2CvzqKagFkvRMV1e9wfNPJ4i6UMpGN9vO6ti1k/v3O6DmA9zR96frPR7arqc+XX19HMW/0m8LD234Or6shOo5qc+wH3qaoLuw5kSm680LqkVXVukmV1bXAdwcnrYzLR68qEQ5fkoTRvqN+vqtGhFfcFel+Vsap+VVV/RTPU7pz26xVVdZ+q6n2PdTuU/E40Jfp3aL++Atyph8PM55tfiGpU7+cIApe0i6sDq4qGtWskDql6YR+ve+uil+eX5IUjj/eet+/Vc497PLxw/hrI87f7rKrqy1V1RVUdC1wwQ0kgVfWk+Ulgkp2S/HOSvl/3AG6+yL4bLLJvydkjKJj9u53j6ut6NYcAj1qg/cvAp5itIU0179+Z0M4RPKrrOKZgpgtRAYfSLNr9KlYVD7s7zRptz+0sKqmxD/Ca9vGLgY+M7NuTnq8lOONDz2+WZDRB32R0u6o+3kFME5dkK+AJwL7ArjTrXu6z6EH98IX2uvCyuWGvSUJzQ/uETiObx0Rw8vqYTOyY5Lg17ayqRy5lMJOWZLvF9lfVz9t/7700EU3c9avqovmNVfXrJDdc6IA+SbI18HHgclYNndw7yRHAo6vqgs6Cm4Akl7BwUhSau8I3WeKQJmmmC1FV1X+3H85eSDPEF+BHwGP6upzJejqn6wCm7JyuA1hPvVnUen2MDi9P8uKq+teuY5qgr9AujbTAdtFcE3sryQE0y1xtTbPG8/7AJ2couf9HmoJoK9s1EwHuApwMPL2zqBbggvJjGjeZ6KO27PLT1rS/qk5cwnAmLsmpNG+c80uDbwncsk9lfheS5CfAHarqqnnt1wNOr6qdFz6yH5J8guYC8Z557X8PPLaqej2PbpYl+dJi+8etSqluJHl1Vb2kfTyTlTOT/AVNb8Tt26YzgA9V1W+6i2oy+rSo9fpI8v2q2q193PvzGZIkV9LMe/zHkQJ3Z1dVn5diu452TeA7tpunza0VvJyYCI5plpOJ0TfTIWjXnnsR8GDgDVX1xk4D2kBJDgf+Ejiwqv7Utt0I+A/g16MFOvooyZlVteBcx8X29UWSGwB/rqo/t9u3A/4WOKeqPtFpcBsoyb1nuRBVknez5iGuVVX7L2U8k7ZYIjELkuxCM0zrs8D3aa7vu9HMZ929qn7cYXgbLMnVwJ9Y1QM/t95sgM2q6npdxTYJs/z7meT1VXVQ+/i5VfUfI/veU1VP6Sy4CWhvwOxN0yt4K5pewadU1badBjZl7ZzyF1TVsukVdGjomKpq19HtecnEqxc4pE8u7jqApZBkZ+ClwL1oKnI9Z+7Dd8+9jGbR8XOTnEtzkd8WeBfwz10GNiELFrVKshHQ2xswI/6bZljMT5PchuYu6X8CD09yr6o6uNPoNsyb6d9yOuvivxZo2xZ4HrPxuznrXgk8t6qOGW1M8liaEv2P7SSqCenzDeox3bqd1pKRx9fq+bSWvx55vB/Njd05d17iWCau7XF/K/DWJNsCjwd+leQM4BNzIxH6KsmdgX9j1fIfbwKOZNXnz2XDHsF1tEAy8d6+JxOzdidtviR3ovmZ3ZFm4vyHqurqbqOavCSbA7dpN1dW1WVdxjMpSV4P3BA4aKTH84bAvwOXV9VzFjt+uUty6tyNpiSvBG5RVc9Osinw3fk3ofpkSKMN2iFAL6H5APfvwLuqqo9zxq+V5HyadWVDk9yuVo6/5+X5hzDaoDeLWq+PJH+z2P4+T2uZN+x1tffRWf7M1vaY7VNVh3Udy4ZI8m3gLaxa/uPFwHuBQ9ricMuGPYJjWiCZ2H8Wk4kZ9QPgPODTwD2BezbFmxozkEgsVPp757lznIHqYi+g6XWf6/EE2I7mTbXXdw1bo3fjdqdZYJequjLJNQsf0hszXYgKrl0q4mU0QwpfCzxz/nzdHpsrxz//8az403ru64veLGq9PsZN9JJ8rKr61ru7UZKb04yImXs898Gl9z29bafKvwE7wf9v787jJR3P/I9/vsTSBCFDD7HvQ9NEC5osiElmYoKIhMnCEJEEsWWsCTPJTGdBwpBkhgwhP2KJYSSyEGJfW1saSTB0BEnEFltE4/v7476Lp6vPqT7ndJ2663nqer9e9fIs5+irXlWn6rme+76vi5nA52w/YvteoNZJYLZIpa7BryV91vahnX6hlEgER67JycRcUyqqGnCxtmfpAMbZP3Q4V/vqYsBGpJGIo0kjnu8iPeeFgTdS/6nNd0o6DniE9PwuBZD0pqJRdccf6bNpMN0k6XxSu4jjSSNmrwBLVm7C1Pq92aAKfsNpb0jeIprR57I2Ta3HWR0LkCxFqpLdutCcUTnXhKl8pwFnkqqhvh84CahrP8uhLFqpkg3wl+q+7RnD/maPxdTQEZK0e6fzts/oVSzd1vSqoVW5iAq2nysdSy9Immj7D6XjmB+SZgDvtv2kpHcA5wD7kxLEv7H9waIBzqc8pfcAYHngNNt35ONTgTVsf69kfPOj6VNDJc1izt6WcxQTa0IFvDyN8AhgvXzobuCrtn9cLqrukHRMp/N1T4Ql3W97zdGea5o6TqWUtGCTZ51Jut32RpX92r1GnUi6ks6FxPpmND4SwTFoWjLR9Is1AEmfJl3MtPrqPUe6mPlWuajGRx5J2plUEv1vbK9QOKT5IukO25Pz9jeBP9r+l7w/x5dJ6C+SLre9bek4wthI2hvYh9QncXo+PAX4CvAd26eUii3Mm6T/BJ5g6KbWf237kyXj65U6JhlKvec+bfuG0rGMB0m/IlUMbd08O4t0zdJ3I2ZNF1NDR6E9mZDUlGTiwdIBjCdJnwemAu9q9XDJhR1OlLSM7X8rGmAX5FGlHUgfpBuT1vLsSJp2UXcLSnpDXne1LVC9eKn9Z1ilNc1cp0h3DutcIW7p0gGMJ0kdLy4bcDFzELBV2xTXK/Io4bVArRNBSefZ/lDe/mq11Y4qzcprrDZNrceZ5v0jfWcf4CRJdwCH2n6qdEBd9jvmLD71+8q+qfn61SFqNxh4HLjd9rMFQhpWjAiOUCWZ2K89mQBuqnMykUtlD/tGqHuxEUm/Bia3V2rKydMdttcuE1l3SDobeDtpbdk5pL5Y99terWhgXSLpKFJfvcdJRWLeatu51cIZtrcsGuB8krRKp/NDrfGpizreiR8NSb/ocLqvpv+MhaRf2v6b0Z6rC3VoSN6kmTKqQVPrsZB0GHDcvKZQSvpb25f2KKyuyaO3nwI+B/wEeK14WM3rUoyYpO1sX1Y6jtFS6jHbbhlS64+9bF/R45CGVfu76T30MdqSCdsPSPoQqZBMbRNBYPsO55pQbMRDleu1/ecGVGWEtHbnKeCXwC9tvyKpMXd4bP+7pMtJa+gubU1xIlVT279cZF2zEDDR9nXVg5K2JN0lrbNGF6KyvfVw5yRt3stYxskzkia31q22SJoM9NVd7THq9DnZpM/QB4DXkj/1YVPrMVoJuFXSvu2fn1V1TAKzZYBNSUW3bqWSCJgdg1QAACAASURBVA6QrwK1SwRtD1Wtt3Xj9zxSC7q+EIngyDU2mRjuDdsgj0ja1vbl1YOStiFNT6g12xvlEva7AT+X9DiwRBMKxbTYvnGIY/eWiGUcnECact7umXyuU1XYftfoqqHzcB5pBLvODgEuzne3Wz3nppAaXH+0WFTds1iu5LcAMKFS1U/AhKKRdYFq1NR6LGzvl6dnn6zUiPzbzDlqVtup2ZI+RWqddCxpBKkxNyZGqY7TeoeVK/YuVDqOqkgER66xycQw5bNf45o3DQY+C/yvpGuZ82JmS9K6utqz/SvgGOAYSZuQ1greIulh21PLRhfmYaLtme0Hbc+UtGrvw+mq55pUdXiUan8BY/taSW8D9gX2yIfvATa3XffRaphznVJ1jVJrv+5OZc6m1reT+q9+ZKgb23Vke4akI4ELSD3pqlV86zw1eytgC9uPlQ6ksEYlwJLWAf5SOo6qWCM4QpLWB/6XtEB+rmTC9t2lYptfeUTzdtIc9L/QdgFT9xLaAJIWJSVHrXUS9wBnNeXLcCh5fcHbbTehYExjSbrP9lrDnKt1iXdJ/2O7Sb2hRkzSQ7brPiI4Iqpnw24kbT7UbIOmGKJE/wNNaGnSImk50sjm6sBn2qcw152khYGPUFnfCZxtu68SifFU13Xmkn7I3EnsMqQlLh/tp2qwMSI4QrbvljSJOZOJq4F9GpBMbEyaVvg+UpL7feDyJk1FyK/RaaXjGA+Sjp7Hj0Qi2N+mS9rb9qnVg5I+wes3nerqrCGqp72mAYWohvqyh3Qz7c09DqekuiYX3wJqd5E5CrVpaj1GNwFfBj7epOsVAEnrARcD1/H698C7gKMkvd/2PaVi67FZpQMYo+Pa9k1q5XKf7ZcKxDOsGBEMc1BqYr0b8G7gMNvDFnqoC0nP0rk8/5I9DqmrJB0yxOHFgb2AN9t+Y49DCqMgaSJwIfASc842WBjYqc5T8IapnNZi23v2LJhxIOmdnc4PyrTYGt+1b0xl0KHUqan1WEha1vYfhzi+ErCr7WMLhNUVuUDaV9orZkp6N3BUp0JVdSBpmu0j83YtK4OOhKTVqMxE68eKvZEIjlDTkwlIH6rAh4BdgNnAF5o8baaJJC0BHEBKAs8Djo81BvUgaWtgUt69u5/KS4fQSY0TwafpMGOi7lVtB0m+ftmFdCN7BeBC258rG9XYSfqV7XWHOdeE1i2vfWbU9fOjk3wt9t/AJqTOAgAbkW727mX7mVKxtYupoSNke4nSMYwXSXuSEsBFgR8AH2pS8iBpMWC27dl5fx1SX7pZti8sGlyXSFoGOJi0nuAMUq+9pjWgbSRJ29i+wvYvJM2y/WDl3AfqPH2y6YWoJM2k8w3CDXscUil1LYzT6Kq2dWpqPRb5YvsDpCU7a5NaXa1me8WigXXHApIWaV8PmOsdxLV7/zuJVItiV9uvwmt1G75Aqtz78YKxzSFGBEeoyclELhZzF9BqXD3Hm6Lud0UlXU26A3NfbkJ+M3AWqf/eLbYPLxrgfJJ0LOnL8BTgm7afKxxSGIVOd0brfqe06YWock+oYdn+Tafz/U7SksPduZa0su2H8nZdG3Y3fWpobZpaj4WkP5O+zz8PXGvbTSmII+nzwObAvq3PkVxF+j+A6ba/WC66+SfpYVKVXgEHMWfF3ibcJOxUBG7YcyVEIjhCTU4mmr7ORdJM2xvk7S8By9jeN1fkurV1rq7yxfZfgJeZM4lvzLTlJqtejLZfmNb9QlWp8fhuwHtpaCGqoUjaCtjN9r6lY5kfbTcpLre97VDn6qr9OQ2KfAPjPNt909R6LCQdCOxKWhP/feBc4LImJIIAkvYDDgUWy4eeB46zfVK5qLpD0jGdzjfgJmFtqoHH8PLILW37vry9O/B92/u3kgmgtolg3RO9EahedG5DatCK7ZdyElVrthcoHUOYLx5me6j9Wsnl3O8ADq8UojpJUiMKUVXlaoz/SFqn9CBpmlrdVUdwl+lwrq6WLh1ACe7DptZjYfsE4ARJq5MSwouAFSQdRlojeG/RAOeT7ZOBk/MUWFrTeSXtbPuCosHNp7oneiNwfa7o/qXqjU9JXyD19ewbcQE5cu3JxGWQkgmg1smEpLUknS7p65JWlPQTSc9JukPSlNLxdcGdko6TdBCwJnApgKQ3lQ0rBABWl3RxbkXQ2m7tr1Y6uG7IhRw2BjYAHgYasQZZ0tqSjpH0K9KakIdIM222bsJdexp8k2KQqQ+bWs8P2w/YnpZn90wBlgR+XDisrrH9bNuazm8UC6aLJP2dpKslPZ4fV0n6+9Jxdcn+pO+7+yVdkB8PAJOB/cqGNqeYGjpCkv4f8HvgEdLo32q2X8jJxFW2JxcNcD5IuhY4k/TheRBwIPBD4O3AvzVg+sgEUiXN5YHT8ihFq1XGGra/VzK+MNiaPDV7iEJU5zWsENWrwDWkZQP352ONWKMEHdfxCDjQ9kqlYuuGplcNVY2aWo+VpB1JN3hn2v5Z6Xh6QdJvG/C3tzewD2nq6/R8eArwFeA7tk8pFVs3SVqDtIQMUvuI/ysZz1AiERyhJicTkm63vVHenmPucvVcCGF85ZEzPERvrDoagEJUO5KmpG0J/BQ4h3QR05SR3Mav4wE+Mdz5Ot+EgSFvMvVtU+uxkPQtUo+264FtgR/a/lLZqMafpIdsr1w6jvkh6R5gK9tPth1/M6nwT93bY6wCPG37T3l/a2BH0nfhyf309xeJYGh01UKIEu+hv+WS0keTppIsQHpfvgyc1IDKcI0d7ayStDiwA2kN5DakGRYX1rGSZpWk/fI6pUaqezGmkVINmlqPhaS7gMm2X1Gq7H6N7U1Kx9UN87huWdv2Ij0OqavUoRdip3N1IekmYCfbj0raCPg58GVSxd7Ztoe9AdVrUSxmhBqeTKwr6U7Sc1kjb5P3mzDFafvSAYTQwUHAVsCmzj0Ec/GDb0s6yHZt14M0JdGbF9vPA2cDZ0tamlQw5jDyeuQa25PU86qpHpz3j9SXhmlqLanvmlqP0Uu2XwHIS3WaUMCopenXLc9ImtyaXdeSK03XvsclMMH2o3n7o6SZhMdLWoDUUqlvxIjgCKnB/aKa/NwAlNp9TLR9XdvxLYHf9+Oc7TA4JN0GbGf78bbjywKX1nnEQtJawJHAU6T1ZaeS1h7/H+lCdHqHX+97kn5MSgAvcgP7dzZhRkgnknamQ9Eb27Wu/Crpu8As4Iueu6n1mrb7pqn1WEh6Abi/tQuskfebcIO+0esflVrsnAWcTqq8D2mN4O6k9avXloqtGzRn27IZwBGt11DSnf303owRwZFbiA7JRJmQuqPuid4InAAcMcTxZ/K5f+htOCHMYaH2JBDSOsEGlHg/ndcLUd1EKkS1EykZ/CZQ60JUwH+R1gh+XdKVpF5ml/TT+o/5tKGkoUaNmtKjtNOoi6l/C5Atbe9RPZBL2X8xr4+su1pPH+ykbf3jlyS9rUnrH21fK+ltwL7AHvnwPcDmtmt9TZ1dIek84HekNjVXAEhaHuir74cYERwhST8iZfQz245vAEyzXdtkQtKzzF0m/HHgF8Bhtp8oEliXSLrF9qbDnHvtrk0IJXQadan7iMygFKLK65P+gZQUbgH8BDjb9mVFA5tPg7KGrqlUo6bW40nSDba3KB3HaDR5/eNoSLrA9s6l4xitPPL+YVKByfNsP5KPbwws108jvDEiOHIT25NAANszJa3a+3C6x/YS7cfyOpc9gP8krXeps079Aif0LIoQhja5w6jLor0OpsuqPVbbn2Ot+69W2X4BOBc4V9KGwBnAx4EFiwYWOpJ0cKfztr/e6XwN1Kap9Tir4+dok9c/jkYt61Tkv7dzhjh+W3W/H25SRCI4cgOVTNh+CviGpI+VjqULpkva2/ap1YOSPsHrc9NDKMJ2k5OFpheiAkDSRFK/xF3Jd4B5fbpTnZ0/3AlJm9q+pZfBjIPjSIUbfkJqsN60i+39ScVi7pfUKlCxMTAD2KtYVL1Xx6lv67Z9Xq5R+Syt/frHUajjazcaxW9SRCI4cgOXTOT1SU14jxwIXCjpI8y5KHlh0nqlEML4aOwaHgBJnyQlf+sAFwD/bPv6slF1j+1p1X1J65FaZOwGPE36HK2zjUnP5X2k74bvA5dXR8/qLFcF3aWtqfWhUSCtFhr92RleU/yzJtYIjlC+43shaZHnXMlEnRe3SvrAEIeXJs1vvrbuvcxackPPSXn3bttXlIwnhFBvkk7j9eShMVNdq/LSh1byNxtYBZhie1a5qLpP0lTSc3w3aW38xYVDmm91amo9nmKta301/bXrhzoATRjt6QnbfwCmtiUTlzQkmWgvdGPgCeBE25cUiKerJG1j+wrbv5A0q9WrLZ/7QN1LhIfQr5peiAq4olUQRtKW1arSTWjGLukGUsXXc4Cdbd8n6cEGJoHLkkYHNwAeBh4rG1HXnEea9fKn3NT6fFJT68nAt4C+aWo9FpI2t33jCH60dktchvjsVN5vRMVeSUsO18dS0sq2H8q7h/UwrBKKT0ePEcERaiUTeXu1SCbqo3rHpf3uSz/cjQlhkFQKUU21XetCVE3/bJF0EfBW4GJSFdTrJT1guxHrOyXtSVrbuSjwA1J1v6YkgXP0K5N0HPCq7UNbTa3rvs6s7e+veNGNbsp/e39NamFyTiUxaoS21+5y29sOda6uJB1Iav0xw/bLHX5uku27ehfZ3BYo+Y/XzHGV7Qvazn2+l4F0m6S9c+NnlJwm6U+S7pRU6z/GTMNsD7UfQhhHtp+y/Q1S8+e6a/Rni+0dSaNktwL/IulBYOnc/6sJvgOsADwLvAf4jqSLW4+yoXVF9T24DXA5QIOmMVefX/GiG92U//beA/wROFXSVZI+I2mZwqF1S/W1a39Otf/sBFYk9al+LL920yRt3/76lU4CIaaGjkaTv/APAL6bt3cjTRtZnTRV5kRS8+c6a5+aNty5EEIPNKgQVeM/W/L6stOB0yUtRxpB+0aevrVS2ejm29alAxhntWlqPUYL5BkGC1S2X7ses/1ksci6oPW3J+kMUlGq/yAlvHVvawIN/+y0/TkASQuT6olMBf4JOEXS07bX6/T7vdSEL+JeafKb9mXbs/P29sCZee3OzyV9rWBc3bJ6vruryjZ5f7VyYYXQbPMoRPWDHoczHgaiPUZLnjZ5MnByLkRSa7avKh3DODuQ15tab1X5nv9r4KhiUXXPUqTR6lbyN6NyztT8b7BSwOjtwLWkwoTXlI2qa5bLfTxV2SbvL1surK6bQFpnvVR+PArM1ZO8pFgjOEKSngauJr1J3563yftb2V66VGzzS9IMUvnsp0jVxLaxfXc+90vbtS5jLOmdnc4PwMVACEVIOr3tUKsQ1ZUNKUTVMRmy/ZtexTJeJG0BvAW42vZjkjYEDgfeXvcRwbwk4kjSd9/XgVNJ3+//B+xle3rB8HqmaevrmkDSLFKLlnNII7lzrDOzPWOIX6sNScd0Om/7X3sVy3iQdAqwPmna+U3AjcCNuUd3X4kRwZHbobJ9XNu59v26ORqYDiwIXFxJAt8JPFAysG6oJnq5Ohy2/1guohAGg+1/Kh3DeBou0cvFOHYj3VirLUnHkmaJ3A4cJulnpEqTXwb2LBlbl5wOnEm6Y38TaQRtJ1Iy+E1gs3Kh9VRt19flqXcfIV10A9xNKmz0l3JRdcUs0o2z9+RHlUlrPuvsibpXVZ6HlYFFgPuAR0jViJ8uGtEwYkRwDJqYTEh6A7BE9W6FpMVJ75Hn8v52rVLpdSJJpGR3f9JaApHurp3UlB6JIfQjSXuTRv/uy3+H/w3sTEqQ9mjAXe0lgX1JI2YXA5cB+wGHAHfY3qHDr/c9SfcAb7X9Yl5/9VtgUlPaR0i63fZGeft+22sOda7p6lqlUdJ6pL+763i9v/MmwJbADq2b2qH/1PU9Nxr5O2990vrAqaTWc08CN9juOCLaS1E1dIRyNc1jJD0O/Bq4V9IfJR1dOrZusP1y+5C17edbSWD21R6H1S0HAVsBm9peJk/j3QzYUtJBZUMLodEOIN3ZhjkLUR1MKkRVd98D1iGt+fgEqT/iB4Ed654EZi/afhFStVfgvqYkgVm1emZ7T7OmVNZsspOAT9ve3fZ/5MfuwKdIa1lrS9Khle1d2s5N631EYbSc3AX8GPgJ6YbFGqTvxb4RI4IjlBey/h3wyVYPQUmrA98GfprLoTeapNtsb1w6jtGSdBuwne3H244vC1xax+cUQh20jbicDdxk+8S8X/s7wpJm2t4gby9Iqs64cit5qrvK2viWd1T3bb+/50F1kaQXgPvJxX7yNnl/dduLl4qtl2r83f4r2+sOc67W9Q0GoEfpy8ALQ50i5VBL9jikrpL0WdLI9BbAbFJPwdZjZj+1cIk1giP3MdqSCdsPSPoocCnQ+ESQ+lZHXag9CYQ0tTeXsQ8hjI9Xc6n6p4BtgX+vnJtQJqSualVhxPYrkh5uShKYtY9qHl8kivFT20RhJEba1Jp0fVNHC0hapH09oKRFqf/1bZNblkFKhmp382EUVgXOBQ60/bvCsXRU9z+UXopkor469UtqQi+lEPpVowtRAZMlPcPrF2YTKvu1v6s90orKki6wvfN4x9NtTajqOg+tptbrSppJmpp2PXB9tcdePzS1HqMzgQsk7dt6LSWtSuq3972CcXVDk1uWDYKjgH2AbXJbodPmcTOmmJgaOkKdhuKbMEw/EpL+x/ZQfcH6mqRXgOeHOgUsajsS+RDGSZMLUYWkxlMLn2XuC+7HSWs9D8v9dGuvran1FvnRV02tx0rSfsChwGL50PPAcbZPKhfV/Ktct4g0e6I1jbIR1y2SjrQ95FpHSZvavqXXMXWTpHNJAw3XkpaV/cZ2X60NbIlEcISanExIOtT21/L2LrbPr5ybZvvIctGFEJqurjfT8hS0TwFrAn1913c81fX1G0qujroHMNX2LvP48VqQtBQp+WutWXoTaWpeY9q7SFoCwPazQ5zb3fYZvY8qjFSuALtbfjxte0rhkOZL2/rxNwA39+tnZCSCofGLkkMI/a3GI0rnktYJXkOf3/UdT038nmjCc6pTU+vx1ITXsonyNN5W8jcbWAWY0oTKxHW6lo41ggGavyg5hNDf6npHcr3KXd//Bm4uHE8pjfqeyOv+m3B9VJum1uOsUe/PJpB0A7AkcA6wc+41+2ATksCstX4c8vTefl0/3oQPujD/YlFyCCGMXrVq6Mupf3BzSFrSdnt/vda5lW0/lHcP62FYXSNpqDXvSwMfBn7Q43C6zvZ725paHwJMktR3Ta3HWVzH9J8/AG8BJgLLkm5WNOZ1sr1g6RhGKqaGhsYvSg4h9LeGFKKqfn723V3fsWhbNnC57W2HOldXkk5vO2TgCeBK25cUCGncSFqRtEZwKrA98GbbbyobVW/Udep50+W1qx8gTQ1di7R29T22B3VmRRGRCIYQQhgXUYiq3qoX0O0X03Fx3f/q1NR6PEk62fZ+peMIw5O0HPAhUlK4su2VCoc0MBYoHUAoT9Ji1V6IktaRdJCknUrGFUKovV0r20e0nXtvLwMJY9LoZQOS9pa0Vt6WpNMk/UnSnZJqPdqZrUpqar2Z7TVsf8z2t23f0ZQkUNIkSWdKmp4fZ0jasPozkQT2P9uP2T7Z9pbAVqXjGSSRCAaAn5K+MJC0JnADsDqwn6SvFIwrhFBvUYiq3paTdLCkQyrbrf1lSwfXBQcAs/L2bsBk0nffwcCJhWLqpqNIBWOOkvTJXMa+MSTtAFwIXAnsmR9XkZrM71AwtDACkraQ9ME8GoikDSWdTeq9F3okEsEAsLTt+/L27sD3be9PKof+vnJhhRBqrtEjSgPgVGAJ4I2V7db+dwrG1S0v224V/NkeONP2E7Z/DixeMK5u+S6wCTAT+Hvg+KLRdN8Xge1sn2b7zvw4Ddgunwt9StKxwGnAzsAlkv4NuJTU5mStkrENmkbdHQpjVr0g2wY4FsD2S5IaMX0khFBEq4R2tXw2eX/RcmGFkbD9r6VjGGevSloeeArYFvj3yrkJZULqqqa3N3nDUO0GbM+qLncJfel9wMa2X5S0NPBbYFKD2kfURiSCAeBOSceR+gytSborg6SBqCgWQhgfdSqhHeYmaX1gDdsX5/1vAEvl0yfbnlEsuO44GpgOLAhcbPtuAEnvBB4oGViXNLq9CfByWxsTACStArxcKKYwMi/afhHA9lOS7osksIyoGhqQNIG0VmJ54DTbd+TjU0kXAd8rGV8IoZ4kLQbMbk2/k7QOaYraLNsXFg0uzJOkHwJftn193r8H+AKwGKkJ9I4l4+uGvG5uCdtPVY4tTro+ei7vb2f7slIxjtUAtDfZEfgaMA24NR+eAhwOHGb7olKxhc4kPQ1cXTn0juq+7ff3PKgBFYlgGJaklYBdbR9bOpYQQv1IuhrYy/Z9uRDVzcBZwHrALbYPLxpg6EjSdNtTKvs32t48b19reyCq+zWhZ2JTSZoMHAKsnw/dAxzXuqEd+lMedR+W7at6Fcugi6mhYQ6SlgV2IVVQW4FUkSuEEMZiyEJUkhYm3cGPRLC/LVHdaSWB2XI9jqWkxs2pbIqc8H28dBxhdEaa6Em6wPbO4x3PIItEMCBpCeADwD8CawP/A6xme8WigYUQ6i4KUdXbo5I2s31T9aCkzYFHC8VUQkyd6kN56vKwr01ML2yE1UsH0HSRCAaAx0hTtj4PXGvb0Uw+hNAFUYiq3g4DzpX0XaBVGGYT0ujuh0sFFUJ2XOkAwriLmzDjLBLBAHAEsCvwLeD7ks4tHE8IoRn2JhWiWhX4W9sv5OPrERdxfc/2zZI2A/YD9siH7wY2t/2HYoH13qzSAYQh/ZPtPUoHEUKdRbGY8BpJq5MSwt1IDT2PBi6yfW/RwEIIjRKFqOqtKa+fpENtfy1v72L7/Mq5abaPLBddmJco4tN8km6zvXHpOJpsgdIBhPIkrSlpS9sP2J6WG9BuCrwX+GXh8EIIDSBpWUmfkXQNcCUwsXBIYRQa+vrtWtk+ou3ce3sZSBiTxSRtLOmtQz1KBxeGJ2nY1iWSVq7sHtaDcAZaTA0NACfQ9iVoe6akA0n9eUIIYdSiEFW