Binary_Prediction_of_Poisonous_Mushrooms/mushroom_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "012ac599-6700-44e8-82da-6916a254a84f",
   "metadata": {},
   "source": [
    "# Binary Prediction of Poisonous Mushrooms"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4f2973b-c9e9-4864-998b-beb9901e5db1",
   "metadata": {},
   "source": [
    "## Table of Contents\n",
    "\n",
    "- [**1. Introduction**](#introduction)\n",
    "\n",
    "- [**2. Exploratory Data Analysis**](#eda)\n",
    "\n",
    "    - 2.1. [Mushroom Type](#mushroomtype)\n",
    "   \n",
    "    - 2.2. [Numerical Features](#numerical)\n",
    "   \n",
    "    - 2.3. [Categorical Features](#categorical)\n",
    "   \n",
    "    - 2.4. [Missing Values](#missing)\n",
    " \n",
    "- [**3. Data Preprocessing and Feature Engineering**](#data-preprocessing)\n",
    "\n",
    "    - 3.1. [Update Categorical Values](#unused-cat)\n",
    "\n",
    "    - 3.2. [Missing Data Handling](#missing-data)\n",
    "    \n",
    "    - 3.3. [Target Enconding](#target-encoding)\n",
    " \n",
    "    - 3.4. [Pre-processing pipeline](#pipeline)\n",
    "\n",
    "- [**4. Modeling**](#modeling)\n",
    "\n",
    "    - 4.1. [Train and Test data split](#data-split)\n",
    " \n",
    "    - 4.2. [Random Forest](#random-forest)\n",
    " \n",
    "    - 4.3. [AdaBoost](#adaboost)\n",
    " \n",
    "    - 4.4. [XGBoost](#xgboost)\n",
    " \n",
    "    - 4.5. [LightGBM](#lightgbm)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "958d086f-a666-4a25-ba20-529ad363370f",
   "metadata": {},
   "source": [
    "<a id=\"introduction\"></a>\n",
    "## 1. Introduction"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "702834ec-ecc1-4a0a-8f80-5d2a7fdcdf1d",
   "metadata": {},
   "source": [
    "I have started working on this project to expand my skill set by doing a comprehensive project. I came across the play groud series hosted by Kaggle and decided to work on it.\n",
    "\n",
    "Find more details on the problem statement [here](https://www.kaggle.com/competitions/playground-series-s4e8).\n",
    "\n",
    "First, I performed comprehensive exploratory data analysis to understand relationships among the most important variables and detect potential issues such as sknewness, outliers and missing values. Then, I handled these issues, cleansed the data and performed feature engineering. Lastly, I built machine learning models to predict mushroom type. By the time I write this notebook, my best model has Mean Absolute Error of 12293.919, ranking 95/15502, approximately top 0.6% in the Kaggle leaderboard."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "59d786e4-7e61-4e96-b976-6b84cf2b0858",
   "metadata": {},
   "source": [
    "#### Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a12fff4d-dad9-4cc5-8d0f-edade620826e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "\n",
    "%matplotlib inline\n",
    "sns.set_style('darkgrid')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b906a28b-0d0d-4b15-801e-7a70463a380b",
   "metadata": {},
   "source": [
    "#### Read train and test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "feb38eb6-6ffd-4a43-9c2a-9fea82cde818",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = pd.read_csv(\"data/train.csv\")\n",
    "test_data = pd.read_csv(\"data/test.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5cbcaad3-07b1-44f1-ba8f-1c609691cf8c",
   "metadata": {},
   "source": [
    "#### Describe data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "3804b090-71c7-4939-9b12-0c7d564313fa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train data's size:  (3116945, 22)\n",
      "Test data's size:  (2077964, 21)\n"
     ]
    }
   ],
   "source": [
    "print(\"Train data's size: \", train_data.shape)\n",
    "print(\"Test data's size: \", test_data.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "5139ef87-0583-410e-b233-731168070cca",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 3 numerical features:\n",
      " ['cap-diameter', 'stem-height', 'stem-width']\n"
     ]
    }
   ],
   "source": [
    "numCols = list(train_data.select_dtypes(exclude='object').columns)\n",
    "\n",
    "# remove \"id\" from the category columns\n",
    "numCols.remove(\"id\")\n",
    "print(f\"There are {len(numCols)} numerical features:\\n\", numCols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ef54fe1d-a365-413b-b4e1-67695f2da93e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 17 categorical features:\n",
      " ['cap-shape', 'cap-surface', 'cap-color', 'does-bruise-or-bleed', 'gill-attachment', 'gill-spacing', 'gill-color', 'stem-root', 'stem-surface', 'stem-color', 'veil-type', 'veil-color', 'has-ring', 'ring-type', 'spore-print-color', 'habitat', 'season']\n"
     ]
    }
   ],
   "source": [
    "catCols = list(train_data.select_dtypes(include='object').columns)\n",
    "\n",
    "# remove predictor \"class\" from the category columns\n",
    "catCols.remove(\"class\")\n",
    "print(f\"There are {len(catCols)} categorical features:\\n\", catCols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5d8de4ae-095d-4035-aba2-b33d523b6b5f",
   "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>id</th>\n",
       "      <th>class</th>\n",
       "      <th>cap-diameter</th>\n",
       "      <th>cap-shape</th>\n",
       "      <th>cap-surface</th>\n",
       "      <th>cap-color</th>\n",
       "      <th>does-bruise-or-bleed</th>\n",
       "      <th>gill-attachment</th>\n",
       "      <th>gill-spacing</th>\n",
       "      <th>gill-color</th>\n",
       "      <th>...</th>\n",
       "      <th>stem-root</th>\n",
       "      <th>stem-surface</th>\n",
       "      <th>stem-color</th>\n",
       "      <th>veil-type</th>\n",
       "      <th>veil-color</th>\n",
       "      <th>has-ring</th>\n",
       "      <th>ring-type</th>\n",
       "      <th>spore-print-color</th>\n",
       "      <th>habitat</th>\n",
       "      <th>season</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>e</td>\n",
       "      <td>8.80</td>\n",
       "      <td>f</td>\n",
       "      <td>s</td>\n",
       "      <td>u</td>\n",
       "      <td>f</td>\n",
       "      <td>a</td>\n",
       "      <td>c</td>\n",
       "      <td>w</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>w</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>NaN</td>\n",
       "      <td>d</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>p</td>\n",
       "      <td>4.51</td>\n",
       "      <td>x</td>\n",
       "      <td>h</td>\n",
       "      <td>o</td>\n",
       "      <td>f</td>\n",
       "      <td>a</td>\n",
       "      <td>c</td>\n",
       "      <td>n</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>y</td>\n",
       "      <td>o</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>t</td>\n",
       "      <td>z</td>\n",
       "      <td>NaN</td>\n",
       "      <td>d</td>\n",
       "      <td>w</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>e</td>\n",
       "      <td>6.94</td>\n",
       "      <td>f</td>\n",
       "      <td>s</td>\n",
       "      <td>b</td>\n",
       "      <td>f</td>\n",
       "      <td>x</td>\n",
       "      <td>c</td>\n",
       "      <td>w</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>s</td>\n",
       "      <td>n</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>NaN</td>\n",
       "      <td>l</td>\n",
       "      <td>w</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>e</td>\n",
       "      <td>3.88</td>\n",
       "      <td>f</td>\n",
       "      <td>y</td>\n",
       "      <td>g</td>\n",
       "      <td>f</td>\n",
       "      <td>s</td>\n",
       "      <td>NaN</td>\n",
       "      <td>g</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>w</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>NaN</td>\n",
       "      <td>d</td>\n",
       "      <td>u</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>e</td>\n",
       "      <td>5.85</td>\n",
       "      <td>x</td>\n",
       "      <td>l</td>\n",
       "      <td>w</td>\n",
       "      <td>f</td>\n",
       "      <td>d</td>\n",
       "      <td>NaN</td>\n",
       "      <td>w</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>w</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>f</td>\n",
       "      <td>f</td>\n",
       "      <td>NaN</td>\n",
       "      <td>g</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   id class  cap-diameter cap-shape cap-surface cap-color  \\\n",
       "0   0     e          8.80         f           s         u   \n",
       "1   1     p          4.51         x           h         o   \n",
       "2   2     e          6.94         f           s         b   \n",
       "3   3     e          3.88         f           y         g   \n",
       "4   4     e          5.85         x           l         w   \n",
       "\n",
       "  does-bruise-or-bleed gill-attachment gill-spacing gill-color  ...  \\\n",
       "0                    f               a            c          w  ...   \n",
       "1                    f               a            c          n  ...   \n",
       "2                    f               x            c          w  ...   \n",
       "3                    f               s          NaN          g  ...   \n",
       "4                    f               d          NaN          w  ...   \n",
       "\n",
       "   stem-root  stem-surface stem-color veil-type veil-color has-ring ring-type  \\\n",
       "0        NaN           NaN          w       NaN        NaN        f         f   \n",
       "1        NaN             y          o       NaN        NaN        t         z   \n",
       "2        NaN             s          n       NaN        NaN        f         f   \n",
       "3        NaN           NaN          w       NaN        NaN        f         f   \n",
       "4        NaN           NaN          w       NaN        NaN        f         f   \n",
       "\n",
       "  spore-print-color habitat season  \n",
       "0               NaN       d      a  \n",
       "1               NaN       d      w  \n",
       "2               NaN       l      w  \n",
       "3               NaN       d      u  \n",
       "4               NaN       g      a  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e5efc1a8-035e-45c7-9bf9-9670d530dc99",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 3116945 entries, 0 to 3116944\n",
      "Data columns (total 22 columns):\n",
      " #   Column                Non-Null Count    Dtype  \n",
      "---  ------                --------------    -----  \n",
      " 0   id                    3116945 non-null  int64  \n",
      " 1   class                 3116945 non-null  object \n",
      " 2   cap-diameter          3116941 non-null  float64\n",
      " 3   cap-shape             3116905 non-null  object \n",
      " 4   cap-surface           2445922 non-null  object \n",
      " 5   cap-color             3116933 non-null  object \n",
      " 6   does-bruise-or-bleed  3116937 non-null  object \n",
      " 7   gill-attachment       2593009 non-null  object \n",
      " 8   gill-spacing          1858510 non-null  object \n",
      " 9   gill-color            3116888 non-null  object \n",
      " 10  stem-height           3116945 non-null  float64\n",
      " 11  stem-width            3116945 non-null  float64\n",
      " 12  stem-root             359922 non-null   object \n",
      " 13  stem-surface          1136084 non-null  object \n",
      " 14  stem-color            3116907 non-null  object \n",
      " 15  veil-type             159452 non-null   object \n",
      " 16  veil-color            375998 non-null   object \n",
      " 17  has-ring              3116921 non-null  object \n",
      " 18  ring-type             2988065 non-null  object \n",
      " 19  spore-print-color     267263 non-null   object \n",
      " 20  habitat               3116900 non-null  object \n",
      " 21  season                3116945 non-null  object \n",
      "dtypes: float64(3), int64(1), object(18)\n",
      "memory usage: 523.2+ MB\n"
     ]
    }
   ],
   "source": [
    "train_data.info(verbose=True, show_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "12d73c73-813f-4288-b545-7327063afd23",
   "metadata": {
    "scrolled": true
   },
   "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>id</th>\n",
       "      <th>cap-diameter</th>\n",
       "      <th>stem-height</th>\n",
       "      <th>stem-width</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3.116945e+06</td>\n",
       "      <td>3.116941e+06</td>\n",
       "      <td>3.116945e+06</td>\n",
       "      <td>3.116945e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.558472e+06</td>\n",
       "      <td>6.309848e+00</td>\n",
       "      <td>6.348333e+00</td>\n",
       "      <td>1.115379e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>8.997847e+05</td>\n",
       "      <td>4.657931e+00</td>\n",
       "      <td>2.699755e+00</td>\n",
       "      <td>8.095477e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>3.000000e-02</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>7.792360e+05</td>\n",
       "      <td>3.320000e+00</td>\n",
       "      <td>4.670000e+00</td>\n",
       "      <td>4.970000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.558472e+06</td>\n",
       "      <td>5.750000e+00</td>\n",
       "      <td>5.880000e+00</td>\n",
       "      <td>9.650000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2.337708e+06</td>\n",
       "      <td>8.240000e+00</td>\n",
       "      <td>7.410000e+00</td>\n",
       "      <td>1.563000e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3.116944e+06</td>\n",
       "      <td>8.067000e+01</td>\n",
       "      <td>8.872000e+01</td>\n",
       "      <td>1.029000e+02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 id  cap-diameter   stem-height    stem-width\n",
       "count  3.116945e+06  3.116941e+06  3.116945e+06  3.116945e+06\n",
