Segmentation Notebook _Final.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Segmentation code for clickstream data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "#from helper_functions import do_pca, scree_plot, plot_components, pca_results\n",
    "from IPython import display\n",
    "#import test_code2 as t\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "#!pip install --trusted-host=pypi.org --trusted-host=files.pythonhosted.org --user helper"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Read Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>year</th>\n",
       "      <th>month</th>\n",
       "      <th>day</th>\n",
       "      <th>order</th>\n",
       "      <th>country</th>\n",
       "      <th>session ID</th>\n",
       "      <th>page 1 (main category)</th>\n",
       "      <th>page 2 (clothing model)</th>\n",
       "      <th>colour</th>\n",
       "      <th>location</th>\n",
       "      <th>model photography</th>\n",
       "      <th>price</th>\n",
       "      <th>price 2</th>\n",
       "      <th>page</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2008</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>29</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>A13</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>28</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2008</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>29</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>A16</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>33</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2008</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>29</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>B4</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>52</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2008</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>29</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>B17</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>2</td>\n",
       "      <td>38</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2008</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>29</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>B8</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>52</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   year  month  day  order  country  session ID  page 1 (main category)  \\\n",
       "0  2008      4    1      1       29           1                       1   \n",
       "1  2008      4    1      2       29           1                       1   \n",
       "2  2008      4    1      3       29           1                       2   \n",
       "3  2008      4    1      4       29           1                       2   \n",
       "4  2008      4    1      5       29           1                       2   \n",
       "\n",
       "  page 2 (clothing model)  colour  location  model photography  price  \\\n",
       "0                     A13       1         5                  1     28   \n",
       "1                     A16       1         6                  1     33   \n",
       "2                      B4      10         2                  1     52   \n",
       "3                     B17       6         6                  2     38   \n",
       "4                      B8       4         3                  2     52   \n",
       "\n",
       "   price 2  page  \n",
       "0        2     1  \n",
       "1        2     1  \n",
       "2        1     1  \n",
       "3        2     1  \n",
       "4        1     1  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read dataset and look at top records\n",
    "\n",
    "df = pd.read_csv('e-shop clothing 2008.csv', delimiter=\";\")\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 165474 entries, 0 to 165473\n",
      "Data columns (total 14 columns):\n",
      " #   Column                   Non-Null Count   Dtype \n",
      "---  ------                   --------------   ----- \n",
      " 0   year                     165474 non-null  int64 \n",
      " 1   month                    165474 non-null  int64 \n",
      " 2   day                      165474 non-null  int64 \n",
      " 3   order                    165474 non-null  int64 \n",
      " 4   country                  165474 non-null  int64 \n",
      " 5   session ID               165474 non-null  int64 \n",
      " 6   page 1 (main category)   165474 non-null  int64 \n",
      " 7   page 2 (clothing model)  165474 non-null  object\n",
      " 8   colour                   165474 non-null  int64 \n",
      " 9   location                 165474 non-null  int64 \n",
      " 10  model photography        165474 non-null  int64 \n",
      " 11  price                    165474 non-null  int64 \n",
      " 12  price 2                  165474 non-null  int64 \n",
      " 13  page                     165474 non-null  int64 \n",
      "dtypes: int64(13), object(1)\n",
      "memory usage: 17.7+ MB\n"
     ]
    }
   ],
   "source": [
    "#Check the number of rows and columns and their types\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 165474 entries, 0 to 165473\n",
      "Data columns (total 14 columns):\n",
      " #   Column                   Non-Null Count   Dtype \n",
      "---  ------                   --------------   ----- \n",
      " 0   year                     165474 non-null  object\n",
      " 1   month                    165474 non-null  object\n",
      " 2   day                      165474 non-null  object\n",
      " 3   order                    165474 non-null  int64 \n",
      " 4   country                  165474 non-null  object\n",
      " 5   session ID               165474 non-null  object\n",
      " 6   page 1 (main category)   165474 non-null  object\n",
      " 7   page 2 (clothing model)  165474 non-null  object\n",
      " 8   colour                   165474 non-null  object\n",
      " 9   location                 165474 non-null  object\n",
      " 10  model photography        165474 non-null  object\n",
      " 11  price                    165474 non-null  int64 \n",
      " 12  price 2                  165474 non-null  object\n",
      " 13  page                     165474 non-null  object\n",
      "dtypes: int64(2), object(12)\n",
      "memory usage: 17.7+ MB\n"
     ]
    }
   ],
   "source": [
    "# Convert categorical variables to string\n",
    "cat_vars = ['year', 'month', 'day', 'country', 'session ID',\n",
    "               'page 1 (main category)', 'page 2 (clothing model)', 'colour',\n",
    "               'location', 'model photography', 'price 2', 'page']\n",
    "df[cat_vars] = df[cat_vars].astype(str)\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "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>order</th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>165474.000000</td>\n",
       "      <td>165474.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>9.817476</td>\n",
       "      <td>43.802507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>13.478411</td>\n",
       "      <td>12.548131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>33.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>6.000000</td>\n",
       "      <td>43.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>12.000000</td>\n",
       "      <td>52.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>195.000000</td>\n",
       "      <td>82.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               order          price\n",
       "count  165474.000000  165474.000000\n",
       "mean        9.817476      43.802507\n",
       "std        13.478411      12.548131\n",
       "min         1.000000      18.000000\n",
       "25%         2.000000      33.000000\n",
       "50%         6.000000      43.000000\n",
       "75%        12.000000      52.000000\n",
       "max       195.000000      82.000000"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check properties of numeric fields\n",
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "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></th>\n",
       "      <th></th>\n",
       "      <th>median_no_of_clicks_per_session</th>\n",
       "      <th>min_no_of_clicks_per_session</th>\n",
       "      <th>max_no_of_clicks_per_session</th>\n",
       "      <th>median_price</th>\n",
       "      <th>min_price</th>\n",
       "      <th>max_price</th>\n",
       "      <th>total_number_of_sessions</th>\n",
       "      <th>most_frequent_day</th>\n",
       "      <th>most_frequent_colour</th>\n",
       "      <th>most_frequent_location</th>\n",
       "      <th>most_frequent_photo_type</th>\n",
       "      <th>most_frequent_price_type</th>\n",
       "      <th>most_frequent_page_number</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>country</th>\n",
       "      <th>page 1 (main category)</th>\n",
       "      <th>page 2 (clothing model)</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">1</th>\n",
       "      <th>3</th>\n",
       "      <th>C30</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <th>P17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>2</td>\n",
       "      <td>6</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">10</th>\n",
       "      <th rowspan=\"3\" valign=\"top\">1</th>\n",
       "      <th>A1</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A11</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>62</td>\n",
       "      <td>62</td>\n",
       "      <td>62</td>\n",
       "      <td>2</td>\n",
       "      <td>22</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A12</th>\n",
       "      <td>3.0</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>5</td>\n",
       "      <td>12</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">9</th>\n",
       "      <th rowspan=\"5\" valign=\"top\">4</th>\n",
       "      <th>P77</th>\n",
       "      <td>18.0</td>\n",
       "      <td>18</td>\n",
       "      <td>18</td>\n",
       "      <td>43</td>\n",
       "      <td>43</td>\n",
       "      <td>43</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>7</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P78</th>\n",
       "      <td>6.5</td>\n",
       "      <td>95</td>\n",
       "      <td>1</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "      <td>3</td>\n",
       "      <td>17</td>\n",
       "      <td>14</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P80</th>\n",
       "      <td>96.0</td>\n",
       "      <td>96</td>\n",
       "      <td>96</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>1</td>\n",
       "      <td>17</td>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P82</th>\n",
       "      <td>17.5</td>\n",
       "      <td>18</td>\n",
       "      <td>17</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "      <td>48</td>\n",
       "      <td>2</td>\n",
       "      <td>20</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P9</th>\n",
       "      <td>3.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>23</td>\n",
       "      <td>23</td>\n",
       "      <td>23</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3141 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                        median_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                    \n",
       "1       3                      C30                                                  2.0   \n",
       "        4                      P17                                                  2.0   \n",
       "10      1                      A1                                                   2.0   \n",
       "                               A11                                                  2.0   \n",
       "                               A12                                                  3.0   \n",
       "...                                                                                 ...   \n",
       "9       4                      P77                                                 18.0   \n",
       "                               P78                                                  6.5   \n",
       "                               P80                                                 96.0   \n",
       "                               P82                                                 17.5   \n",
       "                               P9                                                   3.0   \n",
       "\n",
       "                                                        min_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                 \n",
       "1       3                      C30                                                 2   \n",
       "        4                      P17                                                 3   \n",
       "10      1                      A1                                                  2   \n",
       "                               A11                                                 3   \n",
       "                               A12                                                 5   \n",
       "...                                                                              ...   \n",
       "9       4                      P77                                                18   \n",
       "                               P78                                                95   \n",
       "                               P80                                                96   \n",
       "                               P82                                                18   \n",
       "                               P9                                                  4   \n",
       "\n",
       "                                                        max_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                 \n",
       "1       3                      C30                                                 2   \n",
       "        4                      P17                                                 1   \n",
       "10      1                      A1                                                  2   \n",
       "                               A11                                                 1   \n",
       "                               A12                                                 1   \n",
       "...                                                                              ...   \n",
       "9       4                      P77                                                18   \n",
       "                               P78                                                 1   \n",
       "                               P80                                                96   \n",
       "                               P82                                                17   \n",
       "                               P9                                                  1   \n",
       "\n",
       "                                                        median_price  \\\n",
       "country page 1 (main category) page 2 (clothing model)                 \n",
       "1       3                      C30                                28   \n",
       "        4                      P17                                38   \n",
       "10      1                      A1                                 28   \n",
       "                               A11                                62   \n",
       "                               A12                                38   \n",
       "...                                                              ...   \n",
       "9       4                      P77                                43   \n",
       "                               P78                                48   \n",
       "                               P80                                28   \n",
       "                               P82                                48   \n",
       "                               P9                                 23   \n",
       "\n",
       "                                                        min_price  max_price  \\\n",
       "country page 1 (main category) page 2 (clothing model)                         \n",
       "1       3                      C30                             28         28   \n",
       "        4                      P17                             38         38   \n",
       "10      1                      A1                              28         28   \n",
       "                               A11                             62         62   \n",