9DcDrp2G2h9oP/ectwPEM/VqZVKAq9KcrgbcCSLrc9raVcxe1ztm+tPehDZZIBAPARNsz2w/avlPSKiUCCiE0QhSiqremv34eZnuo/dB/7rcdyV49VZP3ZTqcC+MspoYGgE4V/Cb0LIoQQtMcASxCKkR1hKQ1CscTRqfpr99kSc9IehbYMG+39jcoHVwIDRY3YfpEJIIBYLqkvdsPSvoEqelzCCGMmu0TchPyHfKhi4AVJB0qae2CoYUR6PD6HdaE18/2graXtL2E7Tfk7db+QqXjC/M05PoxSStJ+udeBxNGZTlJB0s6pLLd2l+2dHCDJKqGBiRNBC4EXuL1xG8KsDCwk+3fl4othFBfktYkTT2/rnJsA+BE4J22FywWXJinvE78OuA22y/nY5NIlaU/bHvNkvHNL0mLAbNtz8776wB/D8yyfWHR4MKoSFoW2IX03lwBuND258pGFYYj6ZhO523/a69iGXSRCIbXSNoamJR377Z9Rcl4Qgj1JulHwBHta5AlbQhMs719mcjCSEg6DpgKrAvMJCWF1wPX236yZGzdIOlqYC/b9+WbFjcDZ5H6XN5i+/CiAYaOhilm9OEGFTMKYdxFIhhCCGFcSLrF9qbDnJuZW9WEPidpYdIskanAFvnxtO31igY2n6rvQUlfApaxvW9+vrfG+7O/SfozcxczesD26oVDC/MgaX1gDdsX5/1vAEvl0yfbnlEsuAETawRDCCGMlyhE1QwTgCVJF2pLAY8CNxWNqDuqd8K3AS4DsP0S8GqRiMJoNL2YUZN9BXi8sv8e4BLgF8DRRSIaUNE+IoQQwniZLmlv26dWD0YhqnqQdAqwPvAsKfG7Hvi67aeKBtY9d+bpr48AawKXAkjqdAMj9AnbJwAnSFod2JVKMSPSGsF7iwYYOlne9vWV/WdsXwAgaZ9CMQ2kmBoaQghhXEQhqnqT9FPgr4C7SEngDcBdbsiFg6QJwAHA8sBptu/Ix6eSpq19r2R8YfRyMaN/BD5U92JGTSbp17bXGebcvbZrX5W4LmJEMIQQwriw/QdgalshqkuiEFU92H6vJJFGBacChwCTJD0J3GC7Y+W/fmf7z6Qpau1+C2zZ43BCF9i+S9IXgHtKxxI6elTSZrbnmGIuaXPS1PPQIzEiGEIIIYSOJK1ISo6mAtsDb7bdmCmU0X6gfiQtCewLvAW4mLTGcz/gYOBO2zt0+PVQkKS3AecC3wVahWE2AXYnVX69uVBoAycSwRBCCCHMRdJnSYnfVGA2uXVEfsy0XeuCKtF+oN4k/S/wFGnK8rbAcoCAA2zfXjK2MG+SliMl7uvnQ3cD38wzSUKPRCIYQgghhLlI+jq5d6Dt35WOp9ui/UC9tbX/WBD4HbCy7RfLRhbGStJKwK62jy0dy6CI9hEhhBBCmIvtg21f0MQkMIv2A/U2u7Vh+xXg4UgC60fSspI+I+ka4EpgYuGQBkqMCIYQQghhYFXaD+wGrEXqY3ZRtB/ob5JeAZ5v7ZL6Xb6Qt217yVKxhc5iWnb/iEQwhBBCCANH0prARNvXVY5tAJwIvNP2gsWCC6HBYlp2/4ipoSGEEEIYRCcAz1QP2J4JHAj8pEhEIQyGmJbdJyIRDCGEEMIgmpgTvznYvhNYpUA8IQwE2yfY3hxotfi4CFhB0mGSopl8D0UiGEIIIYRB1KkP4oSeRRHCgJF0oKRNgYdsT8vVX6cASwI/LhvdYIlEMIQQQgiDaLqkvdsPSvoEcGuBeEIYFCuS1uI+JukqSdOAVYHjba9ZNLIBE8ViQgghhDBwJE0ELgRe4vXEbwqwMLCT7d+Xii2EQSBpYdLf3FRgi/x42vZ6RQMbIG8oHUAIIYQQQq/ZNrQm1AAAAzZJREFU/gMwVdLWwKR8+BLbVxQMK4RBMoE0HXSp/HgUmGvdbhg/MSIYQgghhBBC6AlJpwDrA88CNwE3AjfafqpoYAMo1giGEEIIIYQQemVlUvuI3wOPAA8DTxeNaEDFiGAIIYQQQgihZySJNCo4NT8mAU8CN9g+pmRsgyQSwRBCCCGEEELPSVoR2JKUDG4PvNl2p9YuoYsiEQwhhBBCCCH0hKTP8vpI4Gzg+spjpu1XC4Y3UKJqaAghhBBCCKFXVgXOBw6y/bvCsQy0GBEMIYQQQgghhAETVUNDCCGEEEIIYcBEIhhCCCGEEEIIAyYSwRBCCANP0iuSbq88Vh3D/+NNkj7T/ehCCCGE7os1giGEEAaepOdsv3E+/x+rAj+yPWmUv7eg7Vfm598OIYQQRitGBEMIIYQhSFpQ0rGSbpF0p6R98vE3Srpc0gxJMyXtkH/lK8AaeUTxWEnvkvSjyv/vZEl75O1Zkr4qaQawi6Q1JP1U0q2SrpG0bq+fbwghhMES7SNCCCEEmCDp9rz9oO2dgL2AP9neVNIiwHWSLgV+C+xk+xlJfwXcKOli4HBgku2NACS9ax7/5hO235p/9nLgU7bvk7QZ8C1gm24/yRBCCKElEsEQQggB/txK4Cr+FthQ0gfz/lLAWsDDwDRJ7wBeBd4CTBzDv3kupBFGUmPl8yW1zi0yhv9fCCGEMGKRCIYQQghDE7C/7Z/NcTBN71wW2MT2bEmzgEWH+P2XmXMJRvvPPJ//uwDw9BCJaAghhDBuYo1gCCGEMLSfAZ+WtBCApLUlLU4aGXwsJ4FbA6vkn38WWKLy+78B1pO0iKQ3AdsO9Y/YfgZ4UNIu+d+RpMnj85RCCCGEJBLBEEIIYWjfAe4BZki6C/gv0kyas4ApkmYCHwd+BWD7CdI6wrskHWv7t8B5wF35v7d1+Lc+Auwl6Q7gbmCHDj8bQgghzLdoHxFCCCGEEEIIAyZGBEMIIYQQQghhwEQiGEIIIYQQQggDJhLBEEIIIYQQQhgwkQiGEEIIIYQQwoCJRDCEEEIIIYQQBkwkgiGEEEIIIYQwYCIRDCGEEEIIIYQBE4lgCCGEEEIIIQyY/w/OpLapTkT7QAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,5))\n",
    "feature_importance_df_2 = feature_importance(xgb_model_2, X_train)\n",
    "feature_importance_df_2.iloc[:15].plot('bar')\n",
    "plt.xlabel('Feature')\n",
    "plt.ylabel('Importance')\n",
    "plt.title('Feature Importance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bytree=1, gamma=0, learning_rate=0.01, max_delta_step=0,\n",
       "       max_depth=2, min_child_weight=1, missing=None, n_estimators=100,\n",
       "       n_jobs=1, nthread=None, objective='multi:softprob', random_state=12,\n",
       "       reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
       "       silent=True, subsample=1)"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb_model_3 = xgb.XGBClassifier(objective = 'multi:softmax', \n",
    "                              n_estimators = 100, \n",
    "                              max_depth = 2,\n",
    "                              colsample_bylevel = 1,\n",
    "                              learning_rate = 0.01,\n",
    "                              random_state=12)\n",
    "\n",
    "xgb_model_3.fit(X_train, y_train, eval_set=[(val_X, val_y)], verbose=False, early_stopping_rounds=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score - 0.9702938143165597\n"
     ]
    }
   ],
   "source": [
    "y_valid_pred_3 = xgb_model_3.predict(val_X)\n",
    "print ('F1 score - ' + str(f1_score(y_valid_pred_3, val_y, average='weighted')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_3 = xgb_model_3.predict(test_data)\n",
    "y_pred_3 = pd.DataFrame({'CASE_NO': dummy_test_data['CASE_NO'], 'CASE_STATUS': y_pred_3})\n",
    "y_pred_3['CASE_STATUS'] = y_pred_3['CASE_STATUS'].map({0: 'CERTIFIED', 1: 'CERTIFIEDWITHDRAWN', 2: 'WITHDRAWN', 3: 'DENIED'})\n",
    "y_pred_3.to_csv('3_prediction_xgboost.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Feature Importance')"
      ]
     },
     "execution_count": 172,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3sAAAHNCAYAAACnwWKiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3XmYJWV59/Hvj0EUEfcxyiaooKJoVMQ9cSMBo6C4AXGLCzGBCO64BBWjCa6okLgr6qtIXHCMGHCLuwmjsgiIjggCiQFcAiiLwP3+UdXMmaa7p2f6nK6u6u/nuvqaWs7pc9ec06fqrud57idVhSRJkiRpWDbpOgBJkiRJ0viZ7EmSJEnSAJnsSZIkSdIAmexJkiRJ0gCZ7EmSJEnSAJnsSZIkSdIAmexJkiRJ0gCZ7EmSxiLJuUmuSHL5yM9WC/ydD09ywbhinOdrfjjJPyzma84myWuTfKzrOCRJ/WSyJ0kap8dV1c1Gfv67y2CSbNrl6y9En2OXJC0NJnuSpIlL8sAk30ny2ySnJnn4yL6/SnJWksuSnJPkr9vtWwBfBLYabSmc3vI2vfWvbWF8eZLTgN8l2bR93qeTXJzk50leMM+4t09SbYznJ/lNkucnuX+S09rjOWrk8c9K8u0kRyX5vyQ/TvKokf1bJVmV5NdJ1iR53si+1yb5VJKPJbkUeD7wSuCp7bGfOtf/1+j/RZIXJ7koyf8k+auR/ZsneWuS89r4vpVk8/W9R5KkfvKuoSRpopJsDXwBeDrw78CjgE8nuVtVXQxcBDwWOAf4E+CLSU6uqh8k2RP4WFVtM/L75vOy+wF/AVwCXAd8Hvhcu30b4MtJzq6qE+d5GA8AdmzjW9Uex6OBGwE/TPKvVfX1kcd+CrgtsA/wmSQ7VNWvgWOBHwFbAXcDvpTkZ1X11fa5ewNPBp4B3Lj9HXepqqeNxDLr/1e7//bALYCtgd2BTyU5vqp+A7wFuAfwYOCXbazXzeM9kiT1kC17kqRxOr5tGfptkuPbbU8DTqiqE6rquqr6ErAaeAxAVX2hqn5Wja8DJwEPW2Ac76yq86vqCuD+wMqqOryqrq6qc4D3AftuwO97fVVdWVUnAb8DPlFVF1XVhcA3gfuMPPYi4Miq+kNVfRI4G/iLJNsCDwFe3v6uU4D30yR2U75bVce3/09XzBTIPP6//gAc3r7+CcDlwF2TbAI8Gzi4qi6sqmur6jtVdRXreY8kSf1ky54kaZweX1VfnrbtjsCTkzxuZNuNgK8BtK13rwF2orkJeVPg9AXGcf60198qyW9Htq2gSdLm639Hlq+YYf1mI+sXVlWNrJ9H05K3FfDrqrps2r5dZ4l7RvP4//pVVV0zsv77Nr7bAjcBfjbDr53zPZIk9ZPJniRp0s4HPlpVz5u+I8mNgU/TtG59rqr+0LYITvXVrOnPoWlZu+nI+u1neMzo884Hfl5VO25M8Bth6yQZSfi2o+n6+d/ArZNsOZLwbQdcOPLc6ce7zvo8/r/mcglwJXBn4NRp+2Z9jyRJ/WU3TknSpH0MeFySP0+yIslN2kIi2wCb0YxNuxi4pm21+rOR5/4vcJsktxjZdgrwmCS3TnJ74JD1vP5/AZe1RVs2b2O4Z5L7j+0I13U74AVJbpTkycDdabpIng98B/jH9v/gXsBzaP5/ZvO/wPZtF0xY///XrKrqOuCDwNvaQjErkjyoTSDneo8kST1lsidJmqg2ydmbprLkxTStSC8FNmlbuF4AHAf8BtifphVs6rk/Bj4BnNOOA9wK+ChNy9S5NOPVPrme17+WpqDJHwM/p2nhej9NEZNJ+E+aYi6XAG8AnlRVv2r37QdsT9PK91ngNTN0ex31r+2/v0ryg/X9f83DS2i6fJ4M/Bo4guZ9mPU92oDfLUlaYrLusAJJkrSxkjwLeG5VPbTrWCRJ8o6dJEmSJA2QyZ4kSZIkDZDdOCVJkiRpgGzZkyRJkqQBMtmTJEmSpAHq3aTqt73tbWv77bfvOgxJkiRJ6sT3v//9S6pq5foe17tkb/vtt2f16tVdhyFJkiRJnUhy3nweZzdOSZIkSRogkz1JkiRJGiCTPUmSJEkaIJM9SZIkSRogkz1JkiRJGiCTPUmSJEkaIJM9SZIkSRogkz1JkiRJGiCTPUmSJEkaIJM9SZIkSRogkz1JkiRJGqBNuw5gsWx/6BcW9fXO/ae/WNTXkyRJkqRRtuxJkiRJ0gCZ7EmSJEnSAJnsSZIkSdIAmexJkiRJ0gCZ7EmSJEnSAE002UuyR5Kzk6xJcugsj3lKkjOTnJHk45OMR5IkSZKWi4lNvZBkBXA0sDtwAXByklVVdebIY3YEXgE8pKp+k+R2k4pHkiRJkpaTSbbs7Qasqapzqupq4Fhg72mPeR5wdFX9BqCqLppgPJIkSZK0bEwy2dsaOH9k/YJ226idgJ2SfDvJ95LsMcF4JEmSJGnZmFg3zg14/R2BhwPbAN9IsktV/Xb0QUkOAA4A2G677RY7RkmSJEnqnUm27F0IbDuyvk27bdQFwKqq+kNV/Rz4CU3yt46qem9V7VpVu65cuXJiAUuSJEnSUEwy2TsZ2DHJDkk2A/YFVk17zPE0rXokuS1Nt85zJhiTJEmSJC0LE0v2quoa4CDgROAs4LiqOiPJ4Un2ah92IvCrJGcCXwNeWlW/mlRMkiRJkrRcTHTMXlWdAJwwbdthI8sFvKj9kSRJkiSNyUQnVZckSZIkdcNkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgZooslekj2SnJ1kTZJDZ9j/rCQXJzml/XnuJOORJEmSpOVi00n94iQrgKOB3YELgJOTrKqqM6c99JNVddCk4pAkSZKk5WiSLXu7AWuq6pyquho4Fth7gq8nSZIkSWpNMtnbGjh/ZP2Cdtt0T0xyWpJPJdl2gvFIkiRJ0rLRdYGWzwPbV9W9gC8Bx8z0oCQHJFmdZPXFF1+8qAFKkiRJUh9NMtm7EBhtqdum3Xa9qvpVVV3Vrr4fuN9Mv6iq3ltVu1bVritXrpxIsJIkSZI0JJNM9k4GdkyyQ5LNgH2BVaMPSHKHkdW9gLMmGI8kSZIkLRsTq8ZZVdckOQg4EVgBfLCqzkhyOLC6qlYBL0iyF3AN8GvgWZOKR5IkSZKWk4klewBVdQJwwrRth40svwJ4xSRjkCRJkqTlqOsCLZIkSZKkCTDZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBMtmTJEmSpAEy2ZMkSZKkATLZkyRJkqQBmmiyl2SPJGcnWZPk0Dke98QklWTXScYjSZIkScvFxJK9JCuAo4E9gZ2B/ZLsPMPjtgQOBv5zUrFIkiRJ0nIzyZa93YA1VXVOVV0NHAvsPcPjXg8cAVw5wVgkSZIkaVmZZLK3NXD+yPoF7bbrJbkvsG1VfWGCcUiSJEnSstNZgZYkmwBvA148j8cekGR1ktUXX3zx5IOTJEmSpJ6bZLJ3IbDtyPo27bYpWwL3BP4jybnAA4FVMxVpqar3VtWuVbXrypUrJxiyJEmSJA3DJJO9k4Edk+yQZDNgX2DV1M6q+r+qum1VbV9V2wPfA/aqqtUTjEmSJEmSloWJJXtVdQ1wEHAicBZwXFWdkeTwJHtN6nUlSZIkSbDpfB+Y5I7AjlX15SSbA5tW1WVzPaeqTgBOmLbtsFke+/D5xiJJkiRJmtu8WvaSPA/4FPCedtM2wPGTCkqSJEmStDDz7cZ5IPAQ4FKAqvopcLtJBSVJkiRJWpj5JntXtROjA5BkU6AmE5IkSZIkaaHmm+x9Pckrgc2T7A78K/D5yYUlSZIkSVqI+SZ7hwIXA6cDf01TdOXVkwpKkiRJkrQw863GuTnwwap6H0CSFe22308qMEmSJEnSxptvy95XaJK7KZsDXx5/OJIkSZKkcZhvsneTqrp8aqVdvulkQpIkSZIkLdR8k73fJbnv1EqS+wFXTCYkSZIkSdJCzXfM3iHAvyb5byDA7YGnTiwqSZIkSdKCzCvZq6qTk9wNuGu76eyq+sPkwpIkSZIkLcR8W/YA7g9s3z7nvkmoqo9MJCpJkiRJ0oLMK9lL8lHgzsApwLXt5gJM9iRJkiRpCZpvy96uwM5VVZMMRpIkSZI0HvOtxvkjmqIskiRJkqQemG/L3m2BM5P8F3DV1Maq2msiUUmSJEmSFmS+yd5rJxmEJEmSJGm85jv1wtcnHYgkSZIkaXzmNWYvyQOTnJzk8iRXJ7k2yaWTDk6SJEmStHHmW6DlKGA/4KfA5sBzgaMnFZQkSZIkaWHmm+xRVWuAFVV1bVV9CNhjcmFJkiRJkhZivgVafp9kM+CUJG8C/ocNSBQlSZIkSYtrvgnb09vHHgT8DtgW2GdSQUmSJEmSFma+yd7jq+rKqrq0ql5XVS8CHjvJwCRJkiRJG2++yd4zZ9j2rDHGIUmSJEkaoznH7CXZD9gfuFOSVSO7tgR+PcnAJEmSJEkbb30FWr5DU4zltsBbR7ZfBpw2qaAkSZIkSQszZ7JXVecluQC4sqq+vkgxSZIkSZIWaL1j9qrqWuC6JLdYhHgkSZIkSWMw33n2LgdOT/IlmqkXAKiqF0wkKkmSJEnSgsw32ftM+yNJkiRJ6oF5JXtVdUySzYCd2k1nV9UfJheWJEmSJGkh5pXsJXk4cAxwLhBg2yTPrKpvTC40SZIkSdLGmu+k6m8F/qyq/rSq/gT4c+Dt63tSkj2SnJ1kTZJDZ9j//CSnJzklybeS7Lxh4UuSJEmSZjLfZO9GVXX21EpV/QS40VxPSLICOBrYE9gZ2G+GZO7jVbVLVf0x8CbgbfOOXJIkSZI0q/kWaFmd5P3Ax9r1vwRWr+c5uwFrquocgCTHAnsDZ049oKouHXn8FkDNMx5JkiRJ0hzmm+z9DXAgMDXVwjeBf17Pc7YGzh9ZvwB4wPQHJTkQeBGwGfDImX5RkgOAAwC22267eYYsSZIkScvXfKtxXpXkKOArwHU01TivHkcAVXU0cHSS/YFXA8+c4THvBd4LsOuuu9r6J0mSJEnrMa8xe0n+AvgZ8A7gKGBNkj3X87QLgW1H1rdpt83mWODx84lHkiRJkjS3+XbjfCvwiKpaA5DkzsAXgC/O8ZyTgR2T7ECT5O0L7D/6gCQ7VtVP29W/AH6KJEmSJGnB5pvsXTaV6LXOAS6b6wlVdU2Sg4ATgRXAB6vqjCSHA6urahVwUJJHA38AfsMMXTglSZIkSRtuQ6pxngAcR1Mx88nAyUn2Aaiqz8z0pKo6AThh2rbDRpYP3pigJUmSJElzm2+ydxPgf4E/bdcvBjYHHkeT/M2Y7EmSJEmSujHfapx/NelAJEmSJEnjM69kry2y8nfA9qPPqaq9JhOWJEmSJGkh5tuN83jgA8DnaebZkyRJkiQtYfNN9q6sqndONBJJkiRJ0tjMN9l7R5LXACcBV01trKofTCQqSZIkSdKCzDfZ2wV4OvBI1nbjrHZdkiRJkrTEzDfZezJwp6q6epLBSJIkSZLGY5N5Pu5HwC0nGYgkSZIkaXzm27J3S+DHSU5m3TF7Tr0gSZIkSUvQfJO910w0CkmSJEnSWM0r2auqr086EEmSJEnS+MyZ7CW5jKbq5g12AVVVN59IVJIkSZKkBZkz2auqLRcrEEmSJEnS+My3GqckSZIkqUdM9iRJkiRpgEz2JEmSJGmATPYkSZIkaYBM9iRJkiRpgEz2JEmSJGmATPYkSZIkaYBM9iRJkiRpgEz2JEmSJGmATPYkSZIkaYBM9iRJkiRpgEz2JEmSJGmATPYkSZIkaYBM9iRJkiRpgEz2JEmSJGmATPYkSZIkaYBM9iRJkiRpgEz2JEmSJGmAJprsJdkjydlJ1iQ5dIb9L0pyZpLTknwlyR0nGY8kSZIkLRcTS/aSrACOBvYEdgb2S7LztIf9ENi1qu4FfAp406TikSRJkqTlZJIte7sBa6rqnKq6GjgW2Hv0AVX1tar6fbv6PWCbCcYjSZIkScvGJJO9rYHzR9YvaLfN5jnAFycYjyRJkiQtG5t2HQBAkqcBuwJ/Osv+A4ADALbbbrtFjEySJEmS+mmSLXsXAtuOrG/TbltHkkcDrwL2qqqrZvpFVfXeqtq1qnZduXLlRIKVJEmSpCGZZLJ3MrBjkh2SbAbsC6wafUCS+wDvoUn0LppgLJIkSZK0rEws2auqa4CDgBOBs4DjquqMJIcn2at92JuBmwH/muSUJKtm+XWSJEmSpA0w0TF7VXUCcMK0bYeNLD96kq8vSZIkScvVRCdVlyRJkiR1w2RPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBmjTrgPQmLz2Fov8ev+3uK8nSZIkaYPYsidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDNNFkL8keSc5OsibJoTPs/5MkP0hyTZInTTIWSZIkSVpOJpbsJVkBHA3sCewM7Jdk52kP+wXwLODjk4pDkiRJkpajTSf4u3cD1lTVOQBJjgX2Bs6cekBVndvuu26CcUiSJEnSsjPJbpxbA+ePrF/QbttgSQ5IsjrJ6osvvngswUmSJEnSkPWiQEtVvbeqdq2qXVeuXNl1OJIkSZK05E0y2bsQ2HZkfZt2myRJkiRpwiaZ7J0M7JhkhySbAfsCqyb4epIkSZKk1sSSvaq6BjgIOBE4Cziuqs5IcniSvQCS3D/JBcCTgfckOWNS8UiSJEnScjLJapxU1QnACdO2HTayfDJN905JkiRJ0hj1okCLJEmSJGnDTLRlTxqXXY7ZZVFf7/Rnnr5or3XW3e6+aK8FcPcfn7WorydJkqRu2LInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQNksidJkiRJA2SyJ0mSJEkDZLInSZIkSQO0adcBSBq2o5//1UV9vQPf/chFfb23PvWxi/p6L/7kvy3q60mSpP6yZU+SJEmSBshkT5IkSZIGyGRPkiRJkgbIZE+SJEmSBmiiyV6SPZKcnWRNkkNn2H/jJJ9s9/9nku0nGY8kSZIkLRcTS/aSrACOBvYEdgb2S7LztIc9B/hNVd0FeDtwxKTikSRJkqTlZJIte7sBa6rqnKq6GjgW2HvaY/YGjmmXPwU8KkkmGJMkSZIkLQuTTPa2Bs4fWb+g3TbjY6rqGuD/gNtMMCZJkiRJWhZ6Mal6kgOAA9rVy5OcvYgvf1vgkg19UvrTIXWjjo/X9aYBduPev2f14vg27r3rT+P5Rh3fQe+ZQCSTsVHH95Ljhv3+9ciQj2/IxwYeX995fP015GODxT++O87nQZNM9i4Eth1Z36bdNtNjLkiyKXAL4FfTf1FVvRd474TinFOS1VW1axevvRg8vv4a8rGBx9d3Hl9/DfnYwOPrO4+vv4Z8bLB0j2+S3ThPBnZMskOSzYB9gVXTHrMKeGa7/CTgq1VVE4xJkiRJkpaFibXsVdU1SQ4CTgRWAB+sqjOSHA6srqpVwAeAjyZZA/yaJiGUJEmSJC3QRMfsVdUJwAnTth02snwl8ORJxjAGnXQfXUQeX38N+djA4+s7j6+/hnxs4PH1ncfXX0M+Nliixxd7TUqSJEnS8ExyzJ4kSZIkqSMme5IkSZI0QCZ7kiRJkjRPSZ6XZMd2OUk+lOTSJKcluW/X8Y0y2VtGkvxLkpt3HcckJLnvXD9dxzcuSfZM8o0kl7Q/X0/ymK7jGock70ryztl+uo5vIZI8bWT5IdP2HbT4EUlzS7J7ki91Hcc4JHl5khVdx7GYkuyU5H1dxzFuSW6VJF3HsRja+ad7LckjR5Z3mLZvn8WPaKwOBs5tl/cD7gXsALwIeEdHMc2o9x+kcUvyTJo38K7tprOAd1bVR7qLamzOAb6f5DVV9fGugxmz1cCPgEva9dGTQQGPvMEzeibJ84C/Bl5Gc7wAuwL/lGSbqlqSVaA2wOqR5dcBr+kqkAl4EfCxdvldwOgNiGcDRy16RGOU5Oc0f2fXbxpZr6q68+JHNT5JPs+6x7eOqtprEcMZq/Zi7N3AVsDxwBHAh2jewzd0GNo4bUtz7juwqr7ddTDjlORewFtY+/4dTfN98gDgrR2GtmBJDgOOq6ofJ7kx8O/AvYFrkuxfVV/uNsKFS/Ktqnpou/zRqnr6yO7/Yt1zRR+9hbXH8GnWPZ5XA59Z9IjG55qq+kO7/FjgI1X1K+DLSd7UYVw3YLI3ok30DqG5MPsBzcnuvsCbk1RVfbTL+Baqqt6c5OPA25I8B/gX4LqR/X3+o3sR8CTgCuBY4LNVdXm3IY3dC4GHVtWvR7Z9NcmewLdYoiV/56uqjplaTnLI6PoAZJblmdb7aNdp65sATwFeAvxw8cMZu7e0/wZ4H/DcDmMZt7cCBwDfBfZs/z20qnp9A2JUVR3U9vA4KslZ3PDc94POglu499Ecz3eBPYBTgGOAv2ynt+qzpwKvb5ef2f67EtiJ5hh7n+wBW4ws32PaviGcG4Z87rsuyR2A3wCPYt2bY5t3E9LMTPbW9TfAE6rq3JFtX03yRJoEotfJHkBVXZjkCzQfysex9oRX9PgOS1UdCRyZ5E7AvsBXkpwHvLGqTuk2urHJtEQPgKr61QB7tQxtTpiaZXmm9d5p72aSZBPg6cBLaS46/6KqzuwytnGoqq9PLSe5fHR9AKqq/qNdPj7JhUNK9KZU1Q+SvJKmdeHOjLQ80++eHzeuqg+3y2cnObiqXtZlQGN0da2dH+zPgWOr6lrgrCF0cWzN9f3f+3MDwz73HUbTI2kFsKqqzgBI8qc0PemWjKH8sYzLzaclegBU1blDGOuW5B40dwD/G9itqv6n45DGrqrOSfI5mrsqT6e5AziUZO/SJPeuqlNHNya5N3BZRzFpfu6W5DSaO5l3bpdp1+/UXVjjkeRGNN1RX0jTyvz4qlrTbVQT0/cLlOluOW3szKaj6z3v8QFAktvRtGDeCXjk9O/QnrtJkvuwtpXkqtH1nrdaXpXknsD/Ao+g6Skw5abdhDR2t0zyBJreEKN/iwFu0V1YY3OnJKtoz3XtMu36DrM/bemrqn9Lckdgy6r6zciu1TQ9W5YMJ1UfkeT7VXW/Dd3XF233lYOr6qSuYxm3kRa9vYHzaVpiv1BVV3Qa2BgleSjw/2jG03y/3bwrTfeWp1XVt7qKbRySXMbaC+mbAr+f2kXT+tDbGy7tCWFWVXXeYsUyCUkuAK4BjgR+MX1/3xOGJLceWf0a8HBGuiDN1OLeF0k+NMfuqqpnL1owE9KOKf1H4H01sIueJF+bY3dVVW9bLZM8EPgwTdfNI6vq9e32xwBPr6r9OgxvLNbz90dV/dVixTIJbSvXrAbWSwJoilsBL6uq3buOZYrJ3ogkvwdmuhsd4E5VtcUM+3ojyY2r6qoZtj8U2K+qDuwgrLFIch1wGvA54FKm3X2vqrd1Ede4Jfkj4EDW9u0/Ezi6qn7ZXVTjkeRGI4Od1SNJPszsLV69TxhGCtDM1F+6qqr3rbMzSfJHVfW/XcexUElWVtXFM2zfFti3qt7cQVgT53fq0pdkn77fDJuvJCsBZvpb7KP1FbdaSu+ryd6Iod99H9V289gfeDLwc+AzVfWubqPaeEley9zV8l63eNFMRpIPV9Wzuo5jUpL8oKr6XnlsRtNaLWFttcret1ouB0nuOKTv/7kkuSXwRJrzw92raquOQxqr9oLzyTSl0reiKeb1krmf1R/ttASPpHn/HltVf9RxSBstyeOA06b+9trqnE8EzqPppfTzLuMbhyGf9+D6z+NhwN/RdFUNTS+Qd1XV4V3GtlBJfkgzdGGquNXHWKLFrRyzN2LkC2UHRlpOqmpJDbTcWEl2ojnB7UczRcEnaRL+R3Qa2BhU1Wu7jmER3KvrACZscFVmRnwFuD1NEaRjq+oGXR37rh1b81LWfneeAbylqk7vLqqx+Sz9L4E+qySb03SB3x+4D7Al8HjgG13GNS5JtgT2oTm+nWj+Dneoqm06DWyM2i6P+9O8b7em6QHS9yT2DcADAZI8FngazfXLfWhaVP68u9A0Ty8EHgrcfyo5b4fd/EuSF1bV2zuNbmF6U9zKlr0RbRGW99OMg5oq6vHHNOOjnlNVl3YV2zi0XR2/SXMsa9pt5wylC1I7BcErgJ3bTWcAR1Qum7P+AAAgAElEQVTVCd1FNT5JfkxzopsxKer5QPypcV+zdrfte1fcJLegueDcF7gJzc2WY/s83mtKkr1ppif4R9adA/IVwEuq6nNdxTYOSX5YVffpOo5JaKfjeRhwEs1Y568Ca6qq18UTRiW5gmbOslcD36qqGsq5L8kbaVoqfwF8gubGxOohvH9JTq2qe7fLHwTOrqoj2vVBtIitZ/hQVVWvb/K2rV+7V9Ul07avBE7q8/dqknNY94bKm2lueAJLa6y6LXvreifNGKh9q+o6uL4J+u9pJil9RoexjcPUhebXkvw7zYl9EK0pGf6E4wBb01SUm3HcEP0uHw5N+eKbMZDP5HRV9X/Ah5IcQ/N3+E6apK/XSWzrcJoT+rkj205L8lWacbS9TvaArZO8c7adVfWCxQxmzHammSfqLOCsqro2ydDuAr+C5m/un4FPJPlkx/GM03OBn9BU2v58VV01oPcvSW5GU6zrUTTv35SbdBPS2P2cZhqsobrR9EQPmnF7bRXnPvsG6753o+tLajozW/ZGJPlpVe24ofv6JskWNF129qNJED5CM26ht1U6k5zJDSccJ8ltaO7k3r2byMZnyK0LMJw7tbNJ8mCav7mH0UxP8Mmq+ma3UY1HkjOqavqEwFP7zqyqnWfa1xdp5uw8bLb9VXXMIoYzdknuRvPZfCpNF/+7AvccQnGWUSNVm/cDdgReQ3Pu+0mngS1AkhXA7jTH9CiaarGPBratqmu6jG2hkjwbeCVN0bWLqmqPdvt9aLqIP6rL+MZhOZ/Xh37OX0pM9kasJ9lbU1V3WeyYJi3JrWi6gDy1z1+cSc6aLaGba1+fLIOTwmCPL8m5wG9Z201unYuwAXTBPRV43PSxiG3Rq88PoCvSsrkoSXI/msThKcAFVfXgjkOaiHaM6f7AU4Zybk9yY+CxrL2p9JWq2r/bqBYmydbA7YBTR3pc3Z6mxej8ToMbgyRHVdVBs+zrfTXcJNcCv2PdHjtTxcluUlW9bd1LcmRVHdIuH1xV7xjZt6QK6pnsjWi7V/0MeH2N/Mck+Xtgp6p6emfBjVGSXYC7tatnVdWPuoxnHJL8J3BAzTzh+PuqarduIhufJH82n9bXJJ+uqicuRkzjlORWte7EpIOR5D9YW41zegn/qh7PhQWQ5PHAm4A3su4ckIcCL6+q47uKbRySfK+qHth1HIupHcLwsKoaRJEWuL7S6NQN3Z+0XasHqa1B8Piq+kjXsYxTkjvTJOn7ztaboM+GXg13SEZvAk6/IbjUbhA6Zm9dfwd8AFiTZLRAyw9p+sX3Wlsg4nPAtjRz0gXYpe2i9PieF6B5MbAqzQSlN5hwvLOoxmgDutn2tejAL0bGmkwlQ0XzPbVZVfX2+6qqHj7bvgGMW6Cqjk8zF92Lab5HoRn//JTpN2B66ilJtpttZ5+rqyY5rqqe0i4fUVUvh+YORJJXA3/WaYBj0LZ4vYemUuXPab5f7pjks8Dzq+rqLuNbiNlaF6rq0jTzgPU+2UuyFU0X4/2BXWgKQe3baVBjNPRquABJHsHaSs0/Gqli2WeZZXnJsWVvBu2do6kxJmdW1c+6jGdc2gIDVwMvG+kOsQnwT8DmVfV3cz1/qWu7dvwtA5xwfEMstTtKG6sdmH8gTeGdz1bVizsOaWzaVpNBzIW1HCQ5nRlaZIGVwO2qakUngY3BaPfpGe5OD6JrdZLDgTvTJHaXtdu2BI4Gzquqv+8yvoXoU+vChkpyAE2X1K2B49qfzw2h0uiUoVfDbbvhfga4krU34u8HbA48oaou7Cq2hWqHLzycZv7Ar7bLU+eIr01Vkl0KenunfBKSjH4pTn0AbzG1ve/jamgGbd9rKtEDqKrrkrwS6P1cWG1Sd4MiCkkeUlXf7iAkbYS2G8shNNVvP04zP8+vuo1qPDLMubBI8nnWnTR+HVW11yKGM3ZVtcvoepLtgZfTfKe+sYOQxmmuO75DuRu8D7BbVf1+akNVXZbkb4Hv0VTc7qvetC5shKNoJqzev6pWAwyo0uiUoVfDPQr4l6r68OjGJM+gqa66dxdBjcktaBLYqb+70RxhSb2HJnvreusc+4ZQ2v7qmapzVdU1Sa7qIqBxaSuSPYXmDuAXq+qMNJOwvpLmDlLv705vgF6e8JPclqYb4FOBDwL3GcqYmtxwLqzX0cyF1esqjiPe0nUAiyHJjsCrgAfQnC9eUFV/6DaqBbtpW91wE2Dzdjntz+adRjY+140melOq6vIBXFhv0hZa22Rkeeoc0NsW59YdaL4339r23DkO6H2391FV9ccj1XC/nOQSYMshFGdp7VxVT5i+sao+kuRVXQQ0LlW1/Wz72r/DJcNunMtIZp+UO8DH+lyxMsmHacYi/hfNhdh/0xaI6HtxiA0130IuS02S3wEXAx8CLpu+v3o8qXqSi2jmwjqStXNhDWJSZ1h6lcfGra3c+CqaLuJvAj5RVdd2G9V4TCsedANV9YjFi2YyRrpbzXQjbEl1t9pQbaXf65j52GpA3zHb0NwI3A/YgqZr/yu7jWr8hlYNd7Yq9+0Qop/0uRJukvdX1Q3qebSf1X+vqnt2ENaMTPbmIcnuNOPcdu86loVI8rW59vf5pJ7kR7RdVJPcBPglcOehdP8DSHLabLtoTup9L2//Wua+6Hzd4kUzXhnwXFjQ/7FB69OWDz8f+AJwgySvejypepKb97w413oNOSFKsmIoNx7mK8lONNNFvb7rWBYqyUFVddQM2wdRDTfJ24GbAYdU1e/abVsAbweu7Pl35zE0refPGKmDsTPwb8Dh07uudslkb0RbuerdwFbA8cARNK0MAd5QVZ/pMLxFk2T3qvpS13FsiKENTJ9JWyG2aMaxfR64YnR/VZ3XRVzaMBnmXFiz9RoA+j/eOckz59rf5+64SX4GvKqqju06lq4luUdVndF1HBuiPS/8TVV9t+tYFlOSX1TVrBVy+2KI1yqj2mrT/wg8C5i6RtkOOAZ4Zc8r4Yamyu+taKrDPgD4JM3f4791Gdt0JnsjkvwQeCHNgOA9gY/RdAO8wV2XIevjl0+S3wNrplZpKq+tYSCtXlNG+vY/jqba6MeBk4bQOgSQZE/gFaythnsGcERVndBdVJPTVgR8QvV8LqwklwEnM3vLSd/HO1+vrRJLVV3edSzjkGbi+yNp7r7/TVWtWc9TBqun574HAO8CTqXpgTTIuUqnS3J+VW3bdRwL1cfP3MZop5eY6rL5s5nG0PZVW+n+PsAdaaYb+l7HId2Ayd6IGVqHzq6qu3YZUxf6WG67vWCZ1RBbvZI8laZ0+BFV9eau41moJM+jmWbhZcDqdvOuNFODvL+q3ttVbAvVVh6b1QCSvd59Z2yoJH9DcyNii3bT5TR/e//cXVTj095o+TBN0j5asbnXlVQ3RF8/x20Lw/NpKvt+kXXfv952k5vLgFr2rgFmSnymblTffJFDGqskT6PJNT46bfvTgWur6uPdRLZwSd7F2il59qepxnnW1P6l9LdnNc513TLJPiPrm46uL5dunCyxkrHzdIeleDdl3No5a/YFnkBTrvmFwGc7DWp8Xgg8tKp+PbLtq+1F6LeA3iZ7wP1n2b4XTQXZXid7Q5dmcvEHAw+vqnPabXcC3pHk1lX1D50GuEBJ7kqTKHyT5gbSdXM/Y7D6eO6DZhqX+9MUuPo+A3n/5pjSJcBtFjmcSTm9jzcYNsDf0YxTn+4zNJPG9zbZY+1N6enLS44teyPaio6z/YdUVT17EcPpTB+7FWTdiWW/W1UP6