       "mean   1.558472e+06  6.309848e+00  6.348333e+00  1.115379e+01\n",
       "std    8.997847e+05  4.657931e+00  2.699755e+00  8.095477e+00\n",
       "min    0.000000e+00  3.000000e-02  0.000000e+00  0.000000e+00\n",
       "25%    7.792360e+05  3.320000e+00  4.670000e+00  4.970000e+00\n",
       "50%    1.558472e+06  5.750000e+00  5.880000e+00  9.650000e+00\n",
       "75%    2.337708e+06  8.240000e+00  7.410000e+00  1.563000e+01\n",
       "max    3.116944e+06  8.067000e+01  8.872000e+01  1.029000e+02"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "edbdd8e9-706b-4df7-8633-7b123385d3f0",
   "metadata": {},
   "source": [
    "<a name=\"eda\"></a>\n",
    "## 2. Exploratory Data Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af4392ec-3b79-4ebe-90e2-02397575ec28",
   "metadata": {},
   "source": [
    "<a id=\"mushroomtype\"></a>\n",
    "### 2.1. Mushroom Type"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "55e18b4f-1866-47cd-bb76-0883019137dd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='class', ylabel='count'>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(train_data, x=\"class\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fe4f0a69-8180-4ed5-81e2-c586e387c28d",
   "metadata": {},
   "source": [
    "<a name=\"numerical\"></a>\n",
    "### 2.2. Numerical Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "94fd32e6-e39e-454a-a5ad-af1ccdbea3b1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='class', ylabel='cap-diameter'>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(train_data, x=\"class\", y=\"cap-diameter\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "da54b6b6-dfab-491a-b490-6ade0702db7c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='class', ylabel='stem-width'>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(train_data, x=\"class\", y=\"stem-width\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ad8a6a42-dd86-4bb9-affc-77fb383e9aa8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='class', ylabel='stem-height'>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(train_data, x=\"class\", y=\"stem-height\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cc042e99-cee0-4d3f-9837-419b704efcbb",
   "metadata": {},
   "source": [
    "<a name=\"categorical\"></a>\n",
    "### 2.3 Categorical Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b774d126-88d4-43f9-96be-63673ee51082",
   "metadata": {},
   "outputs": [],
   "source": [
    "def print_category_analysis(col_name: pd.Series):\n",
    "    print(train_data[col_name].value_counts().to_string())\n",
    "    print(set(test_data[col_name]) & set(train_data[col_name]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "eaea5428-4232-4c9d-ac32-a163cac4d1b8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cap-shape\n",
      "x            1436026\n",
      "f             676238\n",
      "s             365146\n",
      "b             318646\n",
      "o             108835\n",
      "p             106967\n",
      "c             104520\n",
      "d                 65\n",
      "e                 60\n",
      "n                 41\n",
      "t                 36\n",
      "w                 36\n",
      "g                 34\n",
      "y                 33\n",
      "r                 32\n",
      "a                 30\n",
      "u                 27\n",
      "z                 22\n",
      "l                 21\n",
      "i                 14\n",
      "k                 13\n",
      "h                  6\n",
      "m                  2\n",
      "8                  2\n",
      "3.55               2\n",
      "7 x                2\n",
      "4.97               2\n",
      "3.25               1\n",
      "11.12              1\n",
      "3 x                1\n",
      "4.3                1\n",
      "7.41               1\n",
      "6.21               1\n",
      "8.29               1\n",
      "7.21               1\n",
      "54.78              1\n",
      "2.82               1\n",
      "20.25              1\n",
      "3.52               1\n",
      "3.04               1\n",
      "6.53               1\n",
      "2.63               1\n",
      "3.91               1\n",
      "6.44               1\n",
      "8.3                1\n",
      "7.6                1\n",
      "1.66               1\n",
      "17.44              1\n",
      "4.33               1\n",
      "3.37               1\n",
      "7.43               1\n",
      "9.13               1\n",
      "ring-type          1\n",
      "19.29              1\n",
      "5 f                1\n",
      "2.85               1\n",
      "is s               1\n",
      "4.22               1\n",
      "3.6                1\n",
      "21.56              1\n",
      "6 x                1\n",
      "24.16              1\n",
      "10.13              1\n",
      "b f                1\n",
      "is p               1\n",
      "0.82               1\n",
      "10.46              1\n",
      "2.77               1\n",
      "2.94               1\n",
      "12.62              1\n",
      "5.15               1\n",
      "19.04              1\n",
      "49.21              1\n",
      "19.06              1\n",
      "{nan, '7.21', 'o', 'w', 'k', 'g', '8', 'x', 'a', 'm', 'p', 'ring-type', 'y', '8.3', 'z', 'f', 'r', 'b', 'i', '5 f', 't', 'u', 'h', 'c', 'n', 's', 'd', 'l', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"cap-shape\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "569778bf-d7e7-4a91-bbb3-12918a33b452",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cap-surface\n",
      "t                       460777\n",
      "s                       384970\n",
      "y                       327826\n",
      "h                       284460\n",
      "g                       263729\n",
      "d                       206832\n",
      "k                       128875\n",
      "e                       119712\n",
      "i                       113440\n",
      "w                       109840\n",
      "l                        44859\n",
      "n                           96\n",
      "f                           75\n",
      "p                           59\n",
      "u                           49\n",
      "a                           44\n",
      "m                           42\n",
      "o                           39\n",
      "r                           35\n",
      "x                           33\n",
      "c                           31\n",
      "b                           18\n",
      "z                            8\n",
      "season                       7\n",
      "ring-type                    4\n",
      "class                        2\n",
      "has-ring                     2\n",
      "does None                    2\n",
      "does t                       2\n",
      "spore-print-color            1\n",
      "4.93                         1\n",
      "does-bruise-or-bleed         1\n",
      "10.1                         1\n",
      "2.79                         1\n",
      "spore-color                  1\n",
      "2.11                         1\n",
      "0.88                         1\n",
      "2.81                         1\n",
      "is k                         1\n",
      "7.99                         1\n",
      "2.92                         1\n",
      "2.51                         1\n",
      "10.34                        1\n",
      "is y                         1\n",
      "3.64                         1\n",
      "3.33                         1\n",
      "41.91                        1\n",
      "12.2                         1\n",
      "8.01                         1\n",
      "9.22                         1\n",
      "is None                      1\n",
      "1.14                         1\n",
      "6.49                         1\n",
      "1.08                         1\n",
      "24.38                        1\n",
      "veil-color                   1\n",
      "5.73                         1\n",
      "6.45                         1\n",
      "15.94                        1\n",
      "2.7                          1\n",
      "does l                       1\n",
      "5.07                         1\n",
      "cap-diameter                 1\n",
      "1.43                         1\n",
      "7.14                         1\n",
      "is h                         1\n",
      "0.85                         1\n",
      "6.57                         1\n",
      "12.79                        1\n",
      "4.66                         1\n",
      "14.04                        1\n",
      "23.18                        1\n",
      "3.06                         1\n",
      "16.39                        1\n",
      "4.21                         1\n",
      "11.78                        1\n",
      "8.1                          1\n",
      "does h                       1\n",
      "1.42                         1\n",
      "has h                        1\n",
      "10.83                        1\n",
      "8.96                         1\n",
      "0.87                         1\n",
      "{nan, 'season', 'spore-print-color', 'o', 'w', 'does None', 'k', 'has-ring', 'g', 'class', 'x', 'a', 'm', 'p', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'l', 'd', 's', 'n', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"cap-surface\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "57647800-5a9e-439b-908e-c77dd4154192",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cap-color\n",
      "n                       1359542\n",
      "y                        386627\n",
      "w                        379442\n",
      "g                        210825\n",
      "e                        197290\n",
      "o                        178847\n",
      "p                         91838\n",
      "r                         78236\n",
      "u                         73172\n",
      "b                         61313\n",
      "k                         59888\n",
      "l                         39558\n",
      "f                            53\n",
      "d                            48\n",
      "s                            39\n",
      "a                            30\n",
      "t                            26\n",
      "h                            22\n",
      "m                            14\n",
      "x                            14\n",
      "c                            14\n",
      "i                            13\n",
      "z                            10\n",
      "ring-type                     6\n",
      "season                        6\n",
      "class                         5\n",
      "11.13                         2\n",
      "does-bruise-or-bleed          2\n",
      "3.34                          2\n",
      "2.7                           1\n",
      "8.57                          1\n",
      "17.93                         1\n",
      "2.57                          1\n",
      "11.92                         1\n",
      "3.08                          1\n",
      "2.82                          1\n",
      "4.24                          1\n",
      "17.19                         1\n",
      "20.02                         1\n",
      "7                             1\n",
      "2.9                           1\n",
      "6.36                          1\n",
      "5.91                          1\n",
      "10.56                         1\n",
      "26.89                         1\n",
      "4. n                          1\n",
      "20.62                         1\n",
      "stem-surface                  1\n",
      "20                            1\n",
      "25.98                         1\n",
      "8.67                          1\n",
      "3.11                          1\n",
      "does n                        1\n",
      "5.41                          1\n",
      "10.1                          1\n",
      "2.05                          1\n",
      "13                            1\n",
      "7.72                          1\n",
      "6.76                          1\n",
      "7.15                          1\n",
      "12.89                         1\n",
      "8.83                          1\n",
      "24.75                         1\n",
      "22.38                         1\n",
      "1.51                          1\n",
      "17.94                         1\n",
      "6.9                           1\n",
      "3.57                          1\n",
      "4.89                          1\n",
      "6.2                           1\n",
      "21.53                         1\n",
      "6.41                          1\n",
      "4.98                          1\n",
      "3.95                          1\n",
      "6.59                          1\n",
      "5.25                          1\n",
      "veil-color                    1\n",
      "9.02                          1\n",
      "{nan, 'season', 'o', 'w', 'k', 'g', 'class', 'x', 'a', 'm', 'p', 'ring-type', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'n', 'l', 's', 'd', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"cap-color\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "37c8ed0f-6027-4846-be9d-a30b299e1e95",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "does-bruise-or-bleed\n",
      "f           2569743\n",
      "t            547085\n",
      "w                14\n",
      "c                11\n",
      "h                 9\n",
      "a                 7\n",
      "b                 7\n",
      "x                 7\n",
      "y                 7\n",
      "k                 6\n",
      "s                 6\n",
      "e                 4\n",
      "l                 4\n",
      "p                 4\n",
      "d                 4\n",
      "g                 3\n",
      "z                 3\n",
      "o                 3\n",
      "i                 2\n",
      "n                 2\n",
      "r                 1\n",
      "has-ring          1\n",
      "3.43              1\n",
      "4.42              1\n",
      "2.9               1\n",
      "u                 1\n",