       "                               A12                             38         38   \n",
       "...                                                           ...        ...   \n",
       "9       4                      P77                             43         43   \n",
       "                               P78                             48         48   \n",
       "                               P80                             28         28   \n",
       "                               P82                             48         48   \n",
       "                               P9                              23         23   \n",
       "\n",
       "                                                        total_number_of_sessions  \\\n",
       "country page 1 (main category) page 2 (clothing model)                             \n",
       "1       3                      C30                                             1   \n",
       "        4                      P17                                             1   \n",
       "10      1                      A1                                              1   \n",
       "                               A11                                             2   \n",
       "                               A12                                             5   \n",
       "...                                                                          ...   \n",
       "9       4                      P77                                             1   \n",
       "                               P78                                             3   \n",
       "                               P80                                             1   \n",
       "                               P82                                             2   \n",
       "                               P9                                              4   \n",
       "\n",
       "                                                       most_frequent_day  \\\n",
       "country page 1 (main category) page 2 (clothing model)                     \n",
       "1       3                      C30                                    11   \n",
       "        4                      P17                                    11   \n",
       "10      1                      A1                                      8   \n",
       "                               A11                                    22   \n",
       "                               A12                                    12   \n",
       "...                                                                  ...   \n",
       "9       4                      P77                                     6   \n",
       "                               P78                                    17   \n",
       "                               P80                                    17   \n",
       "                               P82                                    20   \n",
       "                               P9                                      1   \n",
       "\n",
       "                                                       most_frequent_colour  \\\n",
       "country page 1 (main category) page 2 (clothing model)                        \n",
       "1       3                      C30                                        2   \n",
       "        4                      P17                                        2   \n",
       "10      1                      A1                                         8   \n",
       "                               A11                                        3   \n",
       "                               A12                                        2   \n",
       "...                                                                     ...   \n",
       "9       4                      P77                                        7   \n",
       "                               P78                                       14   \n",
       "                               P80                                        7   \n",
       "                               P82                                        6   \n",
       "                               P9                                         4   \n",
       "\n",
       "                                                       most_frequent_location  \\\n",
       "country page 1 (main category) page 2 (clothing model)                          \n",
       "1       3                      C30                                          4   \n",
       "        4                      P17                                          6   \n",
       "10      1                      A1                                           1   \n",
       "                               A11                                          4   \n",
       "                               A12                                          4   \n",
       "...                                                                       ...   \n",
       "9       4                      P77                                          2   \n",
       "                               P78                                          2   \n",
       "                               P80                                          3   \n",
       "                               P82                                          4   \n",
       "                               P9                                           3   \n",
       "\n",
       "                                                       most_frequent_photo_type  \\\n",
       "country page 1 (main category) page 2 (clothing model)                            \n",
       "1       3                      C30                                            2   \n",
       "        4                      P17                                            2   \n",
       "10      1                      A1                                             1   \n",
       "                               A11                                            1   \n",
       "                               A12                                            1   \n",
       "...                                                                         ...   \n",
       "9       4                      P77                                            1   \n",
       "                               P78                                            2   \n",
       "                               P80                                            1   \n",
       "                               P82                                            2   \n",
       "                               P9                                             2   \n",
       "\n",
       "                                                       most_frequent_price_type  \\\n",
       "country page 1 (main category) page 2 (clothing model)                            \n",
       "1       3                      C30                                            2   \n",
       "        4                      P17                                            1   \n",
       "10      1                      A1                                             2   \n",
       "                               A11                                            1   \n",
       "                               A12                                            2   \n",
       "...                                                                         ...   \n",
       "9       4                      P77                                            1   \n",
       "                               P78                                            1   \n",
       "                               P80                                            2   \n",
       "                               P82                                            1   \n",
       "                               P9                                             2   \n",
       "\n",
       "                                                       most_frequent_page_number  \n",
       "country page 1 (main category) page 2 (clothing model)                            \n",
       "1       3                      C30                                             2  \n",
       "        4                      P17                                             1  \n",
       "10      1                      A1                                              1  \n",
       "                               A11                                             1  \n",
       "                               A12                                             1  \n",
       "...                                                                          ...  \n",
       "9       4                      P77                                             5  \n",
       "                               P78                                             5  \n",
       "                               P80                                             5  \n",
       "                               P82                                             5  \n",
       "                               P9                                              1  \n",
       "\n",
       "[3141 rows x 13 columns]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Feature Engineering\n",
    "from scipy.stats import mode \n",
    "df2 = df.groupby(['country','page 1 (main category)','page 2 (clothing model)']).agg(\n",
    "                                                                             median_no_of_clicks_per_session=('order', 'median'),\n",
    "                                                                             min_no_of_clicks_per_session=('order', 'max'),\n",
    "                                                                             max_no_of_clicks_per_session=('order', 'min'),\n",
    "                                                                             median_price=('price', 'median'),\n",
    "                                                                             min_price=('price', 'max'),\n",
    "                                                                             max_price=('price', 'min'),\n",
    "                                                                             total_number_of_sessions =('session ID', pd.Series.nunique),\n",
    "                                                                             most_frequent_day=('day', lambda x: mode(x)[0][0]),\n",
    "                                                                             most_frequent_colour=('colour', lambda x: mode(x)[0][0]),\n",
    "                                                                             most_frequent_location=('location', lambda x: mode(x)[0][0]),\n",
    "                                                                             most_frequent_photo_type=('model photography', lambda x: mode(x)[0][0]),\n",
    "                                                                             most_frequent_price_type =('price 2', lambda x: mode(x)[0][0]),\n",
    "                                                                             most_frequent_page_number =('page', lambda x: mode(x)[0][0])\n",
    "                                                                            )\n",
    "\n",
    "df2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# One hot encoding - to convert categorical data to continuous"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "# Sample code : target encoding\n",
    "#import the libraries\n",
    "import pandas as pd\n",
    "import category_encoders as ce\n",
    "\n",
    "df3 = df2.copy()\n",
    "\n",
    "\n",
    "cat_vars_2 = ['Business_Size_Bins', 'CAGR_Bins', 'Owner_Theater__c','Direct_Flag',\n",
    "            'Final_2_Digit_NAICS_Desc']\n",
    "\n",
    "#Create target encoding object\n",
    "encoder=ce.TargetEncoder(cols= cat_vars_2) \n",
    "\n",
    "#Fit and Transform Train Data\n",
    "df4_1 =encoder.fit_transform(df3[cat_vars_2],df3['Target'])\n",
    "\n",
    "df4 = pd.concat([df4_1,df3[Cont_vars]], axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "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></th>\n",
       "      <th></th>\n",
       "      <th>median_no_of_clicks_per_session</th>\n",
       "      <th>min_no_of_clicks_per_session</th>\n",
       "      <th>max_no_of_clicks_per_session</th>\n",
       "      <th>median_price</th>\n",
       "      <th>min_price</th>\n",
       "      <th>max_price</th>\n",
       "      <th>total_number_of_sessions</th>\n",
       "      <th>most_frequent_day_1</th>\n",
       "      <th>most_frequent_day_10</th>\n",
       "      <th>most_frequent_day_11</th>\n",
       "      <th>...</th>\n",
       "      <th>most_frequent_location_6</th>\n",
       "      <th>most_frequent_photo_type_1</th>\n",
       "      <th>most_frequent_photo_type_2</th>\n",
       "      <th>most_frequent_price_type_1</th>\n",
       "      <th>most_frequent_price_type_2</th>\n",
       "      <th>most_frequent_page_number_1</th>\n",
       "      <th>most_frequent_page_number_2</th>\n",
       "      <th>most_frequent_page_number_3</th>\n",
       "      <th>most_frequent_page_number_4</th>\n",
       "      <th>most_frequent_page_number_5</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>country</th>\n",
       "      <th>page 1 (main category)</th>\n",
       "      <th>page 2 (clothing model)</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">1</th>\n",
       "      <th>3</th>\n",
       "      <th>C30</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <th>P17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">10</th>\n",
       "      <th rowspan=\"3\" valign=\"top\">1</th>\n",
       "      <th>A1</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>28</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A11</th>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>62</td>\n",
       "      <td>62</td>\n",
       "      <td>62</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A12</th>\n",
       "      <td>3.0</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>38</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 67 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                        median_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                    \n",
       "1       3                      C30                                                  2.0   \n",
       "        4                      P17                                                  2.0   \n",
       "10      1                      A1                                                   2.0   \n",
       "                               A11                                                  2.0   \n",
       "                               A12                                                  3.0   \n",
       "\n",
       "                                                        min_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                 \n",
       "1       3                      C30                                                 2   \n",
       "        4                      P17                                                 3   \n",
       "10      1                      A1                                                  2   \n",
       "                               A11                                                 3   \n",
       "                               A12                                                 5   \n",
       "\n",
       "                                                        max_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                 \n",
       "1       3                      C30                                                 2   \n",
       "        4                      P17                                                 1   \n",
       "10      1                      A1                                                  2   \n",
       "                               A11                                                 1   \n",
       "                               A12                                                 1   \n",
       "\n",
       "                                                        median_price  \\\n",
       "country page 1 (main category) page 2 (clothing model)                 \n",
       "1       3                      C30                                28   \n",
       "        4                      P17                                38   \n",
       "10      1                      A1                                 28   \n",