jqmcUvydWBLmvLTnwbWKT4zLUnqnSRnzVYRdq59fdPehf9LmnnazqQZDzxb8Z1e6OM43w2R5Gzg3lV15bTtmwOnVtVO3US2cEn+iWauqxdW1b93HU+Xenruez7wUuDNwHtqQBd1Sf50rv1V9fXFimVS+tqaPF9z/U0lOW0oQ2ymS7JdVf2i6zim2LI3YsilwzfQuV0HsBFGxwrdpLMoJuuONDcj/rr9meo+QLvc24pyrUuT3LuqTh3dmOTezDAVQ98k2ZRmkPpLaCZyflJVnd1pUOPztlnmKxvKmNmanui1G69I0vdWlGto5rS8wfGNGnpC3+pjsYiHAg+qqovmelAfi8/Q3AxbWVVnjm5sKx5e3E1IY/evXQcwYZsn2WKqEueUdrz6Zh3FNDZJHkTTO+cbVXVRknsBh9IUX1syY0pN9kYkObKqDmmXD66qd4zsG8Q8UkluQ9O3+G7tprNo5oy6vpWoqvaZ6blL3FwTy/a+1QvmnsBzIF4MrEryIZquSNCM2Xsm8LTOohqDJAcCBwNfAfaoqnO7jWjsHtt1ABN2YZJHVdVXRje2FZz/p6OYxqKqXj3Phx4B9DLZSzLn2K6pO/BV9cDFiWh8qmq+340fBXrVaklTeGamMbG3AV5Ncy3Tdxcn2bGqftr2+vgg8ESam+7P6nslY+ADwKeSPH+qdkKS7Wm6i3+gw7gWLMmbac59pwAvT3Ii8Fya6qNLqieg3ThHTOsKOLhS/knuDnwVOBH4IU0ydB+a+b8eWVU/7jC8BRnyPEqjkmxG0wXwHu2mM4CPV9VV3UU1PkluD/wta4/vTODoqvpld1EtXNv6cxHN3ejRL92htHyto72p9CfAL6rq++t7/FKX5B7A52jGjo7eiHgIsHcPW0w2WJ+7myU5nXV7QtCurwRuV1UrOglsEfXx/Uuyuqp2nWXfj2oJTVq9sdLMEXyfqvpDkv1pbnr+Gc212Wuq6mGdBjgGbVfjV9BU/IWmuNU/VdW/dBfVwiU5E7hvVV3ZNjCcD9xzKd7MtWVvXZlleSheDxxcVceNbkzyROANNHeTemm+rV497coCXN91ZRXwbdZecD4ceFWSQVxwtkndYV3HMQE7dB3AJCX5N5ppan6U5A40VclWA3dO8t6qOrLbCBemqs5Ick+aloSpGxHfAP56fd0fB6S3d4arapfR9bZl4eXAo4E3dhBSF/r4/m05x74bLVoUk3VNVf2hXX4s8JG2p9WXk7ypw7jGpqreDby77bpJVd1gWEaSZ1b/5iu9cur7v6p+k+SnSzHRA5O96ebqCjiEO3+7VNWTpm+sqk8nWS4nvD52ZZnyLpp5sNbpSpXk0cBRwCM6iWpMksxWpKT3rV/znfqjx8WFdqiqH7XLfwV8qaqe0Z7cv00zj1uvtSf1D3YdhzZekh2BV9FMfvxW4AUjF9paetYkeUxNm2e1rdB8Tkcxjdt17Q2y39BUrXzDyL7NuwlpMmZK8kYcTDPRep/cKcmqdjnADiPrS2raGpO9dd2CpsVkKsEb7Svdx7ti0/1uI/cNSZ9bbLeeqUBCVX25ne+l766j+Tv7OPB54Ipuw+lEX4sLjV4wPwp4HzQn9wEUMJmaNH6uAjS9ngtrns7tOoCN1bbKvoqmVfZNwHOq6tpuo1p0fSw+cwjwhSRPYd3u0w9iOOOED6PpBbECWDXVQ6etRDqUhHY++nhttve09bd0EsU8OGZvGUlyAfC2mXYBh1TVkqkcNCl9HnuZ5Cc0rbNXTdt+E5q5enbsJrLxSXI3YD/gcTTj9T4OnFRV13Qa2CLp6+eznQ/rJOACmtavHarqt+3UBKur6h5z/oIlLsnxwO1p5ob65HxbavtiPoW7+izJtTTjab4A3CDJqyU0+fGGmm/xmb5KcmOaz+bU+LypceqD6T7dVmresqp+M7JtC5pr9Mvb9UFXw+3ruQ+uvwa7S7u6Zil+Nk32RiS5HfBKmjftNJoBpJd2G9X4JHnNXPur6nWLFUtXev6F8mrggcCB06pavZPmgvrw7qIbvyRPpanYdURVvbnreBZDXz+f7Xfn4cAdaArqnNRufwRwv6pasnc85yvJLYB9gH1pWmA/CRzb90q/Qy7cNSXJM+fa38OxQtez+Eyvu7/PW1/PDfPV0wJCm9KM+X02cB7N3+C2wIeAVy2lLuImeyOS/DtNV4Fv0HQR2HII0y0sB0lWzKdbTpLv9bG89pQkBwEvA27abvod8JaqGkI3TpJsTXMx/QSaMQzHAZ+durs5dH084U2X5GYAQ33PkmxC8xl9J/DGqpqpt0RvJPkUcNwshbv2r6reFu6ayTL4fG7P2uIz7xzKuWEuQ/jeXJ+hH2OSo6rqoK7j2BBJ3k5TROiFU+MRk9ycpjvnFVV1cJfxjTLZG5Hk1Kq698j6oO6kJDmuqp7SLh9RVS8f2XdSVf1Zd9EtTJJTaIqXfLfrWBbDXFWt+irJ12m+OI8DPg2s04Wszy0oSe4C/FFVfXva9ocAv6yqn7Xr9xwpdNIrSf6Gprz2FjR3OC+jaZWdaZ6s3knyYJouxg+jmYLhk1X1zW6jWrgkZ1fVXTd0X99M+3xCU/59SJ/P6cVnjllKLQuTNLRrtZn09RiTPA44baQ30mE0ld/Po6kO//Mu41uIJD8FdqppiVSSFcCPl9LQmk26DmCpSXKrJLdOcmtgxbT1vhv94O0+bd/KxQxkAv4aeEeS97VVVAcpyYokt62qy9riF5slOSDJWV3HNgZ3BG5F816eSDNofTVNa/vqDuMahyOBmbqEX8pIpcoeJ3qvphln+fCquk1V3ZqmOuye7b5eSzOP5z8DFwIH0IxL/F2S+ybp3QXYNIMv3DXD5/M2DOTzmeSeST5Bc4PsyzTzfL1/uSR6WvLeQDO/LEkeCzyNptvjKuDdHcY1DjU90Ws3XssSK+poNc51Ta/GCWsrchbQ94m55/rwLakP5oaqqv9M8gDg+cDqJF+kqe44tb+3A/CnJNkXeA/NReZPab5EPwicTDPReq/VsOdK/KOqOn36xqo6ve121XdPB+49OjC9qs5pq+idCvxDZ5GNx7k035F/TjPh8fTxUY/sIKZxuV2SF82wPfT/JuCUIX8+T2Vt8ZndgN2StR/PIZz75qGPlRw31LldB7CRqqp+3y7vA3ygqr4PfD/J33YY1zicmeQZVfWR0Y1JngYsqbHOJnsjBn6xCXDTJPehadHdvF1O+zOE+VxuDdyf5i7S9xlJ9gbi1TTFLta0rQnfBZ5UVZ/vOK7F1se5Em85x74h/O3VTBXIquqKIUy9UFUPn8/jelox733MPnn1+xczkAka8ufz2V0HsNiSPBTYr6oObDc9vct4JiHJ7sDLqmp3gKrap+OQNlbacbK/p5mWZ7TbdF+nGppyIPCZJM9m3alBNqepO7BkmOxtnD5ebAL8D2unXvgl607D8MvFD2d8kjwfeCnwZpo5lHrdUjmLq6tqDUBV/SDJT5dhogf9vIu7Osnzqup9oxuTPJe1J4k+uzDJo6rqK6MbkzyS5ntnuTgC6FWytxyqMDPgz+doJdEhF59pb07vDzwZ+DnNNChAf7u/w/WfwXcDWwHH03yHfIjmPPeGOZ7aF0cCp9AMWTirqlbD9e9n3//2LgQe0L6HU9MLnTD9e2YpMNnbOH282AR4RVV9r+sgJuShwIOq6qKuA5mg6d2tbjm63veqgBugj4n8IcBnk/wl694B3IwldgdwI70A+FySb7Hu8T2EG048O2S9OzcMuXDXiEF/PqcXn0kyiOIzSXaiKYq0H3AJzXQnqapHdBrYeL2VZhzwd4E9238PraqjOo1qTKrqg0lOBG5H0+V4yi+Bv+omqvFIciTwHeDbVfXVruOZi9U4N0KPqyL1Mu6NleTONHcC962eT+oMzpM4pc+f43beuesnB17qJ4gNkWZi2f1Ze4fzTOD/zdR9bqj6+NkcLek+Pf4hlXsf6uezLTDzYOCgqjqn3XYn4B3Af1ZVb8cjtl1sv0nTW2dNu+2cqup7/YTrzfA3N5gKuABJ/pxmGrNPTdv+JOD/etjt/XpppsJ6cPsDTeL3HeDbwKlVtWS6iNuyt7z07q7zhkqyFfBUmpP6LsA/0syJ1XvLJZmbh6u7DmCBatq/Q7EHzbjZk6rqxK6D0bwNtnDXqDap+2DXcUzAkIvP7ENz/v5amnmQj2V41zG3TDI6Hm/T0fWq+swMz+mTw4DHz7D9P4DP07Nu76Pa1tej4Pprz6nE7xCalsybdxfdukz2Nk5fLzZ3SLJqtp1VtddiBjNOSQ6g6eqxNc08bc8BPjekBGm0S1WSV1TVP3Yd0zgl2W6u/VX1i/bfBy5OROOTZrL4zwBXsrYb2ZOTHAE8oe3731tJ/pmmxeQ7wOuT7FZVr+84rC6c23UAG2HohbtIchkzJ66hKd6yZC7KNsJgi89U1fHA8Um2oOluewjNcIZ/AT5bVSd1GuB4fINmWpCZ1ouRsYk9deOqunj6xqq6pH1fey1N6dtdaJK8hwA7A2toanssGXbjHDHfi82+asv1P3e2/VX19UUMZ6ySXE3T1/3FIwOAh9bdY9buVkOQ5HSak9v0svYrgdtV1YpOAhuDJJ+lufnw4WnbnwE8sap6PW4oyY9oWheuTXJT4JtVdb+u4xqnJLeh6TFwt3bTWcAnqupX3UW1cEm+Ntf+gY2PGpwkXwHeOEvxmb8f2vuXZh7dJ9EMz3hU1/Fobkl+AuxcVddM234j4MxaQhOPb6gkX6JpvTsF+B7wvapaknMe27K3ri8wx8Um0NuLzdblfU7o1uMONFW63prk9jStezfqNqSxG/SdmaraZXS9nX/u5cCjgTd2ENI47VxVNyjEUlUfSfKqLgIas6vbiWSpqt9ndKKvAUhyd+CrwInAD2nOEfcHXpnkkVW1pOZU2kBDLtwFQHsD4g/VTjSe5K7AY4Bzq+qznQa3cIMtPjPH+3beUBK9JEdW1SHt8sFV9Y6RfR+uqmd1Ftx4fAZ4X5KDqup3cH3V2HfQ/1bLc4B7ATsCvwIuSXJxVV3SbVg3ZMveHKZdbL6zqt7VaUALlOQrQ/mCnEuSbYGn0HTr3IKmu8cru41q4ZL8lqaLR4CHtcvX63M33FFJdgReBTyAplLZMVMn+75qp8m4wR3MJJsAP6mqu3QQ1tgk+T1N1xVoPp93btenusndq6vYxiHJp4Djquq4adufCOxfVU/sJrKFG2IvgemSfIOmyMdPk9wF+C/g/9F0uTq5qg7tNMAFGnDxmbnet/+qqld0GuAYjP79zVCspfd/m0k2pRk3+lzgPJpzwrbAB2hannt9bgdIcnPggTRdOR9I00D0o6p6ZqeBjTDZm8EQLzZhGF8cG6ot3bxvVR3edSwLleRP59rf91bbJPek+bu7B/Ammi5y13Yb1Xi0JZq3AA4Zubu5BfB24MqqekGX8S1UkjvOtb+qzlusWCZhrgp5fa+eN6SKm7NJcvpUz4EkrwduXVUHJtkM+P70XgVaGpbD+zZteMY6f4tDumZLsjkwdVNzTVVd0WU845TkxjQ9PR7C2oTvoqX0+bQb54gZLjafM5SLzaFrE/S30LQonA68pKourKqfAL1P9GD+yVyST/e0peFU4Hya7tS7AbuN9gbseUL0UpquqOclmUp8tgOOAXrf6jzfZC7Jd6vqQZOOZwJ+t5H7+mCwhbtGjN7VfiTwZoCqurrvRUyGXnxmZHlQ79uITdpxiJuMLP//9u48WrKqPvv49wFFWgWBBDogCDSghEFEQbE1DiD4xqgIBIRXDU7EGCQCcYkY4xATgtIiCJq8IcGIAZwIiBIjhklFVAaBBjSi0AoaBxACCMr0vH/sU3Td4t7qpm/d2vecej5r3cU5Z99L/2rVtH9n7/3bvS++ti8dYqDSaM9Wve/2NlcblfRhSnK3FWV6/yXAPwIH2r69ZmyDkuxN1eXOJsCiDn+pnwycQpna+HLgBErZ5knU1qI0r68dwBx6GnAspQz1lsALKBXX1gAeD/yqWmTjtWbtAFbRBpIOn+a6KFN22uyXlBksXXa1pCXATyjvv3MBJK1TNaoRsL1W7RjmUGeftz5PoKy17HU2r+hr68LUu5cNaWt7tdEbgX8Drhw2MCRpW9vXji+saWLINM7lJA2dX2v7E+OKZS50vBrnlbaf1nfemekPj1QXHnuzgBvbd9WOZRQkXQG8yPavJD2Psl/UIZQk8Pdt/3HVAMekra9NSe8Z1u4Wb/EyIdM4FwBvpRTyOtn2Vc31xcAWtudVmfRHosvFZ7r8vPVIWn1SZ5BJWmj757XjmGvz4Xsvyd4MutbZhG5/qUv6HqUgS+/u2KmUBesCsH3FDH/aOfPhg2VVSXozcCRlfRvAXcAHbH+sXlSzJ+kq2zs0xx8Ffmn7vc35lBsVXdbm12ZXTUrhrq7qevGZrpN0JfBm25fUjmUcmlHZfSj9s9+3vVHlkObcfOh7ZxrngMHOpqROdDYbN9YOYA79D2WaXM/P+s5Nme8/KVpZ9l7Suyjz319g+4bm2iLgeEnr2f7bqgHOzuqSHtXsNbQb8Kd9bZP0OdzW1+ZnbO/XHH/A9hF9befa3qNedLO2bu0A5pqW7+H5sCbaXy12XdvXN8cHUgpbHdIrYgK0Ntkb8rwB0PLnredNwAmSrgLebvu22gGNWjNCuyclwdsRWAt4BQMVxTus+qjaJHUyVqjjnU2AU2dYLAu0e6GsV3LjWEm72/7KXMczFyQdASxZiSkfR6ygfb56DWVj7ofKhdu+QdJ+lPW0bX7/nQ5cJOkW4B7gawDNnfj/rRnYKEja2s1ec5IeY/u3fW27ePk+bq+pEuDs9W+bsTtT32NtX7M3CV5aO4A51OUiJjdRClvdzDzoMM8F29+S9Czgz4DLJH0JeLCvvdW1IiSdRtkq6lxKLYXzKdU4L6wZ16RJsjdVlzubMPwLr+0LZVfWB4BWJnuUvWkul3Sw7Ytn+iXb544xplHydPtC2b6n7Z0W238n6TzK2pNzvXz+/GqUtXttdxrQm555Sd8xwMd657avGXNcozKso9n2TmiXC3f1PBpYOPi5Kek5lFkgbdblIiZfpiSvGwKfoYxafqduSHNiPUrp/l9SRmNb/X03YBvgNuC7wHdtPyCp7Z+ZKyRpI9s/bU7vrRoMSfYGdbazCWD7dbVjmAdaOY0MwPZbJD0dOFHSd4F/YOodwNLc1r4AAB5/SURBVLavS/yJpN1sn9d/UdKulGm6rdY3utV/7fs1YpkDmuF4uvM2eqykHSnJ+YLmWM3PgqqRzd4kVOM8jrI8Y9AdTduwioHz3UGUIiabAXvYvru5vg1lO6LWsn08ZWbVpsD+wMnNlMDTKYlf6z8/Jf0ZZWueYyhrLzuVCNl+mqStKTUV/quZ3bLWBBRn+SZleyVs71I5lhRo6dfceT9qhs7mX6/sVMH5aobS4Q+xfeyw9i7oQoEISS8AzqDsJ9h7A9t2q9clStoW+DzwdcrdTYCdKBuV7lm7dHHMrP99Nfge68h77oJh7W3+bpgPxQPmmqRLbe88Q9tDG3fH/NfcaDkZeKrtLuxD92/A4bZ/UTuWcZD0DMravX2Bm20vrhzSnJB0k+1NasfRk5G9qf4C+LykaTub1aIanSXAlcCXgN/SjTvuE0PSBpQ78IuAXXtlqLvC9rWStqN8EWzbXP4q8KbpRtxjXtlY0kconym9Y5rzJ9YLa2SOnG5ktiO6XLirZ9iUxlaPzHa8+AwAkh4F/CFldG834ELgvRVDGhnbr5a0hqTXsfx771rgtP61z11h+3LKcpS3UdbyddW8GknLyN4ASWsytbN5HXBqFzqbknagDKX/H0oyezpwXtemDQwj6d9tt3KzdUk3An8PnDRJz1nMfxOwR2nrRydnImkfhlc8bP1abkmnA+fbPmng+huB3W2/sk5ks9dMcZyR7R+NK5ZRk7Q7pc/yEsqWEp8CPm/711UDGyFJ2wBnAxezfJDhGZRBhpfbvq5WbKMg6d3D2m3/zbhiGTVJJzDzjZYDba895pBmlGRvQjWbkh4AvAg4wvaMC/TbQNJRtt/ZHLe24uYwkta3/ctprm8C7G/7mAphjYykOxl+h3refHDGZOnyVEdJHx/SbNuvH1swc0TSQuBMSqGE/lk7awB72W5tkZamou+MxWds/7BOZLMn6XxK8aczurglATy0fOjowT6LpBcBf9XmKeIAkv5ymsuPA94A/I7tx485pJFp003OJHt9JqWzKWl9YD/KnOn7KOsRWz1FadiaoS5qnsN9KQn7RsCZtt9WN6qYVMOqOUL7KzpKup0he0K1/fFNCkkvBLZrTq+1fX7NeEZB0hcp04yXDlzfnlKDoM3FZzpP0vdsbz1D23dt//64Y5orktaiFBN6A6W66oe6uFaxmSH4MtufrR1LT9bs9bG9Vu0Y5pKk11OSvDWBzwH7dfGN1lXNB+XelGnGT6ZslbG57Y2rBjYikh4L3Gf7vub8KZTpO8tsn1k1uFiRZ1P2xDod+BbdWw/c2YqVk1C4S9Kuts+3fYGkZbZv7Gvbu+VTVRcOJnoAtpdK2mz84cQjtNrg3qTwUMLQiT66pPWAw4FXAZ8Ant61kVpJqwMvptyA34Oyl+68SfYysten653NZvuIa4DeHP4pT36b705Luhk4ltLJPKw5fkhHOiz3UNYtvAv4um1LusH2osqhjYSkr1JKT1/fTE36NnAqpYT4pbbfUTXAmFHzRddbX/NU4BxKafROVFDt+DTOBxlSuMv2+2rENUpdrhYr6XrbW83Q9gPbW447plh5kt4F7AIc3Ftf2STpHwEua/OaNgBJx1BuUv8T8FHbd1UOaaQkPZ9yA763rvQ5wKK+LVDmhSR7fbre2WxelDOyfdG4Yhk1Se8Z1t6RDsuhlGpkj6OMoHwa+EqHkr2HSqBLej+wnu2DJa0BXJ7y6O0g6TGUpO8Y4H22T6wc0qxJOs/2brXjmAuTULirP1kfTNzbnsh3ufjMpJD0FuDtwGObS78Gltg+oV5Uo9HcTPotcD9TBxhavzyqGWT4MWXP47Ns3ynpRtubVw7tYZLs9UlnM9pA0iJK0ncAsBXwHsqavVZvMCvp6l6ZcEkXA8fYPqs5v8r2DlUDjKGaJO+PKK/LzSgV5k62/ZOacY1C20d/VlbXCnf1dHxkr7PFZyZNs1QD23c25/vYPqNuVDETSccBr6DMmDuNsk/w0vl4A3612gHMM/2Z767AVwBs3ws8WCWiEZK0laSPSzpW0saSviTpLklXSdqpdnyzJekPJX1V0i3Nz0WSXlI7rlGzfYPto5qbDzsBawP/UTmsUbha0hJJhwFbAucCSBq2R1bMA5JOAS4Bnk4ZzdvZ9vu7kOhNiqbo047A9sDNQJfWcy+SdLakL/Qd987n3V34R8L2z102pn4fsKz5eZ/tZyfRaxfbd/YSvcaHqwUTK2T7UMrnx4eAFwD/DawvaT9J86rKaEb2+kj6N+BnwE+Ad1CKX9zddDYvavvIgspm8adQkoPDgEOBL1A2tvxb28+qGN6sSDoIeBNlKsRlzeWdgKOBf7b9T7ViGyVJr6AkQkttf7l2PKMkaQGlUteGlBGhq5rri4EtbH+yZnwxs2aqTm/vq05N1YFuV+OcpnDXZ7pWuKvjSxh27VUVlbR5x4rPTDRJN9nepHYcsXIkPZrlRVpebPt3K4f0kCR7fbre2ZR0pe2nNcdTFm73t7WRpOuA59r+1cD136EUM2l9+WJJHwO2Bb4B7AZ8wfb760YV0X2SrgfeOFN7y5OFzhbumk4zgomn2bO0jbo8RXXSSfqx7SfVjiOmJ+lfbb92hrYFtu8Zc0gz6kRZ11Fpnpijp7n+DUoHu+36p6LeMaStjTSY6AHYvlXqTBX45wE72H6gqRz7NaAzyZ6kpQzf5/KpYw4pVlJTWntG0703W+auNid0K9DqTZtXhsqXwLuBQyjLVyTpfuCEtlc7ZGr11MEvu858+XXVCr73Fo45nHhkZuyTzKdED5LsTTEBnc2tJV1NeTxbNMc05/NuQekjdIekHXqjsT1Npbk7Z/ibtrnX9gMAzfTirn2Rv7R2ALHKLqd8dk73mjTt/3y5ccW/0k4dTmL7HQY8F9i5N82xKXT1D5IOs93mtVGe4Xi685h/8r3XXo+VtCMz3FSxfcWY45lRpnH2kbTpsPbeHiht1eXHJ+m5lG0yPs7UimQHAq+2/fVasY2KpLuBH/ROgS2a807cjGi2O1lo++KB688Bfmb7h3UiixWRtGmbPz9WRNI+DOk4t3ldlKStgHcCt1H2Jz2Jso77h5StiC4b8uetIOk7lG0Ibhm4vj5wbsu3XuitJxXleeutLRVlacO6tWKLldPltfhdJulO4FJmuMlpe9cxhzSjjOxN9WiGdDbrhDQ6Xe6M2f66pGcCBwOvbS5fB+zSoYpkrV93uALHAUdOc/2Opu1l4w0nHoEzKZU4u2rY3XcDrU32KDfIeoW7vkUp3LUXJXH4KNDawl19Hj2Y6EFZt9cUVWizPfuOlwy0DZ7HPDOwFv/9kp6Ztfit8YP5lNANk5G9PpK+CBxpe+nA9e2Bo2y3urPZ3IUYnPJxC3ABZU+lW6sENkaSzrC9T+045pKkS2w/u3Ycj5SkS23vPEPbQ3tgxvyjlm9MPcm6XLirZ1ihki4VMela8ZlJIOkaBtbi235G7bhixdr0vZeRvakWDiZ6ALaXStps/OGMlu21Bq9JWpcyEvaPwL7jjqmCtq8dWhlr1g5gFQ3bT2/B2KKIVfFESR+ZqdH2X4wzmFGTdPiwdtvHjiuWOdDlwl09O0gafGxQpl+19fMS6HzxmUnQ9bX4XXYEgKQ1KdNwoYz2/aZeSNNLsjfVxHU2bd8GfFjSa2rHMiaTMJTd1sd4maSDbJ/Uf1HSG1m+DjPmp3vo9nO0BLgS+BLwW7pV5bDLhbsAsL167RjmUJeLz0yCrQfec1v0vR9bvxa/4y6Q9EHg9ZStawRsIunjwF/Zvq9qdH2S7E01kZ3NZs1CXgtR26HAmZJexdQiO2tQ1hDF/HWr7U/UDmIO7UjZKPePKK/N04Hz3I11EF1fC9x1r2Gg+IztGyS9GjgXSLI3v+X9114fBNYCNrd9J4CktSk3B5dQ9u2eF7Jmr4+khZRCA/cyTWez7YU+JO09zeV1gVdSNh7v/JSPNs2xXlVtf4ySXghs15xea/v8mvHEikn6pu1dVuL3trV97ThimiuSFlMSvxdR1jqfXTmkmGCSrrG93SNti4jZkXQ98OTBm36SVge+Z3urOpE9XEZz+tj+ObB4oLN5Toc6m4MFZgzcChxv+5wK8YyMpLVtT7cmA0lPsv3j5vSIMYY1UpJ2sf3NlfjVVk7JlbSr7fNtXyBpWW9KUtO2d5vL23fdyiR6jU/S4qqdTQGMHYHtgZuBX9SNaPZSuKv17l3FtpgHpnn/ieV7ltr22lUCi5Xh6WZ3NMV25tVIWkb2+vQ6m83x5ulstkd/RTVJ59nebbq2Nht4jK2suDnMwOOb8px15TmcdG0ddZb0emA/SjGPzwGfsd36RG8mfYW7FtuehMJdrSXpAeDX0zUBa9pu+9YSnSbpLOD3KNu3fKrvxnTMc81z9++2Txm4/mpgP9svrxPZw2Vkb6olLL/rfAZT70C/i3bvpYSkg4ALbV/fVHz6F2AfysLS19q+omqAs9NfMGG9IW1t1v84Wl1Bbgaa4Xi682intt5d/GfgGspn5YuBPfqL5s2nL/VRmMDCXa3V8eIznWf7FZKeAOwNnNRUdvw0JfH7Vd3oYgUOAT7X3AzsX/q1gHlWZyDJ3lRd72y+FfjX5vgAYAdKtbUdgeMpm+i21eA0pJna2my15o77an3HD70uO/DFMAnPYbTTC2sHMG4p3BUxHrb/F/i4pE8A+wMfodzQbfOWLpPg87afLmk3YJvm2n/YPq9mUNPJB/lUXe9s3t9XCvalwCnNeoz/asrHttkGzV5Y6jumOV+/Xlgj9QTK3aNegtc/EmvaXyZ9kaSzaUq+N8c055vXCytmQ9JGtn/anLZyDZHti2rHMFdWULjrc2MOJ2Li9BV9+gPg65SCgF+rG1WsBAE0yd28S/D6Zc1eH0m3A1+lPIF/0BzTnD/X9rq1YhsFSVdQSoffRpmOtGuvMp6k79pubQlgSe8Z1m77feOKJVaNpOcPa+9yh7vLJP3Y9pNqxzEbkrYC3kn57DwWOInyHfFD4A22L6sY3qw0e0L16xXuurDthbsi5jtJy4DbgU8B5wP397e3fHlNp0m6mSGjr7bnzchsRvam2rPveMlA2+B5G70buAxYHTi7L9F7PnBDzcBG4FbbJ9YOYq5JWgN4FbBtc+la4DTbv60X1Wj0J3NN1UNs/7JeRDEiXZgC/3HgFGBt4FuUPSH3oiR8HwWeVS+02bH9utoxREywZZQbLC9ufvoZ2HXcAcVKWx14PC34jsvI3gy62tmU9ChgrWYBfu/a4yivhbua891tf6VWjKtiEqo1StoGOBu4mOWLgZ8BPAfYswP7l4lyQ+IQyrpEUe5ynjAJe0B2VUdG9q60/bTm+Ae2t5yurY06XrgrImJOtKnfmZG