      "{nan, 'o', 'w', 'k', 'g', 'x', 'a', 'p', 'y', 'f', 'r', 'b', 'i', 't', 'h', 'c', 'n', 's', 'l', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"does-bruise-or-bleed\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "f04c7997-9424-4ee2-b45e-3dc0a5b50da9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gill-attachment\n",
      "a                       646034\n",
      "d                       589236\n",
      "x                       360878\n",
      "e                       301858\n",
      "s                       295439\n",
      "p                       279110\n",
      "f                       119953\n",
      "c                           74\n",
      "u                           56\n",
      "w                           37\n",
      "t                           36\n",
      "k                           33\n",
      "y                           28\n",
      "i                           25\n",
      "m                           22\n",
      "g                           22\n",
      "b                           21\n",
      "n                           20\n",
      "o                           18\n",
      "h                           18\n",
      "l                           16\n",
      "r                           10\n",
      "z                            5\n",
      "season                       5\n",
      "3.45                         2\n",
      "2.54                         1\n",
      "6.32                         1\n",
      "13.15                        1\n",
      "19.65                        1\n",
      "2.41                         1\n",
      "15.49                        1\n",
      "4.01                         1\n",
      "8.37                         1\n",
      "does f                       1\n",
      "6.11                         1\n",
      "28.7                         1\n",
      "1.51                         1\n",
      "1.37                         1\n",
      "28.15                        1\n",
      "7.09                         1\n",
      "9.88                         1\n",
      "2.67                         1\n",
      "18.21                        1\n",
      "1.48                         1\n",
      "5.93                         1\n",
      "1.32                         1\n",
      "16.27                        1\n",
      "11.26                        1\n",
      "2.79                         1\n",
      "is f                         1\n",
      "8.47                         1\n",
      "3.91                         1\n",
      "has d                        1\n",
      "11.62                        1\n",
      "32.54                        1\n",
      "does-bruise-or-bleed         1\n",
      "4.64                         1\n",
      "4.77                         1\n",
      "p p                          1\n",
      "7.92                         1\n",
      "8.79                         1\n",
      "does None                    1\n",
      "has f                        1\n",
      "ring-type                    1\n",
      "16.33                        1\n",
      "10.85                        1\n",
      "20.07                        1\n",
      "2.82                         1\n",
      "7.86                         1\n",
      "does                         1\n",
      "10.23                        1\n",
      "6.74                         1\n",
      "0.92                         1\n",
      "1                            1\n",
      "is a                         1\n",
      "3.71                         1\n",
      "50.44                        1\n",
      "13.94                        1\n",
      "{nan, 'season', '2.82', '6.11', 'o', 'w', 'k', 'g', 'x', 'a', 'm', 'p', 'y', 'z', 'f', 'r', 'b', '1.32', 'i', 't', 'u', 'h', 'c', 'n', 's', 'd', 'l', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"gill-attachment\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "56a1cd6b-8580-4e41-be69-93081ebbeca7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gill-spacing\n",
      "c              1331054\n",
      "d               407932\n",
      "f               119380\n",
      "e                   24\n",
      "a                   17\n",
      "s                   16\n",
      "b                   12\n",
      "x                    8\n",
      "t                    8\n",
      "p                    7\n",
      "k                    4\n",
      "g                    4\n",
      "l                    3\n",
      "h                    3\n",
      "r                    2\n",
      "6.67                 2\n",
      "y                    2\n",
      "0                    2\n",
      "3.81                 1\n",
      "1.6                  1\n",
      "n                    1\n",
      "1.88                 1\n",
      "does f               1\n",
      "4.09                 1\n",
      "1.36                 1\n",
      "3.24                 1\n",
      "5.55                 1\n",
      "5.42                 1\n",
      "5.7                  1\n",
      "3.62                 1\n",
      "6.4                  1\n",
      "24.38                1\n",
      "12.27                1\n",
      "1                    1\n",
      "cap-surface          1\n",
      "w                    1\n",
      "3.57                 1\n",
      "i                    1\n",
      "0.73                 1\n",
      "4.04                 1\n",
      "5.22                 1\n",
      "3.92                 1\n",
      "9.01                 1\n",
      "class                1\n",
      "4.8                  1\n",
      "2.69                 1\n",
      "3.61                 1\n",
      "55.13                1\n",
      "{'b', 'i', 't', 'x', 'w', nan, 'a', '0.73', 'k', 'p', 'c', 'd', 's', 'e', 'y', 'f', 'g', 'r'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"gill-spacing\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "cd098763-6dd4-4fdb-9f23-3c1f23df8078",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gill-color\n",
      "w                       931538\n",
      "n                       543386\n",
      "y                       469464\n",
      "p                       343626\n",
      "g                       212164\n",
      "o                       157119\n",
      "k                       127970\n",
      "f                       119694\n",
      "r                        62799\n",
      "e                        56047\n",
      "b                        47247\n",
      "u                        45399\n",
      "l                           55\n",
      "d                           54\n",
      "t                           54\n",
      "s                           46\n",
      "x                           35\n",
      "c                           31\n",
      "a                           28\n",
      "h                           24\n",
      "m                           16\n",
      "z                           16\n",
      "i                            9\n",
      "class                        7\n",
      "season                       7\n",
      "ring-type                    5\n",
      "spacing                      4\n",
      "spore-print-color            3\n",
      "5                            3\n",
      "has-ring                     3\n",
      "habitat                      2\n",
      "18.03                        2\n",
      "e y                          1\n",
      "is y                         1\n",
      "1.91                         1\n",
      "does n                       1\n",
      "4.64                         1\n",
      "10.21                        1\n",
      "16.41                        1\n",
      "6.41                         1\n",
      "8.83                         1\n",
      "veil-type                    1\n",
      "does-bruise-or-bleed         1\n",
      "20.44                        1\n",
      "6.4                          1\n",
      "10.07                        1\n",
      "20.6                         1\n",
      "7.59                         1\n",
      "4                            1\n",
      "0.92                         1\n",
      "18.12                        1\n",
      "3.39                         1\n",
      "does w                       1\n",
      "3.45                         1\n",
      "8.06                         1\n",
      "stem-root                    1\n",
      "6.19                         1\n",
      "4.49                         1\n",
      "9.46                         1\n",
      "5.01                         1\n",
      "3.4                          1\n",
      "17                           1\n",
      "8.37                         1\n",
      "{nan, 'season', '4', 'spore-print-color', '5.01', 'o', 'does-bruise-or-bleed', 'w', 'k', 'has-ring', 'g', 'class', 'x', 'a', 'm', 'p', 'ring-type', 'habitat', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'spacing', 'c', 'n', 'l', 'd', 's', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"gill-color\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "a7575923-0c11-4a57-a313-974ebbade412",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stem-root\n",
      "b                    165801\n",
      "s                    116946\n",
      "r                     47803\n",
      "c                     28592\n",
      "f                       597\n",
      "d                        24\n",
      "y                        14\n",
      "g                        12\n",
      "w                        12\n",
      "p                        12\n",
      "k                        11\n",
      "t                        10\n",
      "l                        10\n",
      "n                        10\n",
      "x                         9\n",
      "i                         8\n",
      "u                         8\n",
      "a                         7\n",
      "e                         6\n",
      "o                         5\n",
      "m                         4\n",
      "h                         4\n",
      "z                         2\n",
      "10.87                     1\n",
      "13.03                     1\n",
      "3.49                      1\n",
      "spore-print-color         1\n",
      "18.06                     1\n",
      "20.0                      1\n",
      "1.48                      1\n",
      "3.23                      1\n",
      "15.69                     1\n",
      "16.88                     1\n",
      "5.59                      1\n",
      "2.77                      1\n",
      "20.01                     1\n",
      "3.63                      1\n",
      "7.15                      1\n",
      "{nan, 'o', 'w', 'k', 'g', 'x', 'a', 'm', 'p', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', '1.48', 'h', 'c', 'l', 's', 'd', 'n', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"stem-root\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "c16db19b-ef40-401a-a851-ca56deefca4d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stem-surface\n",
      "s                       327610\n",
      "y                       255500\n",
      "i                       224346\n",
      "t                       147974\n",
      "g                        78080\n",
      "k                        73383\n",
      "h                        28283\n",
      "f                          512\n",
      "w                           49\n",
      "d                           48\n",
      "e                           36\n",
      "n                           34\n",
      "x                           26\n",
      "b                           23\n",
      "c                           22\n",
      "l                           20\n",
      "p                           20\n",
      "r                           20\n",
      "u                           16\n",
      "o                           15\n",
      "a                           11\n",
      "m                            7\n",
      "season                       6\n",
      "0.0                          3\n",
      "z                            2\n",
      "does-bruise-or-bleed         2\n",
      "5.56                         2\n",
      "10.48                        2\n",
      "does s                       1\n",
      "1.03                         1\n",
      "7.45                         1\n",
      "has h                        1\n",
      "1.94                         1\n",
      "does None                    1\n",
      "10.93                        1\n",
      "49.46                        1\n",
      "19.35                        1\n",
      "2.68                         1\n",
      "4.74                         1\n",
      "spore-print-color            1\n",
      "5.35                         1\n",
      "24.12                        1\n",
      "7.23                         1\n",
      "5.81                         1\n",
      "17.46                        1\n",
      "13.1                         1\n",
      "has-ring                     1\n",
      "class                        1\n",
      "5.51                         1\n",
      "5.48                         1\n",
      "4.41                         1\n",
      "3.68                         1\n",
      "5.97                         1\n",
      "1.59                         1\n",
      "25.83                        1\n",
      "3.89                         1\n",
      "4.34                         1\n",
      "6.58                         1\n",
      "12.04                        1\n",
      "13.94                        1\n",
      "{nan, 'season', 'o', 'does-bruise-or-bleed', 'w', 'does None', 'k', 'g', 'class', 'x', 'a', 'm', 'p', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'l', 's', 'd', 'n', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"stem-surface\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "b8dc51f8-9c8d-45b1-a44d-60c6c5461584",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stem-color\n",
      "w                    1196637\n",
      "n                    1003464\n",
      "y                     373971\n",
      "g                     132019\n",
      "o                     111541\n",
      "e                     103373\n",
      "u                      67017\n",