       "                               A11                                62   \n",
       "                               A12                                38   \n",
       "\n",
       "                                                        min_price  max_price  \\\n",
       "country page 1 (main category) page 2 (clothing model)                         \n",
       "1       3                      C30                             28         28   \n",
       "        4                      P17                             38         38   \n",
       "10      1                      A1                              28         28   \n",
       "                               A11                             62         62   \n",
       "                               A12                             38         38   \n",
       "\n",
       "                                                        total_number_of_sessions  \\\n",
       "country page 1 (main category) page 2 (clothing model)                             \n",
       "1       3                      C30                                             1   \n",
       "        4                      P17                                             1   \n",
       "10      1                      A1                                              1   \n",
       "                               A11                                             2   \n",
       "                               A12                                             5   \n",
       "\n",
       "                                                        most_frequent_day_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                        \n",
       "1       3                      C30                                        0   \n",
       "        4                      P17                                        0   \n",
       "10      1                      A1                                         0   \n",
       "                               A11                                        0   \n",
       "                               A12                                        0   \n",
       "\n",
       "                                                        most_frequent_day_10  \\\n",
       "country page 1 (main category) page 2 (clothing model)                         \n",
       "1       3                      C30                                         0   \n",
       "        4                      P17                                         0   \n",
       "10      1                      A1                                          0   \n",
       "                               A11                                         0   \n",
       "                               A12                                         0   \n",
       "\n",
       "                                                        most_frequent_day_11  \\\n",
       "country page 1 (main category) page 2 (clothing model)                         \n",
       "1       3                      C30                                         1   \n",
       "        4                      P17                                         1   \n",
       "10      1                      A1                                          0   \n",
       "                               A11                                         0   \n",
       "                               A12                                         0   \n",
       "\n",
       "                                                        ...  \\\n",
       "country page 1 (main category) page 2 (clothing model)  ...   \n",
       "1       3                      C30                      ...   \n",
       "        4                      P17                      ...   \n",
       "10      1                      A1                       ...   \n",
       "                               A11                      ...   \n",
       "                               A12                      ...   \n",
       "\n",
       "                                                        most_frequent_location_6  \\\n",
       "country page 1 (main category) page 2 (clothing model)                             \n",
       "1       3                      C30                                             0   \n",
       "        4                      P17                                             1   \n",
       "10      1                      A1                                              0   \n",
       "                               A11                                             0   \n",
       "                               A12                                             0   \n",
       "\n",
       "                                                        most_frequent_photo_type_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               0   \n",
       "        4                      P17                                               0   \n",
       "10      1                      A1                                                1   \n",
       "                               A11                                               1   \n",
       "                               A12                                               1   \n",
       "\n",
       "                                                        most_frequent_photo_type_2  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               1   \n",
       "        4                      P17                                               1   \n",
       "10      1                      A1                                                0   \n",
       "                               A11                                               0   \n",
       "                               A12                                               0   \n",
       "\n",
       "                                                        most_frequent_price_type_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               0   \n",
       "        4                      P17                                               1   \n",
       "10      1                      A1                                                0   \n",
       "                               A11                                               1   \n",
       "                               A12                                               0   \n",
       "\n",
       "                                                        most_frequent_price_type_2  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               1   \n",
       "        4                      P17                                               0   \n",
       "10      1                      A1                                                1   \n",
       "                               A11                                               0   \n",
       "                               A12                                               1   \n",
       "\n",
       "                                                        most_frequent_page_number_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                0   \n",
       "        4                      P17                                                1   \n",
       "10      1                      A1                                                 1   \n",
       "                               A11                                                1   \n",
       "                               A12                                                1   \n",
       "\n",
       "                                                        most_frequent_page_number_2  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                1   \n",
       "        4                      P17                                                0   \n",
       "10      1                      A1                                                 0   \n",
       "                               A11                                                0   \n",
       "                               A12                                                0   \n",
       "\n",
       "                                                        most_frequent_page_number_3  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                0   \n",
       "        4                      P17                                                0   \n",
       "10      1                      A1                                                 0   \n",
       "                               A11                                                0   \n",
       "                               A12                                                0   \n",
       "\n",
       "                                                        most_frequent_page_number_4  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                0   \n",
       "        4                      P17                                                0   \n",
       "10      1                      A1                                                 0   \n",
       "                               A11                                                0   \n",
       "                               A12                                                0   \n",
       "\n",
       "                                                        most_frequent_page_number_5  \n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                                0  \n",
       "        4                      P17                                                0  \n",
       "10      1                      A1                                                 0  \n",
       "                               A11                                                0  \n",
       "                               A12                                                0  \n",
       "\n",
       "[5 rows x 67 columns]"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# One hot encoding - to convert categorical data to continuous\n",
    "\n",
    "cat_vars = ['most_frequent_day',\n",
    "           'most_frequent_colour', 'most_frequent_location',\n",
    "           'most_frequent_photo_type', 'most_frequent_price_type',\n",
    "           'most_frequent_page_number']\n",
    "df2[cat_vars] = df2[cat_vars].astype(str)\n",
    "df3 = pd.get_dummies(df2)\n",
    "\n",
    "df3.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>median_no_of_clicks_per_session</th>\n",
       "      <th>min_no_of_clicks_per_session</th>\n",
       "      <th>max_no_of_clicks_per_session</th>\n",
       "      <th>median_price</th>\n",
       "      <th>min_price</th>\n",
       "      <th>max_price</th>\n",
       "      <th>total_number_of_sessions</th>\n",
       "      <th>most_frequent_day_1</th>\n",
       "      <th>most_frequent_day_10</th>\n",
       "      <th>most_frequent_day_11</th>\n",
       "      <th>...</th>\n",
       "      <th>most_frequent_location_6</th>\n",
       "      <th>most_frequent_photo_type_1</th>\n",
       "      <th>most_frequent_photo_type_2</th>\n",
       "      <th>most_frequent_price_type_1</th>\n",
       "      <th>most_frequent_price_type_2</th>\n",
       "      <th>most_frequent_page_number_1</th>\n",
       "      <th>most_frequent_page_number_2</th>\n",
       "      <th>most_frequent_page_number_3</th>\n",
       "      <th>most_frequent_page_number_4</th>\n",
       "      <th>most_frequent_page_number_5</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>country</th>\n",
       "      <th>page 1 (main category)</th>\n",
       "      <th>page 2 (clothing model)</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">1</th>\n",
       "      <th>3</th>\n",
       "      <th>C30</th>\n",
       "      <td>-0.784923</td>\n",
       "      <td>-0.740009</td>\n",
       "      <td>-0.445811</td>\n",
       "      <td>-1.197785</td>\n",
       "      <td>-1.197785</td>\n",
       "      <td>-1.197785</td>\n",
       "      <td>-0.254235</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <th>P17</th>\n",
       "      <td>-0.784923</td>\n",
       "      <td>-0.715043</td>\n",
       "      <td>-0.550030</td>\n",
       "      <td>-0.367561</td>\n",
       "      <td>-0.367561</td>\n",
       "      <td>-0.367561</td>\n",
       "      <td>-0.254235</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">10</th>\n",
       "      <th rowspan=\"3\" valign=\"top\">1</th>\n",
       "      <th>A1</th>\n",
       "      <td>-0.784923</td>\n",
       "      <td>-0.740009</td>\n",
       "      <td>-0.445811</td>\n",
       "      <td>-1.197785</td>\n",
       "      <td>-1.197785</td>\n",
       "      <td>-1.197785</td>\n",
       "      <td>-0.254235</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A11</th>\n",
       "      <td>-0.784923</td>\n",
       "      <td>-0.715043</td>\n",
       "      <td>-0.550030</td>\n",
       "      <td>1.624976</td>\n",
       "      <td>1.624976</td>\n",
       "      <td>1.624976</td>\n",
       "      <td>-0.248735</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A12</th>\n",
       "      <td>-0.686829</td>\n",
       "      <td>-0.665109</td>\n",
       "      <td>-0.550030</td>\n",
       "      <td>-0.367561</td>\n",
       "      <td>-0.367561</td>\n",
       "      <td>-0.367561</td>\n",
       "      <td>-0.232237</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 67 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                        median_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                    \n",
       "1       3                      C30                                            -0.784923   \n",
       "        4                      P17                                            -0.784923   \n",
       "10      1                      A1                                             -0.784923   \n",
       "                               A11                                            -0.784923   \n",
       "                               A12                                            -0.686829   \n",
       "\n",
       "                                                        min_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                 \n",
       "1       3                      C30                                         -0.740009   \n",
       "        4                      P17                                         -0.715043   \n",
       "10      1                      A1                                          -0.740009   \n",
       "                               A11                                         -0.715043   \n",
       "                               A12                                         -0.665109   \n",
       "\n",
       "                                                        max_no_of_clicks_per_session  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                 \n",
       "1       3                      C30                                         -0.445811   \n",
       "        4                      P17                                         -0.550030   \n",
       "10      1                      A1                                          -0.445811   \n",
       "                               A11                                         -0.550030   \n",
       "                               A12                                         -0.550030   \n",
       "\n",
       "                                                        median_price  \\\n",
       "country page 1 (main category) page 2 (clothing model)                 \n",
       "1       3                      C30                         -1.197785   \n",
       "        4                      P17                         -0.367561   \n",
       "10      1                      A1                          -1.197785   \n",
       "                               A11                          1.624976   \n",
       "                               A12                         -0.367561   \n",
       "\n",
       "                                                        min_price  max_price  \\\n",
       "country page 1 (main category) page 2 (clothing model)                         \n",
       "1       3                      C30                      -1.197785  -1.197785   \n",
       "        4                      P17                      -0.367561  -0.367561   \n",
       "10      1                      A1                       -1.197785  -1.197785   \n",
       "                               A11                       1.624976   1.624976   \n",
       "                               A12                      -0.367561  -0.367561   \n",
       "\n",