9PtN1NiV1qrNp+37KVKT+a4Nlmz8AtCrZmxAnAG8eTMQlvQg4kfYXkTgMeC6wc2/bE0mLgH+QdJjtD1eNLmYk6QSmX9csYJ0xhzMXHuw7HtzP80HarcuFuyLmNUlvt/3B5nhf25/tazvK9jvrRRcrMO9H9HoystenKerxh8CfDnY2gf+clM5mG/fCapLyu6droiMbk0r6nu2tZ2hr9ZpLKK87YHfbtwxcXx84t22vyUki6cBh7bY/Ma5Y5oKku4EfUD5PtmiOac4X2X5crdhma2DU8jTgW7aPb85bc+c6oo2yv2x7SVqvLVXQM7I31WsY6GzavqHZIPFcYCKSPdpZeXTpBCQDq0l6zOD6vGZfni68lx89mOhBmUrdlIGPeWpYMiepC+udW30jZQUelLQhZcbHbsDf9bUtqBNSxMTo+pZfndWWRA+60UEcpXQ2Yz47BThD0sG2fwQgaTPKnjyfrBjXqAwry9/Kkv0BwH7A22oHMRu991tHdblwV8R81/Utv2IeSLI3VTqbxbLaAayCz87UIGln25eOM5i5YPtvJb0F+JqkxzaXfw0ssX1CxdBGZQdJg+uhoNzdXHPcwcTItP7utKQ7eXin7BbgAuCIZr/SVrL9RUmbMlC4i5IAvrJ30sbCXREt0PveE7Cg7zsw33sxMlmz10fSA5TO88OagDVtt3p0b5IWAjeVKw9ofm63vVPlkEZK0loAtu+cpu3Atq+RinaRtN5MTcBVtjceZzzjIGld4LXAYtv7Vg5nzmX9UEREOyXZmyBdXwjcTGnsJXj3AZsCO9leVi+q8evCcxntIulGymjXdKN4tr1ozCGNzaS839pYuCsiIjKNc9J0diGwpEsoGx5/Ctin2TPqxklL9Bqtfi6jfWxvXjuGGpq13JPyPZo7wxERLTQpX1JRdHkh8M+BJwILgfWB62n/Y1pVk/q4oxJJgyNbBm6xfVONeEZN0t7TXF6Xsqbtc2MOJyIiYqUl2ZssnV0IbPsVkp4A7A28V9JWwDqSnmn725XDG7eM7MW4fWiaa+tJWgM4wPaV4w5oxF42cG7gVuB42+dUiKeGZbUDiIiIRy5r9qKTJG1AKfl+APAk25tUDmlsJJ1o+y2144iQtBNwrO3n1Y4lpjdJhbsiIiZRkr0J0pTrv8/2fc35U4CXAMtsn1k1uDkkadOu7JMlaTvg7cA2zaVrgQ/ZvrpeVBEz60IBE0kHARc2a4EF/AuwD/Aj4LW2r6ga4Cx0vXBXRMSkW612ADFW/wlsBiBpS+ASYBHwFklHV4xrJCQ9W9IfN6N6SHqqpNOAr1cObSQk7QmcCVwIvL75uYiy0fqeFUOLmJakhXRjDelbWT6N8QBgB8pn5+HA8ZViGpXOFu6KiIis2Zs069q+vjk+EDjd9iHNuprLgXfUC212JB0DvBS4EjhC0peBNwJ/T0mKuuBvgN0HKoxeLel84PPNT8TYSTqBhyd16wGLKYlS293fmxFB+Zw5pdlI/b8kfbBiXKPQ5cJdERETL8neZOn/4t4VOAbA9r2SHqwT0sj8EbCj7d80mx3fBGzXsa0XHjXd47G9rCkBH1HLZQPnvQImh9v+RYV4Ru1BSRsCtwG7AX/X17agTkgj09nCXRERkWRv0lwtaQnwE2BL4FwASetUjWo0fmP7NwC2b5N0fccSPYD7JT3J9o/7L0raFLi/UkwRABcMvi475t2UhHZ14Gzb1wJIej5wQ83AZsv26rVjiIiIuZMCLRNE0gLKlKoNgZNtX9VcXwxsYfuTNeObDUm3A1/tu/S8/nPbLx97UCMm6RXAB4GjKNNuAXaiTL89wvZZtWKLyTZQ5OMM2/vUjmnUJD0KWMv2bX3XHkf5Hr2rOd/d9ldqxbgqJrVwV0TEpEiyF0jaBNjf9jG1Y1lVzR32Gdm+aFyxzCVJOwB/CWzbXLoOWNJL3CNqkPQd2zsOHk+aNlavlPRV4A1NpdEtgW8Dp1Iq/l5qu7VruSMiItM4J5ak9YF9KZXlNqJUeWytlU3m2j7q0CR1f1I7jogBw4p8TJI2Vq/sbOGuiIhIsjdRJK0F7A38X+DJwL8Dm9veuGpg47WodgCrStIXGNKR7sJU1WitYUU+bHvteqGNVRsT3S4X7oqImHhJ9ibLLyhTdN4FfN22Je1VOaZxa2NnrGdJ7QAippMiH63W5cJdERETL8neZDkS2B/4GHC6pE9XjicemdfZfm3tICJiRstqB7AKDqIU7toM2MP23c31bcgNpoiI1lutdgAxPraPs70LsGdz6SxgI0lvl/TkiqGNUxvX1PQ8tXYAEZNI0tv7jvcdaDuqd2x773HGNQq277F9tO23DhR6ugn4vVpxRUTEaKQa5wRpKq0ttH1x37XtgeOB57d5KpaktW3fMUPbQ3vTSdrD9rnjjW40JH2PUlBn2oTV9hXjjShiMgxsLTGl4mYbK3DOZLrCXbbfVjeqiIiYjUzjnCzHUaZyPsT2UkmHUvZua7MLgV5n7Dzbu/W1ndVra2ui13gi8CGmT/ZMKa4QEaOnGY6nO2+VFO6KiOi2JHuTZaHtpYMXbV8tadMaAY1Qf4drvSFtbfYD20noIsZv2NYSbZ8ek8JdEREdlmRvsgyrrrZgbFHMjS53xiKirmFbS6xZL6yRSOGuiIgOS7I3WS6TdJDtk/ovSnojZfPcNttA0uGUzlfvmOZ8/XphjdQR012UtAmwv+1jxhxPxERo83rmFbF9HHCcpEWUpO+hwl3AWba/XzXAiIiYlRRomSCSFgJnAveyPLnbCVgD2Mv2z2rFNluS3jOs3fb7xhXLOKSQQsT4SHoscJ/t+5rzpwAvAZbZPrNqcLPU5cJdERGRZG8iSXohsF1zeq3t82vGEytnhkIKr0whhYi5JemrwBtsX98kR98GTqXsRXep7XdUDXAWJH0ROHJwPbekpwJH2X5pncgiImIUkuxFJ0jaFtjC9tnN+YeBJzTNJ3ZhWwJJ9/DwQgo32F5UObSITpO01Pb2zfH7gfVsHyxpDeDyXlsbSbrU9s4ztC1t82OLiIhsqh7dcTRwS9/5i4FzgAuAd1eJaPSOBB5DKaRwpKQtKscTMSn674ruCnwFwPa9wINVIhqdLhfuioiYeEn2ois2tP2NvvM7bJ9h+5PA79YKapRsH2d7F2DP5lKvkMIRkp5cMbSIrrta0hJJhwFbAucCSBqWKLXFZZIOGrzYkcJdERETL9M4oxMk/bftp8zQ9n3bnUyGJG1HWcO3n+0ta8cT0UWSFgBvBTYETrZ9VXN9MWX6+CdrxjcbXS7cFRER2XohuuOnkp5l+1v9FyXtAvy0UkxzzvY1kv4auK52LBFdZfseylTxQTcBzxlzOCNl++fA4oHCXeekcFdERDdkZC86QdIzgU8D/wr0irE8AziQUrHy25VCGxlJawMHA08EzqasG3oLcDhwte09h/x5RIxAtj2JiIg2SbIXnSFpA0rys21z6Vrgo82d69aT9HngNuASYDdgA8qm8W+1fWXN2CK6LNueREREWyXZi06TtAmwv+1jascyWwPl31cH/gd4ku3f1I0sotuy7UlERLRVqnFG50haX9KfS/oacCGwsHJIo3Jf78D2A8DNSfQixiLbnkRERCtlZC86YRKmWUl6APh175SyB9bdzbFtr10rtohJIGkRsD9lvd5WlD08z7L9/aqBRUREzCDJXnRCpllFxFyRtCWw0PbFfde2B44Hnm979WrBRUREDJFpnNEVmWYVEXPlOOCO/gu2lwKHAl+qElFERMRKSLIXnWD7ONu7AL3tB84CNpJ0hKRObqgeEWOzsEnuprB9NbBphXgiIiJWSpK96ARJh0raGfix7aOaqpU7AWsD/1E3uohouXWGtC0YWxQRERGPUJK96IqNKetnfiHpIklHAZsBH7K9ZdXIIqLtLpN00OBFSW8ELq8QT0RExEpJgZboFElrUEb0FgPPbn5ut71N1cAiorUkLQTOBO5leXK3E7AGsJftn9WKLSIiYphH1Q4gYsQWUKZuPqH5+SnwsLU2EREry/bPgcWSXghs11w+x/b5FcOKiIhYoYzsRSdI+idgW+BO4FvAN4Fv2r6tamAREREREZVkzV50xZMoWy/8DPgJcDNwe9WIIiIiIiIqyshedIYkUUb3Fjc/2wG/Ai6x/Z6asUVEREREjFuSvegcSRsDz6EkfC8Ffsf2sNLpERERERGdk2QvOkHSX7B8RO8+4Bt9P0ttP1gxvIiIiIiIsUs1zuiKzYDPAofZ/p/KsUREREREVJeRvYiIiIiIiA5KNc6IiIiIiIgOSrIXERERERHRQUn2IiJiIkh6QNKVfT+brcL/Yx1Jfz766CIiIkYva/YiImIiSLrL9uNn+f/YDPii7e0e4d+tbvuB2fzbERERj1RG9iIiYmJJWl3SMZIulXS1pDc11x8v6TxJV0haKmnP5k+OBrZoRgaPkfQCSV/s+/+dKOm1zfEySR+QdAWwr6QtJP2npMslfU3S1uN+vBERMVmy9UJEREyKBZKubI5vtL0X8Abgf23vLOkxwMWSzgVuAvayfYek3wW+Kels4B3AdrafBiDpBSv4N2+1/fTmd88D/sz29ZKeBXwM2HXUDzIiIqInyV5EREyKe3pJWp89gKdK+uPm/AnAVsDNwFGSngc8CDwRWLgK/+anoYwUAouBz0rqtT1mFf5/ERERKy3JXkRETDIBh9j+8pSLZSrm+sAzbN8naRmw5jR/fz9Tl0QM/s6vm/+uBtw+TbIZERExZ7JmLyIiJtmXgTdLejSApCdLehxlhO8XTaL3QmDT5vfvBNbq+/sfAdtIeoykdYDdpvtHbN8B3Chp3+bfkaQd5uYhRUREFEn2IiJikv0zcB1whaRrgP9HmfVyKrCTpKXAnwDfA7B9K2Vd3zWSjrF9E/AZ4Jrmv98Z8m+9CniDpKuAa4E9h/xuRETErGXrhYiIiIiIiA7KyF5EREREREQHJdmLiIiIiIjooCR7ERERERERHZRkLyIiIiIiooOS7EVERERERHRQkr2IiIiIiIgOSrIXERERERHRQUn2IiIiIiIiOuj/A5EeaG+Pzxr8AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,5))\n",
    "feature_importance_df_3 = feature_importance(xgb_model_3, X_train)\n",
    "feature_importance_df_3.iloc[:15].plot('bar')\n",
    "plt.xlabel('Feature')\n",
    "plt.ylabel('Importance')\n",
    "plt.title('Feature Importance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Ensemble"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "For my final submission, I applied 3 XGBoost models by varying n_estimators, max_depth, random_state, \n",
    "learning_rate parameters and ensembled the three. I took mode of the three outputs and finally achieved best score \n",
    "of 97.99 on LB. However, XGBoost outperformed other models.\n",
    "\n",
    "The final model was the mode of three XGBoost models.\n",
    "'''\n",
    "\n",
    "Ensemble = pd.DataFrame({'1': y_pred_1.CASE_STATUS, '2': y_pred_2.CASE_STATUS,\n",
    "                         '3': y_pred_3.CASE_STATUS})\n",
    "case_status = np.array(Ensemble.mode(axis=1))[:, 0]\n",
    "y_pred = pd.DataFrame({'CASE_NO': y_pred_1.CASE_NO,'CASE_STATUS': case_status})\n",
    "y_pred.to_csv('ensemble_predictions.csv', index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}