      "p                      54690\n",
      "k                      33676\n",
      "r                      22329\n",
      "l                       9994\n",
      "b                       7365\n",
      "f                        573\n",
      "s                         51\n",
      "t                         40\n",
      "a                         36\n",
      "x                         20\n",
      "i                         19\n",
      "d                         16\n",
      "h                         12\n",
      "m                         11\n",
      "c                          7\n",
      "z                          6\n",
      "ring-type                  3\n",
      "class                      2\n",
      "spore-print-color          2\n",
      "is n                       1\n",
      "6.31                       1\n",
      "12.92                      1\n",
      "3.98                       1\n",
      "20.07                      1\n",
      "7.7                        1\n",
      "22.6                       1\n",
      "6.09                       1\n",
      "17.45                      1\n",
      "3.56                       1\n",
      "3.37                       1\n",
      "4.62                       1\n",
      "2.54                       1\n",
      "39.51                      1\n",
      "18.06                      1\n",
      "3.53                       1\n",
      "4.49                       1\n",
      "1.41                       1\n",
      "e n                        1\n",
      "8.49                       1\n",
      "2.75                       1\n",
      "7.84                       1\n",
      "4.75                       1\n",
      "26.4                       1\n",
      "is w                       1\n",
      "8.32                       1\n",
      "23.59                      1\n",
      "2.78                       1\n",
      "3.13                       1\n",
      "33.52                      1\n",
      "7.33                       1\n",
      "1.75                       1\n",
      "4.33                       1\n",
      "{nan, 'spore-print-color', 'o', 'w', 'k', 'g', 'class', 'x', 'a', 'm', 'p', 'ring-type', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'n', 'l', 's', 'd', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"stem-color\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "78e56118-965b-459d-a574-d118008b21a7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "veil-type\n",
      "u          159373\n",
      "w              11\n",
      "a               9\n",
      "f               8\n",
      "e               8\n",
      "b               5\n",
      "c               5\n",
      "y               4\n",
      "k               4\n",
      "g               4\n",
      "n               3\n",
      "s               3\n",
      "r               2\n",
      "d               2\n",
      "p               2\n",
      "h               2\n",
      "i               2\n",
      "l               1\n",
      "is None         1\n",
      "t               1\n",
      "21.11           1\n",
      "5.94            1\n",
      "{'b', 'w', 'a', 'u', nan, 'k', 'p', 'l', 's', 'c', 'n', 'e', 'y', 'g'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"veil-type\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "c3351b90-696e-4ba4-8c83-a1e96f07aed7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "veil-color\n",
      "w       279070\n",
      "y        30473\n",
      "n        30039\n",
      "u        14026\n",
      "k        13080\n",
      "e         9169\n",
      "g           30\n",
      "p           23\n",
      "r           14\n",
      "o           13\n",
      "s           11\n",
      "a            7\n",
      "t            7\n",
      "d            6\n",
      "i            6\n",
      "h            5\n",
      "c            4\n",
      "f            4\n",
      "l            4\n",
      "b            3\n",
      "z            1\n",
      "8.25         1\n",
      "2.49         1\n",
      "3.32         1\n",
      "{nan, 'o', 'w', 'k', 'g', 'a', 'p', 'y', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'n', 's', 'd', 'l', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"veil-color\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "96df4d8f-6012-4e5a-ab23-58a1882f603d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "has-ring\n",
      "f             2368820\n",
      "t              747982\n",
      "r                  16\n",
      "h                  13\n",
      "c                  11\n",
      "l                  11\n",
      "s                  11\n",
      "p                  11\n",
      "g                   8\n",
      "z                   6\n",
      "e                   6\n",
      "x                   5\n",
      "m                   4\n",
      "y                   3\n",
      "d                   3\n",
      "k                   2\n",
      "o                   2\n",
      "n                   2\n",
      "f has-ring          1\n",
      "i                   1\n",
      "10.3                1\n",
      "w                   1\n",
      "a                   1\n",
      "{nan, 'o', 'w', 'k', 'g', 'x', 'a', 'm', 'p', 'y', 'z', 'f', 'r', 'i', 't', 'h', 'c', 'l', 's', 'd', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"has-ring\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "963cdaeb-3bb9-47d0-a00b-1ecaa5aae32f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ring-type\n",
      "f                       2477170\n",
      "e                        120006\n",
      "z                        113780\n",
      "l                         73443\n",
      "r                         67909\n",
      "p                         67678\n",
      "g                         63687\n",
      "m                          3992\n",
      "t                            98\n",
      "d                            37\n",
      "n                            34\n",
      "x                            32\n",
      "b                            29\n",
      "y                            25\n",
      "s                            18\n",
      "k                            18\n",
      "a                            16\n",
      "h                            15\n",
      "w                            14\n",
      "u                            13\n",
      "c                            11\n",
      "o                             9\n",
      "ring-type                     7\n",
      "i                             4\n",
      "does f                        3\n",
      "season                        3\n",
      "15                            1\n",
      "14                            1\n",
      "2.87                          1\n",
      "sp                            1\n",
      "class                         1\n",
      "spore-print-color             1\n",
      "2                             1\n",
      "11                            1\n",
      "1                             1\n",
      "23.6                          1\n",
      "4                             1\n",
      "does-bruise-or-bleed          1\n",
      "3.12                          1\n",
      "8.25                          1\n",
      "{'does f', nan, 'season', 'spore-print-color', 'o', '1', 'w', '2', 'k', 'g', 'x', 'a', 'm', 'p', 'ring-type', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'l', 'd', 'n', 's', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"ring-type\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "84a4b619-90a2-494f-b565-82298e5ad607",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "spore-print-color\n",
      "k             107310\n",
      "p              68237\n",
      "w              50173\n",
      "n              22646\n",
      "r               7975\n",
      "u               7256\n",
      "g               3492\n",
      "y                 36\n",
      "s                 21\n",
      "c                 16\n",
      "e                 14\n",
      "f                 14\n",
      "t                 13\n",
      "a                 13\n",
      "d                  9\n",
      "l                  8\n",
      "b                  5\n",
      "m                  4\n",
      "h                  3\n",
      "o                  3\n",
      "x                  2\n",
      "i                  2\n",
      "veil-color         2\n",
      "10 None            1\n",
      "9 None             1\n",
      "2.49               1\n",
      "class              1\n",
      "2.62               1\n",
      "season             1\n",
      "9.55               1\n",
      "6.36               1\n",
      "4.58               1\n",
      "{nan, 'o', 'w', 'k', 'veil-color', 'g', 'x', 'a', 'p', 'y', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'n', 'd', 'l', 's', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"spore-print-color\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "783f5d6a-cd2f-4ee2-a8fa-b02287433490",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "habitat\n",
      "d                       2177573\n",
      "g                        454908\n",
      "l                        171892\n",
      "m                        150969\n",
      "h                        120137\n",
      "w                         18530\n",
      "p                         17180\n",
      "u                          5264\n",
      "e                            55\n",
      "s                            52\n",
      "n                            41\n",
      "t                            40\n",
      "r                            38\n",
      "y                            34\n",
      "a                            34\n",
      "k                            31\n",
      "c                            26\n",
      "b                            18\n",
      "o                            12\n",
      "f                            11\n",
      "i                             7\n",
      "x                             5\n",
      "z                             5\n",
      "habitat                       5\n",
      "class                         3\n",
      "spore-print-color             3\n",
      "ring-type                     2\n",
      "18.35                         1\n",
      "3.68                          1\n",
      "2.44                          1\n",
      "1.66                          1\n",
      "17.38                         1\n",
      "veil-type                     1\n",
      "6.75                          1\n",
      "cap-diameter                  1\n",
      "7.37                          1\n",
      "16.46                         1\n",
      "3.11                          1\n",
      "6.63                          1\n",
      "is w                          1\n",
      "5.62                          1\n",
      "7.31                          1\n",
      "10.07                         1\n",
      "5.56                          1\n",
      "is h                          1\n",
      "4                             1\n",
      "2.94                          1\n",
      "does-bruise-or-bleed          1\n",
      "9.28                          1\n",
      "17.1                          1\n",
      "8.09                          1\n",
      "2.25                          1\n",
      "{nan, 'spore-print-color', 'o', 'w', 'k', 'g', 'class', 'x', 'a', 'm', 'p', 'habitat', 'y', 'z', 'f', 'r', 'b', 'i', 't', 'u', 'h', 'c', 'l', 'd', 's', 'n', 'e'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"habitat\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "fb0868e0-086d-4a52-a8d6-4018adf690ea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "season\n",
      "a    1543321\n",
      "u    1153588\n",
      "w     278189\n",
      "s     141847\n",
      "{'s', 'a', 'w', 'u'}\n"
     ]
    }
   ],
   "source": [
    "print_category_analysis(\"season\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a57982a-5acd-4a09-b0f4-e06d91a0b46e",
   "metadata": {},
   "source": [
    "<a name=\"missing\"></a>\n",
    "### 2.4. Missing Values"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b6b3729-339c-420e-a679-e446f8302ef6",
   "metadata": {},
   "source": [
    "Most machine learning algorithms give an error when we train them on data with missing values. Therefore, it's important to identify them before deciding how to handle them (drop features or impute missing value)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "faed8783-7fe6-4fe0-a668-5f517171486d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id                       0.000000\n",
       "class                    0.000000\n",
       "cap-diameter             0.000128\n",
       "cap-shape                0.001283\n",
       "cap-surface             21.528227\n",
       "cap-color                0.000385\n",
       "does-bruise-or-bleed     0.000257\n",
       "gill-attachment         16.809280\n",
       "gill-spacing            40.373988\n",
       "gill-color               0.001829\n",
       "stem-height              0.000000\n",
       "stem-width               0.000000\n",
       "stem-root               88.452732\n",
       "stem-surface            63.551362\n",
       "stem-color               0.001219\n",
       "veil-type               94.884350\n",
       "veil-color              87.936970\n",
       "has-ring                 0.000770\n",
       "ring-type                4.134818\n",
       "spore-print-color       91.425482\n",
       "habitat                  0.001444\n",
       "season                   0.000000\n",
       "dtype: float64"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "missing_data = train_data.isnull().sum()\n",
    "missing_data_per = missing_data / len(train_data) * 100\n",