       "                                                        total_number_of_sessions  \\\n",
       "country page 1 (main category) page 2 (clothing model)                             \n",
       "1       3                      C30                                     -0.254235   \n",
       "        4                      P17                                     -0.254235   \n",
       "10      1                      A1                                      -0.254235   \n",
       "                               A11                                     -0.248735   \n",
       "                               A12                                     -0.232237   \n",
       "\n",
       "                                                        most_frequent_day_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                        \n",
       "1       3                      C30                                        0   \n",
       "        4                      P17                                        0   \n",
       "10      1                      A1                                         0   \n",
       "                               A11                                        0   \n",
       "                               A12                                        0   \n",
       "\n",
       "                                                        most_frequent_day_10  \\\n",
       "country page 1 (main category) page 2 (clothing model)                         \n",
       "1       3                      C30                                         0   \n",
       "        4                      P17                                         0   \n",
       "10      1                      A1                                          0   \n",
       "                               A11                                         0   \n",
       "                               A12                                         0   \n",
       "\n",
       "                                                        most_frequent_day_11  \\\n",
       "country page 1 (main category) page 2 (clothing model)                         \n",
       "1       3                      C30                                         1   \n",
       "        4                      P17                                         1   \n",
       "10      1                      A1                                          0   \n",
       "                               A11                                         0   \n",
       "                               A12                                         0   \n",
       "\n",
       "                                                        ...  \\\n",
       "country page 1 (main category) page 2 (clothing model)  ...   \n",
       "1       3                      C30                      ...   \n",
       "        4                      P17                      ...   \n",
       "10      1                      A1                       ...   \n",
       "                               A11                      ...   \n",
       "                               A12                      ...   \n",
       "\n",
       "                                                        most_frequent_location_6  \\\n",
       "country page 1 (main category) page 2 (clothing model)                             \n",
       "1       3                      C30                                             0   \n",
       "        4                      P17                                             1   \n",
       "10      1                      A1                                              0   \n",
       "                               A11                                             0   \n",
       "                               A12                                             0   \n",
       "\n",
       "                                                        most_frequent_photo_type_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               0   \n",
       "        4                      P17                                               0   \n",
       "10      1                      A1                                                1   \n",
       "                               A11                                               1   \n",
       "                               A12                                               1   \n",
       "\n",
       "                                                        most_frequent_photo_type_2  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               1   \n",
       "        4                      P17                                               1   \n",
       "10      1                      A1                                                0   \n",
       "                               A11                                               0   \n",
       "                               A12                                               0   \n",
       "\n",
       "                                                        most_frequent_price_type_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               0   \n",
       "        4                      P17                                               1   \n",
       "10      1                      A1                                                0   \n",
       "                               A11                                               1   \n",
       "                               A12                                               0   \n",
       "\n",
       "                                                        most_frequent_price_type_2  \\\n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                               1   \n",
       "        4                      P17                                               0   \n",
       "10      1                      A1                                                1   \n",
       "                               A11                                               0   \n",
       "                               A12                                               1   \n",
       "\n",
       "                                                        most_frequent_page_number_1  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                0   \n",
       "        4                      P17                                                1   \n",
       "10      1                      A1                                                 1   \n",
       "                               A11                                                1   \n",
       "                               A12                                                1   \n",
       "\n",
       "                                                        most_frequent_page_number_2  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                1   \n",
       "        4                      P17                                                0   \n",
       "10      1                      A1                                                 0   \n",
       "                               A11                                                0   \n",
       "                               A12                                                0   \n",
       "\n",
       "                                                        most_frequent_page_number_3  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                0   \n",
       "        4                      P17                                                0   \n",
       "10      1                      A1                                                 0   \n",
       "                               A11                                                0   \n",
       "                               A12                                                0   \n",
       "\n",
       "                                                        most_frequent_page_number_4  \\\n",
       "country page 1 (main category) page 2 (clothing model)                                \n",
       "1       3                      C30                                                0   \n",
       "        4                      P17                                                0   \n",
       "10      1                      A1                                                 0   \n",
       "                               A11                                                0   \n",
       "                               A12                                                0   \n",
       "\n",
       "                                                        most_frequent_page_number_5  \n",
       "country page 1 (main category) page 2 (clothing model)                               \n",
       "1       3                      C30                                                0  \n",
       "        4                      P17                                                0  \n",
       "10      1                      A1                                                 0  \n",
       "                               A11                                                0  \n",
       "                               A12                                                0  \n",
       "\n",
       "[5 rows x 67 columns]"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Standardizing\n",
    "\n",
    "\n",
    "con_vars = ['median_no_of_clicks_per_session', 'min_no_of_clicks_per_session',\n",
    "           'max_no_of_clicks_per_session', 'median_price', 'min_price',\n",
    "           'max_price', 'total_number_of_sessions']\n",
    "\n",
    "scaler = StandardScaler()\n",
    "df3[con_vars]=scaler.fit_transform(df3[con_vars])\n",
    "df3.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!pip install --trusted-host=pypi.org --trusted-host=files.pythonhosted.org --user opencv-python\n",
    "#from helper_functions import do_pca,pca_results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using 15 components, we can explain 0.8521121438333593% of the variability in the original data.\n"
     ]
    }
   ],
   "source": [
    "# PCA\n",
    "\n",
    "# Loop Function to identify number of principal components that explain atleast 85% of the variance\n",
    "for comp in range(3, df3.shape[1]):\n",
    "    pca = PCA(n_components= comp, random_state=42)\n",
    "    pca.fit(df3)\n",
    "    comp_check = pca.explained_variance_ratio_\n",
    "    final_comp = comp\n",
    "    if comp_check.sum() > 0.85:\n",
    "        break\n",
    "        \n",
    "Final_PCA = PCA(n_components= final_comp,random_state=42)\n",
    "Final_PCA.fit(df3)\n",
    "cluster_df=Final_PCA.transform(df3)\n",
    "\n",
    "\n",
    "num_comps = comp_check.shape[0]\n",
    "print(\"Using {} components, we can explain {}% of the variability in the original data.\".format(final_comp,comp_check.sum()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# K means\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "\n",
    "!pip install --trusted-host=pypi.org --trusted-host=files.pythonhosted.org --user kneed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a84ca8e7b8>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Elbow Method for K means\n",
    "from sklearn.cluster import KMeans\n",
    "# Import ElbowVisualizer\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "model = KMeans(random_state =10)\n",
    "# k is range of number of clusters.\n",
    "visualizer = KElbowVisualizer(model, k=(2,30), timings= True)\n",
    "visualizer.fit(cluster_df)        # Fit the data to the visualizer\n",
    "visualizer.show()        # Finalize and render the figure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ibaruah001\\AppData\\Local\\Annaconda\\lib\\site-packages\\matplotlib\\cbook\\__init__.py:1377: FutureWarning: Support for multi-dimensional indexing (e.g. `obj[:, None]`) is deprecated and will be removed in a future version.  Convert to a numpy array before indexing instead.\n",
      "  x[:, None]\n",
      "C:\\Users\\ibaruah001\\AppData\\Local\\Annaconda\\lib\\site-packages\\matplotlib\\axes\\_base.py:237: FutureWarning: Support for multi-dimensional indexing (e.g. `obj[:, None]`) is deprecated and will be removed in a future version.  Convert to a numpy array before indexing instead.\n",
      "  x = x[:, np.newaxis]\n",
      "C:\\Users\\ibaruah001\\AppData\\Local\\Annaconda\\lib\\site-packages\\matplotlib\\axes\\_base.py:239: FutureWarning: Support for multi-dimensional indexing (e.g. `obj[:, None]`) is deprecated and will be removed in a future version.  Convert to a numpy array before indexing instead.\n",
      "  y = y[:, np.newaxis]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of clusters based on gap statistic: [29.0]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Gap Statistic for K means\n",
    "\n",
    "\n",
    "def optimalK(data, nrefs=3, maxClusters=15):\n",
    "    \"\"\"\n",
    "    Calculates KMeans optimal K using Gap Statistic from Tibshirani, Walther, Hastie\n",
    "    Params:\n",
    "        data: ndarry of shape (n_samples, n_features)\n",
    "        nrefs: number of sample reference datasets to create\n",
    "        maxClusters: Maximum number of clusters to test for\n",
    "    Returns: (gaps, optimalK)\n",
    "    \"\"\"\n",
    "    gaps = np.zeros((len(range(1, maxClusters)),))\n",
    "    resultsdf = pd.DataFrame({'clusterCount':[], 'gap':[]})\n",
    "    for gap_index, k in enumerate(range(1, maxClusters)):\n",
    "\n",
    "        # Holder for reference dispersion results\n",
    "        refDisps = np.zeros(nrefs)\n",
    "\n",
    "        # For n references, generate random sample and perform kmeans getting resulting dispersion of each loop\n",
    "        for i in range(nrefs):\n",
    "            \n",
    "            # Create new random reference set\n",
    "            randomReference = np.random.random_sample(size=data.shape)\n",
    "            \n",
    "            # Fit to it\n",
    "            km = KMeans(k, random_state =10)\n",
    "            km.fit(randomReference)\n",
    "            \n",
    "            refDisp = km.inertia_\n",
    "            refDisps[i] = refDisp\n",
    "\n",
    "        # Fit cluster to original data and create dispersion\n",
    "        km = KMeans(k)\n",
    "        km.fit(data)\n",
    "        \n",
    "        origDisp = km.inertia_\n",
    "\n",
    "        # Calculate gap statistic\n",
    "        gap = np.log(np.mean(refDisps)) - np.log(origDisp)\n",
    "\n",
    "        # Assign this loop's gap statistic to gaps\n",
    "        gaps[gap_index] = gap\n",
    "        \n",
    "        resultsdf = resultsdf.append({'clusterCount':k, 'gap':gap}, ignore_index=True)\n",
    "\n",
    "    return (gaps.argmax() + 1, resultsdf)\n",
    "\n",
    "score_g, df = optimalK(cluster_df, nrefs=5, maxClusters=30)\n",
    "\n",
    "plt.plot(df['clusterCount'], df['gap'], linestyle='--', marker='o', color='b');\n",
    "plt.xlabel('K');\n",
    "plt.ylabel('Gap Statistic');\n",
    "plt.title('Gap Statistic vs. K');\n",
    "\n",
    "df2 = df[df.gap == df.gap.max()]\n",
    "print('Optimal number of clusters based on gap statistic:', df2['clusterCount'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of clusters based on silhouette score: [4]\n"
     ]
    }
   ],
   "source": [
    "# Silhouette Score for K means\n",
    "\n",
    "# Import ElbowVisualizer\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "model = KMeans(random_state =10)\n",
    "# k is range of number of clusters.\n",
    "visualizer = KElbowVisualizer(model, k=(2,30),metric='silhouette', timings= True, locate_elbow=False)\n",
    "visualizer.fit(cluster_df)        # Fit the data to the visualizer\n",
    "visualizer.show()        # Finalize and render the figure\n",
    "\n",
    "df3 = pd.DataFrame(visualizer.k_values_,columns=['centers'])\n",
    "df3['scores'] = visualizer.k_scores_\n",
    "df4 = df3[df3.scores == df3.scores.max()]\n",
    "print('Optimal number of clusters based on silhouette score:', df4['centers'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAasAAAETCAYAAACIiCl1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzsnXl8VNX5/9+zJJmsJCFhkzUiB1wIElHAURBFauv6rUtrtY7WWr+1rVZbO+2vFdH222nr0lprtbV2urjUYt2r4C4DohhkUzgIYScQIPsySWbm/v44d8IQstyETCbLeb9evMice+45z72zfO455znPYzMMA41Go9Fo+jL2RBug0Wg0Gk1naLHSaDQaTZ9Hi5VGo9Fo+jxarDQajUbT59FipdFoNJo+jxYrjUaj0fR5nIk2oL8ghHAAtwJXo+5bMvAycJeUsrGTcw0gH5gNnCel/F43+h8PbJBSZrQqHwUsllLObuOc7cDlUsqPY8rmAg9LKU/uqg3t2NWj7XWh3ztR74UNcACvAz+RUjb1sh1+YD5woNWhJ6SUD0XfAyCDXrhPQggPcA+wUUq5oJtt3A3kSSm/Y75OBp4ERgAXo74HC4EbpJR/jTkvHdgPvCulvPAYLuOY6Oy7ar5nG6SU93Wz/aXA1VLKg104p93vqcYaWqys80cgBzhXSlllfjGfBB4HrrXSgJTyJeClnjRKSrkXJYKDBiHEFcBlwCwpZYMQwgUsBu4GfpIAkx7s7g9fHPg6SrT/2RONmZ/z/wD1wHwpZVAIAbAT9bn/a0z1LwO1PdHvMXLM39VOmN/VEwbj97Sn0WJlAXNU8zVgpJSyGkBKWSeEuBk406wzCfgDkAmMBNYAV0kpgzHteFAjnQuFEO8CH5jnjwXeBG5CTc3+3ixvBkqA61vZMxl4Dbgd+IQ2RlwWr2s48BgwHPXUvAO4UkpZZo4IPgSmogSg2fw/GRgG/E1K+TOzqQwhxGJgIlAJ3CSl3NzRPRFCLEIJThNwCPCY9+GxGBPHAp9KKc9qZfpI1GgqFWgw2/uOaRdCiAwO38MQ8ALw/4As055pgGHew59IKUNCiEbgRaAQ9V7XAb8Dhpp9PSSlfKKLt7g17d2nIW3ZBdwH1EgpfyaEGAnsBeZJKd8RQlwDXCSlvCrauBDiQeB0YIIQIh94wur1xo6+Y9rLAV5FvW/fkVJGYg6/DlwqhBgtpdxtll0H/BOYbJ6fDPwKmIO6h58A35NSVgshLqSNz5M5Uv8F6nN/MpAEfEtKuVwI4QYeMNsygF9KKZ9rZfN4OvmutqpvAPnRUVLMLEgQJcQnABGgGPgW8Bfz1HeEEF80jz2M+qwmAc9IKf/PtGMZsBEYb96bN6SUGebIdTzqczwO2ANcI6UsFULMQIltMrDVPH67lPLd1rYPNvSalTWKUD+a1bGFUsp9MV+Wb6K+cDNRP0YTgC910u7xwFyUIFyA+lLPMssKpZRFqC/t1OgJQoiTgVeAG6WUz1uw/UkhxJroP9TTZZSvAB9IKWcBBain59gnzw1SyimoH/s7gOuklKcBM4EfCyHyzHpjgAeklNOAp4B/dHRPhBBjgNuAGWZ7S4EzpJQfSimnme18E/WU/s02rulvqB/7fUKID4QQ9wNjpZQfmcfvAVzAFNQP9Zmoe/sQShhPAU5D/VD/wDwnGXhZSilQP86LAa/5HswBfiCEmNnOPf5+7D02/53SRr327lN7dv0H9bkA+AKwj8NP9RcDR/xQSym/D3wM/FBK+aDV621LqFAPL++iBOOeVkIF6uHlWZQwIIQYi3oo2RBTx4t6WCiSUhaixNYnhLDR8efpDOB+KeWpKMH4P7N8kXn/ioAbgHlt2G3lu2qFy4BM872aYZYVSCmjD47nSCl3od7DJ0ybTgfOE0JcadYZDdwrpZwElLZq/yzgCinlZNSD0c1CCCfqPf+ZlHIq6v2b1gWbBzRarKwRofN79SPggLmW8kdgFGqdoiNellJGzC/WFiAXWA+EgQ+FEPcCz0kpV5j1U4B3gDVSyrcs2v61qACYX7wboweklL8DVgghbgceQf0wxdq8zKxnABcBRUKIhainWxuQbtZbF2OjHzjNHC20d0/2AGuB1UKI+8zreSHaqRBiIupLe62UclPrC5JSVkkpz0c9wT+OejJ/VQjxK7PKecBfpJRhKWWTlHKO+WR6AWrdyDDXGR/lsBi0XC8wCfUg8YQp8O+hRnGntnOPH4y9x+a/9W3Ua+8+tWdXABhtjoC/APwcmG+OWOYA/23HnihWr7ctLkWNGJ4BnhVCJLVR5+/ANebf15qvY7kQuAT4xLyPlwInWvg87ZBSrjH/Xo36XoASxz8IIZ5EiVJbU75WvqtWCAAnmTMgXuC3UsotsRXM6cU5wL3m9a1EjbCiAhNCzZ60xbsxgvoJ6hpPAZBSvmb+/w5Hiv+gRouVNT4EpgghMmMLhRDHCSFeFUKkAk+jpvF2AA+ivmS2TtptiPnbAGxSykoOPwGHgX8JIb4dU+9SYLoQ4svHckGm/b9CjUIOAH9CjXBiba4166WjvlDTUdf1Q9STdbRuuFXThnm8zXtiPqXPQU39HQIeFEL82uxrGGq6yiulfK8du+8UQsyWUpZIKf8ipbwW9SN8i1klZNoQrT9GCDEU9XmPDYZpR03dHHG9qGmmqlYiP5Mj12e6Q3v3qU27zPv0CvBF1Gjjz6ipoyuAFVLKztaHrF5vW/xVSvln4Huo6dP7W1eQUq4CHEKIacBVqNFiLA7g1ph7eDpwuYXP01HfC7O/x1A/6G8AC4B15nplLFa+q62xmXWSY65tG2o24Jfm9b8phLiojeuzAbNbfU6iI8FGKWWojf7au8YQR/9mtP7MDFq0WFlAqsXRJ1FP2lkA5v+PAIeklA2oL889Usp/maedgfowdwlzLv8t1I/R3ain1eg0RKOUcjlqCuSPQogR3b8qMG3+rZTyH0AZaoqpLZtPQH1hfyqlfBk1TZkSU7fQ/MECNa8fkFLW0849EUIUop4YN0opf4kSshnmWtOrqFHRkx3YnYaaTsqNKTsF9cMHav3vOiGEXQiRgprSmwMsAb4jhLCZ5TehfvhaI4EGc10Ic9pyA+pp/lho7z51ZNd/gDuB9VJ5Or6N+gG1MqVl9XrbohFAqjXXK1D385o26v0D9f5tllKWt9N/shDCjhLbX9L556lNhBArgFOllH7zWrJR05UtWPyuxnIANUUKynsw2tf/oh5Olkopf2Rey3TzcBj1MFGNGk3dbp6TDSxHjSa7w0agUQjxBbO901Gfax1tHC1WXeHbwGeoabM1qCe4zzg8rfYT4HkhxHqUk8B7qCezrvIa8CmwQQjxMcqDaFFsBXNK6xnUAvqxcA9wnxBiHcpLMUDbNq9DPeFvEkJsRE3hfBZTdyOwUAixFrWWcp1Z3uY9kVKuRU3pfGxe4w2oL/x3UaPKy4QQn8Sss7XmXpQgrRBCbBRCbAbcQHStYBHKcWMt6gn+v1LK/6BGCcNQU63rUaL0i9aNm6JwCXCjeW+WotYRlrdzH9tas3q0jXrt3aeO7HoTNX0aFZklKIeYl9uxJRZL19sZUsrPUZ/zx2LENso/gbNR05qtuRfYjnoPPkONGu6g889Te9wJ3COE+AS1nrZISrm9jXqdfVdj+R5qanE1ao0zurb0d5R4fiaEKAaGoNaQAP4NvGeuH18NzDQ/4x8CT3fyoNUu5ijsy8Dd5jXegVqnrO9OewMNm04RotFoNH0DIcRvgPuklPvNEf1alGNHZYJNSzjadV2j0Wj6DjuAt4QQ0TW8G7VQKfTISqPRaDR9Hr1mpdFoNJo+z4CYBiwuLs5E7ffYhHKB1Wg0Gk3nJKH2K/6zqKioJtHGdMSAECuUUD2SaCM0Go2mH/PHRBvQEQNFrDYBjB49mrS0tETb0m02b97MpEmTEm1GXBno16ivr/8z0K8x9vrq6+vZvXs3mL+hfZmBIlbNAGlpaWRmZnZWt0/T3+23wkC/Rn19/Z+Bfo1tXF+fXz7RDhYajUaj6fNosdJoNBpNn0eLFbBoyVoWLVmbaDM0Go1G0w4DZc2q2yxaspZ7lq5reb1wQWECrdFoNBpNWwxqsWotVNG/tWBpNAOD0sotlBxYQ22wggxXDgX50xiZ3Z340ppEM2jFqrVQRdGCpdEMDEort7B219str2uC5S2vtWD1PwblmlV7QhXlnqXr9BqWRtPPKTmwhkgkTF1jFY3N9RhGpKVc0/8YtCMrjUYzsKkNVlDXVEVzKEgjUNdow+lIprG5nsZQPSnO/htAYDAyKMUqOsXX3ujqrvOn6mlAjaafk+HKIRwJ0WhX2eqbwg2Ewo0YRoRNez+gcOy5AESMMHabo2V9a2/jToKfb9PrW32MQSlW0L5gaaHSaPo39Y3VhCLNFORPY+2ut0lNzgAglQwikTAjs49ndO6UlvrF21+nqv4gNcFDJDtdgKHXt/ogg1as4GjB8sw4XguVRtOPqWus4qOSV4gYIc6adBWFY+aZ3oCVZLiyjxotGUYEGzYO1e4mHGmmoakajCRAhSMqObBGi1UfYVCLFSjB+nRfJc+t28mk/KxEm6PRaLpJbbCCVdteoTHUgBgxk2Sni5HZEzsUG5vNzmkTvsiBml00hRqob6oiZDTFtKmT9PYV4ipWQogzgF9JKecKISYCfsAANgC3SCkjQoiFwJeAEHCblPKj9urGy86HLjud59bt5P2S/fzo3JPj1Y1Go4kTNcFyVpW8QlM4yJSRsxmX17XvcVbqUGqC5TSFgzSH6ohEwtjtDjJc2XGyWNNV4ua6LoS4E3gccJlFDwA/lVKeBdiAS4QQ04E5wBnAV4A/tFc3XnYCjMhKZVJ+Fsu3HSAciZsmajSaOFDdcIiPTKE6cdSZXRYqgIL8aQCkONNIsrnAZjuiXJN44rnPaivwPzGvi4D3zL9fA84D3MBSKaUhpdwJOIUQ+e3UjSvuCcOoaWxm7d6KeHel0Wh6EIfdgd1m56TjzmLs0JO61cbI7IkUjpnH0IxROG2pDEnNo3DMPL1e1YeI2zSglPI5IcT4mCKblNIw/64BhgBZwKGYOtHytup2yubNm7tt7xh7PQBPvVeMMXlot9s5VoqLixPWd28x0K9RX1/8qA2XURneQZNRT5ItjRzHODIcw8gwTqBsWz1l247NNhcTKEiZgFFlsLe6ir0MzPeyP35Ge9PBInZ+LROoBKqJut0cWd5W3U6ZNGlSt5OmDZ1Qy6KVz7O9OYWioqJutXGsFBcXJ6zv3mKgX6O+vvihwicVk4IdRziZusZD1KUYTBp3Qo+OgN768AXsWXXMmngprqT0Hmu3rxD7HtbU1BzTQ35v0pvhlj4RQsw1/74AWAYsBxYIIexCiLGAXUp5sJ26cWVcTjpjstNYVrIfwzA6P0Gj0fQq0TBJoXAT1cFDhCPNhCLNcQifZNAYqqO8rrSH29UcC70pVncAi4QQHwDJwGIpZTFKiD4AngNuaa9uvI2z2WycVTCcg3WNbCqrjnd3Go2mi9QG1XpysLkODIN0Vw7JTlePu5e77MoDsEKLVZ8irtOAUsrtwEzz780oz7/Wde4G7m5V1mbdeOMuGMZTq7fxfsl+pgy3tEym0Wh6iQxXDjXBckKRZmw2OymOVLO8Z93LU2yZ1NuT9MiqjzHoNwXHcnbBcACWbd3Pt2ZNSrA1Go0mloL8aazZ+RaRSAinI1ltaqHn3cttNhs5aSM4WLuLxuZ6UpL6bsDbeOfr8ge8duARoBBoBG70uH1bYo7/APgqys/g/zxu3/M91nkrBmWKkPaYPCyLvPQUlpWU6XUrjaaPMTJ7IhOHT8dhT8LpSCbTlRs39/Lc9JEAfXp0Fc3XVRMsx4iJZ1hauaXzk61zKeDyuH2zAC9wf/SAP+DNBr4HzALOB37bkx23Ro+sYoiuWz2/fic7KuoYn5uRaJM0Gk0MBfmnkp85jiRHMukp8YsukZ81huZIIxmunLj1cayUHFhDc7iRxuZ6MlKyWzYy93A8QzfwOoDH7VvpD3hPizlWB+wA0s1/cY2ooEdWrTi7YBgA75fsT7AlGo2mNQ67k+y0YXEVKoBM11DEiDPIdOXGtZ9j4WDNHmoaVIioUKS5pbyHHU6ygKqY12F/wBs7yNkFfAasBh7qyY5bo8WqFWe1rFuVJdgSjUbTmmBzXa9P0fe1JQHDiLCpdCXB5lpsNhuZrqFqDc+khx1OWu+FtXvcvpD59wXASGACMBa41B/wnt6TnceixaoVU0dlk+VKYpkeWWk0fQrDMFi2+V98WPJir/R3oGYnyzY/S1n19l7pzwqhcDOf7HiD7QfXkZ02nKzUPJJihAp63OFkOfBFAH/AOxNYH3OsAmgAGj1uXxAVvCFuQ169ZtUKh93O7PH5vL5pL/uqGxiRlZpokzQaDVDfVE04EiItuXe2lTjtydQ1VlJet5fhQyb0Sp+dYbPZCIbqGZo+imlj53OwdleH+bp6gOeB+f6AdwXK//J6f8B7O7DF4/a95A94zwNW+gPeCBAA3ujJzmPRYtUGZxcM5/VNe1m2rYwrCscl2hyNRgPUBFUY0UxX78TuHJKaj93m6BMegaFwE05HMg67k9PGX4DTkYTd5ug0X9ex4nH7IsDNrYo3xRxfCCyMmwEx6GnANjjLdLJYtlVPBWo0fYXqBlOsUnvH6cFud5CdNlzluQoFe6XPtthXVcK7m56iom4fAMlOF3abI2H2JAo9smqD08YMxeV0sKxEO1loNH2F3h5ZgdpvVV63l4q6fQwfMr5X+oxu9K0JVmAYEZrCQdKSswhFmjo/eQCjR1ZtkOx0MGt8Huv3VVBe35hoczQaDUqsUpxppDh7bx25tzcHRzf6VjccojZYQUVdKQ1N1RTkF5KfObZXbOiraLFqB/eE4RgGLN+mR1caTaIxDIOTj5vDlFGze7XfIWnDGJN7IvmZY3qlv5IDazCMCDXBQzSF6nE4kshy5bGvqqRX+u/LaLFqh5Z1Kz0VqNEkHJvNRl7maEYMKejVfh12Jycd5yYvc3Sv9FcbrMAwDAwjQrIzlSxXHna7o8cjy/dHtFi1w8xxeTjtNgJarDSahBMx4hrJp8+Q4crBbneQlZpHeko2NjOEUk9Hlu+PaLFqh/SUJE4bM5Ti3YeobWzu/ASNRhM3Nux+j3c3PaVyWfUy9U3VrNz6Ipv3fRTXfkortzI8S40cbTZ7i1BBz0eW749oseqAswqGE4oYrNxxMNGmaDSDGrWGE+xV54ooKc40qhoOcLB2T9z6qKjbx7rd77C7YiOnjJ5LpisXG/a4Rpbvb2jX9Q5wFwzjN+98yrKS/Zw3aWSizdFoBiURI0xtYyVZrqHYbL3/fO2wOxmSmk9l/f6Wzbk9SUNTDZ/sWAqGwSmj5zI04ziOy9H59FqjR1YdcOb4fGw27WSh0SSS2mAlhhHp1f1VrYm6sFfU7+vRdkPhZlbvWEJTOMiUUbMZmnFcj7Y/kNBi1QE5aSlMHZnDhzsO0hgKJ9ocjWZQEt0MnJWaSLEaBUB5bc/ttzIMg3W7VfLEMbknMnboST3W9kBEi1UnnFUwjGAozMe7DiXaFI1mUJKIyBWtyU4bjg17j24ObgzVU9NQztD0UUwZNavH2h2oaLHqhJb8VjpliEaTEIZljqMgf1pCEyE6HUkU5BcyJndyj7XpSkpn5sRLmTZ2/qCM9ddVetXBQgiRAvwVKEAl9boFGAr8DggBS6WUi4QQduARoBBoBG6UUm7pTVujnNWSObgM77mJsECjGdzkZowiN2NUos3ghBEzeqSdqoYDOO3JpKcMSYh3Y3+lt0dW3wRqpZQzge8CDwOPAlcDbuAMIcR04FLAJaWcBXiB+3vZzhaGZ6YyKT+LFdsOEAp3bWPioiVrWbRkbZws02g0/Y2GplpWb3+dlVtfJBQe3IFpu0pvu66fCLwGIKWUQogZwH4p5VYAIcQS4FxUquTXzXorhRCnWWl88+bNcTF6SpadzQeaeebtFUzJtfYk9Kd1ZTy+Qe3P2rt3LzdNHWbpvOLi4m7b2V8Y6Neor6/nqI+UUx7aSo5jAumOvF7rt61rNAyD/aENGBiMTJpqua3acBmV4R00GXU0Gw3YcTAs6UTWrlnf+clxoj9+RntbrNYAFwohXgDOAIYAW2OO16CmCLOAqpjysBDCKaUMddT4pEmTyMzM7GGT4TIjhxe3LudQylCKiqZ0Wn/RkrUtQgXw+IaDjBo1ioULCjs8r7i4mKKiomO2ty8z0K9RX1/PsrXsE+r2b2PyuMkMy+qdRKgdXeOKLbuoCR5i2omFOOyd/3yqKOrFpBh2mhqbIRQiKSmZycefxKicxGz0jb2+mpqauD3k9zS9PQ34BGqt6h3gImAtkB5zPBOoNOvEqo69M6GKJ2e3rFt17mSxaMla7lm67qjye5au01OCGk0X6QuegLHkpo/EMCJU1ltzuCo5sAYMqGuqojkUxOlIJi15CNsOromzpQOP3h5ZzQACUsrvm1N7xwOThRDHAyXAAmARMBolZs8KIWYCiRsvA+NyMxibk86yrWUYhnFEzK5Y2hOqKNFjnY2wNBqNorrhEEmOFFxJ6Z1X7gVy00ey/eA6yutKLW3grQ1WECFMUyiI3e4kw5WLzWbrN1HU/QHvUc5uHrdvi3lsGvDbmOozgUs9bt/r8bClt8Xqc+BeIcQPUCOobwBjgScBB8ob8EMhxCpgvhBiBWADru9lO4/irIJhPFm8jY37qzhxhI6ArNHEm1CkmfqmKnLTR7b7gNjb5KSPAKC8di8M77x+hiuHmmA5WakqVJTdDBfVj6KoXwq4PG7fLH/AOxPl7HYJgMftWwPMBfAHvFcAe+MlVNDLYiWlPAic16p4L0qRY+tFgJt7yy4ruCcosXq/pKxNsWoOR3DYbdhtEDHabuOu86fqUZVGY5HaYDnQd6YAAZIcKWS58qhsKCMcCbW7bhWOhNi4dwWjcgSy9IOj6vWjKOpuTGc3j9u30h/wHuXs5g9401EzYmfH0xC9KdgiZ5ubg/+4fNNRa0/Fuw5x+oP/ZeHraxmRmcpV045eCB6ZlcrP5lv3INJoBjtOezJjck/stcSHVhkz9EROGH4aRjs5tkKRZlZvf53dFZuoC5ZTOGZef46ifpSzmz/gba3Q3wD+7XH74pqeQkddt4gYlkVakoMN+6rYsE+tPd057yTuWbKO+9/7jHDE4BtnTOTXFxWRnZqMGHZ4/eqkEUP4dF8Vf/t4K9ef3m8+pBpNQslw5XDSce5Em3EUHUWxaA43Ubz9NSrr9zMscxxTRp2Jw+7sT+LUmqOc3TxuX2tnt68Bl8fbkE7FSgiRA/wa5QxxOXAfcIeUsiLOtvUp7lm6jvrm8BGvHw5sory+iQm5GTx2xUzOjUkjEjvd940zJjLlVy/y//77CV+eOpYsV8+mGNBoNImnKRTk423/pTp4kJFDjueUMXMHQhil5ZjObuaa1RHObv6AdwiQ4nH7dsXbECvTgH8GVqHCItUCpcA/42lUX6M9L7/y+iZOHzuUtT+48AihirJwQSELFxQyOjudH807mf01QX7xRkIdGzWafoFhRFi17VVKDvTN7R6b933Ee/JpIhH1AGsYBqt3LKE6eJDROZOZOuacgSBUAM8DQX/AuwJ4EPi+P+C93R/wXmwenwRs7w1DrEwDTpBS/kkI8b9Syibg/wkh+uYnKA505o7+0c5D3PfuZ506Ttwx90Se+HALv1u2iRtnnsAJ+Vk9bapGM2Cob6rhUO0eUpxpiTalTcKRZhqaaqhqOEBO+ghsNhtixEzKarYzafjpfcZ78VjxuH1tObttijm+CuUxGHesjKxCQoghgAEghDgB6FqQPA2pSU5+fVERzeEIP3ip/4U60Wh6k+oGtVbflzwBY4kYEarqD/D2xn/wvnyG0sot5KQPR4w4Y8AIVV/DysjqLuBdYKwZJmkWcEM8jepLREdM7Y2uuuKO/uWpY5lz/HBe+Ww3SzbtZcHkxEeS1mj6Ii2RK1ITlxakPUort7Dj0KeEI82EI82UVm6lrlE5zPVjR4o+j5WRVSkwH/g6KlzSVCnlq3G1qo+xcEEhd51/tNt5V/dN2Ww2Hrz0NOw2G7e/uIrmLkZx12gGCzV9cI9VlJIDa7Db7DjsSQCkJKVhs9lUaCVN3LAysvqXlHIKMKgEqjWtR1jd3eBbOCqXb848gcc+2MwjyyW3nt15YFyNZrBR3XCIFGdan8z3VBtUjtAZrhwiRpgkR4pZ3j9CKPVXrIjVZ0KIu4APgYZooZTy/bhZ1UeJFadjiURxzxcK+dea7Sxasparp08gP8PVE+ZpNAOCiBEmN30ETlME+hrREEoOuxNHzE9oPwqh1C+xIla5wDnmvygGMC8uFvVxeiJcUl6Gi7sXTOW2Fz7mZ6+t4dErZnZ+kkYzSLDbHBSO7btpuQvyp7F219ttlmviR6diJaU8B0AIkQk4pJR6rNsD3Dxb8KcPPufxDz/n5tmTmHZc31tI1mg0RxN1oig5sIbaYCUZrmwK8qdp54o4YyWCRQHwDCqChU0IsQO4Ukr5ebyNG8gkOew8cMlpfOFPb3HbC6uYe/xwSkvLeGzg5u3TaCyx/eB6mkNBJgybhtN0YuhrjMyeqMWpl7HiDfgY8Gsp5VApZS7wS1RUC80xMl+M4qKTRrOspIx731jP4xsO6gSNmkHPngrJtoPrcQyMCBCaHsKKWOVJKRdHX0gpn0WtY2l6gHE5RyaV0xmFNYOZiBGmtrFSRSm36aQQmsNY+TQ0CiGmR18IIYqA+viZNHhYtGQtDwfkUeVasDSDldpgJYYRIdOln4c1R2LFG/A24DkhRDkqa28ucFVcrRoEdBZzMHpMJ2vUDCaikSuyUvveZmBNYrHiDbhSCDEJFV3XDmyXUtbE3TKNRjPoaAmz5MpLsCWaeOAPeE8BovFlt3jcvg1Wz7XiDXgl8DMp5SlCiONRm4S/I6V8sdsWa3o05qBGM1Cw2xy4ktLJdOUk2hRND+EPeG2oyO23ATXATiAEjPcHvFnA74DHzAjv7WJlGvCnwHkAUsqt5prVUkCL1THSnmDZbXDZKWMTYZJGk1AmjTidSSNOT7QZmp5lMfAGcIbH7Ts+jwWQAAAgAElEQVRin66ZvPE6VN6sSzpqxIpYJUsp90dfSCnLhBDdioEvhEgC/gaMB8LAN1EK60dFxdgA3CKljAghFgJfMo/fJqX8qDt99nVaC9ZXTh3PM59s57qnlrPytgtIcWr3XY1G06/5usftq2vrgMftqwIe8ge8f+msEStiFRBCPA08iRKUrwAfdMXSGL4IOKWUs4UQ84FfAEnAT6WU7wohHgUuMTcezwHOAMYAzwEzutlnnycqWHv37uWxa84iI8XJ4yu3cPfra/nlhdM7OVujGRhUNxzkUO1ehg8ZT1qyTk46UIgKlT/gzQWme9y+N/0B70+AUwGvx+3b2p6YxWJFrG4Bvgt8C2gG3gP+2E27NwNOIYQdyDLbm2m2CfAacD4ggaVSSgPYKYRwCiHypZQHOmx88+ZumpV4LswD8oZRXFzMNWOcvLYhid+88ykTHXVMG9Y3s6V2l+LigZ18Ul9f96gI7aAivI3hzpNJdyTWwUK/h3HhaeANf8ALcDnwIPA4R8adbRcr3oCNQognpZT3CSHOBk4Bujs3VYuaAtwE5AEXAmebogRq8W0ISsgOxZwXLe9QrCZNmkRmZmY3TUs8xcXFFBWpeEtPDx/H3D8s5ZerD/LJHReS6eqbYWe6Suw1DkT09XWfT3aUE6rOZIY4k9TkjLj0YYXB9B7W1NT05kN+jsftu88f8P4e8Hvcvn/4A95brZ5sxRvwj0CyEOJ+4J+ohbJZwDXdMPb7wBIp5Y+FEGOAt4HkmOOZQCVQbf7dunzQcOaEYfzwnBP51dufcsdLH/OnK2cl2iSNJq7UBMtJcqTgSkrvvLKmV/AHvHbgEaAQaARu9Lh9W2KOXwAsNF+uBm7xuH3GUQ0p7P6Atwi4FJjjD3inYW12T51soc7pwI3AlcATUspvAMJqB62oAKrMv8tR61WfCCHmmmUXAMuA5cACIYRdCDEWsEspD3azz37L3QsKKRyVw18+3MLLn+5KtDkaTdwIhZupb6oywyx1y39LEx8uBVwet28W4AXujx7wB7yZwG+ACz1u30xgO2rGrD1+ZNa/3+P2lQCPArdbNcSKqjlQonYJcLMQIg3o7qPPg8ATQohlqBHVT4CPgT8LIZKBjcBiKWXYrPOB2fct3eyvX5PsdPD3q89kxoP/5aZnV7Luh/k6UaNmQNKX09gPctzA6wAet2+lP+A9LebYbGA9cL8/4C0AHve4fe0u1XjcvreAt2JedymRnxWx+jtQCiyXUn4ohPgMFYm9y0gpa1EjtNbMaaPu3cDd3elnIHHyyBx+fsE07nxlNTcvXsni6+boJ0/NgKK0cguf7llGRd0+7DYH2WnDdPqNvkMWh2fDAML+gNfpcftCqFHUOcA0lD/CMn/A+4HH7TtiEcwf8EZQnuRRmlFbl1xAtcfts7QDvNNpQCnlA8AIKeVlZtHZUsrfWWlc0zPcNmcKc44fzgvrd/H3j0sAFVtQB7vV9HdKK7ewdtfbhCLN5KSNIBwJsXbX25RWbun8ZE1v0Np/wG4KFSgnuFUet2+fx+2rBd5HCdcReNw+u8ftcwB/Qm0ATvW4femogcvi1vXbw9LilpQyEvP3oFs7SjQOu52/fmU2hfe9wq3Pr2LNnnIeWrap5bgOy6Tpr5QcWEMkEgabDbvNjg1bS7keXfUJlgMXAc/6A96ZqGm/KMXAyf6ANw/lADeTjnMdnuFx+/43+sLj9j3nD3h/atUQy54YmsQyLjeD3146g2/8a8URQqWjs2v6K8HmOvZXbScYqsPlTCct5fBG4NrgoHL+7cs8D8z3B7wrUFk3rvcHvLejgtC+5A94fwwsMes+20lg2jp/wHs98CxqVu9alKOdJay4rg+TUpa1KjtTSrncaieanmF7edvB7rVgafoTjaEGth1Yw85Dn9EcbsRus+NwHLmPMMOVnSDrNLGYwWVvblW8Keb4M8AzFpu7BngYeAi1hvUGSrAsYWVktVsI4TXXrqL8HtBxgHqRRUvWcu8b69s9rgVL0x/YU7GZT/cEiBghXEkZTB41i70Vnx/lNFSQf9TSh6af43H7dqCmFLuFFbEqAc4UQpwGXC+lbAS0O5pGo2mX0sotlBxYQ22wgrSUIUwcNp2R2RPJdOWS5Ejh+GFnMDpnMna7g2GZY826lWS4sinIn6bXqwYg/oB3AfBzVALfFg3xuH0FVs63Ilb1UsovCyHuBT4QQlyGioSu6UV0/itNfyHq4WcYEYLNdZTXlVJVf4Ci8TAyeyJzJn8Fu+1wxLaR2RO1OA0Ofo/aBLyBI13ZLWFFrGwAUsqfCSHWooLO6rwVCaA9wRqR6eK2s6ckwiSN5ihKDqwhFG6mJngIw4hgs9mJGOEWD79YodIMKg563L5XunuyFbF6IvqHlHKxEOJz4Nfd7VBzbLQWrOnH5bJ6TzmXPPEOr910LqlJ2sFTk1gq6vZTEzyIYRikJmfiSkrHZrNrDz/NMn/A+wAqIkYwWuhx+963crKVX7bHhBBTpZTrhBBXo3KQfK1bpmp6hNjpvp/OP4Wr/xlg8dodXPX393nOM5ckh5WQjxpNz1NZX0Z9UzWGYZDuyiHFmdpyTHv4DXqiKaBPjSkzgHlWTrYiVv8EtgkhUoFFqPBLflR6D02CiBWsf1x9JlUNTbz62R5ueGYFf/vqmdjt2gdG0/ukp2QzNOM46horSXYeGcdSe/gNbjxu3znQEgDX0TrFfWdYeQSfIKX8EfA/wONSynuB4V01VBM/kp0OnvPMYda4fJ5avY3vv7gKw+jy+qVG022aw40AJDmSmTflWmZM+KKKoI6dTFcuhWPmaSeKQY4/4C3wB7wfoaKzl/gD3k/8Ae8JVs+3MrJyCiHygMuA/xFCjABSOzlH08ukpyTx8o3ncM4jS3k4IMlNS2kZfUVjCGpvQU082F+1jfW732P6+AXkpo/EZrNpDz9NWzwG/Nrj9i0G8Ae8V6LCM821crKVkdVvgA+BV6WUG1DBCu/tlqmauJKTlsJrN51LwdAM7lm6jt8v28iiJWu5Z+k67lm6Tge+1fQ4pZVbWbPzTQwioEfzmo7JiwoVgMftexa158oSVtLaPwU8FVM0hSOz+2r6ECOz0ljyrfM4++El3PbCx0cc01EuND3JnorNrN/9Lk57EkXjv0hOul4d0HRIoz/gne5x+1YDmFmD662ebCU24EWoXccZqD1XDiANyO+WuZq4UzA0k4tPGs1jH3x+1DEtWJqeYFf5Rj7ds4wkRwqnjf8iQ9L0z4GmU24DnvMHvOUoLckFrrJ6spU1qweBbwJ3AL9ApTnubqZgTS+waMnaNoUqihYsTVeJhk/a27iT+s1bqW+sJsnhYsaEL5GVqrP7ajrHzDQ8CZiEWoLa7nH72o7O3QZW1qwqpZTvACuBIaZnoCW/eI1G0/+Jhk9SqecN6horCUeamZA3VQuVxjKmQ8Vqj9v3KWr67zN/wHuJ1fOtiFWDEGISsBGYK4RIRq9Z9WkWLijkrvOntntcxxHUWKU51Mine5ZR11hFZX0ZYaMJALvdQWmVzuar6RI/Bc4D8Lh9W4Ei1N5dS1iZBvwpas3qWsALfAv4S5fN1PQq7cURzEpxcsPp2qVYc2Rk9AxXTku0c8OIsHn/Kg7V7qG64SDltXvVCTYbtpjnWx0+SdNFkj1u3/7oC4/bV+YPeC1HL7DiDfgeKngtwAwhRI6UsqLrdoIQwgN4zJcuYBrKx/53qEjuS6WUi4QQduARoBBoBG6UUurHuC7SWrDmTRzO21v284U/vcl7tywgL8PV0emaAUx0ag/DIBRppqx6BxV1+1sio5dVbae+uYactBEYRoRwJITDnkRtbW1LGzp8kqaLBPwB79PAk6gwS1cBH1g92Yo34GhUZse5QBPwphDi+1LKA121VErpR4VqQgjxB1SQ3EeBL6PyZr0qhJgOjAdcUspZQoiZwP2A5blNzWFip/vuOn8qd768mgfe+4wvPf42b948n0xXUgdnawYqJQfWYBgGtcHylugTRrLREhn91HHn40rOwGlPOixsrdDhkzRd5Bbgu6jZuWbUnt1HrJ5s6ywsjxBiKfAC8A+Uu+E3gHOllN2ODWgmcrwPuBj4UEo5xSy/FbUeNhL4SEr5jFm+R0p5XHvtFRcXu4Fl3bVnMGEYBj//sJSXSyopGp7Gb+eOJUUHvh10lDS+S7NRT8howo4Tpy0Fu82JDTsFKXOOql8bLqMyvIMmo55kWxrZjnFkOIYlwHJNnDirqKgo0LrQH/AeNcvlcfu2xBx/CDgTiHr1XeJx+6ra68Qf8I4HTgKWAGM8bt82qwZaWbPKl1LGqt+DQojrrHbQDj9BLaxlAdUx5TVAgVkee8FhIYRTStlh0sdJkyaRmZl5jKYljuLiYoqKiuLez+JTI1z59/d5ccMu7vu0jme/fjbOXhKs3rrGRNFfrq9s3Soq6mpIcbjIcuW1pJXPdOVSdEL79veX6zsWBvo1xl5fTU0Nmzdv7qj6pYDL4/bN8ge8bc1yTQcWeNy+g5316w94r0L5QKQCs4EP/AHvDzxu3z+t2G3lF+ojIcRXoi+EEBcCH3dQv0OEENnAZNMdvhqIVZdMoLKNcntnQqWxjtNh56lrzmLexBG8uGEX3/r3Sh34dhARjoSw25zY7Q4yU3JbhAr01J7mKNyo/FN43L6VwGnRA+ao6wTgT/6Ad7k/4L2hk7Z+hBKpGo/bV4ZKFfJjq4a0K1ZCiIgQIozaEPyUEKJWCFENvIQKattdzgbeBJBSVgNNQojjhRA2YAFqOm858EXTjpnA+mPoT9MGriQH/7l+LqeNGYp/1VbufHk1hmGwaMlaHUNwgOOwO5k35VrOKLiYIWn5OjK6piOOmuXyB7zRGbl0VKr6a4AvAN/2B7zt75mBcOwmYI/bVwpErBrS7jSglDJe80IC5UwR5WaUd4gD5Q34oRBiFTBfCLECtU52fZxsGdRkupJ49cZ5zPnDEh547zPW7DnE21taPEv1XqwBRm2wAoMIma6hJDlTmJBfyIR8/R5rOuSoWS6P2xed5aoHfudx++oB/AHv26i1rSP3yxzmU3/A+x0gyR/wTgO+DayxaogVb8B8lHLGxgacIKX8utVOYpFS/qbV65XAzFZlEZSIaeJMXoaL1286j6m/eekIodIhmQYWjc31fLz9NULhJs4SVx2RwVej6YDlwEXAs+aaVews1yTgGX/AOx01S+cG/tZBW7eg1qwaUJ7gb6HC+FnCyujpX6j9UNeghn2X04Whm6bv88RHW6huPHpJUKcVGRiEws0U73idYHMtE/KnaqHSdIXngaA/4F2BihP7fX/Ae7s/4L3Y4/ZtRM2KrUTtxf27GUqpTTxuXx2w0OP2zUDtsXoXqLNqiBVvwFFSynlCiPuA/wC/Bo7edKHpl0TzXbWHHmH1bwwjwtpdb1HdcJDROYKC/FMTbZKmH+Fx+9qa5doUc/zXKE3oFH/AexcwxR/w/gglbp8C5wO3WjnfysgqGq1CAoVSykNWGtZoNInFMAw27l3BgZqdDM0YzYnHuY/w/NNoeplLgBuAq4EnPW7ffNQeLUtYEau3hRD/BpYCdwghHkXNOWoGAJ0Fvb1jzhQ9qrJAaeUWln++mJLG91j++WJKKxMfHawpHGR/9Q4yXbmcOvY87DZHok3SDG7sHrevAbgQeNV0fbecbspqPqshUsodQoivAnOAe7plqqZP0l7QW4DAtgPUN4VIS7byURmcRMMRGYaBQYSaYHlLeKLedgU/OjhtIcOHjMfp0IkSNAnnTX/AuwHlRfg+airwJasnWxlZLZNSbgWQUq6WUj4opdzbLVM1fZbWI6yfzT+Fa4oK+HDnQa59KkA4on1q2iMaZ686eJDGSDWRSLilvDeJimZF3T5CkWZqguVsLF1BRd2+XrVDo2kLj9v3Q9T+2VnmWth3PW7fj6yeb+Vxea0Q4lrgI2Km/6SUO7tqrKZvEzvdt3BBIU2hMLsr63hh/S5+9Mpq7rv4tA7OHrzUBitoDNUTDjdjYFDXVEWmK7fXU2hs3reKusYqGkP12LCTnTYMm83WEpxWo0kE/oD3CcDncfs2e9y+Ft3wuH1rzOMnAT/wuH0d7qe1IlZnmP9iMVAx/DQDjFjBSnY6WOyZg/v3r/PgexspyM3k226RQOv6JqnJmZTXlWKz2bEDzaEgTaEgQzNG9Ur/dY1VlBxYw56KzYCB3e4gLTmrxZlC553SJJifAb/1B7wjgQCwGxV1fTxwjvn69s4asZLPasIxmanp1+SkpfDKjfOY/dDr3PrCKsbmpnPhiaMTbVaf4vhh0zlUuxebzUZTpBm7E5yO5F6Js1deV8pHJS8DkJKUitOeTLIz9QivP513SpNIPG7fHuAKf8BbgNpgPBk14NkCfM3MGtwpViJYnAB8h6MjWJzdTds1/YwJQzN58RvnMO+RpXz1H+/z7rcXUDRmaKLN6jMclzOJmcdfTMmBNewN7mRk9tiWrLs9QWunifzMcYzPO4Vkp4vstOEMzxrPyOyJRCJh1u1+56jzdXBaTV/A4/aVoBLtdgsrDhZPoyKhn4qK4zQW2NDdDjX9k9PH5vGPr7lpaA5z8V/eYWdF3aAPemsYBrvLJREjzMjsiZx5wuUUpMzhzBMuZ2T2RPZXbeNgze5j6iPqNFHTUE5TOMieis18vO3VFm9Du83OqePOZ8SQAkblnEDhmHlkunJ1cFrNgMPKmlWylHKhECIJWA38mWNIEaLpv1x2yljuu6iIO14q5vQHX+VAXWPLscG4F2tPhWTDnvepbSxn8shZRxxraKplza63SHak4p50BUnddB0vObCGcCRMXWMFoXATAEmOFGoa2t6bPzJ7ohYnzYDEiljVCyFSgM1AkZQyIIReZB+s3Hr2FJ75ZDurdh3+sRyMIZmaQ43IfR/hsDsZn3f0purU5AyOzz+VLWXFyNKVnDy6e7PmVfUHqAkeJBIJk+R0kZqUidOR1JKKXqPpT/gD3nTgeFRA3DQzXqAlrEwD/hN4GXgV+K4Q4jVgT3cM1fR/7lm67gihii0fTFOCn5d9THM4yPHDinAltb0Jv2DYNDJdueyu2MSh2u59ZVxJ6USMCKnJWWS6cnE6kgDtNKHpf/gD3nOBtcCLwHBghz/gPd/q+Z2KlZTyYeDLUsoDwFzgT6hUx5pBhpWgt4NBsKobDrHz0KekJQ9hfN7J7daz2xycPHoONmxs2PM+oXBzl/uaMmo22anDSE3OOKJcO01o+iH/h0ojUulx+/ahEvH+puNTDtOpWAkhcoBrhBA/QwUhPAX4Qfds1Wj6Nyo47HIAThx1Zqfx9oak5jMhv5CGphp2ln9mqY+Kun18vO2/hMJNjMyeyKnj5munCc1AwG6KFAAet8/aF8LEyprVC0AZKpy70TXbNAOJjmIIAgzLSOGG0wf2j6jNZmNCfiFZtXnkZVrbb3b8sOm4kjIYnTu507qHavdQvH0JhhGhsr6MvMzR2mlCM1DY7Q94LwQMf8CbjUrGaDkSkhWxypVSzumudZqBRXuCVTQ6l+Ld5cx+6DVevnEe047LTYR5vcKwrHEMyxpnub7D7mTs0BM7rVdWvYM1O9/EwODUcfMti6FG00/4Fmqf1RigBJUp+CarJ1sRq/VCiCIpZXH37NMMNFoL1l3nT+Wu86fywLufcecrq5nzhyU8fe3ZfHHKcYk0s8cprdxCTvrIdh0qOiNihNla9gkAJww/Ms5iaeVW1u16B5vNTtH4BeRlaKHSDCw8bl8Z8NXunt+uWAkhtqGm/dKAq4QQe4AQKoqFIaXUsQEHMa2D3gLccc5JjMvN4LqnlnPJX97hof+Zwf/OPrzNYdGStezdW8ZjRb1u7jFTG6xg3a53yXBlM3vil7uVxDBiRNhb+TnBplqGZY1jSGo+AMHmWtbtfgeH3UHR+AvISR/R0+ZrNAnHH/BeDvwYyIkt97h9lrSko5HV3O6b1T5CiB8DFwPJwCOonCZ+lDBuAG6RUkaEEAuBL6EE8jYp5UfxsEfTfdraV3V54ThGZ6dx6RPv8J3nPmLboVp8X5rOvW+saxmJjVqytl/tyTIMg42lKzCIcMLwGd3Otuu0J3HycWcT+Hwxb336N9JSlDt6Qf40Thk9l7TkLLLThvWw9RpNn+F+4FpgR3dOblespJTdarAjhBBzgdmoVMZpKK/CB4CfSinfNbMQXyKE2IFK8ngGan7zOWBGT9ujiQ8zx+Wz4nsXcOGf3+b+dz/j1c92s6msuuV4f9tEvL96G4dq95CXOYb8zLHH1FZTqIFQuJHG5nqaw40YhsHaXW9TOGaeFipNn8PM5vsIUAg0Ajd63L4tbdR5FXjR4/Y92kFzW4CAmcuqy/R2+tcFqJ3LzwNZwA+Bb6JGVwCvAecDElgqpTSAnUIIpxAi39zr1S6bN2+Om+G9RXHxwFkafPjskVz7Wv0RQhXlnqXr2Lt3LzdN7bs/0LXhMirC26kJ78OGjaT6fFYfWt3peR29h7ubVhE27IQjEULhIBWhMpJsaXy08U1GJ1f1pPlxYyB9RttjoF9jF67vUsDlcftm+QPemajR0SWt6vwcsOJRdT/wjj/gfQ81YwaAx+2zlHm+t8UqDxgHXAhMQKU0tpuiBFADDEEJWWyYhGh5h2I1adIkMjMze9rmXqO4uJiion64oNMOi5asZV99qN3jj284yKhRo/rkCEsFkC0m0lSPo8mGKzmD5uQDjBpzSodu5J29hwfXryaFLEJhF8HmOlKTM3HYndiwU3RK33/vB9pntC0G+jXGXl9NTU1nD/lu4HUAj9u30h/wHuEZZK5DRVADjc74KbAJCKN8H7pERw4WF0opXxFCfL2t41LKv3e1M5QAbZJSNgFSCBFETfNFyURFeK82/25drtH0CtGU9CnOVCJGmNSkzJbyY9nzlOHKoSZYjtORTEZMcFsdPknTR8kCYof8YX/A6/S4fSF/wHsycDVwOXCXhbaSPG7fDd01pKOR1QzgFVQmx9YYQHfEKgDcKoR4ABgJpANvCSHmSinfBS4A3kHNbf5aCHEfMBo1+jrYjf40CaSzTcR3nT+1z42qwpEQOw5t4GDNblKS0nDYnWSkHBaSY826W5A/rSW9R+tyjaYP0nrgYPe4fdHpkq8DxwFvo7L+NvkD3u0et+/1dtp6wx/wfgc1UmuKFsamuu+IjhwsFpp//kBKeUTkUiHERVYab6PNV4QQZwMfoUI93QJsA/4shEgGNgKLpZRhIcQy4IOYepp+SHuClZrk4LJTjs1ZoScxDIN9VSVs3vcRDc01hCJNpBip0Mrz71hHQNFRmUqmWEmGK7tHEzVqND3MclR232fNNav10QMet+/O6N/+gPduYF8HQgWH91jdEVNmAMfsuh7lTSHEfCnlQSHECOBh4ERUJPYuI6W8s43ioyJkSCnvBu7uTh+avkVrwfrSlON4deMezvvjGyy9+byER7uorC9jU+kHVNbvx2azMz5vKukpQ/h0z7Kj6vbECEiHT9L0I54H5vsD3hWodabr/QHv7cAWj9v3Ulca8rh9E47FECti9XPgDSHE34E7UW6M3d6FrBmcRAVr7969PHb9PJ74cAs3/fuDFsGaPnpor9jROkX88KwJbClTnlHDs8YzacQZpKcMAdS+KD0C0gxmTDfzm1sVb2qj3t3tteEPeO/2uH13+wPev9JGfFmr61idipWU8jkhRDVqr9PF5tqSRtNlFi4opLhYTXffcMZEHHYb3/jXCuY/+iav33QuM8bmxbX/aIp4wzAAg5pgOTUN5WSnD+eE4TMYmjHqiPp6BKTR9AhRP/l3j6URK+GWQA3/bMALQohyAB1uSXOsXDfjeBx2G9c/vYLzH1OCdca4/Lj1t7XsE4LNdTQ01ZLsTCE9JRtsEI40HyVUGo2mZ/C4fdElo1Eet++Xscf8Ae//WW2n18MtaTSxXFNUgMNm4+tPLWfBY2/x32/OY/aEYS1JHHvCWzBiRNhb8Tm7KySRSAhsNmw2h3oUsx27h59Go2kff8DrA4YBF/sD3hNiDjmBmcBPrLTTabglIUQuMF1K+aYZ12864O2u4RpNa746fQIOu41rngxwwZ/f4vKp4/Cv2tpy/FgEq6JuH+t3v0d9UxV2m52kpHRSkzKw2w8nTdR7nDSauPIcyinvXA5HKwIVxeJeq41YcbB4GuVgAXAF8CDwOG3vv9JousWV08bjsNu46m/vHyFU3YkjaBhGS7DZZGcqweZaxuROISMlm42lHxxVX+9x0mjih8ftWwWs8ge8L3jcvm7HFLMiVjlSyvuEEL8H/FLKfwghbu1uhxpNe2worWwzFXVHghXr3ZeekkNuxkgq6vZx4qgzyUkfQXrKEM6ZfA1JzhRAiZf28NNoep9jESqwJlZ2IUQRKqDhHCHENIvnaTSWWbRkbbuRLqBtwYp692FAc7iRygbJzkMbSHflUFG/ryUvVFSoQHv4aTT9FbuFOj8CfgPcL6UsAR4Fbo+rVRpNGyiX88OUHFhDOBKiOniImuAhwuFmkpwuhqTm6ak9jaaP4Q94r2ujzHJ0Iiv7rN4C3op5PdOydRqNRTqLIwiwYvsBNh+oZlJ+FqCy9zaG6qlprMcwkhiRNRSnI4nG5oZesVmj0XSOP+C9DRUQ92Z/wDsu5lASKhDuH6y006lYCSGuQ+UhOSIVsZTS0fYZGk33aE+wvusWfH6whtc37aXwNy+xcP4Ibp0zhwxXDtvKq9lX7SAYgpDRyPjcJO3dp9H0LT4HTuPwft0oQcBjtREra093AXOllBu6Yp1G0x1aC1Y0MrthGCxeu543N75JZW0JV/x1J8fn5ZJqryM6m729vBaASwrnJcR2jUZzNB6371XgVX/A+y+P23dUqCarWBGrvVqoNK1pHWOvJ73qYp0oFi4opDnUyOdlH5Pp+JSLT0xnw/6hLN8R5rXNNUzOy2fmmCqGpjVzqD6Jlzelsj9Yx8IFPbluVV0AABzQSURBVGKKRqM5RvwB7yset+9C4DV/wNtWbMAei7peLIRYDCxFDduAbidf1AwAWrzwTGqC5S2ve0KwSiu3cF7B59QGK3hjQzFN4UYcdgdpyUM4cdRsPt5XTnWjGnltOpjBpoMZR5zfnb1ZGo0mbjxp/n8lUNbdRqyI1RBUWvlZMWXdTb6oGQBEs+g2h5uob6wiyZlCijPtmLPogimEO99umdmub66hrrESMWImp447D7vNAZQf4xVoNJpe5F5/wPtv4DGP2ze9u41Y8Qa8vruNawYmtcEKDMOgrrGCSCRMuKmZYFMt9Y1VlFZuZfiQCdhtVnZFHMYwDGoby1m94w1qggfJcuVhtztIcaSSlJZMsLnGFKrOPQfz0lM4eWQ2kYiB3W476nhPxh3UaDSd8j7QCNj8AW84ptwGGB63z5KzXkdR11+RUl7YKvp6Czrq+uAlw5VDdcMhkp2pYIDDkURjcx0RI4Lc9yEjhqgca7Fhj6JrXHsbdxL8fFvLGld9UzWllVsprdxCbWMFVfX7sdlshCIhku0OsIEdx1HBZtsTrGmjcli/r5Ir//Y+px6Xyz0XTOOCyaNa7Gi9+VgLlkYTX8x8VTf4A94XPW7fJf+/vTsPb6u8Ez3+1WJbdiIbJ7aTmDgkgfgtkBUTlqJCKARIKReYtvcyQ+kYHobbodNpodOpO2VL6ExdCnRu71yWNnTUdQqXNi1tCTFtSMABkkYkMVl4ndVZnM12Esu7ljN/SHIUW5IVx9p/n+fhwTrn6Jz31XH807uc3zva88RqWf1d8P+LRntykZ1mls9ny4HVFOUXD24rsBZSPekKiovKMAVbVbuOuejoPkxRfjEHOz4KBgxjcIzL6/OwvbURAwOTycyk4ulYzHl4fQODwSUk0nT0aDMHm493snTVFl7evI/blq/mmunlLFsyn7W7j54RqGRsS4jkOZdABbGD1eJg8tpoZMwqB/kNP4ZhMHfq9ext2xIzx17vgJsT3YfZd7wJv+Ej31rIgNHDgDePfKuN/R3bmFE+j6L8EiaVTCfPUjBs8kZItIwUQ2cOAlSXF/OLz3+Cuhtm89jKzby27SA3PP9mxPdLwBIiM8QKVrGyqssEixy1v30rHx1+n4sqarhm1mdjHju36nourLiMVVuX0+/ppt/Tjc/w4fUPkI+Nrr6Tw84RCnhnk2w2WqCZM6WUFfddzwOvvMdL63dFfb8ELCEiczbWmYHngHkExp3ur3XU7wrb/yUCD/YawLJaR/0fElWWWOtZRZ1YoZQqHO0FlVKbgFD23b3Ai8D/IbC2SYPWeqlSatgHpLWO/tdGJEXPQCfNRzaSZ7ExbeKlcb1nXEEJk4ovoLO3Ha9/gN6evsHuw2iZJsY62ez5JUWjfq9MxhA57g7AVuuov9rZWHcVgWxGtwM4G+vKgAeB+YAN2O5srPtjraM+0uIJ52zEKVtKqduUUluUUruVUnuUUi1Ay2guppSyAWitFwX/u5dAYty/ARzAlUqpywh+QFrrqwks9PjMaK4nxo5hGGw/tA6/4eXiKVeTb7XF/d6Z5fMxmUzkWQowhf3KJSvZ7OM3z+Oxm+bGPGbjgXZ+++F+PD7/4LbQZIxlDU2DQUuIHOMA3gCoddS/TyBtEsHXbcC8Wke9B5gMnExUoIL4nrP6PoHJFl8D/pVAIBk3yuvNA4qUUg3Baz8BFGitdwMopVYRWE1yCsEPSGv9vlLq8sinE8ly+NRu2roOMHH81LNu9YR37bndXdhtE5K+jlS02YOfuvh8jnf18fqOQ7y+4xCT7DbuqZlJj8fLc+uaB4+TrkKRo4o53RMG4HM21llrHfVegFpHvdfZWPcPwFLgB4ksSDzB6qTW+i2l1DVAidb6G0qp7aO8Xg/wNIGVhmcBK4HwOcluYCYRPiCllFVr7Y118ubm5li7M4LL5Up1EYbxG14ODKzHj4/z+ov5oOODUZ3HxgxmFsyATmjtPEUrya3rp8ugdXYZy7e2AXD/7DIemFsClLDrZAmv7T7J63tP8fSayL/eyxqaaG1t5YG5FTGvk473cCxle/0g++t4FvXrBOxhr82hQBVS66j/D2dj3Q8JpFO6vtZR/9YYFfMM8QSrXqVUNbADWKSUWg3kj/J6zcAurbUBNCulTgETwvbbCQSvIoZ8QCMFKoDq6mrsdvtIh6Utl8tFTU1NqosxjGEYTD1Vhtc/QNWEi8/pXKmu44s1UBlhHKoG+F83wKMrN/Fvf4qeCnP51jYqKyujtrBSXb9Ey/b6QfbXMbx+brd7pC/564DbgFeCY1YfhnY4G+sU8B3gM4CHwPwCf6STjIV4gtUjwLeBewiMH/1v4KVRXu8+YA7woFKqkkBQ6lZKXQjsAW4m0JycSvADUkqd8QGJ5DOZTEw578JUF2PMxOrKs5pHzrzh8yesW16IdLMCWOxsrHuXQMaJe52NdQ8Du2od9a85G+u2AO8RmA24stZRvzZRBYkn3dJapdQOrXW/Uuo64FKt9V9Geb2XAKdSqpFA5e4jEIl/AVgIzAZcr5T6C4HnvAY/oFFeT5wDn9/LrqMuZpTPO6sJFZksnkUgf7VpH9UVxfz1gulYwoLb0lVbaG09xovZ+6Vc5JhaR70f+OKQzR+F7V9KoIGRcPEsvviPBObRXwaUAz9VSn1fa/3Ds72Y1nqAwMy/oa4aclykD0gk2e5jH7C3bQsmk5nqyQtTXZykiRawHrr2Yvq8Ppav38Xf/nIdT63eyhM3z+fOOVWDswYh0M0oEzGEGFvxdAM+AFwJoLVuUUrVAOuBsw5WInN09raz9/gWbHnjmVmRnCnm6SRaKieAf7r+Up5saOKnG/fwuZ+sZbLdxhH34Oo5MnNQiASIJ1jlERg4CxkgQmJbkT0Mw8+2Q29jYHDp+Z/Aas5LdZFSIlIqJ4DpE8bz0l0f558/eSmfc65l29FTw94rAUuIsRVPsPotsFop9QqBIPUZ4HcJLZVIiVBm9OPuA/R7eji/tJpye1Wqi5VSsYLNrzbtixioQiRgCTF2Rpz6pLX+BoGHvRRwIfADrfWjiS6YSK5QAtnO3nZ6B9z4DR+dfe0cPilZrs7FiZ7+kQ8SQowonpYVWutXgVcTXBaRQqHVf00mE8WFE/H5fZhN5jFZ/TdbxTNz8IX3duI34JHFc6iwn06pKTkHhTg7cQUrkf3cfScwDD8mkxmLOQ9LcJxq6KKH4kzRAtaji+egKkp47I3N/L91GudfdvO1RZfw8HWX8Oza7bIApBBnSYKVwDAMfH4P7r4Oim0TMZtPrzIdLTO6OC3WzMHPzJ3G8vd38eSbgantT7+1jR7P6ZW9ZVxLiPhIsBLsa2vC6/dggmEr9CYrM3qmCwWb1tbWMwJPvtXCgw7FFxbO5PaX3mLN7qPD3isBS4iRSbDKca0nd6KPrKfYVsa8qk9y6ISOe9FDcabHb56HyxU5heUza7ZHDFQhyxqaMAyDJ26J/OVAxrhErpNglcPaug7y4cG1WM35XD5jCXbbBKaXzUl1sXLW8+82U5Rv5e6amWcsGBlaVytEApbIRRKscpTX76Fp/2oALrvgJuy2CSO8Q5yLkWYOXjyphN1tbr75x0186/XN3DBrMl9YeCHbj5zkO38+nQVeugxFrpJglaOs5jzmX7CYAW8vE8ZXpro4OSFawApNyDjR08/Lm1v42cbdvNl8mDebD0c8jwQskYtGXg9BZBWPtx+v3wPAhHFTmFwyM8Ulyi2P3zyPx26aO/g6fOZgaVEBX/x4Nev+cQkPXlMd8zzLGpoGx7GGWrpqS9R9QmQqaVnlEJ/fi6tlJYZhsHDGrVgto11DU5yLaDkHw5WNG3lJljW7jnDXgumoipLBbTK+JbKVBKsMFMrh19V3gvG20piz9kLHuvs66Pf0gglmlM0dfOhXpMZIQWSkMS6LycTbe45xyXdfY35lKXctmEFrZw8/eGdwqSHpLhRZRYJVhgnl8IPAw7ynetvYvP/PAMMC1uCxBnQPnKTf04PVUsDE8ecPe55KpJ9YY1wPX3cJr207wMub97Hqo1bq/vhBxHOMFLBkSrzIFBKsMkwohx+Ax9dPV18HAO80v0K5vQqrpQCrOZ+5VdcHjjWgs68dr68fizkPu62UfW1NnF8ae0xEpIdY2THurpnJ3TUzqfuDi++9tT3qOZY1NOH3GyxdcuYzXNJlKDKJBKsM09F1GLPZisVswWwyk2e1YRgGYGAx5+Px9tHj7wSgq+8EBn68vn7MZgt22wRMJrPk+8swI41xFeaN/M/46TXb2NnmZsnF53PLxyp5bp0+I1BJl6GIxNlYZwaeA+YRWNfw/lpH/a6w/Q8BdwVfvh5c5j4hJFhlkGOdLfQMdAIGJYUVWC352C2B56PstglcM+uzAMHgBeNtpXT2tlNcWIbFnDfY9Sf5/jJPrCAy0vhWzdQJHO/u5+XN+3h5876o54mny7C19Rgv1sRZaJEN7gBstY76q52NdVcBzwC3Azgb62YCdxNYSd4A3nE21q2oddRHX4bgHMjU9QzR3nWIzfv/RFF+CeMKzoMhQ07hOfxMJhMmk4mZ5fMxmUxYLflnjFFJvr/sM3RKfMhjN81lw0O3sudbd9L09du4cdbkmOeJNiU+1GW4fGubTIvPLQ7gDYBaR/37wOVh+w4At9Q66n21jno/gVXl+xJVkJS0rJRSFYALWAx4ASeByLwV+JLW2q+Uehy4Nbj/q1rrDakoazo42XOMD1pWYWBw9UV34PH1BWcDxs7hF9oWz7Ei88Ua3zKZTFw6+Tw+PqOCP+08EvM8v3DtoXycjU9fOpVppeOGjW1Jl2FOKQbCl8P2ORvrrLWOem+to94DtDkb60zA94BNtY765kQVJOnBSimVB7wI9AY3PQs8orVeo5R6AbhdKdUCXEegeVkF/BpYmOyypgN3XzuufSvx+33Mn3YjZfapwPCZf9FMOe8iCU45ZKTxrZG6DCeNt7G7vYsvr9jAl1dsoGK8jWNdw78sS8DKGZ2APey1udZRP5it2dlYZwN+DLiBBxNZkFS0rJ4GXgC+GXxdA6wN/rwSuAnQQIPW2gD2K6WsSqlyrfXxpJc2xby+wO/F7KnXMalkRopLIzLBaJ/hCrXE9p/o5o/bD/L9tdvZ3d4V9TwyLT4nrANuA14Jjll9GNoRbFH9Dlhd66j/bqILktRgpZSqBY5rrVcppULByhQMShCIziUEmp7tYW8NbY8ZrJqbE9YCTRqXyzVsm91QHNnj5gjD92WiSHXMJplQv0+XQevsMpZvbQPg/tllfLrMO1j2K2xw/RRbzGAFsGX3fjZu9Ax7bu+HTccGz93a2soDcysSUIvEyYR7eC7Oon4rgMXOxrp3CYyU3+tsrHsY2AVYCPSAFTgb65YEj/9mraP+vbEuLyS/ZXUfYCilbgTmAz8Fwn+L7cBJhjc9Q9tjqq6uxm63j3RY2nK5XNTU1NDv7eWj1nf5WOXHKbAWprpYYypUx2yVSfV7sQYqY7R+QvujdRkC/G73SXZ0+vnrBdP5m5oZXFRWzNJVWwYDFcDyrW1UVlZmTAsrk+7haITXz+12x/ySH5w48cUhmz8K+3nkvGBjJKnBSmt9behnpdQaAh/C95RSi7TWa4AlwFsEovZTSqmngamAWWvdNvyM2cfjG8C1dyWdfW0UF5Yxozwz/oGLzDTaLsNv3TibhdPK+OUHe/n9toMsbWhiaUMTlcWFtHb2DjuPdBmKc5UOz1l9DfiRUiof2AG8qrX2KaXeAd4jML3+S6ksYKKF8ve19rewf/NazGYLF5YvYHrZ8KnIQiRbrFmGt11ahbvPw4qt+3lyVRN7Os5+jEsyaYh4pCxYaa0Xhb28LsL+J4AnklSclAnl7zMMgwGjC8+AQb61kAnjpkj+PpE2QgGktbV1WDCx2/L4wuUXsre9K2aXIcDbu4/y1q4jXDmtjKJ866imxUsrLDelQ8sqp4Vy/bn7OvAZXgqt4xhXcB5727ZQWTorxaUT4rTHb56Hy+WNuR+iT4sHWLP7KGuefxOr2UTFeNtZdxlKKyx3SbBKAcMwONFzhN4BN119JwAosBbi9fgYX1CKyWSS/H0iI8WaFv/gNYp39x1n3d5j/GrTXg6dGh6oQpY1NOHx+fn2pxYMbpOHk3ObBKsEibTmVLn9Ag6f3Mn+ju24+zqwmvMpKiihu/8kBXlFDPT5JH+fyHixxrhun13F7bOrGJdvHbHLsP7PW1m54xCXT5vIkc5e/rD90LBjpNswd0iwSoDwNacATvUc591dKyiwFmG15GHCzOSSmUybeCl9A900HVw97BySv09ksnPNpHHFtIlYzWY+ONjB5tYTMa81lt2Gkqw3fUmwOgsjrdBrGAZe/wD6yHoGvH3kWfIxmcz4DR/9nm78ho8FUxYzdcLHsOWNC7xpHJhMgbErt7sLu22C5O8TWeFcM2kAeHx+vvybDfzo/Z0xz/VfH+wlz2LmimllLKyaSElh/ll3G4YfX7lqi7TE0owEqziFWkt+w4/f78Xd18GWA6vpGejkRPdRej1u+jzd+PweOrpaASguLMNqycdqyWe8bQL51kIumjT8K1sof5+r00XNLPlKJ3JHrC5DgDyLmRc+dxVTigujtsKKC6zsbHPz6MrAZCWTCSYWFdDW3T/s2Hinz8t4WPqRYBWnPcc30+/tpbv/JBgGpcGp5fvbd9Dv7SbPUkBRvh1b3ni8Pg8eXz9mk2Xw/flWG3ZbaQprIER6GqnLMHx7tFbY4c4e1re0sWF/Gy9v2se+E91Rr7esoYmfbdzNVReUU2G38WHrCVbvOhrxuFhlkrGw5JJgFQfD8HOss4XeATcmk4mC/PGD+/q9Pdx4SS1WS/7gtqFjViEyDiVEZPH8wY/VCptSXMQdc6Zxx5xpFFgtI07e2NvRzd6O6AEtZFlDEz6/n2VLFpyxfTRjYfEcJ6KTYDUCj2+ApgOrGfD2YTZbsdsmYDGf/tjsttIzAhXIOlJCJMq5tMJCHrtpLnU3zOZ4Vx9PNjSxfP2uiMeFPLV6GxsPdnDjrCl8ctZkfvvhfp58czD5+FmNhcU6TsQmwWoEzUc2cNy9n4qS6fQNuDGZzlxcOVprSdaREiIxRtMKCwlvjU09bxwv/s+rqSwpihrYFlZNpGvAy6qPWln1UWvU6431WJi0xIaTYDWC6skLseWNY0b5PI6e2iOtJSEyxEiTN6IdFxJ+fOupHv7hNxv43dYDUa+3rKGJ//pgL4tVJReUjmN9Sxu/+XB/xOPCrzuUtMQik2A1hGEYtLR/SFF+CRXFF5BnKeDCikB/tbSWhMgs8XQbhu+LFtgqS4qYV1kaM1gB7Gxzs7NNj1iuZQ1N7Gl38+SSBUwtKcJsDiQDkFyJ0UmwCuPze9l26B1aT+6kKL+EMnsV5iHdfkKIzBLvH/FYyXrD90frMnx08Ry+cu3FtJzo5qnVW3l5c0vM6/3ctZefu/ZSmGdhVlkxHp+PHcc6hx0nuRIDcj5YhR70PdXbRr+nG7PZSoV9GgsuuEkClRA5ZrTJesNbYqVFBfzynmtRFdEXrrxjdhVzppSij59i53E3Ww+fwOM3Ih4bul5LRxf/fudCim2BCV259mxYTger0BRzr28Ad98JDMNHvrWQqomXnM4wIYQQYcZyLCzkiTc2nzHDMJKfbNzDTzbuobq8mDyziW1HTw07JpsDVk4Hqz3HN4MBPQNuDMNHUUExNut4Wto+ZGqpSnXxhBBpaqzGwkKeuGU+JpMpakvsr+ZUMWOinU0HO1i37xj9Xn/Ua2ZrwMrpYNXVdwJMML7gPHx+L3nWguB2WZ5DCBHb2Y6FjfSeeFti8bTCxoqzsc4MPAfMA/qB+2sd9buGHFMOvAvMqXXU9yWqLDk9KDM+mP7IbLYMBqrAdlmeQwgxdh6/eV7cz4c9dtPcwdcRuwxvmX/GMUNFa72N0h2ArdZRfzVQBzwTvtPZWHcz0ABMGqsLRpPTLauZ5fMlLZIQIq2MRa7EMeQA3gCoddS/72ysu3zIfj9wI+Aay4tGktPBStIiCSHS0bnmShxDxUD4TA6fs7HOWuuo9wLUOurfBHA21o31dYdJarBSSlmAHwEK8AH3AibACRjAVuBLWmu/Uupx4FbAC3xVa70hEWWSB32FEJkq3vGwc9AJ2MNem0OBKtmS3bK6DUBrfY1SahHwLIFg9YjWeo1S6gXgdqVUC3AdcCVQBfwaWJjksgohRNpL8Ky/dQT+br/ibKy7CkjOzI4IkjrBQmv9W+CB4MsLgKNADbA2uG0lgf5PB9CgtTa01vsBq1KqPJllFUIIwQqgz9lY9y7wfeAhZ2Pdw87Guv+R7IKYDCP6U9OJopT6CXAn8FnAqbWuDG7/JHAf8BHQrrV+Prj9beA+rXXEXP4ul8sBvJOMsgshRBb6RE1NTWOqCxFLSiZYaK3/Vin1DWA9UBi2yw6cZHg/aWh7TNXV1djt9pEOS1sul4uamuxe1j7b6yj1y3zZXsfw+rndbpqbm1Ncovgke4LFPcBUrfV3gB4C0x43KqUWaa3XAEuAt4BdwFNKqaeBqYBZa90W49R5AD09PYksflK43e5UFyHhsr2OUr/Ml+11DNUv7G9mXsoKE6ekdgMqpcYB/wlMJvDh1AM7CMwQzA/+/Hdaa59S6gkCwcsMPKS1jtpEdblcf0/gKWshhBBn78GamprnU12IWFIyZjXWXC6XHfg8gbEuT4qLI4QQmSIP+Bjw85qamrRuTmZFsBJCCJHdcjo3oBBCiMwgwUoIIUTak2AlhBAi7UmwEkIIkfYkWAkhhEh7Ob1ESDpRSm3idCr+vVrre1NZnrGilLoS+K7WepFS6iIiZNhPZfnGwpA6Xgb8HtgZ3P281vrl1JVu9JRSecCPgelAAfBtYDtZdA+j1PEg2XMP417pIlVljJcEqzSglLIBaK0XpbgoY0op9c/APUB3cNOzDMmwTyBRZsaKUMfLgGe11s9Ef1fG+DyBHJ33KKUmApuAzWTXPYxUx2Vkzz2Ma6ULMuAeSjdgepgHFCmlGpRSq5VSV6W6QGNkN/BXYa8jZdjPdJHqeKtS6m2l1EtKqcxNVgn/H3g07LWX7LuH0eqYFffwLFa6SHsSrNJDD/A0cDPwReAXSqmMb/VqrX/NmRlFTFrr0FPobqAk+aUaWxHquAH4utb6WmAP8HhKCjYGtNZdWmt38I/1q8AjZNk9jFLHrLmHAFprb3Cli/9LoI4ZeQ8lWKWHZuDnwfW7moF2YEqKy5QI4f3icWXSz0ArtNau0M/AglQW5lwppaoIJJf+mdb6l2ThPYxQx6y6hxBY6QKoJjB+FWmli7QnwSo93Ac8A6CUqgSKgcMpLVFibAr2m0MgSXE2rkG2Sil1RfDnGwBXrIPTmVJqEtAAfENr/ePg5qy6h1HqmE338B6l1DeDL89Y6SK4LWPuYcZ3NWWJlwCnUqqRwAyd+7TW3hSXKRG+BvxIKRXKsP9qisuTCH8P/IdSagA4wunxgkz0L0Ap8KhSKjSu8xXgB1l0DyPV8WHg37PkHv4G+M/gArZ5wFcJrnSRafdQEtkKIYRIe9INKIQQIu1JsBJCCJH2JFgJIYRIexKshBBCpD0JVkIIIdKeBCshEkgptUgptSbV5RAi00mwEkIIkfbkoWAhkkQp9RXgTuBTWuueVJdHiEwiwUqIJFBK1QKfQQKVEKMiwUqIxJtNIIHoXVrrrlQXRohMJGNWQiSem8CaV99TSo1LdWGEyEQSrIRIvBat9e+BNQRWoRVCnCUJVkIkz9eBu5VSl6W6IEJkGsm6LoQQIu1Jy0oIIUTak2AlhBAi7UmwEkIIkfYkWAkhhEh7EqyEEEKkPQlWQggh0p4EKyGEEGnvvwEhRD7qn3sBAwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of clusters based on calinski harabasz: [2]\n"
     ]
    }
   ],
   "source": [
    "# Calinski Harabasz Score for K means\n",
    "\n",
    "# Import ElbowVisualizer\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "model = KMeans(random_state =10)\n",
    "# k is range of number of clusters.\n",
    "visualizer = KElbowVisualizer(model, k=(2,30),metric='calinski_harabasz', timings= True, locate_elbow=False)\n",
    "visualizer.fit(cluster_df)        # Fit the data to the visualizer\n",
    "visualizer.show()        # Finalize and render the figure\n",
    "\n",
    "\n",
    "df3 = pd.DataFrame(visualizer.k_values_,columns=['centers'])\n",
    "df3['scores'] = visualizer.k_scores_\n",
    "df4 = df3[df3.scores == df3.scores.max()]\n",
    "print('Optimal number of clusters based on calinski harabasz:', df4['centers'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of clusters based on DB score: [4]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Davies Bouldin score for K means\n",
    "\n",
    "from sklearn.metrics import davies_bouldin_score \n",
    "\n",
    "def get_kmeans_score(data, center):\n",
    "    '''\n",
    "    returns the kmeans score regarding Davies Bouldin for points to centers\n",
    "    INPUT:\n",
    "        data - the dataset you want to fit kmeans to\n",
    "        center - the number of centers you want (the k value)\n",
    "    OUTPUT:\n",
    "        score - the Davies Bouldin score for the kmeans model fit to the data\n",
    "    '''\n",
    "    #instantiate kmeans\n",
    "    kmeans = KMeans(n_clusters=center,random_state =10)\n",
    "\n",
    "    # Then fit the model to your data using the fit method\n",
    "    model = kmeans.fit_predict(cluster_df)\n",
    "    \n",
    "    # Calculate Silhoutte Score\n",
    "\n",
    "    score = davies_bouldin_score(cluster_df, model)\n",
    "    \n",
    "    return score\n",
    "\n",
    "scores = []\n",
    "centers = list(range(2,30))\n",
    "\n",
    "for center in centers:\n",