    "missing_data_per"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "f9d833a9-3fd6-4e8e-b0eb-a26e5fcec603",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='None', ylabel='None'>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp8AAAGsCAYAAACb7syWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8fJSN1AAAACXBIWXMAAA9hAAAPYQGoP6dpAACAk0lEQVR4nOzdeVRV9f7/8SczmpIIKmgOVy0yuybK4IATDpRTDqBmkSOCpjmV8jUzp8zMMC85Ipp69aooJmJmOeaAOKQNGtmgIaAMigaUAuec3x9dzy+uVpoKB3w91nItzt6f/dnvvT8rerH32ftjZTKZTIiIiIiIFAPrki5ARERERB4cCp8iIiIiUmwUPkVERESk2Ch8ioiIiEixUfgUERERkWKj8CkiIiIixUbhU0RERESKjcKniIiIiBQbhU8RERERKTa2JV2AyB+5fDkHo7Gkq3iwWVmBi0tFLl3KQXOhlTyNh+XQWFgWjYdluDEOf0XhUyyYFZr91TKYTOgXugXReFgOjYVl0XiUDrrtLhbLysqqpEsQERGRe0zhU0RERESKjcKn3BNpaWl4enqSlpZ207rY2Fj8/f1LoCoRERGxNPrOp9wT1atX58SJEyVdhoiIiFg4XfmUeyIlJQUPDw9SUlL44YcfCA4OxtPTk27dunH69OmSLk9EREQshMKn3FOFhYWEhoby6KOPcvjwYSIiIti5c2dJlyUiIiIWQuFT7qmMjAwuXLjAhAkTcHBw4NFHH2XQoEElXZaIiIhYCIVPuac+//xznJ2dcXR0NC+rVatWCVYkIiIilkThU+4pLy8vLl++TF5ennnZxYsXS7AiERERsSQKn3JPVa1alX/84x/MnDmTX3/9lZ9++only5eXdFkiIiJiIRQ+5Z6ytrZm6dKlZGRk0KJFC4YOHUr79u1LuiwRERGxEFYmTZ4tFio7O4/CQmNJl/FAs7ICV9eKZGXlaL5kC6DxsBwaC8ui8bAMN8bhr+jKp1gs/V0kIiJS9ih8isVS+BQRESl7FD5FREREpNgofIqIiIhIsVH4FItlZWVV0iWIiIjIPabwKRZL4VNERKTsUfgUERERkWKj8ClmiYmJeHh4lHQZIiIiUoYpfIqIiIhIsVH4fECdOnWK4OBgPD098fPzY/78+Te9V3P37t3069eP5s2b89RTT/HCCy9w7tw5AHJzcxk7diy+vr60bNmSIUOG8MMPPwCQnp7O0KFD8fHxoXXr1owcOZKMjIziPkQRERGxQAqfD6ArV64wePBgfH19SUxMZO3atcTGxpqDJcDFixcZPXo0w4YNIyEhgb1792IymViwYAEAy5cvJzc3l3379rFnzx6qVKnC3LlzAYiIiMDNzY2DBw/y0Ucf8csvv7B06dKSOFQRERGxMLYlXYAUvz179uDg4MBLL72ElZUVtWrVYsWKFXz11VfmNpUrV2bbtm3UqlWL3NxcLl68iLOzM+np6QA4OjqSlJTEhx9+SMuWLZk1axbW1r/9LePg4MDRo0fZtm0bzZs3Z9myZeZ1IiIi8mBTIngAZWZm4u7uXuRVRnXr1sXNzc382c7Ojvj4eFq3bk2XLl2IiIjg0qVL5lvzISEhDBkyhI0bNxIQEMAzzzzDJ598AsDkyZPp3Lkz0dHRtGnThl69enHs2LHiPUgRERGxSAqfDyA3NzcuXLhQ5DueO3fu5MKFC+bP27dv59///jerV69m3759REVF8cQTT5jXf/vtt/j7+7Nx40YSExPp1asXY8eOJScnh9OnT9O3b1+2bt3KoUOHaNq0KSNHjizWYxQRERHLpPD5AGrbti2FhYUsXryY/Px8kpOTmTVrFtevXze3ycnJwdraGkdHR0wmE5999hkffvghBQUFAMTExDBhwgQuXbpEhQoVqFChAuXLl8fe3p7FixczY8YMcnNzcXJyoly5cjg7O5fU4YqIiIgFUfh8ADk5OREdHU1CQgJ+fn4EBwfTr18/6tSpY27Ts2dPWrRoQZcuXWjWrBmLFi1iwIABnD17lvz8fMaNG0ft2rXp0qULTZo0ITY2loULF+Lg4MD06dMxGo20b98eb29vvvjiC+bPn19yBywiIiIWw8r0v+/XEbEQ2dl5FBYaS7qMB5qVFbi6ViQrKwf9pih5Gg/LobGwLBoPy3BjHP6KrnyKxdLfRSIiImWPwqdYLIVPERGRskfhU0RERESKjcKnWKzfv4dUREREygaFT7FYCp8iIiJlj8KniIiIiBQbhU+5IxkZGfzyyy8lXYaIiIiUUgqfpUR4eDjh4eEALF68mKFDhxZ7DVlZWQQEBHD58uVi37eIiIiUDbYlXYDcubCwsBLZ77Vr13TVU0RERO6KwucdOnXqFLNnz+brr7/moYceIigoiJdffplNmzaxdu1aUlNTyc/Px8fHh7feeovKlSsTGRnJN998g42NDfv376dy5cqEhobSt2/fP9zPrl27iIiIIDU1FV9fXwDz/OiRkZEcOXKE1atXYzKZiIqKYuvWrVy4cAErKytat27Nm2++iaOjI+Hh4ZQvX57U1FSOHDmCq6sr06dPZ+fOncTFxWFvb8+YMWMICgoCMM/zfuLECcqXL0/37t156aWXsLGxoWvXrgB07dqVWbNm0blzZ7Zt28bixYtJS0ujdu3ajBs3Dj8/PwCCg4OpUaMGiYmJmEwm4uPjqVChwv0cHhEREbFwuu1+B65cucLgwYPx9fUlMTGRtWvXEhsbS1RUFDNnzmTq1KkkJiayfft2zp07x6pVq8zb7tq1iyZNmnD06FGmT5/OjBkzSEhIuOV+fvzxR0aPHk1oaCjHjh0jKCiI/fv337Lt9u3bWbVqFZGRkRw7dox169Zx4MABtm7dam6zadMmQkJCOH78OI0aNWLIkCHUqVOHhIQEQkNDmT59Ovn5+fzyyy8MHDiQRx99lM8++4y1a9dy6NAhIiMjsbGxIT4+HoD4+Hg6d+7Mvn37eOONN5gyZQpHjhxh1KhRjBo1iu+++86870OHDrFu3Tri4uIUPEVERETh807s2bMHBwcHXnrpJezt7alVqxYrVqygW7duxMfH06hRI65evUpGRgaVK1cmPT3dvK2HhweDBg3Czs4OPz8/AgIC2LJlyy3389FHH/Hkk0/SvXt3bG1t6dChA+3atbtl29atW7Nx40bq1KnD5cuXyc7OplKlSkX23axZM7y8vLC2tqZZs2aUL1+e4OBgbG1tadeuHfn5+WRlZbF3717y8/MZN24cDg4OuLu7M3r0aNasWXPLff/73//mueeew9vbGxsbG9q1a4e/vz/r1q0rUl+1atVwcnL6O6dcREREyhjddr8DmZmZuLu7F3n/ZN26dcnPz2fu3Lls3bqV8uXL4+HhQW5ubpHpIevUqVOkL3d3d7755huOHTtGSEiIeXloaCjp6elUr169SPtatWqRnZ19U00mk4l58+axZ88eKleuTIMGDSgoKCiy70qVKpl/trGxKRIEbxyL0WgkNTWVy5cv4+3tXaT/goICLl26dNO+b9zK/89//mNeZjAYaNasmflz1apVb9pOREREHlwKn3fAzc2NCxcuYDKZzKFt586dJCUlcfDgQbZu3Yqrqytw80NBv78SCZCSkoK7uzteXl6cOHGiyLoFCxawd+/eIssuXryIg4PDTTXNnTuXtLQ0du/ebb6t3a1btyJtbvdl7W5ubtSqVYuPP/7YvCw3N5dLly5RuXJlUlNTb2rfo0cPhg0bZl6WlpaGo6PjHe9bREREHgy67X4H2rZtS2FhIYsXLyY/P9/8cM66deuwtbXFzs6OwsJCtmzZwv79+ykoKDBve/LkSbZs2YLBYGDfvn3s2rWL3r1733I/3bt358yZM2zYsIHCwkIOHDjAp59+esu2ubm5ODg4YGNjw/Xr11m+fDlnzpwpsu/b1a5dO/Ly8li2bBn5+fn8/PPPTJw4kbFjx2JlZWUOv7m5uQD06dOHVatW8eWXXwLw1Vdf0atXL/N3Q0VERET+l8LnHXByciI6OpqEhAT8/PwIDg6mX79+xMfH4+7uTrt27WjVqhVxcXH079+fM2fOmLdt0KABu3btolmzZsyePZt33nkHT0/PW+6nZs2aLF68mDVr1tC0aVMWLlxIx44db9l2zJgxXLt2jRYtWuDv78/Jkyd59tlni+z7dlWoUIEPPviAxMREWrduTYcOHbC2tmbRokUAuLq60rFjR/r27ct//vMfnn76acaNG8ekSZNo0qQJo0ePZuDAgQQHB9/xvkVEROTBYGX6/ZcD5b74/auR5PZlZ+dRWGgs6TIeaFZW4OpakaysHPSbouRpPCyHxsKyaDwsw41x+Cu68ikWS38XiYiIlD0Kn2KxFD5FRETKHj3tXgxGjRpV0iWIiIiIWARd+RSLpdc0iYiIlD0Kn2KxFD5FRETKHoVPERERESk2Cp8PoNjYWPz9/Uu6DBEREXkAKXyKiIiISLFR+CwBp06dIjg4GE9PT/z8/Jg/fz4mk4mNGzfSq1cvfH198fT0JDQ0lMuXLwO/vah+xIgRjBo1isaNG+Pv78/69ev/cB+FhYVMnTqVli1b4uvrS//+/Tl+/HiR9XPnzqVt27Y0adKEyZMnU1hYCPw2febkyZPp1KkTjRs3plWrVixevNi8rb+/P++//z4BAQF4enry/PPP8/333990fN7e3nTq1IkPPvhAr00SERERQOGz2F25coXBgwfj6+tLYmIia9euJTY2lqioKGbOnMnUqVNJTExk+/btnDt3jlWrVpm33bVrF02aNOHo0aNMnz6dGTNmkJCQcMv9bNmyhRMnTrB9+3YOHTqEt7c306ZNM69PT0/HycmJnTt3smHDBuLj4/n4448BmDt3LikpKWzcuJETJ04wefJk5s2bx08//WTefv369bz33nskJCRQr149wsLCKCgoID09nQEDBvD0009z6NAhFi5cyNq1a/80KIuIiMiDQ+GzmO3ZswcHBwdeeukl7O3tqVWrFitWrKBbt27Ex8fTqFEjrl69SkZGBpUrVyY9Pd28rYeHB4MGDcLOzg4/Pz8CAgLYsmXLLffj6OhoDpBnz55l9OjRxMXFmddXqFCBkJAQbG1tqV+/Po8//jjJycnAb+8lfe+996hQoQIXL17EwcEBgIyMDPP2Q4YMoUGDBjg6OvJ///d/XLhwgc8//5y4uDjq1avH888/j52dHfXr12fIkCGsWbPmfpxOERERKWX0kvlilpmZibu7e5HXCNWtW5f8/Hzmzp3L1q1bKV++PB4eHuTm5ha5XV2nTp0ifbm7u/PNN99w7NgxQkJCzMtDQ0PNVyJjYmKIiIjAxcWFsLAwnnvuOQAefvjhIjXY2dlhMBgAuHTpEm+++SanT5/mkUce4cknnwTAaPz/86zXrl3b/HO5cuWoVKkSmZmZpKamcurUKby8vMzrjUYjNjY2d3PaREREpIxQ+Cxmbm5uXLhwAZPJZA5/O3fuJCkpiYMHD7J161ZcXV0BCAsLK7Lt76+CAqSkpODu7o6XlxcnTpwosu7s2bM0bNiQHj16cO3aNT7++GMmTpxYJBT+kdGjR+Pv7090dDS2trZkZ2ezYcOGP6wlLy+P7Oxs3N3dcXNzw9fXl+joaPP67Oxs8vLybuPsiIiISFmn2+7FrG3bthQWFrJ48WLy8/NJTk5m1qxZrFu3DltbW+zs7CgsLGTLli3s37+fgoIC87YnT55ky5YtGAwG9u3bx65du+jdu/ct97Nnzx5GjhxJSkoKjo6OVKpUCVtbWypWrPiXNebk5ODo6IiNjQ2XL19m5syZAEVqWbFiBT/99BO//vorb731FnXr1sXT05Nu3bpx8uRJ4uLiKCwsJCMjg7CwMGbPnn2XZ05ERETKAoXPYubk5ER0dDQJCQn4+fkRHBxMv379iI+Px93dnXbt2tGqVSvi4uLo378/Z86cMW/boEEDdu3aRbNmzZg9ezbvvPMOnp6et9zPiy++SNu2benXrx+NGzfmnXfeYd68ebi5uf1ljW+99RYfffQRTZo0oVevXlSrVo0nnniiSC1NmzblpZdeomXLlmRmZrJ06VKsra2pUaMGy5YtY/369bRo0YJnn32WunXrKnyKiIgIAFYmvQOnVIiMjOTIkSOsXr26pEvB39+fkSNH0qtXr/u6n+zsPAoLjX/dUO4bKytwda1IVlYO+k1R8jQelkNjYVk0Hpbhxjj8FV35FBEREZFio/ApFksX5UVERMoePe1eSowaNaqkSzDbvXt3sexH4VNERKTsUfgUi2VtbY21rs1bBBubPx4Io9GE0ag/FERE5PYofIrFcnZ+qKRLkP/6s7EwGIxcufKLAqiIiNwWhU+xWDM2HCIp7XJJlyF/4h9VH2Zm/1ZYW1spfIqIyG1R+BSL9VPWzySlKnyKiIiUJfpGnQBw9OhR2rZti6enJ2vXri3pckRERKSM0pVPAWDLli00aNCARYsWlXQpIiIiUobpyud9cOrUKYKDg/H09MTPz4/58+djMpnYuHEjvXr1wtfXF09PT0JDQ7l8+bfbypGRkYwYMYJRo0bRuHFj/P39Wb9+/R/uo7CwkKlTp9KyZUt8fX3p378/x48fByA2NhZ/f/8i7YODg4mMjAQgPDycl19+mWeeeYZmzZrx8ssvs3nzZj777DM8PT3Jz8/n888/58UXX8TPz49//vOf9OrVi5MnT5r7O3jwIIGBgXh6euLv78+///1v87pDhw4RGBiIl5cXXbp0IS4u7l6dWhERESnlFD7vsStXrjB48GB8fX1JTExk7dq1xMbGEhUVxcyZM5k6dSqJiYls376dc+fOsWrVKvO2u3btokmTJhw9epTp06czY8YMEhISbrmfLVu2cOLECbZv386hQ4fw9vZm2rRpt13n/v37mT9/Pp988gn/+te/6NatG926dePEiRMYjUaGDx9OQEAAn332GYmJidSqVYs5c+YAcPbsWcLCwujXrx9Hjx7lX//6FxEREezfv5+kpCSGDx/OsGHDSExMZMaMGcyaNYv9+/ff3YkVERGRMkHh8x7bs2cPDg4OvPTSS9jb21OrVi1WrFhBt27diI+Pp1GjRly9epWMjAwqV65Menq6eVsPDw8GDRqEnZ0dfn5+BAQEsGXLllvux9HRkZSUFDZu3MjZs2cZPXr0HV1hbNy4MY899hhOTk43rbOzs2P9+vX079+f/Px8UlNTqVSpkrnWbdu20bBhQwIDA7G1teXJJ59k7dq1NGzYkHXr1tG+fXs6deqEjY0NTZo0oU+fPqxZs+YOz6SIiIiURfrO5z2WmZmJu7s7VlZW5mV169YlPz+fuXPnsnXrVsqXL4+Hhwe5ublFZvGpU6dOkb7c3d355ptvOHbsGCEhIebloaGhhIWFUVBQQExMDBEREbi4uBAWFsZzzz13W3VWrVr1D9fZ2NiQmJhISEgIv/zyC/Xr18fW1tZca0ZGBtWrVy+yzeOPPw5Aamoqhw8fxsvLy7zOYDBQq1at26pLREREyjaFz3vMzc2NCxcuYDKZzAF0586dJCUlcfDgQbZu3YqrqysAYWFhRbb9/VVQgJSUFNzd3fHy8uLEiRNF1p09e5aGDRvSo0cPrl27xscff8zEiRPx8vLC2tqa/Pz8Iu2zs7OLfP59OP5fX3zxBTNmzGDdunU8+eSTACxfvpyzZ88Cv4Xiffv2Fdlm06ZNuLi44ObmRs+ePZk+fbp5XUZGhqbKFBEREUC33e+5tm3bUlhYyOLFi8nPzyc5OZlZs2axbt06bG1tsbOzo7CwkC1btrB//34KCgrM2548eZItW7ZgMBjYt28fu3btonfv3rfcz549exg5ciQpKSk4OjpSqVIlbG1tqVixIvXq1SMrK4vDhw9jMpnYsmULP/zww20fQ05ODtbW1jg6OprrWrVqlTnQdunShdOnT/Phhx9iMBj4+uuvmT17Nra2tgQGBhIfH8+BAwcwGo2cO3eOF154geXLl9/FWRUREZGyQlc+7zEnJyeio6N56623WLFiBeXKleP555+nT58+hIeH065dOxwcHHjiiSfo378/hw8fNm/boEEDdu3axcyZM3F1deWdd97B09Pzlvt58cUXSU9Pp1+/fuTm5lKjRg3mzZuHm5sbbm5uDB8+nPDwcPLy8ujQoQMBAQG3fQwtW7akf//+PP/88xiNRh555BGCg4N59913ycrKolatWixdupR3332XGTNm4OLiQnh4OH5+fgBEREQQERHB6NGjKVeuHF27dmXcuHF3d2JFRESkTLAy6X6oRYiMjOTIkSOsXr26pEuxGJpe0/LdmF4zOzuPwkJjSZdT5llZgatrRbKyctBv7pKlsbAsGg/LcGMc/oqufIrFer1Pi5IuQW6DwWDUvO4iInLbFD7FYmVn55V0CQI4Oz/0p2NhNJoUPkVE5LYpfFqIUaNGlXQJFsdoNGLUndwSdeOlCAaDUbeyRETkntDT7iIiIiJSbBQ+xWL92btIRUREpHRS+BSLpfApIiJS9ih8ioiIiEixUfiUOxIeHk54eHhJlyEiIiKllMKniIiIiBQbhU8Lc+rUKYKDg/H09MTPz4/58+djMpnYuHEjvXr1wtfXF09PT0JDQ7l8+bfZfyIjIxkxYgSjRo2icePG+Pv7s379+j/dz8GDBwkMDMTT0xN/f3/+/e9/m9ft3LmTXr160aRJEwICAvjggw8w/sE7j2JiYujSpQtNmjShW7duxMXFmdcFBwebpxRt27Ytubm59+AMiYiISGmm93xakCtXrjB48GCCg4OJjo7m4sWLBAcHU65cORYuXMiqVato1KgRFy9eZMCAAaxatYoxY8YAsGvXLsLDw4mIiCAxMZGwsDBq1apF8+bNb9rP2bNnCQsL44033qBHjx4kJSXx4osvUrt2bezs7BgzZgxz5syhU6dOfPvtt4wYMQKAgQMHFuknNjaW2bNn8/777+Pj48ORI0cYOXIk5cqVo2PHjgAcOnSImJgYypUrR4UKFe7r+RMRERHLpyufFmTPnj04ODjw0ksvYW9vT61atVixYgXdunUjPj6eRo0acfXqVTIyMqhcuTLp6enmbT08PBg0aBB2dnb4+fkREBDAli1bbrmfbdu20bBhQwIDA7G1teXJJ59k7dq1NGzYkNjYWNq3b0/nzp2xtbWlYcOGDBs2jHXr1t3Uz6ZNm+jbty/NmzfHxsaG5s2b07dv3yJtW7duTbVq1XBycrr3J0xERERKHV35tCCZmZm4u7sXecVQ3bp1yc/PZ+7cuWzdupXy5cvj4eFBbm4upt9NOVOnTp0ifbm7u/PNN99w7NgxQkJCzMtDQ0PJyMigevXqRdo//vjjAFy6dIkGDRoUWffII4+Qmpp6U71ZWVnUrFnzpra7d+82f65ateptHr2IiIg8CBQ+LYibmxsXLlzAZDKZA+jOnTtJSkri4MGDbN26FVdXVwDCwsKKbPv7q6AAKSkpuLu74+XlxYkTJ4qsW7RoEfv27SuybNOmTbi4uFCjRg2Sk5OLrDt//jxVqlS5qd5HHnnkL9vqXZ0iIiLye7rtbkHatm1LYWEhixcvJj8/n+TkZGbNmsW6deuwtbXFzs6OwsJCtmzZwv79+ykoKDBve/LkSbZs2YLBYGDfvn3s2rWL3r1733I/Xbp04fTp03z44YcYDAa+/vprZs+eja2tLb1792b37t1s374dg8HA6dOniYqKumVfgYGBrF+/noSEBAwGA4cPH2b9+vV/uF8RERERXfm0IE5OTkRHR/PWW2+xYsUKypUrx/PPP0+fPn3MT407ODjwxBNP0L9/fw4fPmzetkGDBuzatYuZM2fi6urKO++8g6en5y33U6tWLZYuXcq7777LjBkzcHFxITw8HD8/PwDmz5/PggULmDRpEs7Ozjz33HNFbt3f8Mwzz5Cbm8vMmTNJS0ujWrVqTJgwgR49etyX8yMiIiKln5Xp918clFIpMjKSI0eOsHr16pIu5Z7Kzs6jsPDWr3iS4mFlBa6uFcnKykG/KUqexsNyaCwsi8bDMtwYh7+i2+5isfR3kYiISNmj8CkWS+FTRESk7NF3PsuAUaNGlXQJIiIiIrdFVz5FREREpNgofIrF0jtCRUREyh6FT7FYCp8iIiJlj8KniIiIiBQbiwifkZGRBAcHF+s+Y2Nj8ff3v2f9HTt27A9f6m4J/P39iY2NveW6lJQUPDw8SElJKfZ9i4iIyIPFIsJnWXCrOdRFREREpKgSCZ+ff/45vXv3pnHjxvTr16/IFbedO3fSq1cvmjRpQkBAAB988AFG42+z3JhMJlatWkVAQABeXl7079+fr7/+2rztjh076NKlC02bNuWZZ55h4cKFf1pHYWEhb7/9Ni1atKBDhw4sW7bM/G7J8PBwXn75ZZ555hmaNWtGcnIyHh4eJCYmmrf//dXTxMREPDw8zOsiIyNp06YNPj4+9O7dm127dpnXnTp1iuDgYLy9venUqRMffPDBn77TMjs7m9dffx0/Pz98fX0JDQ3l3LlzwP+/ajl79my8vb2ZNm3aH/Zz6tQpevXqhY+PD0OGDDH38b+ysrJ45ZVXaNmyJX5+fkyZMoXc3Nzbqt9kMrF48WL8/Pzw8vLi7bffxmAw/GFNIiIi8mAp9vCZnZ1NaGgoAQEBHD16lFdffZWdO3cCcPjwYcaMGcPQoUM5cuQIERERrFixglWrVgGwdu1aVqxYwfz580lISKBXr14MGjSIrKwsrl27xquvvsqUKVM4fvw47777LlFRUXz55Zd/WEt6ejrW1tbs3buX9957j6ioKLZs2WJev3//fubPn88nn3xCrVq1bvsYDx8+zPr164mJiSExMZGgoCBee+01CgoKSE9PZ8CAATz99NMcOnSIhQsXsnbtWtavX/+H/b388sskJyezefNm9u3bR926dRk4cGCRQJiXl8fBgwcZO3bsH/azc+dO3nrrLfbv388jjzxCaGgohYWFRdoYjUZGjBiBtbU1O3bsYOvWrWRkZDBlyhTzOfuz+jdt2sTKlStZsmQJhw4dws7OjosXL972uRMREZGyrdjD5969eylXrhwhISHY2dnRtGlTevfuDfx2JbF9+/Z07twZW1tbGjZsyLBhw1i3bh0Aa9asITQ0lMcffxw7OzsCAwOpV68ecXFxADg6OrJx40YSEhKoV68ex48fp1GjRn9Yi7OzM+PGjcPe3p4nn3ySvn37mvsCaNy4MY899hhOTk53dIwODg5cvXqVDRs2cPr0aYKCgkhISMDOzo64uDjq1avH888/j52dHfXr12fIkCGsWbPmln2dP3+eI0eO8Prrr1OlShUcHR155ZVXKCwsZN++feZ2PXr0wN7e/k9rHTx4MB4eHjg4OBAeHk5KSspN4fzrr7/m1KlTvPHGG1SoUAFnZ2cmTpzItm3byM7O/sv6t2zZQp8+fWjYsCH29vaMHj0aZ2fnOzp/IiIiUnYV+wxH6enpuLu7F3mNTq1atfjmm2+4dOkSDRo0KNL+kUceITU1FYDU1FTefvtt5s6da15fWFjIk08+iaOjI//5z39YuHAh48ePJzc3l4CAACZPnsy+fft44403zNvcuDXt7u6OjY2Nebm7u3uR2+NVq1b9W8fo6elJZGQkq1evZtmyZTg6OhIcHMzw4cNJTU3l1KlTeHl5mdsbjUZzHV26dCEtLQ2A6tWrM3PmTABq1qxpbm9jY4O7uzupqak89dRTN9X6v31s27bNfC5vKFeuHJUqVSI9Pb3ItikpKRgMBtq0aVPkmOzt7Tl//vxf1p+RkYG7u3uRWqtXr37H51BERETKpmIPn25ubqSmpmI0GrG2/u3C643bsjVq1CA5OblI+/Pnz1OlShXzti+//DJdunQxr09OTqZSpUrk5uaSkZHBu+++C8A333zDuHHjWLx4MRMnTqR79+5F+o2NjSUzMxOTyWQOwufPn6dGjRrmNv/7nklra2sKCgrMn7Ozs295jGlpabi4uBAdHU1+fj4JCQmMHDmShg0b4ubmhq+vL9HR0UX6ycvLAzAHxRsyMjLMx/noo48CYDAYSEtLM5+X/631f/v4374AcnNzyc7OLnK88Ns5dnR0JDEx0Rwo8/PzOX/+PLVr1+bQoUN/Wr+bmxvnz583rzOZTEX2KyIiIg+2Yr/t7u/vj8lkIjIykvz8fL7++mtiYmIA6N27N7t372b79u0YDAZOnz5NVFSU+bZ8nz59WLRoET/88APw23cyu3TpwtGjR8nLyyMkJIStW7diMpmoWrUq1tbWf3rLNzMzk0WLFpGfn8+JEyeIiYmhX79+f9i+Xr167Nixg8LCQpKTk9m4ceMt23311VcMHTqUpKQk7O3tcXFxAX67zd+tWzdOnjxJXFwchYWFZGRkEBYWxuzZs2/ZV9WqVWnTpg0zZ84kMzOTa9euMXfuXAwGA+3atfvrE/47y5cv58cff+TXX3/lzTffpEGDBjz55JNF2jRq1IjatWsze/Zs8vLyuHbtGrNmzWLgwIEYDIa/rD8oKIgNGzZw4sQJCgoKWLRoEZmZmXdUp4iIiJRdxX7l08nJiejoaKZOncqKFSuoXbs2AQEBnD17lqeeeor58+ezYMECJk2ahLOzM8899xwhISEADBw4EJPJxIgRI8jIyKBatWpMmTKF9u3bA/Cvf/2L9957jylTpuDo6Ejnzp0ZOHDgH9Zy492Wvr6+VKlShQkTJvzpuz/feOMN3nrrLXx8fKhTpw6BgYG3/K5mQEAA586dY/jw4WRnZ+Pi4sKkSZPMt8iXLVvG3LlzmTlzJjY2NrRt25bXXnvtD/c7Z84c5s6dS8+ePfnll19o3LgxK1euNF/xvV0dOnQgLCyM7OxsvL29Wbhwofnq8w22trYsWbKEt99+m06dOnH9+nUaNWrEihUrcHBwoEaNGn9af9euXcnOzmbs2LFcvXqVp59+ushbAEREROTBZmX6s3f8iJSg7Ow8CguNJV3GA83KClxdK5KVlYN+U5Q8jYfl0FhYFo2HZbgxDn9FL5kXi6W/i0RERMoehU+xWAqfIiIiZU+xf+dT5HZZW1tjrT+P7prRaMJoVJAXERHLoPApFsvZ+aGSLqFMMBiMXLnyiwKoiIhYBIVPsVgzNhwiKe1ySZdRqv2j6sPM7N8Ka2srhU8REbEICp9isX7K+pmkVIVPERGRskTfqCth586d+9PPZUFZPCYRERH5exQ+b0NcXJx5Ss/ExETzS9NTUlLML6r/O3bv3s2QIUPMn9esWcPrr79+1/X+vsaSdvr0abp27VrSZYiIiIiFUPi8Dd27d//D+dLvxpUrV4q8Tujy5bJ3izknJ4eCgoKSLkNEREQshMLn75w+fZrnnnsOT09Pnn32WRYtWoS/vz+xsbF/Ou3mHzGZTCxdupRu3brh5eWFt7c348eP59q1ayQmJvLGG2+QlpaGp6cnmzdvZsmSJRw7dgwvLy8AfvjhB0JDQ2nbti2NGjWic+fO7Nmzx9z/qVOnCA4OxtPTEz8/P+bPn18kzEZHR9OxY0caN27Myy+/bJ6KMzIyktGjRzNx4kSaNGlC69at2b59OwsWLKBFixb4+PiwcOFCcz9ZWVm88sortGzZEj8/P6ZMmWLuKzExEX9/fxYtWkSrVq3w8fFh1KhR5Obmcv78efPUqJ6enpw4ceLOB0VERETKFIXP/8rNzWXo0KE0a9aMxMRE5syZw4YNG+6qz+3bt7Nq1SoiIyM5duwY69at48CBA2zduhVfX1+mTZtG9erVOXHiBD179iQ0NBQvLy+OHTsGwKhRo3jsscf49NNPOXbsGH5+fkydOhX47arp4MGD8fX1JTExkbVr1xIbG8v69evN+09NTSU+Pp4dO3Zw8uTJIvPQ79ixg3bt2nH8+HG6d+/O+PHjyc3NZd++fcyaNYv58+eTmpqK0WhkxIgRWFtbs2PHDrZu3UpGRgZTpkwpsp/09HQ+/fRTYmJiOHHiBGvXrqVmzZpERUUBcOLECTw9Pe/qfIqIiEjpp/D5X7t378bGxoZRo0Zhb2+Ph4cHQ4cOvas+W7duzcaNG6lTpw6XL18mOzubSpUqkZ6eflvbL1myhFGjRmEymUhNTcXJycm87Z49e3BwcOCll17C3t6eWrVqsWLFCtq2bWveftSoUTg4OFCtWjW8vb1JTk42r6tfvz5PP/00VlZWtGzZEoPBQFhYGHZ2duarvGlpaXz99decOnWKN954gwoVKuDs7MzEiRPZtm0b2dnZ5v5eeuklHB0dqV27Nr6+vpw9e/auzp2IiIiUTXrV0n9dvHiR6tWrY/27KXVq1qx529vHxcXxxhtvmD9PmzaNdu3aMW/ePPbs2UPlypVp0KABBQUFtz1tZFJSEiNGjCAzM5N69epRuXJl87aZmZm4u7tjZWVlbl+3bl0AfvrpJwCcnZ3N6+zs7DAYDObPlSpVMv9845gffvjhIp+NRiMpKSkYDAbatGlTpDZ7e3vOnz9v/lylSpUi+9LUmCIiInIrCp//Vb16ddLS0jCZTOZAl5aWdtvbd+/ene7duxdZduM7nbt376ZChQoAdOvW7bb6S09PZ/To0bz//vvmK5E7duzgk08+AcDNzY0LFy4UqXfnzp3k5ubi7u7+l/3/PrT+GTc3NxwdHUlMTMTGxgaA/Px8zp8/T+3atTl+/Pht9SMiIiICuu1u5u/vj8lkYvHixeTn5/Pjjz8SHR19V33m5ubi4OCAjY0N169fZ/ny5Zw5c8b89LeDgwO//vorhYWF5s+5ubmYTCby8vIwGAyUK1cOgO+//54FCxYAv4W/tm3bUlhYaK43OTmZWbNmcf369buq+X81atSI2rVrM3v2bPLy8rh27RqzZs1i4MCBRa6k/hEHBwfgt6feRURERBQ+/6t8+fIsXLiQXbt24ePjw7hx42jZsiV2dnZ/u88xY8Zw7do1WrRogb+/PydPnuTZZ5/lzJkzAHh7e+Pi4oK3tzfffvst7dq148qVKzRt2hRXV1cmTJjAq6++StOmTRk9ejS9e/fGzs6OM2fO4OTkRHR0NAkJCfj5+REcHEy/fv3o27fvvTolANja2rJkyRKysrLo1KkTfn5+JCcns2LFCnOw/DOPPfYYTZs2pVWrVuzbt++e1iYiIiKlj5VJX84DIDs7mx9//JGmTZual61evZpt27axbt26EqzswaW53e/ejbnds7PzKCw03vH2Vlbg6lqRrKwc9Jui5Gk8LIfGwrJoPCzDjXH4K/rO538ZDAYGDBjAggULaNOmDSkpKaxdu5YePXqUdGkPrNf7tCjpEsoEg8GI0ajfxiIiYhkUPv/L1dWV9957j7lz5zJmzBicnJzo2bNnkekvpXhlZ+eVdAllgtFoUvgUERGLofD5Ox06dKBDhw4lXYbIPWVtbYW19e293eCP2Njo6+GWxMbGWn9UiEippfApFsvZ+aGSLkH+S2NhWZydH8JgMHLlyi8KoCJS6ih8isXSA0cit3bjQTJrayuFTxEpdRQ+xWL9lPUzSakKnyIiImWJvshVQuLi4ujSpQsAiYmJeHh4AJCSkoKHhwcpKSklWV4Rx44dw9PTs6TLEBERkTJA4bOEdO/enW3btpV0GbfFy8uLEydOlHQZIiIiUgYofN5np0+f5rnnnsPT05Nnn32WRYsW4e/vT2xsrHnO9juRm5vL2LFj8fX1pWXLlgwZMoQffvgBgMjISEaMGMGoUaNo3Lgx/v7+rF+/3rztDz/8QGhoKG3btqVRo0Z07tyZPXv2mNefOnWK4OBgPD098fPzY/78+ZhMpltemY2JicHf35+mTZsyaNAgLl68aO5n27ZtBAQE4OXlxZAhQ3j99dcJDw//u6dQREREyhCFz/soNzeXoUOH0qxZMxITE5kzZw4bNmy4qz6XL19Obm4u+/btY8+ePVSpUoW5c+ea1+/atYsmTZpw9OhRpk+fzowZM0hISABg1KhRPPbYY3z66accO3YMPz8/pk6dCsCVK1cYPHgwvr6+JCYmsnbtWmJjY4uE19/bu3cvH374ITt27CArK4uFCxcCcOLECSZOnMjEiRM5fPgw/fr1IzY29q6OWURERMoOPXB0H+3evRsbGxtGjRqFtbU1Hh4eDB06lOjo6L/dp6OjI0lJSXz44Ye0bNmSWbNmYW39//+G8PDwYNCgQQD4+fkREBDAli1baN68OUuWLKFatWqYTCZSU1NxcnIiPT0dgD179uDg4MBLL72ElZUVtWrVYsWKFZQvX56ffvrppjpCQkJwcnICwN/f33xbftOmTXTq1Ml8Vbdjx456d6qIiIiYKXzeRxcvXqR69epFwmHNmjVve/u4uDjeeOMN8+dp06YREhKCvb09GzduZPr06dSsWZPx48fTqVMnAOrUqVOkD3d3d7755hsAkpKSGDFiBJmZmdSrV4/KlStj+u8kuJmZmbi7u2Nl9f9fRl63bl2AW4ZPV1dX88+2trbmfi5cuMATTzxRpG3NmjXJysq67eMWERGRskvh8z6qXr06aWlpmEwmc6hLS0u77e27d+9O9+7diyxLSkrC39+fgQMHkpOTw9q1axk7diyHDx8GMF/JvCElJQV3d3fS09MZPXo077//vvmq5I4dO/jkk08AcHNz48KFC0Vq3blzJ7m5ubi7u992zTVq1LjpGNPS0rC3t7/tPkRERKTs0nc+7yN/f39MJhOLFy8mPz+fH3/88a5uuQPExMQwYcIELl26RIUKFahQoQLly5c3h7uTJ0+yZcsWDAYD+/btY9euXfTu3Zu8vDwMBgPlypUD4Pvvv2fBggUA5Ofn07ZtWwoLC821JicnM2vWLK5fv35H9QUFBfHpp5+yf/9+cw03Aq6IiIiIwud9VL58eRYuXMiuXbvw8fFh3LhxtGzZEjs7u7/d57hx46hduzZdunShSZMmxMbGsnDhQhwcHABo0KABu3btolmzZsyePZt33nkHT09P6taty4QJE3j11Vdp2rQpo0ePpnfv3tjZ2XHmzBmcnJyIjo4mISEBPz8/goOD6devH3379r2j+v75z38ybdo0pk6dire3N2vWrKF58+Z3dcwiIiJSdliZbnxZT+657OxsfvzxR5o2bWpetnr1arZt28a6devu+f4iIyM5cuQIq1evvud9366zZ89iNBqpV6+eedmoUaOoW7cuY8eOvaO+hi78mBNnM+51iSKl3uM1KrNmTFeys/MoLDSWdDkPJCsrcHWtSFZWDvq/aMnTeFiGG+PwV/Sdz/vIYDAwYMAAFixYQJs2bUhJSWHt2rX06NGjpEu7b77//numTZvG2rVrqVWrFomJiezfv9/8BP6dqO3qxK/5hfehSpHS7R9VHy7pEkRE/jaFz/vI1dWV9957j7lz5zJmzBicnJzo2bMnQ4YMKenS7puOHTvy/fff8+KLL3L16lVq1KjBjBkzaNKkyR339XqfFvehQpGywWAwYjTqEo+IlD667S4WKzs7r6RLEMDZ+SGNhQW5MR5Go0nhswTpNq9l0XhYBt12l1LPaDRi1NfZStSN174aDEb9QrcAGg8RKQv0tLtYrN+/8F5ERETKBoVPsVgKnyIiImWPwqeIiIiIFBuFz1IuLi6OLl26AJCYmIiHhwfw27SaHh4epKSk3PN9BgcHExkZec/7FRERkbJP4bOU6969O9u2bSvpMkRERERui8JnKXH69Gmee+45PD09efbZZ1m0aBH+/v7Exsbi7+//t/rcunUrXbt2xdPTk2eeeYaPPvrIvC4mJsY8hWe3bt2Ii4u7ZR9Go5GlS5fSoUMHmjZtSmBgIPv37zev9/f3Z8qUKbRs2ZIePXpg1OPrIiIiDzS9aqkUyM3NZejQofTt25eVK1dy9uxZwsLC7uqBnMTERCZNmsT7779Pq1atOHDgACNGjOCxxx7jyy+/ZPbs2bz//vv4+Phw5MgRRo4cSbly5ejYsWORfhYsWMDGjRtZuHAhHh4efPLJJ4wYMYI1a9bQqFEjAL788ku2b98OgLW1/t4RERF5kCkJlAK7d+/GxsaGUaNGYW9vj4eHB0OHDr2rPj/88EM6depEmzZtsLa2pnXr1qxdu5Zq1aqxadMm+vbtS/PmzbGxsaF58+b07dv3lvPRb9q0iWHDhtGwYUNsbW3p3Lkz/v7+bNy40dwmICAAJycnnJyc7qpmERERKf0UPkuBixcvUr169SJXDWvWrHnb28fFxeHp6Wn+FxcXR0ZGBtWrVy/SrlGjRlSsWJGsrKyb+n/kkUdITU29qe/baVu1atXbrlVERETKNt12LwWqV69OWloaJpPJfKs9LS3ttrfv3r073bt3L7LsyJEjN/WxfPlyGjduzCOPPEJycnKRdefPn6dKlSo39V2jRg3Onz9/U9vfB069r1NERERu0JXPUsDf3x+TycTixYvJz8/nxx9/JDo6+q767NmzJ59++ikHDhzAaDSyf/9+IiMjqVixIoGBgaxfv56EhAQMBgOHDx9m/fr19O7d+6Z+goKCWLp0KadOncJgMLB9+3Z2795Nz54976o+ERERKZt05bMUKF++PAsXLmT69OksWbKEOnXq0LJlSxISEv52n02bNuXtt9/m7bffJjU1lRo1ahAREcGjjz7Ko48+Sm5uLjNnziQtLY1q1aoxYcIEevTocVM/gwYNwmg0MnbsWDIzM6lduzYRERH4+PjcxRGLiIhIWWVlMplMJV2E/Lns7Gx+/PFHmjZtal62evVqtm3bdsuHgMqK7Ow8Cgv1aqaSZGUFrq4VycrKQb8pSp7Gw3JoLCyLxsMy3BiHv6Lb7qWAwWBgwIAB7Nu3D/ht9qK1a9fSrl27Eq7s/tLfRSIiImWPbruXAq6urrz33nvMnTuXMWPG4OTkRM+ePRkyZEhJl3ZfKXyKiIiUPQqfpUSHDh3o0KFDSZchIiIicld0211EREREio3Cp1gsvR9URESk7Plb4fPy5ct88MEHvPnmm+Tm5rJnz557XZeIwqeIiEgZdMfh89SpUzz99NN8/PHHbNy4kezsbEaPHs2mTZvuR31lSk5ODpcvXy7pMopdRkYGv/zyS0mXISIiIhbgjsPnW2+9RXh4OOvWrcPW1paaNWuyYMGCu55x50HQsWNHvvvuu2LZV2RkJMHBwX97e09PT44dO/aX7VJSUvDw8CAlJeWW67OysggICHggQ7eIiIjc7I7D55kzZ3j22WeB/39btFWrVqSnp9/bysqg7Ozski7htp04cQIvL6+77ufatWu66ikiIiJmdxw+K1euzI8//lhk2Y8//oirq+s9K6q0i4yMpE2bNvj4+NC7d2927dpFQEAAACEhIURFRQFw6NAhAgMD8fLyokuXLsTFxZn7CA8PZ/r06YSGhuLp6UnHjh1JSEhgxowZeHt707JlS2JiYv60jry8PCZPnoyfnx++vr7MmzfPvC4/P5/58+fTvn17fHx8CAkJ4aeffjKv9/DwIDExEfgtNI8dO5amTZvSvn17Vq9ezRNPPFHkaufWrVt55plnaNy4MQMHDiQ9PR2DwUDXrl0B6Nq1Kx999NFdnlkREREp7e44fPbv35/Q0FA2bNhAYWEhH330EaNHj6Zv3773o75S5/Dhw6xfv56YmBgSExMJCgritddeIz4+HoCoqChCQkJISkpi+PDhDBs2jMTERGbMmMGsWbPYv3+/ua9NmzYREhLC8ePHadSoEUOGDKFOnTokJCQQGhrK9OnTyc/P/8NaTp8+jbe3N/v372f+/PksWbKEEydOADBv3jz27t3LBx98wP79+3nqqacYPHgw169fv6mfV155hZycHHbt2kVMTAx79uzBYDAUaXPq1Ck2bNjAvn37uHr1KgsWLMDGxsZ83PHx8XTu3Pmuz6+IiIiUbnccPl988UWGDBnCypUrMRqNzJ8/n549ezJw4MD7UF7p4+DgwNWrV9mwYQOnT58mKCiIhIQE7OzsirRbt24d7du3p1OnTtjY2NCkSRP69OnDmjVrzG2aNWuGl5cX1tbWNGvWjPLlyxMcHIytrS3t2rUjPz+frKysP6zl0Ucf5dlnn8XKyopmzZrh6upKcnIyJpOJdevWMW7cOGrWrImDgwMvvfQSBQUF7N27t0gf6enpHDhwgEmTJlGpUiUqV67MpEmTbtpXWFgYFStW5OGHH6ZVq1YkJyff3YkUERGRMulvzXD0/PPP8/zzz9/rWsoET09PIiMjWb16NcuWLcPR0ZHg4GCGDx9epF1qaiqHDx8u8r1Kg8FArVq1zJ8rVapk/tnGxgYnJyfz5xvftzUajSxevJglS5aY1924rf/77QHs7e0xGAxcvnyZX375hdGjR2Nt/f///igoKCA1NbXINhcuXADgkUceMS+rWbPmTcf9+33Z2dnddGVUREREBP5G+DQYDOzYsYNz585hNBqLrBs5cuQ9K6y0SktLw8XFhejoaPLz80lISGDkyJE0bNiwSDs3Nzd69uzJ9OnTzcsyMjKKzGd+u++5DAsLIywsrMiyhISEP2zv7OyMg4MDy5cvp3HjxublP/74I9WqVSvStnr16sBvYfkf//iH+WcRERGRv+OOb7u/8cYbvPbaa+zfv5/ExETzvyNHjtyP+kqdr776iqFDh5KUlIS9vT0uLi7Ab4HP3t6enJwcAAIDA4mPj+fAgQMYjUbOnTvHCy+8wPLly+97jdbW1gQGBvLuu+9y8eJFjEYjmzdvpmvXrkUeOgKoWrUq7dq145133uHq1atcvXqVOXPm3Pa+HBwcAMjNzb2nxyAiIiKl0x1f+dyzZw+rVq3in//85/2op9QLCAjg3LlzDB8+nOzsbFxcXJg0aRJPPfUUffv2Zfz48QwcOJCxY8cSERFBREQEo0ePply5cnTt2pVx48YVS50TJ04kMjKS/v37c+XKFWrWrMm//vUvnnjiiZvavvnmm0yZMoW2bdvi7OxMjx492LNnD3Z2dhQUFPzpflxdXenYsSN9+/YlPDyc55577n4dkoiIiJQCVqbf3+e9Dc2bN+fAgQPY2Njcr5rEwhw8eJCmTZvi6OgIwLfffkuPHj04efKk+crm/ZCdnUdhofGvG8p9Y2UFrq4VycrK4c5+U8j9oPGwHBoLy6LxsAw3xuGv3PFt965du2o2owfM22+/zaJFiygsLCQ3N5dFixbRokWL+xo8Ae7w7yIREREpBe74tvupU6f4/PPPWbRoEZUrVy6ybteuXfesMLEc7777LjNnzqRZs2ZYW1vTqlWrO/re59+l8CkiIlL23HH4DAoKIigo6H7UIhbq0UcfZeXKlSVdhoiIiJQBdxw+e/bsCcClS5dITU2lSpUquLu73/PCRERERKTsuePwmZuby8SJE9m9ezcmkwkrKyuaN2/Oe++9V+Ql6CJ367f3nOrWu4iISFlyxw8cvfvuu+Tl5REfH88XX3zBli1bMBqNvPPOO/ejPnmA3e5L9kVERKT0uOPwuWfPHt59913q1auHg4MDjz32GO+88w47d+68H/WVeTk5OVy+fLmky7gr586dK+kSREREpJS44/D566+/UrFi0Xc4OTk53TTVptyejh078t1335V0GUUsXryYoUOH/uH64OBgIiMjAdi9ezdDhgy55ToRERGR/3XH4fOpp55i/vz55tfgmEwm5s+frxmP/qbs7OySLuEmYWFhLFu27LbaXrlyRa9EEhERkdt2x+HzlVdeYePGjbRu3Zp+/frRunVr4uPj+b//+7/7UV+ZERkZSZs2bfDx8aF3797s2rWLgIAAAEJCQoiKigLg0KFDBAYG4uXlRZcuXYiLizP3ER4ezvTp0wkNDcXT05OOHTuSkJDAjBkz8Pb2pmXLlsTExNxy/2fOnKFBgwbmsJuSkoKHhwdr1641t+nVqxcbN24kMjKS4OBg8/KYmBjat2+Pp6cnEydO5NdffwUgMTGRN954g7S0NDw9PUlPTwfgp59+YvDgwXh7e9O+fXs+/vjje3gmRUREpDS74/D52GOPsWPHDkaPHo2/vz/h4eF89NFH1KtX737UVyYcPnyY9evXExMTQ2JiIkFBQbz22mvEx8cDEBUVRUhICElJSQwfPpxhw4aRmJjIjBkzmDVrFvv37zf3tWnTJkJCQjh+/DiNGjViyJAh1KlTh4SEBEJDQ5k+fTr5+fk31fDYY49Rs2ZNDh48CMCBAwdwdHTk0KFDAKSnp/Ptt9/i7+9fZLuEhASmT5/OzJkzOXr0KE899RRfffUVAL6+vkybNo3q1atz4sQJqlWrBvw2Hef48eNJTEykV69e/N///d9fzgEvIiIiD4Y7Dp8AlSpVIjAwkGHDhtGlSxcqVKhwr+sqUxwcHLh69SobNmzg9OnTBAUFkZCQgJ2dXZF269ato3379nTq1AkbGxuaNGlCnz59WLNmjblNs2bN8PLywtrammbNmlG+fHmCg4OxtbWlXbt25Ofnk5WVdcs62rdvz2effQb8Fj779etHYmIiBoOB3bt307Rp05tmrYqLi6NTp040b94cW1tb+vfvzxNPPPGnx9u5c2caNmyItbU1nTt35pdffuHSpUt/59SJiIhIGXPb7/n09/f/01ffWFlZ6Yn3P+Dp6UlkZCSrV69m2bJlODo6EhwczPDhw4u0S01N5fDhw3h5eZmXGQwGatWqZf5cqVIl8882NjZF3q16Y3yMRiOLFy9myZIl5nVRUVF07NiRkSNHkp+fz+HDh9m2bRsffvghp06dKvI1gN9LT0+nYcOGRZbVrFnzT4/39zXeCNiFhYV/uo2IiIg8GG47fI4aNeqWy0+ePMn69ev/8mrYgywtLQ0XFxeio6PJz88nISGBkSNH3hTq3Nzc6NmzJ9OnTzcvy8jIKPJAz+2++zIsLIywsLAiy4xGI9bW1qxbtw53d3eqVatGixYt2LlzJ0ePHuXNN9+8qR83NzfOnz9fZNnFixd59NFHb6sOERERkd+77dvuPXv2vOlfdnY2mzZt4rnnnmPdunX3s85S7auvvmLo0KEkJSVhb2+Pi4sLAM7Oztjb25OTkwNAYGAg8fHxHDhwAKPRyLlz53jhhRdYvnz5PanD2toaf39/Fi5ciJ+fHwAtW7Zk1apVPP744+bvbP5e79692blzJ3v27KGwsJDNmzfzxRdfmNc7ODjw66+/6sqmiIiI3JY7nl4T4Oeff2bixIkcO3aMd955h2eeeeZe11WmBAQEcO7cOYYPH052djYuLi5MmjSJp556ir59+zJ+/HgGDhzI2LFjiYiIICIigtGjR1OuXDm6du3KuHHj7lktHTp0YP369bRs2RKAVq1a8euvv9KxY8dbtm/atClz5sxh9uzZjB07lmbNmpm3BfD29sbFxQVvb2/9ASIiIiJ/ycp0hy9pPHnyJGPHjsXZ2Zn58+f/5ff/RP6u7Ow8Cgs1eUFJsrICV9eKZGXloNe5ljyNh+XQWFgWjYdluDEOf+WOnnZftmwZwcHBtG/fnnXr1il4yn2ll9eLiIiUPbd92z0sLIx9+/bxwgsv0KlTpyLf+7vB29v7nhYnDzaFTxERkbLntm+7P/7443/ekZUV33zzzT0pSgR+u+0uJc/Z+SGNhQXReFgOjYVlKS3jYTSaMBrL5sWV273tfsff+RQRERGRv8dgMHLlyi9lMoDebvj8W0+7ixSHGRsOkZR2uaTLEBERuSf+UfVhZvZvhbW1VZkMn7dL4bMUyMnJoaCg4KapL0uLjIwMKlSoQPny5e9ou5+yfiYpVeFTRESkLPlbc7tL8erYsSPfffddSZfxt2RlZREQEMDlywqRIiIiovBZKmRnZ5d0CX/btWvX+OWXX0q6DBEREbEQCp8WJDIykjZt2uDj40Pv3r3ZtWsXAQEBAISEhBAVFQXAoUOHCAwMxMvLiy5duhAXF2fuIzw8nOnTpxMaGoqnpycdO3YkISGBGTNm4O3tTcuWLYmJifnTGgYPHkzv3r3x8fHh6NGjZGdn8/rrr+Pn54evry+hoaGcO3fOvM23335LSEgIPj4+tG7dmqlTp5KTk4PBYKBr164AdO3alY8++ug+nDUREREpTRQ+LcThw4dZv349MTExJCYmEhQUxGuvvUZ8fDwAUVFRhISEkJSUxPDhwxk2bBiJiYnMmDGDWbNmsX//fnNfmzZtIiQkhOPHj9OoUSOGDBlCnTp1SEhIIDQ0lOnTp5Ofn/+HtSQkJPDKK6+wZ88ePD09efnll0lOTmbz5s3s27ePunXrMnDgQHJzc8nOzubFF1+kfv36fPbZZ2zatImzZ88yYcIEbGxszPXHx8fTuXPn+3sSRURExOIpfFoIBwcHrl69yoYNGzh9+jRBQUEkJCRgZ2dXpN26deto3749nTp1wsbGhiZNmtCnTx/WrFljbtOsWTO8vLywtramWbNmlC9fnuDgYGxtbWnXrh35+flkZWX9YS01a9akefPmPPTQQ1y4cIEjR47w+uuvU6VKFRwdHXnllVcoLCxk37597Nq1Czs7O1555RUcHR2pUqUKr7/+Ort37yYzM/O+nS8REREpnfS0u4Xw9PQkMjKS1atXs2zZMhwdHQkODmb48OFF2qWmpnL48GG8vLzMywwGA7Vq1TJ/rlSpkvlnGxsbnJyczJ+trKwAMBqNLF68mCVLlpjX3bitX7VqVfOyGyH191Op2tjY4O7uTmpqKiaTierVq2NjY2Ne/8gjj5hrdXV1vfOTISIiImWWwqeFSEtLw8XFhejoaPLz80lISGDkyJE0bNiwSDs3Nzd69uzJ9OnTzcsyMjKKTEV5I2D+lbCwMMLCwoosS0hIKLJ9jRo1AEhOTubRRx8Ffgu7aWlpVKlSBTs7O9LS0jAYDOYAmpycDECVKlU0RaaIiIgUodvuFuKrr75i6NChJCUlYW9vj4uLCwDOzs7Y29uTk5MDQGBgIPHx8Rw4cACj0ci5c+d44YUXWL58+X2pq2rVqrRp04aZM2eSmZnJtWvXmDt3LgaDgXbt2tGmTRsA5s6dy7Vr18jMzOTNN9+kWbNm1KhRAwcHBwByc3PvS30iIiJSuih8WoiAgAAGDx7M8OHDady4MaNHj2bSpEk89dRT9O3bl/HjxzNv3jyeeuopIiIiiIiIwNvbmxdeeAF/f3/Gjx9/32qbM2cONWvWpGfPnrRo0YJvv/2WlStXUqlSJSpWrMiKFSs4c+YMbdq0oWvXrtSoUYP58+cD4OrqSseOHenbty//+c9/7luNIiIiUjpobnexWEMXfsyJsxklXYaIiMg98XiNyqwZ05Xs7DwKC40lXc49p7ndpdSr7erEr/mFJV2GiIjIPfGPqg+XdAkWQVc+RURERIqJwWDkypVfMBrLXvzSlU8p9bKz80q6BAGcnR/SWFgQjYfl0FhYltIyHkajqUwGzzuh8CkiIlKCFEbu3o03BBoMRnQ/1/LptruIiEgJKsu3YYvLjdu9WVk5Cp8lSLfdpdSbseEQSWmXS7oMEZH75h9VH2Zm/1ZYW1spfMoDQ+HzHsrJyaGgoIDKlSuXdCllwk9ZP5OUqvApIiJSlugl8/dQx44d+e6770q6jL/lgw8+wNvbG29vb5KSkkq6HBERESmjFD7voezs7JIu4W9bu3YtI0aM4OjRozz++OMlXY6IiIiUUQqff0NkZCRt2rTBx8eH3r17s2vXLgICAgAICQkhKioKgEOHDhEYGIiXlxddunQhLi7O3Ed4eDjTp08nNDQUT09POnbsSEJCAjNmzMDb25uWLVsSExPzhzWkp6czdOhQfHx8aN26NSNHjiQjI8Pcd3h4eJH2Hh4eJCYmAuDv78+UKVNo2bIlPXr0oHnz5iQnJxMREcGLL74IwMaNG+nVqxe+vr54enoSGhrK5cv//xb4ypUr6dixI56envTq1YuEhAQATCYTq1atIiAgAC8vL/r378/XX399t6dcREREygiFzzt0+PBh1q9fT0xMDImJiQQFBfHaa68RHx8PQFRUFCEhISQlJTF8+HCGDRtGYmIiM2bMYNasWezfv9/c16ZNmwgJCeH48eM0atSIIUOGUKdOHRISEggNDWX69Onk5+ffso6IiAjc3Nw4ePAgH330Eb/88gtLly697eP48ssv2b59O6tWrSIhIYHq1aszbdo0Vq1axZdffsnMmTOZOnUqiYmJbN++nXPnzrFq1SoAYmNjWbhwIXPmzOH48eM899xzDB8+nCtXrrB27VpWrFjB/PnzSUhIoFevXgwaNIisrKy7OOsiIiJSVih83iEHBweuXr3Khg0bOH36NEFBQSQkJGBnZ1ek3bp162jfvj2dOnXCxsaGJk2a0KdPH9asWWNu06xZM7y8vLC2tqZZs2aUL1+e4OBgbG1tadeuHfn5+X8Y2hwcHDh+/Djbtm0jLy+PZcuWMXny5Ns+joCAAJycnHBycrpp3WOPPUZ8fDyNGjXi6tWrZGRkULlyZdLT0wHYvHkzffv2xdPTE2tra4KCgli+fDmOjo6sWbOG0NBQHn/8cezs7AgMDKRevXpFrvqKiIjIg0tPu98hT09PIiMjWb16NcuWLcPR0ZHg4GCGDx9epF1qaiqHDx/Gy8vLvMxgMFCrVi3z50qVKpl/trGxKRIErf77xlyj0cjixYtZsmSJeV1UVBSTJ09myZIlREdHEx4ezuOPP87kyZOL7O/PVK1a9Q/XWVtbs2rVKrZu3Ur58uXx8PAgNzeXG6+EzczMpHr16kW2adKkifm43377bebOnWteV1hYyJNPPnlbdYmIiEjZpvB5h9LS0nBxcSE6Opr8/HwSEhIYOXIkDRs2LNLOzc2Nnj17Mn36dPOyjIwMfv9O/xsB86+EhYURFhZWZNnJkyfp27cvo0aN4vLlyyxYsICRI0dy+PBhrK2tuX79urnt77+reTv7/uCDDzh48CBbt27F1dXVXMMN7u7uXLhwocg28+bNo3v37ri5ufHyyy/TpUsX87rk5OQiQVtEREQeXLrtfoe++uorhg4dSlJSEvb29ri4uADg7OyMvb09OTk5AAQGBhIfH8+BAwcwGo2cO3eOF154geXLl9+TOhYvXsyMGTPIzc3FycmJcuXK4ezsDEC9evU4duwY6enpXLt2jQULFtx20AXIzc3F1tYWOzs7CgsL2bJlC/v376egoACAXr16sX79er788kuMRiObNm1izZo1ODs706dPHxYtWsQPP/wAwP79++nSpQtHjx69J8ctIiIipZuufN6hgIAAzp07x/Dhw8nOzsbFxYVJkybx1FNP0bdvX8aPH8/AgQMZO3YsERERREREMHr0aMqVK0fXrl0ZN27cPalj+vTpTJs2jfbt25Ofn8+TTz7J/PnzAejbty9fffUV3bt3x97engEDBtx0m/zPDB48mDNnztCuXTscHBx44okn6N+/P4cPHwagW7du/Pzzz7z66qtkZmZSv359oqKiqFy5MgMHDsRkMjFixAgyMjKoVq0aU6ZMoX379vfkuEVERKR009zuYrGGLvyYE2czSroMEZH75vEalVkzpivZ2XkUFhpLupxSS3O7WwbN7S6lXm1XJ37NLyzpMkRE7pt/VH24pEsQKXa68ikiIlKCDAYjV678gtGo/x3/XbryaRl05VNKvezsvJIuQQBn54c0FhZE42E57tVYGI0mBU95oCh8isUyGo0Y9RWoEnXjJQkGg1FXEyyAxsNyaCxE/j69akks1p28HkpERERKB4VPsVgKnyIiImWPwqcFycnJueVsRCIiIiJlhcKnBenYsSPfffddSZfxl/z9/YmNjS3pMkRERKQUUvi0INnZ2SVdgoiIiMh9pfBZAiIjI2nTpg0+Pj707t2bXbt2ERAQAEBISAhRUVEAHDp0iMDAQLy8vOjSpQtxcXHmPsLDw5k+fTqhoaF4enrSsWNHEhISmDFjBt7e3rRs2ZKYmJg/rWPlypV07NgRT09PevXqRUJCAvDbU+ZLly6lQ4cONG3alMDAQPbv33/LPq5du8acOXNo06YN3t7eBAcH8+WXX5rXe3h4MHPmTHx9fQkLC7ur8yYiIiKln8JnMTt8+DDr168nJiaGxMREgoKCeO2114iPjwcgKiqKkJAQkpKSGD58OMOGDSMxMZEZM2Ywa9asIiFw06ZNhISEcPz4cRo1asSQIUOoU6cOCQkJhIaGMn36dPLz829ZR2xsLAsXLmTOnDkcP36c5557juHDh3PlyhUWLFjAmjVrmD9/PomJiQwePJgRI0YUCZU3TJ06lQMHDrBq1SoOHjxIhw4dGDhwIGlpaeY2ycnJ7N27lzlz5tzjsykiIiKljcJnMXNwcODq1ats2LCB06dPExQUREJCAnZ2dkXarVu3jvbt29OpUydsbGxo0qQJffr0Yc2aNeY2zZo1w8vLC2tra5o1a0b58uUJDg7G1taWdu3akZ+fT1ZW1i3r2Lx5M3379sXT0xNra2uCgoJYvnw5jo6ObNq0iWHDhtGwYUNsbW3p3Lkz/v7+bNy4sUgf169fJz4+nvHjx1O7dm3s7e0ZMGAAdevWNYdpgK5du1KuXDmcnJzu4ZkUERGR0kgvmS9mnp6eREZGsnr1apYtW4ajoyPBwcEMHz68SLvU1FQOHz6Ml5eXeZnBYKBWrVrmz5UqVTL/bGNjUyTc3XhNkdFoZPHixSxZssS8LioqiszMTKpXr15kn02aNAEgKyuLmjVrFln3yCOPkJSUVGTZ1atXKSgo4JFHHrmpbUpKivlz1apV//iEiIiIyANF4bOYpaWl4eLiQnR0NPn5+SQkJDBy5EgaNmxYpJ2bmxs9e/Zk+vTp5mUZGRmYfjeVxu2+BzMsLOym71u6u7tz4cKFIsvmzZtH9+7dqVGjBufPny+y7vz58zeFSFdXVxwcHDh//jz16tUzL09OTsbf3/+O6xQREZGyT7fdi9lXX33F0KFDSUpKwt7eHhcXFwCcnZ2xt7cnJycHgMDAQOLj4zlw4ABGo5Fz587xwgsvsHz58ntSR69evVi/fj1ffvklRqORTZs2sWbNGpydnQkKCmLp0qWcOnUKg8HA9u3b2b17Nz179izSh7W1Nb179yYiIoKffvqJ/Px8Vq5cyffff0+XLl3uSZ0iIiJStujKZzELCAjg3LlzDB8+nOzsbFxcXJg0aRJPPfUUffv2Zfz48QwcOJCxY8cSERFBREQEo0ePply5cnTt2pVx48bdkzq6devGzz//zKuvvkpmZib169cnKiqKypUrM2jQIIxGI2PHjiUzM5PatWsTERGBj4/PTf1MmDCByMhIBg4cyJUrV/Dw8CA6Opp//OMf96ROERERKVusTL+/jytiQbKz8ygsNJZ0GQ80Kytwda1IVlYO+k1R8jQelkNjYVk0Hpbhxjj8Fd12F4ulv4tERETKHoVPsVgKnyIiImWPwqeIiIiIFBs9cCQWy9raGmv9eWQRbGw0EJZE42E5NBaWRePx14xGE0Zjyd5Z1ANHIiIiIg8Ig8HIlSu/3JcAersPHOnKp1isGRsOkZR2uaTLEBERKRP+UfVhZvZvhbW1VYle/VT4FADOnTtHnTp1SrqMIn7K+pmkVIVPERGRskRfjiij4uLizLMMJSYm4uHh8YdtT58+TdeuXYurNBEREXmAKXyWUd27d2fbtm231TYnJ4eCgoL7XJGIiIiIwqfFmzBhAuPHjy+ybMyYMUybNo3k5GTCwsLw9fWlXbt2zJs3j/z8fABiY2Px9/f/y/7Pnz9PSEgIAJ6enhw/fpwnnniCzz//3NwmKyuLhg0bkpycTHh4OJMmTeLFF1+kcePGPPPMM+zcubNI21deeYWWLVvi5+fHlClTyM3NvRenQkRERMoAhU8L16dPH3bu3GkOcD///DO7d++ma9euDBw4kEcffZTPPvuMtWvXcujQISIjI++o/5o1axIVFQXAiRMnaNq0KS1btmTLli3mNnFxcXh6elKrVi0ANm/eTL9+/Th27BihoaGMGTOGH374AaPRyIgRI7C2tmbHjh1s3bqVjIwMpkyZco/OhoiIiJR2Cp8WzsvLC3d3d7Zv3w5AfHw8devWJT09nfz8fMaNG4eDgwPu7u6MHj2aNWvW3PU+e/fuzccff2y+irp582Z69+5tXt+2bVs6d+6Mra0tPXr04Mknn+Sjjz7i66+/5tSpU7zxxhtUqFABZ2dnJk6cyLZt28jOzr7rukRERKT009PupUBQUBBbtmwhKCiIzZs3ExQURGpqKpcvX8bb29vczmQyUVBQwKVLl/6wr8WLF7NkyRLz5xtXPX/P39+fN954g3379lG9enVSU1MJCAgwr//fp+Ld3d3JzMwkJSUFg8FAmzZtiqy3t7fn/PnzODs73+mhi4iISBmj8FkK9OzZk/fee49Dhw7x7bff0rVrVz777DNq1arFxx9/bG6Xm5vLpUuXqFy58h/2FRYWRlhYWJFliYmJRT7b29vTrVs3tm3bRvXq1XnmmWcoX768eX16enqR9ikpKfj7++Pm5oajoyOJiYnY2NgAkJ+fz/nz56ldu/bfPn4REREpO3TbvRSoXLky7dq1Y/LkyXTq1ImHH36Ydu3akZeXx7Jly8jPz+fnn39m4sSJjB07Fisrqzvq38HBAfjtqfcbAgMD2b9/P59++im9evUq0v7TTz/l0KFDFBYWsnHjRs6cOUPXrl1p1KgRtWvXZvbs2eTl5XHt2jVmzZrFwIEDMRgMd38iREREpNRT+Cwl+vTpQ2pqKoGBgQBUqFCBDz74gMTERFq3bk2HDh2wtrZm0aJFd9z3Y489RtOmTWnVqhX79u0D4PHHH6dWrVpYW1vTtGnTIu29vLyIiorCx8eHtWvXsnTpUmrWrImtrS1LliwhKyuLTp064efnR3JyMitWrDAHXBEREXmwaW53+UMjR46kUaNGDBs2zLwsPDwcgNmzZ9/3/Wt6TRERkXvnxvSa2dl5FBYa73n/mttd/rbz58+TlJTEoUOHmDp1aonV8XqfFiW2bxERkbLIYDCW6LzuoPApt/D++++za9cuJk2ahKura4nVkZ2dV2L7lv/P2fkhjYUF0XhYDo2FZdF43B6j0VTi4VO33cViXbqUg/He3xWQO3DjFkpWVg76TVHyNB6WQ2NhWTQelkG33aXUs7a2xlqPxFkEGxsNhCXReFgOjYVlsdTxsISrjZZEVz5FRERE7iODwciVK7+U+QCqK59S6ulpdxERKe1uPGFubW1V5sPn7VL4FIv1U9bPJKUqfIqIiJQllvnlCLljcXFxdOnSBfhtukwPD4/7sh9/f39iY2PvS98iIiJS9il8lhHdu3dn27ZtJV2GiIiIyJ9S+LQwEyZMYPz48UWWjRkzhmnTppGcnExYWBi+vr60a9eOefPmkZ+fD0BsbCz+/v63vZ+VK1fSsWNHPD096dWrFwkJCQAYjUaWLl1Khw4daNq0qXmO91u5du0ac+bMoU2bNnh7exMcHMyXX35pXu/h4cHMmTPx9fUlLCzsTk+FiIiIlEEKnxamT58+7Ny5k9zcXAB+/vlndu/eTdeuXRk4cCCPPvoon332GWvXruXQoUNERkbe8T5iY2NZuHAhc+bM4fjx4zz33HMMHz6cK1eusGDBAtasWcP8+fNJTExk8ODBjBgxokiovGHq1KkcOHCAVatWcfDgQTp06MDAgQNJS0szt0lOTmbv3r3MmTPn758UERERKTMUPi2Ml5cX7u7ubN++HYD4+Hjq1q1Leno6+fn5jBs3DgcHB9zd3Rk9ejRr1qy5431s3ryZvn374unpibW1NUFBQSxfvhxHR0c2bdrEsGHDaNiwIba2tnTu3Bl/f382btxYpI/r168THx/P+PHjqV27Nvb29gwYMIC6desSHx9vbte1a1fKlSuHk5PT3Z0YERERKRP0tLsFCgoKYsuWLQQFBbF582aCgoJITU3l8uXLeHt7m9uZTCYKCgq4dOnSH/a1ePFilixZYv4cFRVFZmYm1atXL9KuSZMmAGRlZVGzZs0i6x555BGSkpKKLLt69SoFBQU88sgjN7VNSUkxf65ateptHrWIiIg8CBQ+LVDPnj157733OHToEN9++y1du3bls88+o1atWnz88cfmdrm5uVy6dInKlSv/YV9hYWE3fd/S3d2dCxcuFFk2b948unfvTo0aNTh//nyRdefPn78pRLq6uuLg4MD58+epV6+eeXlycnKR755aWVnd/oGLiIhImafb7haocuXKtGvXjsmTJ9OpUycefvhh2rVrR15eHsuWLSM/P5+ff/6ZiRMnMnbs2DsOeL169WL9+vV8+eWXGI1GNm3axJo1a3B2diYoKIilS5dy6tQpDAYD27dvZ/fu3fTs2bNIH9bW1vTu3ZuIiAh++ukn8vPzWblyJd9//735lU8iIiIi/0tXPi1Unz592LFjB7NmzQKgQoUKfPDBB8yePZtly5ZhNBrx9fVl0aJFd9x3t27d+Pnnn3n11VfJzMykfv36REVFUblyZQYNGoTRaGTs2LFkZmZSu3ZtIiIi8PHxuamfCRMmEBkZycCBA7ly5QoeHh5ER0fzj3/8466PX0RERMomze0uFkvTa4qISGl3Y3rN7Ow8CguNJV3OfXW7c7srfIqIiIjcRwaDkStXfinzc7vfbvjUbXexWNnZeSVdggDOzg9pLCyIxsNyaCwsiyWPh9FoKvPB804ofIrFMhqNGMv2HQqLd+NZNoPBiO6RlDyNh+XQWFgWjUfpoqfdxWLpNU0iIiJlj8KnWCyFTxERkbJH4VNEREREio3Cp4iIiIgUG4XPUiQlJQUPD48ic6cXl8WLFzN06NBi36+IiIiULXraXW7L/84PLyIiIvJ3KHyWQlu3biUuLo4LFy7QuHFj3n77bapWrUpUVBRbt27lwoULWFlZ0bp1a958800cHR357rvvmDp1KmfOnKFChQr4+Pjw+uuvU6FChZv6T0lJoX379gwaNIhNmzbRtWtXKleuzJEjR1i9ejWxsbHExMTQsGFD4uPjsbKywt/fn6lTp2JnZ4fBYCAyMpKYmBhMJhMdO3bk22+/pU+fPvTq1asEzpiIiIhYCt12L4VOnTrFhg0b2LdvH1evXmXBggVs376dVatWERkZybFjx1i3bh0HDhxg69atAEybNo3mzZtz5MgRNm3axOnTp4mJifnT/eTl5XHw4EHGjh1707rPP/8cFxcX9u/fz5IlS/joo4/45JNPAIiOjiYuLo6VK1eyd+9enJycOHHixL0/ESIiIlLq6MpnKRQWFkbFir9NX9WqVSu+/PJLJkyYQJMmTXBzc+Py5ctkZ2dTqVIl0tPTAXBwcGD//v3Uq1eP5s2bs2XLFqyt//xvjx49emBvb4+9vf1N6xwdHQkLC8PKyopGjRrh4eHB2bNnAdi4cSPDhg2jfv36AIwZM4bNmzffy1MgIiIipZSufJZClSpVMv984za3yWRi3rx5+Pj40L9/f9asWUNBQQGm/0718N577/HUU08xb948mjdvTnBwMN999x1paWl4enqa/02ZMsXcd9WqVf+wBhcXlyLv4bSzszPv68KFC9SoUcO8zsbGhurVq9+rwxcREZFSTFc+y4i5c+eSlpbG7t27zd/j7NatG/DbNJWnT59m1KhRTJo0iQsXLvDWW28RHh7Opk2bbrolfuNp+r/7kvfq1auTlpZm/mwymbhw4cLf6ktERETKFl35LCNyc3NxcHDAxsaG69evs3z5cs6cOUNBQQHW1tbMnDmT9957j+vXr1O5cmUcHBxwdna+L7X07duX5cuXc/bsWfLz81mwYAEZGRn3ZV8iIiJSuih8lhFjxozh2rVrtGjRAn9/f06ePMmzzz7LmTNngN9uu//www/4+fnRokULcnJymDFjxn2pZcCAAfj7+9OvXz/atm3LlStXcHNzw87O7r7sT0REREoPK9ONL+qJ3CNffPEFNWrUwNXVFfjttnuzZs2IiIigZcuWt91PdnYehYXG+1Wm3AYrK3B1rUhWVg76TVHyNB6WQ2NhWTQeluHGOPwVXfmUe27r1q1MmDCBnJwcCgsLWbFiBQCNGze+o370d5GIiEjZo/Ap99yYMWNwdXWlY8eO+Pj4sGfPHqKjo3nooYfuqB+FTxERkbJHT7vLPVehQgXmzJlz1/1YW1vzF68ivW1GowmjUWFWRESkpCl8isVydr6zK6V/xmAwcuXKLwqgIiIiJUzhUyzWjA2HSEq7fNf9/KPqw8zs3wprayuFTxERkRKm8CkW66esn0lKvfvwKSIiIpZDDxyJWU5ODpcvK+yJiIjI/aPwWYosXryYoUOH3rf+O3bsyHfffXff+hcRERHRbfdSJCws7L72n52dfV/7FxEREVH4tGApKSm0b9+eQYMGsWnTJrp27cr333/P6tWriY2NJSYmhoYNGxIfH4+VlRX+/v5MnToVOzs7DAYDkZGRxMTEYDKZ6NixI99++y19+vShV69eN+0rICAAgJCQEEaNGkVsbCzPPvtskcDbrVs3BgwYgLW1NevWrePxxx8nPj6e8uXL079/f4YPH46VlRX5+fksWrSIuLg4cnJyeOqpp5g8eTK1a9cutnMnIiIilkm33UuBvLw8Dh48iK1t0b8VPv/8c1xcXNi/fz9Llizho48+4pNPPgEgOjqauLg4Vq5cyd69e3FycuLEiRN/uI8dO3YAEBUVRUhICL169WLLli3m9V9//TUpKSk888wzwG9TaJYrV46EhAQWLVrEypUr2bhxIwDz5s1j7969fPDBB+zfv5+nnnqKwYMHc/369Xt6XkRERKT0UfgsBXr06IG9vT1OTk5Fljs6OhIWFoadnR2NGjXCw8ODs2fPArBx40aGDRtG/fr1sbe3Z8yYMVSpUuWO9pmcnMxXX30FwIcffsjTTz9tnqWoUqVKvPLKKzg4OPDPf/6Tvn37EhcXh8lkYt26dYwbN46aNWvi4ODASy+9REFBAXv37r03J0RERERKLYXPUqBq1aq3XO7i4oKVlZX5s52dnXlKygsXLlCjRg3zOhsbG6pXrw7AsWPH8PT0NP9bvHjxTX1XqVKFVq1asWXLFgoKCoiPjy9yu75GjRrY2dmZP7u7u5ORkcHly5f55ZdfGD16NF5eXnh5eeHt7c3Vq1dJTU29uxMhIiIipZ6+81kK/D5g3q7q1auTlpZm/mwymbhw4QIAXl5ef3oL/obevXszbdo0WrZsScWKFfH29javy8jIwGQymWtLSUmhevXqODs74+DgwPLly2ncuLG5/Y8//ki1atXu+DhERESkbNGVzzKqb9++LF++nLNnz5Kfn8+CBQvIyMj4023s7e3Jyckxf27bti0Gg4F//etfNz2klJmZydKlSykoKODLL78kJiaGoKAgrK2tCQwM5N133+XixYsYjUY2b95M165d+emnn+7LsYqIiEjpoSufZdSAAQPIzMykX79+2NjY0LlzZ9zc3IrcKv9fffv2Zfz48QwcOJCxY8diZ2dH9+7dWbVqFYsWLSrStkqVKqSkpODn58dDDz3E6NGj6dy5MwATJ04kMjKS/v37c+XKFWrWrMm//vUvnnjiift6zCIiImL5rEw3viQoZcoXX3xBjRo1cHV1BX677d6sWTMiIiJo2bLlbfezatUqPvvsM5YtW2ZeFhsby/vvv8/u3bvved2/d6/nds/OzqOw0HgPKntwWFmBq2tFsrJy0G+KkqfxsBwaC8ui8bAMN8bhr+jKZxm1detWfvzxR+bPn0+5cuVYtWoVQJHvYf6ZzMxMLly4wMqVK5k0adJ9rPSPvd6nxT3ry2AwYjTqN5KIiEhJU/gso8aMGcP06dPp2LEj+fn5NGzYkOjoaPOrkv7K3r17mTlzJs8++yzt27e/z9XeWnZ23j3ry2g0KXyKiIhYAN12F4t1L8On/H3Ozg9pLCyIxsNyaCwsy4MwHpZ+IeV2b7srfIqIiIiUAgaDkStXfrHYAKrvfEqpd68eOBIRESntbjw8a21tZbHh83YpfJZR586do06dOiVdxl35KetnklIVPkVERMoSvWS+lJgyZQpTpky5rbZr1qzh9ddfvy91hIeHEx4efl/6FhERkbJPVz5LienTp99228uXdbVQRERELFOZD5+RkZFs3LiRX3/9lZo1azJixAgqVKjAq6++Su/evVm7di0A3bt359VXX8Xe3h6j0ciyZcvYsGED2dnZ/OMf/2D06NG0atUKAH9/f/z8/Ni1axdVqlQhNjaWb775htmzZ5OUlISzszP9+/dnwIABfzgvu7+/P7169WLr1q1kZGTwxBNPMG3aNOrXr09iYiITJkzAy8uLffv2MWzYMH788UcAZs+eTWRkJN999x329vbs3buX8uXL8+yzzzJ+/Hg2b97MkiVLMBgMeHl5cezYsVvu/+DBg8ybN48ffvgBZ2dnBg8ezAsvvADAzp07WbhwIefOnaNKlSo899xzvPjii1hb33yhPCYmhg8++IALFy5Qo0YNQkJC6N69OwDBwcHUqFGDxMRETCYT8fHxVKhQ4e4GVEREREq1Mn3b/fDhw6xfv56YmBgSExMJCgritddeo7CwkPT0dM6ePcuuXbtYv349e/fuZeHChQAsWLCANWvWMH/+fBITExk8eDAjRozgyy+/NPf95Zdfsn37dlatWkVmZiYDBgzg6aef5tChQyxcuJC1a9eyfv36P61v/fr1vPfeeyQkJFCvXj3CwsIoKCgA4OLFi9StW5eEhAT69+9/07affPIJfn5+JCYmMmPGDKKiojh58iQ9e/YkNDT0T4Pn2bNnCQsLo1+/fhw9epR//etfREREsH//fg4fPsyYMWMYOnQoR44cISIighUrVphfUv97sbGxzJ49m8mTJ3P06FEmTZrEtGnT+PTTT81tDh06xLp164iLi1PwFBERkbIdPh0cHLh69SobNmzg9OnTBAUFkZCQgK2tLVZWVrzxxhtUqFCBOnXqMHToUOLi4gDYtGkTw4YNo2HDhtja2tK5c2f8/f3ZuHGjue+AgACcnJxwcnIiLi6OevXq8fzzz2NnZ0f9+vUZMmQIa9as+dP6hgwZQoMGDXB0dOT//u//uHDhAp9//rl5fWBgIHZ2drcMbXXq1KFHjx7Y2NjQpk0bqlSpwrlz527rvGzbto2GDRsSGBiIra0tTz75JGvXrqVhw4bExsbSvn17OnfujK2tLQ0bNmTYsGGsW7fupn42bdpE3759ad68OTY2NjRv3py+ffsWadu6dWuqVauGk5PTbdUmIiIiZVuZvu3u6elJZGQkq1evZtmyZTg6OhIcHEyTJk14+OGHcXZ2Nrd1d3cnIyMDgKysLGrWrFmkr0ceeYSkpCTz56pVq5p/Tk1N5dSpU3h5eZmXGY1GbGxsAOjSpQtpaWkAVK9enW3btgFQu3Ztc/ty5cpRqVIlMjMzqVKlyk37+F832txgZ2eH0XjzvOXHjh0jJCTE/Dk0NJSMjAyqV69epN3jjz8OwKVLl2jQoMFNx56amnpT3390nn4/5/ufHYOIiIg8eMp0+ExLS8PFxYXo6Gjy8/NJSEhg5MiRREZGkpOTw6+//kq5cuUASElJMQeyGjVqcP78+SJ9nT9/vkiQ+v13Od3c3PD19SU6Otq8LDs7m7y832ZauBE2/1d6err557y8PLKzs3F3d6ewsPCmffxdXl5enDhxosiyRYsWsW/fviLLNm3ahIuLCzVq1CA5ObnIuvPnz98UduG3oPlXbe/FMYiIiEjZUaZvu3/11VcMHTqUpKQk7O3tcXFxAeDMmTMYDAbefvttrl+/zo8//kh0dDSBgYEABAUFsXTpUk6dOoXBYGD79u3s3r2bnj173nI/3bp14+TJk8TFxVFYWEhGRgZhYWHMnj37T+tbsWIFP/30E7/++itvvfUWdevWxdPT866P28HBgdzcXP5o8qouXbpw+vRpPvzwQwwGA19//TWzZ8/G1taW3r17s3v3brZv347BYOD06dNERUXRu3fvm/oJDAxk/fr1JCQkYDAYzN+xvVVbERERESjjVz4DAgI4d+4cw4cPJzs7GxcXFyZNmkTdunUBePjhh2nfvj0A/fr1Y+jQoQAMGjQIo9HI2LFjyczMpHbt2kRERODj43PL/dSoUYNly5Yxd+5cZs6ciY2NDW3btuW111770/qaNm3KSy+9RFpaGt7e3ixduvSWT5TfqXbt2vGf//yHpk2bsnfv3pu+b1mrVi2WLl3Ku+++y4wZM3BxcSE8PBw/Pz8A5s+fz4IFC5g0aRLOzs4899xzRW7d3/DMM8+Qm5vLzJkzSUtLo1q1akyYMIEePXrc9TGIiIhI2fRAzu2emJjIiy++yLfffltiNfj7+zNy5Eh69epVYjVYOk2vKSIi8psb02tmZ+dRWHjzMx6WQHO7S6n3ep8WJV2CiIiIxTAYjKV+XndQ+BQLlp2dV9IlCODs/JDGwoJoPCyHxsKyPAjjYTSaFD5LK19f3xK95Q4UeR2R3JrRaOQWb4+SYnTjZQUGg5EH7ws6lkfjYTk0FpZF41G6lOmn3UVERETEsih8isXSO0JFRETKHoVPsVgKnyIiImWPwqf8Lbc7jzzA9evXuXjx4v0rRkREREoNhc8yJCUlBQ8PD1JSUu542/DwcMLDw/9wfZcuXYiLiwNgzZo1vP7667fdd//+/Tl06NAd1yQiIiJlzwP5tLvcud/PT3/58p29+D07O/telyMiIiKllK58lkFbt27lmWeeoXHjxgwcOJD09HRMJhNLly6lW7dueHl54e3tzfjx47l27Zp5u8uXLzN8+HC8vb3p0aMHn332mXmdv78/sbGxbN68mSVLlnDs2DG8vLwA+OGHHwgNDaVt27Y0atSIzp07s2fPHgAGDx5MWloab7zxBtOnTy/eEyEiIiIWR+GzDDp16hQbNmxg3759XL16lQULFrB9+3ZWrVpFZGQkx44dY926dRw4cICtW7eatztw4AA9e/YkISGBgQMHMmLECJKTk4v03bNnT0JDQ/Hy8uLYsWMAjBo1iscee4xPP/2UY8eO4efnx9SpUwFYvnw51atXZ9q0aUyZMqXYzoGIiIhYJoXPMigsLIyKFSvy8MMP06pVK5KTk2ndujUbN26kTp06XL58mezsbCpVqkR6erp5u3bt2tGpUydsbW3p0aMHTz75JB999NFf7m/JkiWMGjUKk8lEamoqTk5ORfoVERERuUHf+SyDKlWqZP7Zzs4Og8GAyWRi3rx57Nmzh8qVK9OgQQMKCgow/W4qiEceeaRIP+7u7rcVIpOSkhgxYgSZmZnUq1ePypUrF+lXRERE5AaFzwfE3LlzSUtLY/fu3VSoUAGAbt26FWmTkZFR5PP58+dp2LDhn/abnp7O6NGjef/99/H39wdgx44dfPLJJ/ewehERESkrdNv9AZGbm4uDgwM2NjZcv36d5cuXc+bMGQoKCsxtdu3axb59+ygoKGDDhg388MMPNwVUAAcHB3JzczGZTOTl5WEwGChXrhwA33//PQsWLAAgPz8fAHt7e3JycorhKEVERMTSKXw+IMaMGcO1a9do0aIF/v7+nDx5kmeffZYzZ86Y27Rv356oqCh8fHzYsGED0dHRVKtW7aa+2rVrx5UrV2jatCmurq5MmDCBV199laZNmzJ69Gh69+6NnZ2due/AwEDmzZvHK6+8UmzHKyIiIpbJyqQv54mFys7Oo7DQWNJlPNCsrMDVtSJZWTnoN0XJ03hYDo2FZdF4WIYb4/BXdOVTLJb+LhIRESl7FD7FYil8ioiIlD0KnyIiIiJSbBQ+RURERKTYKHyKxbKysirpEkREROQeU/gUi6XwKSIiUvYofIqIiIhIsVH4FBEREZFio/ApIiIiIsVG4bMMiIyMpE2bNvj4+NC7d2927doFwKlTpwgODsbb25tOnTrxwQcfmN+dmZ+fz9tvv80zzzyDp6cnzZs3Z8aMGeb1R48epVevXnh5edGxY0fefPNNCgsLAcjOzub111/Hz88PX19fQkNDOXfuHAApKSl4eHgQExODv78/TZs2ZdCgQVy8eLH4T4yIiIhYHIXPUu7w4cOsX7+emJgYEhMTCQoK4rXXXiM1NZUBAwbw9NNPc+jQIRYuXMjatWtZv349ACtXrmT//v2sXLmSEydOsHDhQtatW8fhw4cBmDBhAsHBwRw7dowVK1bw8ccfm0Ptyy+/THJyMps3b2bfvn3UrVuXgQMHkpuba65r7969fPjhh+zYsYOsrCwWLlxY/CdHRERELI7CZynn4ODA1atX2bBhA6dPnyYoKIiEhAQ++ugj6tWrx/PPP4+dnR3169dnyJAhrFmzBoA+ffrwwQcfUKVKFTIyMrh27RoPPfQQ6enp5n63b9/Onj17qFSpEvv27SMgIIDz589z5MgRXn/9dapUqYKjoyOvvPIKhYWF7Nu3z1xXSEgITk5OuLq64u/vb74yKiIiIg8225IuQO6Op6cnkZGRrF69mmXLluHo6EhwcDCZmZmcOnUKLy8vc1uj0YiNjQ0Av/76K9OnT+fo0aO4ubnxxBNPYDKZMBqNwG9XRiMjI5k2bRqZmZm0atWKqVOnkpWVBUDNmjXN/drY2ODu7k5qaipPPfUUAK6urub1tra2mipTREREAIXPUi8tLQ0XFxeio6PJz88nISGBkSNHMmLECHx9fYmOjja3zc7OJi8vD4DJkyfz8MMPc+DAARwcHDAajXh7ewNw/fp1vv/+e6ZOnYqtrS1nz55l8uTJzJo1i8mTJwOQnJzMo48+CoDBYCAtLY0qVaoU89GLiIhIaaPb7qXcV199xdChQ0lKSsLe3h4XFxcAmjZtysmTJ4mLi6OwsJCMjAzCwsKYPXs2ALm5uTg4OGBtbU1ubi5z5swhNzeXgoICrKysGDduHMuXL6ewsJAqVapga2uLs7MzVatWpU2bNsycOZPMzEyuXbvG3LlzMRgMtGvXriRPhYiIiJQCCp+lXEBAAIMHD2b48OE0btyY0aNHM2nSJHx8fFi2bBnr16+nRYsWPPvss9StW9ccPidPnkxSUhI+Pj48/fTT5Obm0qpVK86cOYO9vT2LFi1i165d+Pr64u/vT5UqVXjllVcAmDNnDjVr1qRnz560aNGCb7/9lpUrV1KpUqUSPBMiIiJSGliZ9GU8sVDZ2XkUFhpLuowHmpUVuLpWJCsrB/2mKHkaD8uhsbAsGg/LcGMc/oqufIrF0t9FIiIiZY8eOBILZsLKqqRreLDdOP8aB8ug8bAcGgvLovGwDLd7/nXbXURERESKjW67i4iIiEixUfgUERERkWKj8CkiIiIixUbhU0RERESKjcKniIiIiBQbhU8RERERKTYKnyIiIiJSbBQ+RURERKTYKHyKiIiISLFR+BSLcenSJUaMGIGXlxe+vr68+eabFBYWlnRZD4ykpCQGDRqEj48PLVu2ZMKECVy+fBmAL774gqCgIDw9PfH39ycmJqaEq30wGAwGgoODCQ8PNy/TWBS/K1euMGHCBHx9ffH29mbEiBFkZGQAGo+ScOrUKZ5//nm8vLzw8/Nj5syZ5OfnAxqPUsMkYiFeeOEF0/jx402//PKLKTk52dSlSxdTVFRUSZf1QPj1119NLVu2NM2fP990/fp10+XLl00hISGm0NBQ05UrV0w+Pj6mf//736aCggLToUOHTJ6enqYvvviipMsu89577z3T448/bpo4caLJZDJpLErICy+8YHrppZdMV69eNeXk5JhGjhxpGjZsmMajBBgMBlPLli1NK1euNBkMBtOFCxdMAQEBpvfff1/jUYroyqdYhJ9++okjR47w6quvUq5cOWrWrMmIESNYs2ZNSZf2QEhLS+Pxxx/npZdewt7eHmdnZ/r27cvRo0f55JNPqFSpEs8//zy2trY0b96cbt26aWzus4SEBD755BM6depkXqaxKH5ff/01X3zxBbNnz8bJyYkKFSowY8YMXnnlFY1HCbh69SqZmZkYjUZMJhMA1tbWlCtXTuNRiih8ikX47rvvqFSpEtWqVTMvq1evHmlpafz8888lWNmDoW7duixbtgwbGxvzsh07dtCwYUO+++47HnvssSLt69evT1JSUnGX+cC4dOkSr732Gu+++y7lypUzL9dYFL8vv/yS+vXrs2HDBjp27Iifnx9vv/02VapU0XiUAGdnZwYOHMjbb7/NP//5T9q0aUOdOnUYOHCgxqMUUfgUi5CXl1fkf7KA+fMvv/xSEiU9sEwmE/PmzWPPnj289tprtxwbR0dHjct9YjQaefXVVxk0aBCPP/54kXUai+J39epVvv32W86dO8fmzZv58MMPSU9PZ+LEiRqPEmA0GnF0dOT111/n5MmTxMfH88MPP/Cvf/1L41GKKHyKRShfvjy//vprkWU3Pj/00EMlUdIDKTc3l5dffpmtW7fy73//Gw8PD8qVK8e1a9eKtLt27ZrG5T5ZsmQJ9vb2BAcH37ROY1H87O3tAXjttdeoUKECrq6ujBkzhn379mEymTQexezTTz9lx44d9O/fH3t7ex599FFeeukl/vOf/+i/j1JE4VMswqOPPsqVK1fIysoyL/vhhx9wc3OjYsWKJVjZgyM5OZnevXuTm5vLxo0b8fDwAOCxxx7ju+++K9L2+++/59FHHy2JMsu8LVu2cOTIEby8vPDy8iI+Pp74+Hi8vLw0FiWgfv36GI1GCgoKzMuMRiMADRo00HgUswsXLpifbL/B1tYWOzs7/fdRiih8ikWoU6cOTZs2ZdasWeTm5nL+/HkWLlxIYGBgSZf2QLh69SoDBgygSZMmREdHU7lyZfO6jh07kpWVxQcffEBBQQGHDx9m69at9O7duwQrLrs+/vhjPv/8c44dO8axY8fo2rUrXbt25dixYxqLEtCiRQtq1qzJpEmTyMvL4/Lly8ybN48OHTrQtWtXjUcx8/PzIzMzk8WLF2MwGDh//jyLFi2iW7du+u+jFLEy3XhcTKSEZWVlMX36dBITE7G2tqZHjx688sorRR6CkftjxYoVzJ49m3LlymFlZVVk3YkTJ/jqq6948803OXPmDJUrV2bEiBH06tWrhKp9sNx4x+fs2bMBNBYlID09ndmzZ3P06FGuX7+Ov78/r732Gk5OThqPEnDo0CHee+89fvzxRypWrEj37t3Nb+rQeJQOCp8iIiIiUmx0211EREREio3Cp4iIiIgUG4VPERERESk2Cp//r737CYWtj+M4/pHnaDYK00RGNtRYyEoaG8okZVaGiCIW/mRhIQ0RKRQ1C7OSkilFWbCwsFAUs7BQU8rGgqmpsTBEbBSaZ/HUqefe7r25985P6v1a/n6n0++7e3fOmQYAAADGEJ8AAAAwhvgEAACAMcQnAAAAjCE+AQAAYAzxCQD4EI/Ho4GBAX37HyW7u7tqaGj4pFMB+CqITwDAhx0fH2ttbe2zjwHgCyI+AQAf1t3drXA4rFgs9sNrLi8v1d/fr5qaGtXV1Wl2dlbPz8+S/ntK2tnZqfn5eXm9XtXW1mpqakqvr6+SpHQ6rY2NDTU1Nam6ulpdXV26uLgwMhuAzCI+AQAf1tjYqI6ODo2Ojurx8fG7/YeHB/X09Ki8vFwnJyfa2dlRPB5XMBi0r4nFYnI6nYpGo1pdXdX+/r4ODg4kSVtbW4pEIgqHwzo9PVUgEFBfX5/u7u5MjQggQ4hPAMBvGR8fV0FBgSYmJr77/vPw8FCWZWlsbEwOh0Mul0vT09M6OjpSKpWSJDkcDg0NDcmyLFVVVcnj8Sgej0uSNjc3NTg4qIqKClmWpba2NpWVlWlvb8/4nAD+rn8++wAAgK8pJydHy8vLamlp0fr6uvLz8+29+/t7FRcXKzs7214rKSmRJCWTSUmS0+lUVlaWvW9Zlh2xyWRSS0tLCoVC9v7b25sqKyszOhOAzCM+AQC/rbS0VHNzcwoGgwoEAva62+3Wzc2N3t/f7QBNJBKSJJfLpevr65/et6ioSCMjI/L7/fZaIpFQXl7e3x8CgFG8dgcA/JHm5ma1trZqe3vbXquvr5ckhUIhvby8KJVKaWFhQV6vV263+5f3bG9v18rKiq6uriRJ0WhUfr9fZ2dnmRkCgDE8+QQA/LHJyUmdn5/r6elJkpSbm6tIJKLFxUU7RH0+3/9+cPQzvb29SqfTGh4e1u3trQoLCzUzMyOfz5exGQCYkZX+9itxAAAAIEN47Q4AAABjiE8AAAAYQ3wCAADAGOITAAAAxhCfAAAAMIb4BAAAgDHEJwAAAIwhPgEAAGAM8QkAAABjiE8AAAAYQ3wCAADAmH8BqqgzOR9bGuMAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.barplot(y=missing_data_per.index,\n",
    "            x=missing_data_per)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "078232f9-c28f-4487-8e79-e4ffc8d68b7d",
   "metadata": {},
   "source": [
    "<a name=\"data-preprocessing\"></a>\n",
    "## 3. Data Preprocessing and Feature Engineering"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e3267465-dedb-4e35-9e7d-3e927223f68e",
   "metadata": {},
   "source": [
    "<a name=\"unused-cat\"></a>\n",
    "## 3.1. Update Categorical Values"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8544f858-bc38-4fa5-abbd-cb4de954076d",
   "metadata": {},
   "source": [
    "For kaggle problems, model is evaluated on test data. It would make more sense to update columns in train data to contain unique values from test data and set all other value to NaN."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "f07e83d7-64fe-4dc1-b1e7-8efc2f78c59a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Update train data to contain values from test data\n",
    "# for col in catCols:\n",
    "#     if col != \"class\":\n",
    "#         vals = test_data[col].unique()\n",
    "#         train_data.loc[~train_data[col].isin(vals), col] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "2d6c075d-b3fe-4f72-b728-86efcfdbec07",
   "metadata": {},
   "outputs": [],
   "source": [
    "replacements = {\n",
    "    '5 f': 'f', '7 x': 'x', 'is s': 's', 'is p': 'p', \n",
    "    '3 x': 'x', 'b f': 'b', 'does n': 'n', 'does w': 'w',\n",
    "    'is n': 'n', 'e n': 'e', 'is w': 'w', 'f has-ring': 'f',\n",
    "    'does f': 'f', 'is h': 'h', 'has f': 'f', 'e y': 'e', 'is y': 'y'\n",
    "}\n",
    "\n",
    "# Function to clean and correct errors in a column using vectorized replace\n",
    "def clean_column(column):\n",
    "    return column.replace(replacements, regex=True)\n",
    "\n",
    "# Apply the cleaning function to all object columns in the DataFrame\n",
    "for col in catCols:\n",
    "    train_data[col] = clean_column(train_data[col])\n",
    "\n",
    "for col in catCols:\n",
    "    test_data[col] = clean_column(test_data[col])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4eb21b7b-d466-4416-b3d0-37714156615d",
   "metadata": {},
   "source": [
    "<a name=\"missing-data\"></a>\n",
    "## 3.2. Missing Data Handling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "a579b0ce-b574-48c3-b226-0aea442c6e74",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.impute import SimpleImputer"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0984324b-6634-432e-a46f-0ba05097e99d",
   "metadata": {},
   "source": [
    "<a name=\"target-encoding\"></a>\n",
    "## 3.3. Target Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "827a1239-483d-4a6d-acc6-4086693c4739",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import TargetEncoder"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a3992a2-53f5-4907-b611-3cef88d61658",
   "metadata": {},
   "source": [
    "<a name=\"pipeline\"></a>\n",
    "## 3.4. Pre-processing pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "9cc41539-984b-4dc2-8b0a-72764c77814a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.compose import ColumnTransformer\n",
    "from sklearn.pipeline import Pipeline\n",
    "\n",
    "cat_process_pipe = Pipeline([\n",
    "    (\"categorical_imputer\", SimpleImputer(strategy=\"constant\", fill_value=\"None\")),\n",
    "    (\"target_encoder\", TargetEncoder(target_type=\"binary\", random_state=22)),\n",
    "])\n",
    "preprocessor = ColumnTransformer(\n",
    "    transformers=[\n",
    "        (\"numerical_imputer\", SimpleImputer(strategy=\"mean\"), numCols),\n",
    "        (\"categorical_processing\", cat_process_pipe, catCols)\n",
    "    ]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d049781-af73-4c16-8cce-62bed338a9a1",
   "metadata": {},
   "source": [
    "<a name=\"modeling\"></a>\n",
    "# 4. Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5625c0af-6633-4241-8688-a4db2b8a9046",
   "metadata": {},
   "source": [
    "<a name=\"data-split\"></a>\n",
    "## 4.1. Train and Test data split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "eb758fe2-3080-4c45-b266-6cc8ac0093a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split the data into training and test sets\n",
    "from sklearn.model_selection import StratifiedShuffleSplit\n",
    "\n",
    "split = StratifiedShuffleSplit(n_splits=1, test_size=0.2, random_state=22)\n",
    "\n",
    "for train_indices, test_indices in split.split(train_data, train_data[\"class\"]):\n",
    "    train = train_data.loc[train_indices]\n",
    "    test = train_data.loc[test_indices]\n",
    "\n",
    "y_train = train[\"class\"]\n",
    "y_test = test[\"class\"]\n",
    "X_train = train.drop(columns=\"class\")\n",
    "X_test = test.drop(columns=\"class\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7265de3-9095-4381-afa0-7a03bdb91c79",
   "metadata": {},
   "source": [
    "<a name=\"random-forest\"></a>\n",
    "## 4.2. Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "9ef73c48-7a1f-4ff1-b2ec-deed284741cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a pipeline\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rf_pipeline = Pipeline([(\"preprocessor\", preprocessor),\n",
    "                        (\"randomforest\", RandomForestClassifier(warm_start=True, n_jobs=-1, random_state=22, verbose=2))])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "3a965983-679b-48e0-b4ad-128746d6671b",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_param_grid = [{\"randomforest__n_estimators\": [600],\n",
    "                  \"randomforest__max_depth\": [60]}]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee0515a9-e64e-4a1e-94ed-f0f1746fbe9a",
   "metadata": {},
   "source": [
    "#### Best public score in kaggle with Random Forest is 0.98437. Let's try another ensembling methods for classification and analyze their scores."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "117266b8-5d9f-41ff-821b-b8f572a57613",
   "metadata": {},
   "source": [
    "<a name=\"adaboost\"></a>\n",
    "## 4.3. AdaBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "1a59a0cc-d231-43cc-a04c-ca52e6b7f613",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a pipeline\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "\n",
    "# Define the base model\n",
    "base_estimator = DecisionTreeClassifier()\n",
    "\n",
    "ab_pipeline = Pipeline([(\"preprocessor\", preprocessor),\n",
    "                        (\"adaboost\", AdaBoostClassifier(estimator=base_estimator, random_state=22))])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "34d06f28-ed5d-4400-bb17-1496052e9ee9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ab_param_grid = [{\"adaboost__n_estimators\": [300, 400, 500],\n",
    "#                   \"adaboost__learning_rate\": [0.01, 0.1, 1, 10],\n",
    "#                   \"adaboost__estimator__max_depth\": [1, 2, 3, 4]}]\n",
    "ab_param_grid = [{\"adaboost__n_estimators\": [300]}]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d5f1bdd7-5cae-4054-8ba6-7b76fe92eaa9",
   "metadata": {},
   "source": [
    "### Select the pipeline and parameters for selected model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "497380cd-79dd-4250-8590-febb01618555",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Random Forest\n",
    "# pipeline = rf_pipeline\n",
    "# param_grid = rf_param_grid\n",
    "\n",
    "# AdaBoost\n",
    "pipeline = ab_pipeline\n",
    "param_grid = ab_param_grid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "ccb7a26e-3a57-41be-8d55-0169d241f7d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "grid_search = GridSearchCV(pipeline, param_grid, cv=3, scoring=\"accuracy\", return_train_score=True, n_jobs=-1, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "1f109937-cb63-4597-819a-233284eaec18",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_and_test_model(search_model, X_train, Y_train, X_test):\n",
    "\n",
    "    grid_search.fit(X_train, Y_train)\n",
    "    \n",
    "    # Create a variable for the best model\n",
    "    best_model = grid_search.best_estimator_\n",
    "    print(best_model)\n",
    "    \n",
    "    y_pred = best_model.predict(X_test)\n",
    "\n",
    "    return best_model, y_pred"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f36dd67-e89a-471d-ab5d-6d36e935adee",
   "metadata": {},
   "source": [
    "### Uncomment and run following code to identify best model hyperparemeters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "5ab48b5d-4023-4541-96ec-2c2536e109cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# from sklearn.metrics import matthews_corrcoef\n",
    "\n",
    "# best_model, y_pred = train_and_test_model(grid_search, X_train, y_train, X_test)\n",
    "\n",
    "# accuracy = matthews_corrcoef(y_test, y_pred)\n",
    "# print(accuracy)\n",
    "# best_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "194a0216-7ff2-4eed-a477-73e1a1a34ed9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 1 candidates, totalling 3 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\cmodi\\anaconda3\\envs\\dev\\Lib\\site-packages\\sklearn\\ensemble\\_weight_boosting.py:527: FutureWarning: The SAMME.R algorithm (the default) is deprecated and will be removed in 1.6. Use the SAMME algorithm to circumvent this warning.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pipeline(steps=[('preprocessor',\n",
      "                 ColumnTransformer(transformers=[('numerical_imputer',\n",
      "                                                  SimpleImputer(),\n",
      "                                                  ['cap-diameter',\n",
      "                                                   'stem-height',\n",
      "                                                   'stem-width']),\n",
      "                                                 ('categorical_processing',\n",
      "                                                  Pipeline(steps=[('categorical_imputer',\n",
      "                                                                   SimpleImputer(fill_value='None',\n",
      "                                                                                 strategy='constant')),\n",
      "                                                                  ('target_encoder',\n",
      "                                                                   TargetEncoder(random_state=22,\n",
      "                                                                                 target_type='binary'))]),\n",
      "                                                  ['cap-shape', 'cap-surface',\n",
      "                                                   'cap-color',\n",
      "                                                   'does-bruise-or-bleed',\n",
      "                                                   'gill-attachment',\n",
      "                                                   'gill-spacing', 'gill-color',\n",
      "                                                   'stem-root', 'stem-surface',\n",
      "                                                   'stem-color', 'veil-type',\n",
      "                                                   'veil-color', 'has-ring',\n",
      "                                                   'ring-type',\n",
      "                                                   'spore-print-color',\n",
      "                                                   'habitat', 'season'])])),\n",
      "                ('adaboost',\n",
      "                 AdaBoostClassifier(estimator=DecisionTreeClassifier(),\n",
      "                                    n_estimators=300, random_state=22))])\n"
     ]
    }
   ],
   "source": [
    "_, predictions = train_and_test_model(grid_search, train_data.drop(columns=[\"id\", \"class\"]), train_data[\"class\"], test_data.drop(columns=\"id\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "349b3da7-53cd-4182-a0b4-b6dc213a50f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "submission_df = pd.DataFrame(test_data[\"id\"])\n",
    "submission_df[\"class\"] = predictions\n",
    "submission_df.to_csv(\"data/mushroom_classfication_adaboost_20240825.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "51abffd1-d2df-4eb8-9ee4-f893036c1f31",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}