    "    scores.append(get_kmeans_score(cluster_df, center))\n",
    "    \n",
    "plt.plot(centers, scores, linestyle='--', marker='o', color='b');\n",
    "plt.xlabel('K');\n",
    "plt.ylabel('Davies Bouldin score');\n",
    "plt.title('Davies Bouldin score vs. K');\n",
    "\n",
    "\n",
    "df3 = pd.DataFrame(centers,columns=['centers'])\n",
    "df3['scores'] = scores\n",
    "df4 = df3[df3.scores == df3.scores.min()]\n",
    "print('Optimal number of clusters based on DB score:', df4['centers'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Silhouette Score: 0.2732\n",
      "Calinski Harabasz Score: 887.8355\n",
      "Davies Bouldin Score: 1.2291\n"
     ]
    }
   ],
   "source": [
    "# K means\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics import silhouette_score \n",
    "from sklearn.metrics import calinski_harabasz_score\n",
    "from sklearn.metrics import davies_bouldin_score\n",
    "# Fit K-Means\n",
    "kmeans_1 = KMeans(n_clusters=4, random_state= 10)\n",
    "# Use fit_predict to cluster the dataset\n",
    "predictions = kmeans_1.fit_predict(cluster_df)\n",
    "# Calculate cluster validation metrics\n",
    "score_kemans_s = silhouette_score(cluster_df, kmeans_1.labels_, metric='euclidean')\n",
    "score_kemans_c = calinski_harabasz_score(cluster_df, kmeans_1.labels_)\n",
    "score_kemans_d = davies_bouldin_score(cluster_df, predictions)\n",
    "\n",
    "\n",
    "print('Silhouette Score: %.4f' % score_kemans_s)\n",
    "print('Calinski Harabasz Score: %.4f' % score_kemans_c)\n",
    "print('Davies Bouldin Score: %.4f' % score_kemans_d)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def gap_stat_kmeans(data = cluster_df,nrefs=3,k=4):\n",
    "    # Gap Stat\n",
    "    refDisps = np.zeros(nrefs)\n",
    "    # For n references, generate random sample and perform kmeans getting resulting dispersion of each loop\n",
    "    for i in range(nrefs):\n",
    "\n",
    "        # Create new random reference set\n",
    "        randomReference = np.random.random_sample(size=cluster_df.shape)\n",
    "\n",
    "        # Fit to it\n",
    "        km = KMeans(k, random_state =10)\n",
    "        km.fit(randomReference)\n",
    "\n",
    "        refDisp = km.inertia_\n",
    "        refDisps[i] = refDisp\n",
    "\n",
    "    # Fit cluster to original data and create dispersion\n",
    "    km = KMeans(k)\n",
    "    km.fit(cluster_df)\n",
    "\n",
    "    origDisp = km.inertia_\n",
    "\n",
    "    # Calculate gap statistic\n",
    "    score_g = np.log(np.mean(refDisps)) - np.log(origDisp)\n",
    "    return score_g\n",
    "\n",
    "score_kemans_g = gap_stat_kmeans(cluster_df,3,4)\n",
    "print('Gap Statistic Score: %.4f' % score_kemans_g)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a855be5080>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Inter cluster distance map\n",
    "\n",
    "\n",
    "from yellowbrick.cluster import InterclusterDistance\n",
    "\n",
    "\n",
    "# Instantiate the clustering model and visualizer\n",
    "\n",
    "visualizer = InterclusterDistance(kmeans_1)\n",
    "\n",
    "visualizer.fit(cluster_df)        # Fit the data to the visualizer\n",
    "visualizer.show()        # Finalize and render the figure"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Heirarchical Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x27c388392b0>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Elbow Method for Heirarchical Clustering\n",
    "from sklearn.cluster import MiniBatchKMeans\n",
    "# Import ElbowVisualizer\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "from sklearn.cluster import AgglomerativeClustering\n",
    "\n",
    "# define dataset\n",
    "# define the model\n",
    "\n",
    "\n",
    "model = AgglomerativeClustering()\n",
    "# k is range of number of clusters.\n",
    "visualizer = KElbowVisualizer(model, k=(2,30), timings= True)\n",
    "visualizer.fit(cluster_df)        # Fit the data to the visualizer\n",
    "visualizer.show()        # Finalize and render the figure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh0AAAFlCAYAAABP+VrWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Xl8U2XWB/Dfzc3epHsppYUCLWURClJABJEBFFAWGcFBRdQRXMbXBVxRX0VQEFFn3BnFhXcQgRFxVJhRBpBFBISyla1QCrSF0r1pkma/z/tHyG26pEnbLF3O9/PxI0nucnKTJifPch6OMcZACCGEEBJgklAHQAghhJCOgZIOQgghhAQFJR2EEEIICQpKOgghhBASFJR0EEIIISQoKOkghBBCSFBQ0tEGHDlyBLNnz8aUKVMwefJkzJ07F2fPngUAZGVl4YknngAALFiwAJ9//jkAoHfv3igvLw9KfA888IB4rm+++QZr1qxp8jF27NiBmTNnYurUqZg0aRKefPJJXLlyxd+herVx40ZkZGTgtttuq/Xfc889ByD41/jUqVO46aabcPvtt6OgoKBFxyovL0d6ejoWLlzY4rgKCgpw7bXXtvg4/pSfn4/HH38cAFBUVIQ777yzRcc7d+4cHn/8cUyZMgVTp07FPffcg4MHDwJo+fPfsWMH3nvvvSbv99577+Ff//pXs89bl8Viwbvvvotp06bhtttuw5QpU/Dpp5/CVUlh9uzZ+Omnn5p1bL1ej3vvvbfJ+7l/ppH2RxrqAEjjrFYrHn74YXzxxRe45pprAADff/89HnzwQWzbtg0DBgzA+++/H9IY9+zZI/47MzMTvXr1atL+RUVFeP7557Fx40YkJiYCAFasWIF58+Zh3bp1fo3VF0OGDMEnn3wS9PM2ZNu2bbjuuuuwZMmSFh9rw4YNGDduHDZt2oT58+cjMjLSDxG2HpcvX8b58+cBAPHx8S167+Tm5uK+++7DG2+8gVGjRgEA9u7di0ceeQRr166FSqVqUaxZWVnQ6XRN3u/JJ59s0XndMcbw6KOPokePHli/fj0UCgUqKirw8MMPo7q6GvPmzWvR8XU6HbKyspq8X2v4TCOBQ0lHK2cymaDX61FdXS3eN3XqVGg0GjgcDhw8eBCvvfYaNm3aVG/fDz74AEePHkVlZSXmzJmDWbNmAQA++ugjbN68GTzPo0ePHnj55ZcRFxeH2bNnY9asWZg4cSIA1Lp97tw5LFmyBJWVlXA4HJg9ezZmzJiBF154AQBw3333Yc6cOdi+fTv27NkDpVKJWbNmYcWKFdiyZQsEQUBiYiIWLlyI+Pj4WnFWVFTAZrPVeo733Xcf+vTpI97+5JNP8N1330EqlSI5ORnLli2DVqtt9LlEREQgNzcXd911F6ZNm4YlS5bgzJkzsNlsuP766/Hcc89BKm3Zn8C7776LrKwsCIKAefPmYcyYMR6v8dGjR/HFF1/g66+/BgBMmDABkyZNwhNPPIErV65gxowZ2LVrFyQSZwPkDz/8gLVr18LhcMBsNuOdd97x+fnOnj27VpyCIGD9+vVYuHAhqqur8c9//hMPPfQQAMDhcGD58uXYvn07tFot0tPTce7cOaxevRoXL17Eiy++CJ1Oh7i4ODDGMHXqVAwbNkw8ts1mw7Jly7B3717wPI/09HS88MIL0Gg0GDt2LCZPnox9+/ZBp9Nh7ty5OHToEE6cOAGpVIoVK1YgPj4eRUVFWLx4MQoLC2Gz2TBp0iQ88sgjKCgowKxZs5CSkoJLly5h9erV2LhxI7Zt2waz2QyTyYTnn38eY8eOxf/+7/+iqKgIc+bMwaJFizBlyhQcPHgQY8eOxUcffYT+/fsDAObNm4dhw4bh7rvv9vj+XLlyJaZPny4mHABw/fXX45133oFSqax1bT/44ANUVFTglVdeqXd7y5YtWLFiBTiOA8/zeO655yCXy7Fu3To4HA5otVrMnz8f33zzDdauXQtBEBAZGYmXX34ZKSkpWLBgASorK5Gfn48//OEPKCsrQ69evTBnzhwMGDAADz30EPbs2YPi4mLMnTsXd999d6Ovp7sDBw4gNzcXn376KXieBwBERUVh+fLluHTpUq1tCwoKMGXKFBw+fLje7ZKSEjz//POoqKgAAIwePRrz5s3DCy+8ALPZjNtuuw0bN27EhQsXGvwM2b9/P5YsWQK1Wg2j0YjnnnsOb775JjZt2oQFCxZAo9EgOzsbV65cQe/evfHmm28iLCwMO3fuxNtvvw2JRIK+ffvit99+w9dff42kpKQm/Q2TIGOk1fviiy9Yeno6Gzt2LHvmmWfYN998w6qrqxljjO3bt49NmjSJMcbY888/zz777DPGGGNpaWns888/Z4wxduLECda/f39mtVrZhg0b2MyZM5nRaGSMMfb++++zBx54gDHG2D333MP+85//iOd13bbZbOzWW29lx48fZ4wxVlVVxW655RZ2+PBh8VxlZWX1Yvjuu+/YvHnzmM1mY4wxtm7dOjZ37twGn+Mbb7zBrrnmGnbLLbewl156iW3atEncb+vWrWz8+PGssrKSMcbY0qVL2ccff+z1ubzwwgvi8RcsWMD+8Y9/MMYYs9vt7JlnnmGffvppvTi+/fZbNnjwYDZ16tRa/23YsKHBa/zJJ58wxhjLzs5mw4YNY2VlZR7jMplMbPDgwUyn07H8/Hw2cuRINnPmTMYYY1999RVbuHBhvXjef/99tmjRIsYYa9LzrWvHjh1sxIgRzGazsX//+99s1KhR4vVdu3YtmzVrFjObzcxisbAHHniA3XPPPYwxxv70pz+xNWvWMMYYy8nJYQMHDmTffvsty8/PZ4MGDWKMMfbee++xxx57jFmtVuZwONiCBQvYyy+/zBhjbMyYMWzp0qWMMcY2b97M+vTpw06dOsUYY+zRRx9lK1asYIwxNnv2bLZt2zbGGGNms5nNnj2bbd68meXn57O0tDR24MABxhhjBQUFbPbs2cxkMjHGGNu0aRObPHkyY6z230Ld+FzXsLKykg0bNoxVVVU1+v6cPHky27Fjh8fr6X5899eo7u1x48aJfye7d+9mH3zwQb1t9u/fz+6++27xb3r37t1s4sSJjDHn++2+++4Tj133/bd69WrGGGNZWVmsf//+zGw2N/p6uvv888/ZE0884fE5MlbzGeD+fOs+/w8//FB8vY1GI5s3bx6rqqqqtU1jnyH79u1jffr0YQUFBYyx+p9pM2fOZBaLhVmtVjZt2jS2YcMGVl5ezoYNGya+lzZu3MjS0tJYfn5+o8+HhB61dLQBf/7zn3HHHXfgwIEDOHDgAFauXImVK1diw4YNje43efJkAEDfvn1htVphMBiwa9cu3H777VCr1QCAe++9F3//+99htVo9HufChQvIy8vDiy++KN5nNptx8uRJDBo0yON+v/zyC7KysjB9+nQAzl/bJpOpwW0XLFiAhx9+GL///jsOHDiA5cuXY/Xq1VizZg327t2LiRMnIiIiAgDE1pUnn3yy0ecyZMgQ8fg7duxAVlaWeM3MZrPHuJvSvXLXXXcBANLS0pCSkoLDhw97vMYSiQQjRozAnj17UFFRgZkzZ2L9+vXQ6/XYvn075s6d2+i5vL127s+3rrVr12LKlCmQSqUYN24cFi5ciJ9++gmTJ0/Gzp07cdttt0GhUAAAZs6cidWrV0On0+HYsWP46quvAAApKSkYPnx4g3HNnz8fMpkMgLOF7H/+53/Ex8ePHw8A6Nq1K2JjY8UWrG7dukGn06G6uhoHDhyATqcTxzlUV1fj9OnTSE9Ph1QqFd9niYmJWL58OX788UdcvHgRR48ehdFobPS6TZ8+HTNmzMCCBQuwadMmjB07FlqtttH3J8dxEASh0eP6YtKkSXjssccwevRojBw5Eg8++GC9bXbs2IGLFy/WGoNSVVWFyspKAEBGRobH448bNw4AcM0118BqtaK6utrj61mXRCIRx260xKhRo/DQQw+hsLAQI0aMwNNPPw2tVlur+6ixz5CUlBQkJCSIXasNHV8ulwNw/p3pdDocPHgQKSkp4nvpj3/8I15//fUWPxcSeJR0tHKZmZk4fPgw5s6dizFjxmDMmDF46qmnMHnyZOzZswdRUVEe93V1HXAcB8DZhysIgngbcH7Q2u128bb7h5DNZgMAsRn4+++/Fx8rLS2FVqttNHZBEMQmX8A5PqWhfuxt27ahsrIS06dPx4QJEzBhwgTMnz8fo0ePxsmTJ8HzfK2Yq6qqUFVV5fW5uL6cXY+99957SElJEY/hvm9zubpCXOeQSqWNxnXTTTdh165dqKqqwty5c5Gbm4utW7fizJkztbosGtKU5+vu0qVL2LlzJ06cOIEtW7YAAOx2O1atWoXJkyfX62JyPSdXk7v7e8J1n7e4XO8dAOIXBgAxMam7P2MM69atE8dKlJeXi2MM5HK5GOOJEyfw6KOP4v7778fIkSMxdOhQLFq0qMHn7ZKYmIh+/fphx44d2Lhxo/jF19j7c9CgQThy5IjYXeby4Ycfolu3bhg8eLB4H8dxDf7dAMD8+fMxffp07NmzBxs3bsQXX3xR78eCIAi47bbb8Oyzz4q3i4uLxSTb0+sKQEws3P/GPb2edQ0cOBD/93//B4fDUet1PXbsGFavXo233nrLp+eYnp6Obdu2Ye/evdi3bx/uuOMOrFy5staYocY+Q44cOdLoc3TvznLFwfN8vYTJ0/MkrQu9Sq1cdHQ0VqxYIY6aB4CSkhIYDAakpaU1+XijRo3Ct99+K46fWL16NYYOHQq5XI7o6GgcP34cAJCTk4Ps7GwAQI8ePaBUKsUPjMLCQkyePFnclud58cvP/d833HADNmzYAIPBAMA58t41C8RdWFgY/vrXvyInJ0e8Lz8/HzzPo1u3bhgxYgT++9//isf54IMPsGrVqkafS1033HADVq1aBcYYrFYr/vKXv4i/4Fviu+++A+D8MszLy8PAgQMbjWvs2LHYu3cvTp06hfT0dIwcORLvvfcebrzxxga/0N015fm6W79+PTIyMrB7925s374d27dvx8aNG3Hy5EkcOnQIo0ePxg8//ACr1Qq73S4+J41Gg8GDB2Pjxo0AnK/J3r176yVro0aNwtq1a2Gz2SAIAtasWYORI0f6fA01Gg0GDRqEL7/8EoAzIbzrrruwbdu2etseOHAA/fv3x5///GcMGzYM27Ztg8PhAOB877l/Gbr705/+hJUrV8JkMoktB429P+fMmYNvvvkGv/76q3iMXbt2YfXq1bXGGgHOcRAnTpwAYwwGgwG//PILAGdiN3bsWJhMJtx1111YuHAhsrOzYbVa6/2dbN68GcXFxQCcrVL33Xefz9evLk+vZ13XXnstevbsiTfeeAMWiwWAMxF4/fXX642LCA8Ph81mE/9GN2/eLD729ttv4+OPP8ZNN92El156CampqTh79iykUikcDgcYY14/Q5pq8ODBuHDhAk6fPg0A+Pnnn/32Q4IEFrV0tHI9evTARx99hL/97W+4cuUKFAoFtFotli5dip49e6KkpKRJx5sxYwYKCwtxxx13QBAEJCcn4+233wYA/OUvf8GCBQuwc+dO9OzZU2yul8vl+Pjjj7FkyRJ89tlnsNvtePLJJ8UP74kTJ2L27Nn44IMPcOONN2LZsmUAgAcffBBFRUX405/+BI7jkJCQID7mbvjw4Xj55Zfx/PPPQ6/Xg+d5xMXFYeXKlYiIiMDo0aORk5MjdmWkpqbitddeg1qt9vhc6nrppZewZMkSTJkyBTabDSNGjPDYnXHw4EHcdtttte7jeV788nWXn5+PadOmgeM4/PWvf0VkZGSj11ir1SIlJQUqlQo8z2PUqFF46aWXxC6I5r52nlitVmzYsAFLly6tdX/37t0xadIkrFq1Cu+++y7Onz+PadOmQa1WIykpSWxxePPNN/HSSy/h66+/Rnx8PJKSkuoNpPzLX/6CN998E9OmTYPdbkd6ejpefvllr8/H3dtvv43XXnsNU6ZMgdVqxeTJkzF16tR604QnT56MLVu24JZbboEgCBgzZgx0Oh0MBgNSU1OhUCgwY8YM/O1vf6u139ixY7Fo0aJa3Rt33HGHx/dncnIy/v73v+Pdd9/Fm2++CUEQxB8AaWlpteKaOnUqdu/ejfHjxyM+Ph7Dhg0TWxxefPFFPPPMM5BKpeA4DkuXLoVcLsfw4cPxzDPP4LXXXsPLL7+MBx98EA888AA4joNGo8GHH37Y7C/Q22+/3ePrWdf777+Pv/3tb7j99tvB8zwEQcC0adMwZ86cWttptVo8++yzePDBBxEdHS0ONgecg74XLFiAyZMnQy6Xo3fv3pg0aZI4qHjSpElYs2aNx8+Q/fv3N/k5RkZG4q9//Suef/55SCQS9O/fH1KptMWzikjgccwfnXqEkDbr119/RVlZmZhovf7661AoFHj22WexYsUKjB8/HikpKdDr9Zg6dSpWrlyJ1NTUEEdNPGns9WwvDAYDPv74Yzz++ONQqVQ4ceIEHn74YezevZtaO1o5aukgpIPr1asXPv/8c3z22WcQBAF9+vTBq6++CsDZIjJ//nxIJBI4HA48+OCDlHC0co29nu2FRqOBTCbDjBkzIJVKIZVK8e6771LC0QZQSwchhBBCgoIGkhJCCCEkKCjpIIQQQkhQtOkxHXa7HWVlZVAqlTRHmxBCSLsnCALMZjNiYmJavIxDKLS9iN2UlZW1eOVNQgghpC2qu45VW9Cmkw5XvYCkpKRGK9q1d2fOnGlWobCOiK6V7+haNQ1dL9/RtfJd3WtVXV2NgoKCevVy2oo2nXS4ulTUarXXktztXUd//k1B18p3dK2ahq6X7+ha+a6ha9VWhxS0zagJIYQQ0uZQ0kEIIYSQoKCkgxBCCCFBQUkHIYQQQoKCkg5CCCGEBAUlHX6y6OejWPTz0VCHQQghhLRabXrKbGux6OejWLzlmHh74YSBIYyGEEIIaZ0o6WihugmH69+UeBDSdhRW5iC35AgM5gpolFHoGTcICZGpoQ6LkHaHko4WqJtwuFDiQUjbUViZg6P528XbenO5eJsSD0L8i8Z0NJOnhMNl8ZZjNMaDkDYgt+QIGGOw2E1gjNW6nxDiX5R0BFBumR4Wu8PrdjQIlZDQMZgrYLIZYDRXwGwzuN1fGcKoCGmfqHulmVxdJ421dnyVeR4bs/IwOqUzJvROwPjeXZAWFw6O48RtaBAqIaGlUUbBZNMDAHiJzO3+yFCFREi7RUlHC3hKPF4Y1x839OyE/2YXYkv2Zfzn1CX859QlAEByVBhuvpqAHMwrw/JfToj70VgQQoKvZ9wgFFVdBIBaPwh6xg0KVUiEtFuUdLTQHQOTayUdr4xPF5OGiX0SAQD5FUZsOXMZW7ILse1MIT7bl4PP9uU0eDxKPAgJroTIVHQO74H88lMQmACtMppmrxASIJR0tNBn+88CAGakd0O/zpENJgtdo8Iw57pemHNdLzgEAY98sw9f/H7O4zEp8SAkuGS8HBHqOESqO2F4yrRQh0NIu0VJRwuYbQ7840AuOmmUWD3rBsilvNd9eIkESZFhQYiOEOKzq90qbpNXCGmTGBOw99z3qDAWQsLxGNlrOsJVsfW22XpyFbpG90OfhOHi/ZXVxdh89CPMvO5/IXUb3+RPNHulBTYcu4gKkxV/HpbiU8LhsnDCQLwyPt3j4+5dNISQwHON32BMCHEkhLRMXtlJOAQbJg18FBndb8GB85vrbXPo4hZYbKZa91ntZhw8vxm8JLBtEZR0tMBn+5xdK3Ou69XkfT0lHpRwEBJ8MZpEKGVhsDksoQ6FkBYpqrqAxKjeAIBO4d1QZrhU6/ELpVngOA6JUWnifYwx/JazEYOTJ9SawRUIlHQ006kiHXbnFmNcr85IidU26xh1E49p/btSwkFIiMh4BSUdpM2zOcyQ80rxNsdxEJizXlSF8QpyS47g2m4319rnSN5WdI3ug2hNl4DHR2M6msnVyvHQ9WletmzcwgkDcaakCusOX8B1ybHedyCE+N3uM/+E0eIsBiYwARKOfo+RtknGK2slz4wxSDhn9/+54kOotlThp6yVMFgqwHNSaJRRyC05ArU8HGeuHIDJasB/j3+OW9IfCUh8lHQ0g9nmwD8OnkMnjRJTr0lq8fGeHXMN1h2+gILKaj9ERwhpCodgh9FSCbU8HEN63AoOnPedCGmlOoUnI7/8FHrEpaO4Kg9RYZ3Fx4b0uFX89+GL/4VKrkVSVG8kDXlWvP+bA8twc/85AYuP0vlm+PbYRZRXW3H/0KYNIPUkKUINALhURUkHIcHm+lUYroqFWl67YjAhbU1yzDXgJTJsPvoxDpzfhKE9JuPEpd3IKzsZ6tAAUEtHs4gDSIf7p3hQTJgCCqkEl6ilg5CgcyUdzjEdVvAcD4mk5T8mCAkFjpNgROofa90Xqe5Ub7trk2+udx8A3DF0QUDicqGWjiY6VaTDrqsDSFNjw/1yTI7jkBihRoGOkg5Cgs3usAIA8stPYdvJVSgzXg5xRIS0X5R0NJGrlePBFg4grSspQo0rehNsDqoTQEgwuVo61PLwWrcJIf5HSUcTuAaQxmkUuM0PA0jddYlQgzHgSpXJ+8aEEL+RS1XoEtkL0WEJAACb3RziiAhpvyjpaIKNWXlXB5Cm+mUAqTvXYFLqYiEkuCLVnZDedQw6R6YAoJYOQgKJko4mWLn3DABgrp8GkLpLiqSkg5BQchVUsl0d40EI8T+aveKj0wEYQOouMcK5CNxlSjoICarLlTnQVRejc4SrpYO6VwgJFGrp8JFrCfu5w5u+zoovEiNUAEAFwggJslJ9Pi6WHYeMl6Nvwohaa1IQQvyLWjp8YLY58H8HnANIp/XvGpBzuJa7L9AZA3J8QkjDXFNmFTI1kmP7hzgaQto3aunwQSAHkLp01irBSzgqEEZIkLkGjkol8hBHQkj7R0mHD2qWsPf/AFIXXiJBglZFpdAJCTKbwwIZrwDHcTiatw37z/0Q6pAIabco6fAiu1iHneeKMDa1M3rF+X8AqbvECDUu6UwQBBbQ8xBCatgcVkh5ZytHtbUKlaZiMEZ/g4QEAiUdXqwUK5AGZgCpu8RINWwOASVGGj1PSLDIpQqoZFoAzmXBGRMgMEeIoyKkfQrYQFJBEPDqq68iOzsbcrkcr7/+OpKTk8XHV61ahc2bNwMARo8ejcceewx6vR7PPvssDAYDbDYbFixYgGuvvTZQIXpltjnwjwO5AR1A6k4sEFZZjXitKuDnI4QAI3vNEP8tu9riYXOYwUs0oQqJkHYrYC0dW7duhdVqxfr16/H0009j2bJl4mP5+fn44YcfsG7dOqxfvx6//vorTp8+jS+//BLDhw/HV199hTfeeAOLFy8OVHg++S4rD2XVFtw3xD9L2HuT6Frinmp1EBISMl4BgKqSEhIoAWvpyMzMxKhRowAAgwYNwvHjx8XHOnfujM8++ww87/wit9vtUCgUuP/++yGXO39pOBwOKBSKQIXnE9cA0kDV5qiLkg5Cgstmt6DUUACtMhoaZRRkUqV4PyHE/wKWdBgMBmg0Nc2TPM/DbrdDKpVCJpMhOjoajDEsX74c/fr1Q48ePcRtS0pK8Oyzz+LFF1/06Vxnzpzxe/wXqyzYca4IQ+LVqMo7i8w8v5+iHmOxs0bHwdPnkKk0NGnfzMzMQITULtG18l17v1ZmQYfLtsOI4LsiRpoCo6MUgqDC6VNnoZAUNvl47f16+RNdK9+1p2sVsKRDo9HAaKwpdCUIAqTSmtNZLBa8+OKLCAsLw8KFC8X7s7Oz8dRTT+G5557DsGHDfDpXWloatFqt/4IHsO4H54v81M0ZyLi2u1+P7Ul0mR7YehEOdQQyMjJ83i8zM7NJ23dkdK181xGuVYk+D/oLOUiNT0PPToNadKyOcL38ha6V7+peK71eH5Af2sESsDEdgwcPxq5duwAAR44cQVpaTWlhxhgeffRR9O7dG4sXLxa7WXJycvDkk0/inXfewejRowMVmlcWu7MCaWyYAtMGBH4AqUsXV/cKFQgjJChci7u5BpASQgIrYC0dN998M/bs2YM777wTjDEsXboUX375Jbp16wZBEPD777/DarVi9+7dAICnnnoKn376KaxWK5YsWQLA2VqyYsWKQIXo0cZjzgGkz/yhHxRBGEDqopDyiNMoaKVZQoJErEZ6dQCp2WbAxdITiFR3QnxEj8Z2JYQ0Q8CSDolEUm/2SUpKivjvrKysevuEIsFoiFiBNEgDSN0lRYQhu0QHxhg4jgv6+QnpSOxXkw7XrBW7w4bzpUfRNbovJR2EBAAt+OZm0c9HUWZ0DiAdkxqPtABXIG1IYoQahy+VQ2e2IVJFTb6EBFJNS4e81v9pyiwhgUFJx1WLfj6KxVuOibeDNU22rkSxQJiRkg5CAiyl02AkRfWFSu6caVdTp8MayrAIabeoDDrqJxwAcOJKZUhiSYq8mnTQuA5CAk7GK6BRRoKXOH9/8RIpJJwUNgctRUBIIHT4pKOhhAMAlm49jkU/Hw16PFQgjJDgsdrNsAu2WvfJeAV1rxASIB066fCUcLgs3nIs6IlHIk2bJSRoDpzfhB2nvqp1n1IWBp6jnmdCAoH+sloZcdE3aukgJOCcy9rXXm7h+tRpIYqGkPavQ7d0LJwwEK+MT/f4+Cvj07FwwsAgRuQ2kJSSDkICzu6wUGEwQoKoQycdgOfEIxQJBwBolTJEKGW4TEkHIQElMAF2wSbOWHExWnQoqrpAM1gICYAOn3QA9ROPUCUcLokRahTQmA5CAsp+Namo272SX34Khy9ugdESmhlshLRnNKbjKvckI5QJB+BMOk4W6WC02BCmkIU0FkLaK1udaqQuNdVJaQYLIf5GSYebUCcbLq5aHZeqTEiLo6SDkECQ80r0TxwNlbz2CtWupMNKSQchfkfdK61QUkQYAKrVQUggyaQKJEX3RoymS+37xaqklHQQ4m+UdLRCXSJUAEDjOggJAepeISRwKOlohZIiXS0dxhBHQkj7dbkyB3vObkCpoaDW/dTSQUjg0JiOVihJLIVuCnEkhLRfJqseenM5GBNq3a9RRmFE6nQoZWEhioyQ9ouSjlbIfaVZQkg+VTA+AAAgAElEQVRgiFNmJbVnr/ASKcJVMaEIiZB2j7pXWqFotRxKKU8DSQkJIE9TZgFn4TAqDkaI/1HS0QpxHIekSDWVQickgMSkQ1q/DPovp1Zj/7nvgx0SIe0eJR2tVFKEGsUGM6x2R6hDIaRdsjfS0kHL2xMSGJR0tFJdItRgDCisosGkhARCVFgCOkf0hITj6z1GSQchgUEDSa8qrMxBbskRGMwV0Cij0DNuEBIiU0MWj/sS98nRmpDFQUh7lRqf4fExGa+AwBxwCHbwEvqYJMRf6K8JzoTjaP528bbeXC7eDlXi4SqFTgXCCAk+91odlHQQ4j/01wQgt+SIx/tDlXS4ps1erqKkgxB/Y4zh1OU90Kpi0TW6T73HpW5JB9XrIG0JYwL2nvseFcZCSDgeI3tNR7gqtt42W0+uQtfofuiTMBxWuxm7stfB5rBAYA4M7TEJncKTAxIfjekAYDBXAACMFp34b+f9oVvauqZWByUdhPibXbAir/wkSqouNvh4QkRP9OsyEnKpKsiREdIyeWUn4RBsmDTwUWR0vwUHzm+ut82hi1tgsdWMFzxxaTcSIlNxS/rDuKHXHdgXwJlb1NIBZwVCvbkcdsEKh2B3uz8yZDGJ3Ss0bZYQv3MNEpXy9afLAkC0pgui6ywER0hbUFR1AYlRvQEAncK7ocxwqdbjF0qzwHEcEqPSxPv6Jd4AnnOmAwITAtqlSC0dAHrGDQIAcJAAjIExVuv+UOikUUIq4XCZkg5C/M5VjbSh6bKEtGU2hxlyXine5jgOAnOWXqgwXkFuyRFc2+3mWvsopCpIeRmqrXrsPrMOGd0nBiw+aulAzWDRA+f/DYPDCrUiHGnxQ0M6e4WXSJAQrqKWDkICwFtLR2V1EU5d3ouu0X2RFN07mKER0iIyXllrujdjTJwWfq74EKotVfgpayUMlgrwnBQaZRSSonqjwngFO7O/xpDuk9A5omfA4qOk46qEyFT0is/AxbITuLbbzfUG3oRCUkQYDuSXwiEI4CXUKEWIvzRWAh1wNjHrTMWIsSYGMyxCWqxTeDLyy0+hR1w6iqvyEBXWWXxsSI9bxX8fvvhfqORaJEX1RmV1EX45vQZ/6H1XwLsVKelwI7vaJNVaigIlRqqx9yJDscGMhHB1qMMhpN1gjEHOKz0OFKXl7UlblRxzDS5X5mDz0Y8BACN7zcCJS7uhVcagW0y/BvfJvPATHIIN+3N/BADIpUqM63dfQOKjpMNNmCIS0WEJkHCt47K4L3FPSQch/pMQmYKEyBSPj9ckHeZghUSIX3CcBCNS/1jrvkh1p3rbXZtcM64jUAlGQ1rHt2sr4e2DKNjcl7gf0pWW2iYkWGqSDlpplhB/ooECrVii2NJBg0kJ8afK6mKU6PNqTZF3x0ukkHBS6l4hxM8o6XBjs1twofQYij0UDAo2KoVOSGBcLM1C5oWfGk0qOkf0QHRYQhCjIqT9o+4VN3bBhtOF+5AQkRKwErBNIbZ0UCl0QvzK2+wVAEjvOiZY4RDSYVBLhxu5tHXNXukS7hxZf4laOgjxK5vDAo6TNLisPSEkcCjpcOPsx+VhbSVJh1zKI16rpAJhhPiZzWGFjFeA4ziP2xRWnsPpwr1i9VJCSMtR0lGHXKqEzd56psklRahxSVctlmYnhLSczWHxWgK91JCPC6VZsLaizwNC2jpKOuqQ8YpWNTe/S4QaJpsDFSb6tUWIPzDGYPch6XA9bm1FnweEtHU0kLQOGa+EkekgMEer6O9NclviPlpNi1MR4g+j0maCofHWQ1eFYupeIcR/KOmoY0j3W8Bxkkb7eoPJfYn79C5RIY6GkLaP4zioFeFet5NdXQyuNbV8EtLWUfdKHRIJ32oSDgBIjAgDQAXCCPEXgTlgc1i9jpOi9VcI8T9KOuqw2s2oMBa1msFjiRE0bZYQf6o0FmHbyVXIKTrY6HYyXgmpRE6DuAnxI0o66iiszMH+3O9Rbrwc6lAAAEmRzpaOAp0xxJEQ0j641lORSRsfIxWrTcJN19yP5Nj+wQiLkA4hYGM6BEHAq6++iuzsbMjlcrz++utITq6p8rlq1Sps3rwZADB69Gg89thjMJvNePbZZ1FWVoawsDC8+eabiI6ODlSIDXJ9EAWrSbWwMge5JUdgMFdAo4xCz7hBSIhMFR9PdBUI05mCEg+pz9trRNoW19+2VEIDswkJtoC1dGzduhVWqxXr16/H008/jWXLlomP5efn44cffsC6deuwfv16/Prrrzh9+jTWrl2LtLQ0fP3115g2bRo+/vjjQIXnkWvEejC6Vworc3A0fzv05nIwMOjN5Tiavx2FlTniNmEKGSJVclyilo6Q8OU1Im2LXfBeAh0AGBNQZriECmNRMMIipEMIWNKRmZmJUaNGAQAGDRqE48ePi4917twZn332GXieh0Qigd1uh0KhqLXPjTfeiL179wYqPI/kvKsUeuCTjtySIwAAi90Es81Y736XpAg1LfoWIq7Xwu6wotpSBdcsy7qvEWk7bHZX0iH3uu2B85tx5sr+QIdESIcRsO4Vg8EAjUYj3uZ5Hna7HVKpFDKZDNHR0WCMYfny5ejXrx969OgBg8EArVYLAAgLC4Ner/fpXGfOnPFb3DZmgt6qx3njORgve/9QaonLljwADCah0nluiQAA0OsNyKzKFLfTcjbozDbs3ncAalnDeWJmZmaD95P6mnKt6r1GFgcknLTea9Retcf3Van9HPQOPbJPnYVcUtjotkaLCRbDZWRW+nYd2uP1ChS6Vr5rT9cqYEmHRqOB0Vjz610QBEilNaezWCx48cUXERYWhoULF9bbx2g0Ijzc+1x6AEhLSxOTlZayO6yoPHkScdpYZHTP8MsxPTGfPQ+9uRw2gzO50oRpwHEctMpoZPSqOXe/c1bsLcxBp55p6N0pot5xMjMzkZER2Fjbi6Zeq7qvkVqlhoxX1HuN2qP2+r6qMCZCZypBYmSa18GkhuwzEAQHMvp6vw7t9XoFAl0r39W9Vnq93q8/tIMtYN0rgwcPxq5duwAAR44cQVpamvgYYwyPPvooevfujcWLF4PneXGfnTt3AgB27doVkjclL5FhSPdb0bvz8ICfq2fcoFrT8RgTxPvdJbpVJSXB5f4aSXm5OA6g7mtE2o6osM7oHjvAa8IBOMd42agiKSF+E7CWjptvvhl79uzBnXfeCcYYli5dii+//BLdunWDIAj4/fffYbVasXv3bgDAU089hbvuugvPP/887rrrLshkMrzzzjuBCs8jjuMQq00KyrkSIlNhd9hw8MJ/ADBEqOManBkhJh1UICzoEiJTUW2twsHz/wEvkUGrjKbZKx2IjFdAYHY4BDt4CRVwJqSlAvZXJJFIsHjx4lr3paSkiP/OyspqcL/3338/UCE1CWMsKJVJu8b0RdeYvo1u4yqFTlVJQyNGk4RuMf2glocjMbo3YjXBSUpJYBy68DMAYHD3CV63rSmFbqGkgxA/oOJgDTh04WdsOf45hKvdHYFmd1hRWV1cawaLO9eib5R0hEakuhOGp0xDsT4P2YX7Qh0OaSGdqRgGS6VP26bFD8MNve6AXKoMcFSEdAyUujeA4zgwCLA5LFBIVQE9l666BCcv74HOVIw+Cdeje+yAetvQmI7Q4zgOarkWJqs+aK1gJDBsDiuUMo33DQGfFoYjhPiOko4GuAqE2ezmgCcdZcbL0JmKAXguSBapkkMt56mlI0Ryi49AKddAJdNCby6H3WH1aRBiKFEV1YY5BDsE5oDUS2Gwuvtw4CCR8AGMjJCOgbpXGuBqSg1GKXSrvaa8uaeCZBzHISkijJKOEBCYA2eLDiKv7DhUcuevY5PNt/oxoUJVVD1z/U37UhgMAC5VnMF/T3yByzq6doT4AyUdDXAlHdYgVCV1Tzrc/11XYoQKxQYzLHZHwGMiNaotVWAQEKaIgkrurAVjsrbupMNTtVSqouocPwV4L4HuIpXIr+5Hy9sT4g+UdDTAvXsl0Ny7VBpb7yUxwrna7GVq7Qgqg6UCAKBRREElu5p0tPKWDoPZGbPVbobVbhbrjBjMvg2ebM8kHI/OET0Roerk0/biApB2SjoI8Qca09GAcFUsUjoNhlYVE/BzWe0m8BIpJBzfaMuKa9psga4aPWL8U32VeOf6AtcoIxGl7oxRaTOhlIWFOKrGaZRR0JvLYbLp4RDsiFJ3vnp/ZIgjCz21IhyDut3k8/aubhgrtXQQ4heUdDRAq4yGVhkdlHPZBRvkUhX6J45utJ+Zps2GhvHq1EqNIgpSXg6pj2MBQqln3CAcydsGh8MGABCYAJ7jqYpqM7haPal7hRD/oKQjxG7sPdOnaoddXEkHTZsNKodgh1QiE6dYOgQ7LPZqqOWtdyplQmQqDJZKHLrws3PmBcdhYNexNHsFQJnhEkr1BUiK7o0whfeWH9fYj2AMKiekI6CkowEOwY6jedsQpoxE787XBfx8roTD1ffeUA2IJCqFHhKDu0+AwBzia7Lv3L9QbdXjpn73t+paHXJeiQh1HACga3RfSjiuqjBewfnSo4jVJvmUdPASKdI6X4cwRf2FFgkhTUdJRwMkHI9i/UVEBnj2it1hRZW5DGp5OC6UZuFi6XGM7DWjwb539zEdJLgkXE19BpXcWavD5rC06iqVOlOJ+O/GZkV1NK7F25rSTdYzbmCgwiGkw6HZKw3gOA5yXhnwJlW9uRy/5/6IvLITkHCSq1VQG0504sKUkPESmr0SREZLJYp052Fx+9IWZ7C08mmzZpsRUokMAGrF39HV1Olo3cXdCGmvKOnwQCZVNjqF1R9cv0DlUmVNbRAP55RIOHQJV1Ep9CC6ojuPw3n/ha66WLxPrNXRyqfNDus5GaP73A2FVEUtHW5cA0JlEt+TjtOFe/Fbzkax+5MQ4uyqvFh6HBfLTqDCeMXn/ah7xQMZr4DRUhnQdTZqkg4VcPXzzOrw/AWRFKHGvrxSOAQBvITyxUBzr9Hh0lYKhAHO97BcqmoTsQaLq6WjKd0r1ZYqVJlKYRes1EJCOjTGGLKv7MfJy79CxisQpoiEhJPAYK6A1WFBvy4j0bvzMHCc5+8nSjo8cLU8BHKdDVezt5xXgV3NOhotEBaphuMCQ5HeLM5mIYFjNFdCwknF8udA2+heqTAWQWB2RKrjcU3ijeDQege8BpuEk0AhVTfph4TUbQYLJR2kI9tx+iskRPbCpIH/U29dMqvdjJziTGw/tRrj+t3n8Rg+JR0//vgjcnJy8Mgjj+Dnn3/GtGnTWhZ5GxCujIVDcEBggSs77ioGJpeqwJgAwPP6K4DbarO6ako6AowxAQZLBTSKqFpZu1oRjgFJf0B4EArHNdf5kiMo1l/EH/rMQqTat8qbHcXQnpObvI9cStNmCQGAG9JmeqwnJZcq0a/LSPSKH9roMby20b/99tvYuXMntmzZAofDgW+//RbLli1rXsRtSGp8Bob2uBUKWeC+3N3HdKjkWvSMuxaxmiSP2yfREvdBY7IZIDAHwurMJJJKZEiMSoNW2TqTDsYYKk3FUMrCxMqpjDExqSVNJ9bqoFLopINzJRwWWzUuV54FABzL/wW/nFqDKlNZrW088Zp0/Prrr3jrrbegUCig0Wjw5ZdfYteuXS2NnQBIix+GoT0mQS5VQS5VIq3zUMRqu3rcPjHS+SVySWcMVogdlnsl0rbEbDPCajchQuWs0XGx9Di2HP8MJfqCEEcWegITUFx1EXpzWZP2k1KBMEJq2Zm9FuWGQlyuPIsLpVnoFtMXv+V869O+XpMOydUBi64+UKvVKt7XnhktlcgtPgJddYn3jZtJrQhHjCYRkkYG3birKYVOsxECLU7bDWP6zkbX6L71HjtduBdbT6xqlVNRdSbnTJsIdTwAgJfIwMBoBgucScOhiz8jp+hQk/bTKCLRSZvcquuyEBJMVrsJ/ZNuRF7ZSaTGZyCl02Cfk3Kv33YTJ07EvHnzoNPpsGrVKtxzzz2YPLnp/aJtjcFSiTNFv6PcWBiwcwh1mrxPXNqNI3lbPW4vjumopJaOYFBcbYFqiF2wtsrBpK4k2dXSIb862KuxWVEdhb2ZNTpiNIkY3H0CYjSJgQiLkDaHgaHUUIC8spPoGt0HZYbL9b7PPPE6kHTOnDn47bff0KVLFxQWFuLxxx/HmDFjWhx0ayd3LW8foKqkAhOw5fhniNV2xZDutwBwfmEYLJ6XH08IV4HjaNG3YNCZSqCUaeqN0AZqz2BpbQM19ZZyAECEKhYA3Oq/UNJBhcEI8Y+M7rfg4Pl/45rEUdAqY7Dp6EcY1mOST/t6TTpmzJiB7777DqNGjWpxoG2Ja3XJQBUIcyUzPFfzEsikSghmu8cF4GS8BJ21KiqFHmCMMfye+yPU8nCM7DWj3uM1BcIMwQ7Nq4zkiTDZDGIdCoXU2ToW6EJ3bUFzanQ497PiQukxhCki0CWyVyBCI6RN6RKZii5u6zlNHvg/Pu/rNemIjY3FwYMHkZ6eDrm89S/r7S8ycZpcYD6saxUGu8q9Kql7bQh3SRFqHCusCGjRso7ObDPAIdg9DiJ1vTYma1Uww/IJx3FQX02KgJr3VGscfxJsrnVXmtrSwZiAc8WH0Cm8OyUdpNVjTMDec9+jwlgICcdjZK/pCL/a8um+zdaTq9A1uh/6JAyH3WHD7jPrYLIZIeMVGJV2h7iytrtVv75Qq+oPx/GQcBwcgh0yXoG7r3/Va3xek46srCzcc889te7jOA6nTp3yevC2zPXBZA3QiHXXL0/35nv3Lh0VGk46ukSocSC/DGVGC2I1NLAtEFyVSD2tQlrTvdK6WjoM5go4BDu0qmhxkTpeIkVqpwxaJRXNH9PhmgJIs1dIW5BXdhIOwYZJAx9FcVUeDpzfXK9Y16GLW2Cx1fwQyb6yD5HqzhiTfDNyS47iaN52XJcytd6x77/hDQDA3pzv0Cm8O3rGDQLHcbhQmoVLFWd8is9r0rFv3z6fDtTeSDgJZLwCDsEekONb3Gp0uPjS/+6+xD0lHYFhMF+dLqtsuKVDysuRHDMAWg+Ph8qF0iwUVJzGiNTba/2ySY3PCGFUrUdCZCoi1fFi0ugrjpNAKpGLSQshrVlR1QUkRvUGAHQK74Yyw6Vaj18ozQLHcUiMSqu1T//E0QCApKjeOJa/rdFzlOjzcX3qH8Xb3WMH4Fj+dp/i85p0mEwmfPjhh9i7dy8cDgeGDx+OJ598Emp1+6+IOabvPbWWNfenhrpXwhSRiA7rAkkD4zlcXEvcX9JVY1BidEBi6+gaWnOlrr5drg9WOD7TmUog4XiPyVJHJ+MVkKmaN4hUJlXASsXBSBtgc5jFVnPA2TMhMAckHI8K4xXklhzBmD6zcCSvJrGw2S3ij14ZL/c6BkzKy3G26CC6x6YDjOFcySFx/Jg3XpOOxYsXQ6VSYenSpQCAf/7zn1i4cCHeeustn07QlgUq4QCASHU8Ujtl1Kps2TmiJzpH9Gx0vy5uLR0kMIyWCnCQQK0ID3UoPnMIdhjM5YhQd6r3vs0tOYLiqovISJ4YsHWE2gK7YIOE432ui+NOxitgsHmeWUZIayHjlbW6Ahlj4mfCueJDqLZU4aeslTBYKsBzUmiUUZBJFeI+Noe11o/hhtyYNhP7zn2P/bk/gAOHLpGpGJU206f4vCYdJ06cwA8//CDefuWVV3Drrbf6dPC2zmTVw2QzIFLVCRKJfxOQSHWnZk23FAuEUSn0gBmQNAYmq77RpLNIdx755aeR1nlYq1iHpcpUCgaGCFX991S1pQqV1UWw2Ks7dNKRlf8LiqouYGzfe5tc6EsuVUEqMYi/GAlprTqFJyO//BR6xKWjuCoPUWGdxceG9Kj57j588b9QybVIiuqNSmMRCspPI07bFQUV2YgP797oOTTKKNx0zf3Nis9r0sEYQ1VVFcLDnb/6qqqqwPMd44/ubFEmLleewY2974RaHvhfvTaHFQXlp6CWhyM+okeD27i6V6ilI3DCFBFeB16abUaUGvKRaElrFUlHTSXSuHqPudYPcnbpddyul+ZOmQUg1tIhpLVLjrkGlytzsPnoxwCAkb1m4MSl3dAqY9Atpl+D+/RJGI7dZ77Bv4+tgITjcWPvuxo9x6WKMzh0cQus9mowVnP/jKHPeY3Pa9Jx//33Y8aMGRg7diwAYPv27XjooYe8Hrg9cK0uabWb/Z50HC/YCavDgsHJ48X7BOZA9pX9iA/v7jHpSIyoGdNB/M8u2CCBxGvLlliro5VUJXVVIo1soKXD1b/b0afN2hxWSCWyZnWvENJWcJwEI9wGeQJosFX92uSbxX9LeTnG9J3l8zn2n/sBQ3tOQqQ6HhyaVrrBa9Ixffp0DBgwAAcOHIAgCPjwww+Rlpbmbbd2QRbAqqTlxsJ6M2PkfE2S44lKJkW0Wk5JR4AUlJ9GduE+XJs8Hp3Ckz1uV1MgrHUkHQO6/gHdY9PFuNyJpdA7fNJhERdvayqT1QCjpQJaVWyDVWoJ6UgUMnWD61L5wmvKn52djRUrVmDWrFkYMWIEFi1ahNzc3GadrK1x9fsGYklrq91Ub7AOd3WartVLkpMUEUbL2weIwVwBBuaxOJuLeyn01kDC8YhQxzVYME4hrr/SsauS2h0Wr8tue3JFdw4HL/wHuupiP0dFSNsTH94Dv+duwqWKM7iiyxX/84XXpOPll1/GH//obKpJSUnBo48+ipdeeqllEbcRNQXC/Pth7RDssAu2BkcIy3klbF6mKyVGqqG32FBltvo1LuJWGEzecGEwFykvg4xXtoqkw2wzwmjRgbl3rrpRyjWICesCVQMVBjsKxgTYBVuzWzpktLw9IaJSQz7KjZeRVbADR/K2iv/5wqc6HTfeeKN4e+TIkR1iuizgViHUz+tWuLpPGhpBL5MqYbTqGi1z7r7Efbiy45SmDzTGGIyWSqjlET7NVorVJIKh4S/6YMovP4VzxYeQ0X0i4rTd6j2ulodjaM/2vzJ0YxgY+nYZWat+QVPUJB0du7UolAorc5BbcgQGcwU0yij0jBuEBLf1P0jwTBzgHNdps1sgQGhSl6PXpCM6Ohpr167F1KnOkqibN29GTEzoR+sHg1YVi6E9JiPMz/UaXH3rDb1QcqkKEo6H3WH1OL3RfYn7vvFU3tpfrHYTbA4LosISfNp+YLdxAY7INzqTcxBpuKr+zBXiJOF4JMdc0+z9a5IOal0MhcLKHBx1q3ipN5eLtynxCD69uQw7T6+F3lwOBgaNIhJ/6DOr3hovDfGadLzxxhtYtGgRli9fDrlcjiFDhmDJkiV+Cby1k/FyxGi6+P24EokEsdqutQqDuQzqNg4cJI0u5pZIBcICoqYSaeNdK60JYwy66mKoZNpGf23klZ2EzWFBSqdrgxhd+yEuAEmr9YZEbskRCExAZXURVDKtOOYqt+QIJR0h8FvOd+ifNBrdYwcAAM6XHMOes9/ilvSHve7rNeno0qULPvnkEwCAXq/HlStX0LlzZy97tS+MCeD8OM1Oq4zxOO/fl8JDrlodlynp8KswRSSuSbwR4Q0kgw0xWipRWHkOsdokRKrjAxxdw0w2PWwOC2I0iY1ul19+EiarvsMmHRXGImRf2YtuMf1rLcntKxrTEVoGc4Xz2jMGk7VKTDpc6ySR4LLYjGLCAQA94tJ9XnvF6zfpN998gwULFqC8vByTJk3CE088gb///e/Nj7aN2XH6a/yWszFo57PZLagwFsFsM3rchlo6AkMpC0PX6D4NFthqSLW1CjnFmSgzXA5wZJ6J9Tm8VLeVS9WwC7aALWDY2pltBlRWFze7pUIuVWF4yjT0TrjOz5ERX2iUUWDM0cD9badVsj2RSKS1FpIrNRSA52W+7ettg7Vr1+Kpp57Cpk2bMG7cOPz444/YsmVL86NtYziO8/tCT0VVF3Cu+HCDxZqK9RexP/d7lOjzPO4vrjTbwmmzi34+ikU/H23RMdrSef2tNUybFSuRNlAUzJ04bbaDdg+Iy9o3swy8hOMRqe4EZQeeARRKPeMGiQkzx0ngKoPZM25QKMPqsIb1mIJfTn2FHw9/gB8Ov49fTn2F63pO8Wlfr90rANCpUyfs3LkT9957L6RSKSyWjtPEKOeV0NvKG51N0lRFuvO4XHm2wb7ImuXtPX85hCtl0CikLSoQtujno1i85Zh4e+GEgc0+Vls4ry92Za9HuCoGg7rd5NP2yqtNvCabIZBhNSqlUwbitN28DuByva8s9mqvNUjaI5vgHAAqa+aUWcA5fsbB7JBKfPtFR/wnITIVw3pOwdkrB2F1mKGl2Ssh1Sm8G27PeAY6UykABo0iyueE3mvSkZqaiocffhgFBQW4/vrrMW/ePKSnp7c05jZDxishMIfzw4bzz4dNzbL29afvyXnnL9LGpuZxHIekCHWzk466X/yufwc6AQjVeX1hsZtQbdU1aRCpVCKDPMS1OpyDnRsfzwG4VyXtmC0d4rorkuYnHb/lfAuzzYhx/e7zV1ikCbpG9212FUziX+dLjuFo/jZMGzwfVaYyfHforxieMhXdfJgh5jXpWLp0KQ4fPoxevXpBLpdj6tSptep2tHc1o9YtkMr9lXSYwUukDf5ikvnQ0gE4x3WcLq6C2Va/n7Mxdb/4XQKdAITqvL4yWpwD0sKa2EeskmtRZS7za0uYr2wOCwTmgEKq9rqtnFdBwknhEGxBiKz1cVUVbm5FUgDgJTLYHJaQvNYdnUOww+awgOekuFKVC5VMi1htUqjD6rCO5W/HhP5zAQDhqhhMGfQ4tpz43D9Jh1QqxdChQ8XbroXfOgq52/orKvinWdrqqF8CXTxfE5IOoGkLv3n64ncJVAIQqvM2hcFcDgDQKJq2CqtKroXJqofVYQ76mhxXKs/hxOVfkZ40Bl2iejW6bWJUGpKiewcpstZHq4xGrLarx787X7g+C+yCrUXJC2k6nakEv+f+iISIFBTqziEhIoWSjhByMP+eMiIAACAASURBVEetdZ5Ucg3goSJyXT6N6WgOQRDw6quvIjs7G3K5HK+//jqSk2svoFVeXo4777wTP/74IxQKBfR6PebPnw+TyQSZTIa33noLcXGhLXgUp+0GuVTVYFdIczDGYLGbEK5suA/euQqm1Pv6K5E1SUeYXyLr2AxXWzo0yqYlHQO7jvXrdOqmqLxaFEyriva6bUf/ZZ4c2x/Jsf1bdAzp1UTD1oI1XEjzGM3OGjoxmiQUVV2A0aILcUQdW3x4MnaeXouenQYB4HCh5CjiGlkg013APi23bt0Kq9WK9evX4+mnn8ayZctqPb5792488MADKC0tFe/buHEj0tLSsGbNGtx66634/PPPAxWez2K1SUjpdK3fRq07BOdAtMZ+FQ/pPhEDkkY3epzECGeq0ZRpswsnDMQr4z2Px3llfHpAWhtCdd6mMFz9UAtrYmGwUCUcAKCrLoaEkyLMh9YZgQmoMF5BJS1Y1myuVaftVKsj6Nx/FIQpImCwVHpca4gE3vCUaYjRJCK7cD/OFh1AtCbR59krXj8xrVYrVqxYgeeeew4GgwEffvghrFbvpYAzMzMxatQoAMCgQYNw/Pjx2ieWSPDll18iMrLmQz4tLQ1Go7M+hcFggFQasIaYkJHyMozrdx+uTR7vcZtoTRdolY3/ek2McCYtl5o4bbZbVMPtIlIJhzsG+papNsdLNw1AcgPnbg0JBwB0juiJ5JgBTZ6ZYHfYUGooQJWp1PvGfmR32GCwVCBCFQuJD4kPYwL25/6As0UHgxBd63Ou+DAulh73vmEjZG4tHSS4XElHmCISYYpICMzeaC0jEli8RIrk2P7onTAcf+hzN7rF9AMv8e372utWixcvRnR0NE6ePAme55GXl4cXX3wRb7/9dqP7GQwGaDQ1rQM8z8Nut4uJxMiRI+vtExUVhT179uDWW2+FTqfDmjVrfHoSZ86c8Wm75rAKRpQ5ziFMEodw3rc1OfzBlcV7ahbXlztnwBw6ewFjh3RGZmam12PuyK/Cgl8LEC7ncXNyOL496/x1P66rFtvy9Zjyyc9YNaEn1DL//3r/8EgRLlYY0T1cjgtVzqQ1PVaFybF2n2L3J8/nkyOzsGmxWIVqFNh+h1aSgDhZ8MZMmIRK6G16SKojkVnpW8xGSzUshjxklvv+HIP92gTKBcuv4Dk5Si82P2EwC1WQs044e/oCLnBXGtymvVyvYGjKtSqw5oIxhmNHslBur4TeocfBo3uhlnjvWmwPWtv76nzJURzN3w6HYMet6X/B5qMfY2iPST5VPPaadJw4cQLfffcddu3aBZVKhTfffBNTpnhvRtFoNGKrBeAc4+Gt5eLDDz/E3Llzceedd+L06dN4/PHH8eOPP3o9V1paGrRardftmqPKVIbfck4jPiYK/bpktPh4JqvBOTVTGe2xi+V4wS4UVJzG6N53e6yp0M1gBn46D5vC+XhGRuOx7ci5gpf/eRoqmRQ/PXITrkuOEwt0LZwwEE9/fxDv7jqFFWfNWHPPDX4dA/D98Xz84+RJ9IrV4vf5t+K1Lcfw152nkNolzmvc/paZmenXczoEO3QnTiEmLBwZPYP3XM6XHIXhihYDumYgITLFp32M2WfhEGzI6OtbnP6+VqHCGEPJ8UxEquORkRK459NerlcwNOVa2R02lJzMRExYF2T0zMDlyghk5VegZ1IyEqPSAhxp6NW9Vnq9PqA/tH2RVbATk9IfxX+y/g6VXIOp1z6BLcc/8ynp8PqTluM4WK1W8UuooqLCpy+kwYMHY9euXQCAI0eOIC3N+5sjPDxcTB5iYmJqJS2hIvfzQk8l+jwcOL8ZZYYCj9u4mqkaq9URG6aAnJfgks77NTpUUIZpX+yAwIAN94/GdcnOwbkLJwwUuzaWTR6M65PjsP7IBfz9N/+9oc+V6vHntXugkvH45v7RCFfKsXxKBsLkUpwrDV19C3eFleeQeeGnZo134CVSyKWqoBcI6xLVC9d2uxnRTViQUC5VwWo3d7i+cPvVacItKQxGQofjOAxOnoCeV7/QOof3wM39H+gQCUdrxXGSWsXA1PJwAL79UPWadNx7773485//jJKSEixZsgTTp0/Hvffe6/XAN998M+RyOe6880688cYbeOGFF/Dll19i27ZtHvd58skn8a9//QuzZs3CY489htdee82nJxFIrsFjVj/144qFwXjPA0l9mTbLcRwSI9ReS6GfKanCrSu3wWC14R93j8T43g1/Scl4CdbOHoXYMAXmf38Qv+e1fIyCyWbHHf+3EzqzDR/PuA4DEqLE2FNjtcgp1beKL8DK6uJGy857o5ZrYbIZwJjgx6gap5CqER/Ro0nTdBVSFRhYhxuT4Hq+LU06jBYd9p/7ARdKPU//Jv7HS6ToFJ4sFsGTSHifFsYkgROp7oRTl3+DwASUGS7jt7MbER3m2w8gr90r06ZNQ//+/bF//344HA6sWLECffr08XpgiUSCxYsX17ovJaV+M/D27TUr08XHx2PlypW+xB00vEQKCSf1W0uHaypsY/UC5GKi433a7J7zJbALDX9xX9JVY+InW1FisOCj6dfhT4O6N3q8rlFhWD3rBty6chvu/McuHHxqEqLVzf+gfnzj7zh6uQIPDu+Fe4fUfu1TYrU4+v/svXeYXPV59/09bXrd3ntRl5CQBGoUATJgYz8GG2KIiZ28ee08kLxP/CR+bRMwGBv8xIS8SWyCHSN3THDDpgmEQEgUgVaspJVWZXvVlmk7/dT3j9kzu6st086ZmZ2dz3VxoZ2Z8zu/nT3le+7ffX/vERfGvCGUWdLrb3E5qba01zNmuAPjCPOBtPTmEEQekiSBjrPBksyMK2lQsRLw5UC074oCZa6uwCUYtdaUx8kTPwuZsfnDHvjDbpTEWaaZR1muavwUTg0eAkUyeOfib1Bua8LW6lvj2jZmpOP+++9HU1MT7r77bnz+85/HqlWrcO+9K8sGWEPrYgqAeFnKAl1mxpV0fkO42VRaDRAlCY7Q/M6hzkAYH3v6IPpdfjzysY340o74QpE3tVbgwRs3oN/lx72/egfiIoImFs8c68L+D7qxuaoA//qprfPebyqMLKNdnJxKanwl8Ydd0DHGqA9DosgmOUE2PUssDt8wDp7djwHHmYS2qy/eiJ3Nt0+HQlcOgiiAJplo1DJZ8u3tM0P7wOs4fO5ZcMJM1eTZkXdwov8AeCF2JWUe5WEoDTbV3IBPbLoP1676HMqsDXE/BC0a6bjvvvvQ2dmJsbEx7N27N/q6IAgoKytLfdbLiAJjOcQF2ionQzyiQ34ijRVdkV1JxwNzra19YQ4f/9EhnB3z4G93r8LXb1if0By/ceN6vNs3gZc7h/F/3uzA/7s3se3bh524/3cfwK7X4L8/vwc6Zn4otLEocqPumvRid0NpQuMrCS+wCHF+FJqSdzesKVyLKvsq6DTpsWmTc08SFQ8rTWzI2I2luGHtF1JeyqPzJbMZwRdygRPCc8rZTVorHL4h+MMeWA2ZNZBcibQPHIQnMIEtdTfjldNPw2YoxYjrArY33hZz20VFx+OPPw63241vf/vbeOCBB2Y2oGkUFhYqM/Nlwobq6xQbK8wHwVDaJdckDRoLmkq2oMC4dIluVVR0zEQ6WF7AHT85jGMDk7h7Sz2euO3KhCtRKJLEz+/ehSv/5SX80ysnsb22GNc1xSc03UEWn/npYYR4Af997x7UFy5cVdQ0LToynUwaNR1K0P58NjomvZ6wnmkn0ljt7BdClARIkhR3TX0ukWpFFkmQoKf7r+RJD6IkIMBOwWoomfP3k038/Kw7LzoywKCjEzdv+BLOjryDhuIrsLX+Fvyp/d/j2nbR5RWTyYSqqipUVFSgsrIy+l9paSm+8Y1vKDb5lcbm2n3YUnfzkp/RMUY0lW6JWZlQOW2F/kJXJCdBEEXc++w7eP3CKG5ZXYkf37kDJJnchbbYpMOvP78HJAF87udHMBKH86kkSfjCs++gx+HD1/auw61rFo8eNBdHnrq7Miw6CBAoMdfCZkj8Bj4bXuDSYlYkSRI8gXEYNJa4W0nLuPyX8FrHj9Ez/pFKs8tO/GEPHL7hOeH5ZGEonSLj5ImPQHgKEqR5TsFR0RFyZ2JaKx4JImiKwZCrE1X2VkiSGPdS16KPO9/4xjcwODiIjo4OXLx4Mfo6z/PwerOj1DFdOP2jcPpGUFXQmnKioJJJaK92DgMA3r/kxzdfbcekP4z/bu/HzrpiPPf5PWCo1Ey+rq4rxj9/Ygv+1wvH8blfHMHBL90Ieokxv/fmWfzxzBCubyrDwx9b2mW03KyHnqHQ7cjssWQ1FGNz3b6UxhBEHgfP7keBsQLbGj6u0MwWJsBOgRdZFBtqEt5WzkkIx8gVyjWGXefRM9GObQ2fiBk9jEWhqVKxpdY8sfFHI5ELiw45UpknvZTbmvGHE0+CJhmUWevxyukforpgTVzbLio6vvzlL2N4eBjf/va3cd9990VfpyhqwSqUXMbhG0b3+AnYjWUpiQ5JEiFKYlyh7VODb0KSRGys2bvg+w8fOIlnPuiO/vyt108DADaU2/HHv7oeBo0y4fP7d6/C0d5x/PbUAB54pR2Pf3zzgp873D2Gr7/8ESosevzynl2gyKUFD0kSaCycKZtdzg3JKJKGljYgyKkvoDzBSD6HLYmlFS0diYyxwsoSHXJkQgmfjnVVe1IeI0/8zNifz13+1NIG0CQTFSV50svW+luwunwHDFoLCILE9obbUBinZ9Cid4aqqips374df/zjH1FRUYFAIIAtW7agpKRkTr+UlUC0hDXFsllf2I3XzzyDzpF3Y37WExiHwze84HtLtYrf21IGm165DpgEQeC/7rwazUVm/PObZ/BCx+C8z4xOBfBnP38bJAH8+vN7UGKOrwS2sciMqRCHSX/m1sg7R95NuApkIfQaM0KsH6LKXh0FxnKsq7oGRebqhLelKQ0IkCkfx8sNXiGfjjzpx24sQ33xRlj0c/MICYLAtoZPYFucTcbyKMPRC89Hc8pMOlu075MsOFz+MRy98PySY8R8HH755Zfx1FNPIRgM4rnnnsNdd92Ff/zHf8QnP/nJVOe/bJBLWJdyCI0HuXIlnosfQ+vgZz3zogBLCQ4AePJwJ8xaRtEmahadBs//xTW4+v97BV949h0c//tb8fPjPQAijdw+9/MjGPOG8MRtW7CzPv4n8KZZFSzFpvT7RvACh35HBwqMFagpXJvSWHrGBDfGEOYCi1rXK4GOMaHKnlyPF4IgIuXfK2x5RY50JFsSPZtJ7xCc/hHUFK5NewLxSqTAWL7okphFX5Tm2eS5ovYmfNDzIoLcFEosdTBqrCAJCr6wC6Oebhg1VmytX3qJOeai/49+9CM8++yzMJlMKCwsxO9//3v88Ic/VOyXWA7IkY5Us9bjKZeN7lMum82STPn15XZ8//bt8IQ47P73V/HIa6fwyGuncO33D+DtnnHcvqEGf7dndUJjymWzmfLqkEOzZl3ylSsyM14d6v0ukiSlHElZmaIjDIIgQRGpLzk6/SPomWhPmydLnsWRJAlhPgCOz45r5ErAqLXiutV3Y3fLZ2FgzPAEJ+AKXIKeMWFPy124bvU9MOmWXgmJeRaSJDmnW2xJSQnIGOv1uYZcJcCmeHDLYe2l3EhlZi/pzBYpcgRjsWiHmq3i793aiH8/0omPhl3R197rn0SBQYP/uvPqhPMyZIOwTJXNyk6kl68XJ0NUdKjYg8UTnMAHPS+iuXQL6ouT+xvXF22EIHLLPo8mETghDIbUKPL70lGDsJW1RJUJgqwPHw28hip764KRyPGpPnw08Dpay65CffGGDMxw5WLWFWJN5a6kto0pOpqbm/GLX/wCPM+js7MTv/rVr+KyQc8lZAEgiFyMTy5NOBrpiEN0LLGks5jwUFNwAJGlndmCQ8YZYPHk4c6E9z17eSUTRO3PFYh0FJoqsaH6OtgN6hnneYITECU+utyXDBX2ZgVntDzYXLsv5XNXJu9Kmj58YRemgpNgzQtbnRumKwHzyaTLi5ghiwcffBBjY2PQarX4+te/DpPJhIceeigdc8sadIwJN637q5Qz1+WwdjxNuky6AhSZqhY1EXto30Y8eNOMuk+H4Fgql+SR107h4QMnExqz2maEliYzFumQa/xTMQaTMWgsqLA1RyMeauAJJF+5spIx6WyKGUhp8qIjbfijkciFw/VGTV50LEdiRjoMBgO+8pWv4Ctf+Uo65pOVEAQBIs62vUtRbmuEQWuBlo6dgFZha0KFrWnJz8giY2RkRFXBoRYkSaBhumw2E1AkDYPGumyan3mCE6BJJiWvl0ueHvROnERL2bZo185cRpIkSJIIklSmKymdFx1pwyc/FCySI0CSFAwaS96rI0NwAgtvyAG7oQy8yMXdUDGm6Fi1atW8tdDi4mK8/fbbyc10mSL7/9uNyYfPC02Vil/oH9q3EW1t8xu+KY1auSSNhWZ0jnngDIRT6mibDIt5oCTLh70vg+UD2Nl8h6LjApET3B92o8BYAYJIPqeKFzh4ghMrJhEyzPvx1rlfocLWrEg7A7mFgaRyaXSemQiGUbN4YqJRa8OEd2Be7lsedRlxd+G9rt9DkkTcsvFv8MKJJ7Gn9S5U2mM3Fo0pOs6dOxf9N8dxOHjwINrb21Ob8TLk5OAbCLI+3LD2L9KyP5YPYdDZCZPOjlJLXVr2GQs1cklm53Vsq1nePgqCyMEXckGUxGj9ulJMTdfGp2rXPru9/UpAjkhQZHwdMGNh1hXgpnV/qchYeZbGF3bBoLEsGaWSRYc/7MmLjjRyou8Abt7wJRw88wwMGjNu3vB/4/C5Z+MSHQldGRmGwc0334z3338/6ckuVxhKB15kU7JAfr/7BZwZPhrXZ3mRw8WxDzHm6U16f2qgdC5JppJJPcEJDDo7Fe2XoteYIUFSpQeLQWPFqvKrUZKiAJXziVaKK6mSbqRA6k3j8sSHKAkotdSjzNqw5OcqbM24ovameTbpedRFggTDrPw1myH+TuExIx1/+MMfZnYkSbh48SJoeuV1qJydtS7bSSeCIPJwB8bifuJSygVVDWaLjFRzSaIt7ifS69Ux5ulDz8RH2Fp/q2ImTwZG9urwzjkhlUCvMaGuaH3K48iRjjAXu4FfLsCp4EbqDoxDksSUllrzLA1JUHEl7lv0hfPcSvOoj1FjwaCzEwCBMB/EudH3Fk34vZyY6uHYsWNzfrbb7fjXf/3XpCa6nImWsPLJiQ5ZPMRTuQJEkhxJggKbpX4ASiWuRiMdaW78JmfGK1G5IjPj1aH876KUr4Z8HGfrcaU0aoiOE32vgqa02NN6p2Jj5kkNSRJTynXKkxhXN30aH/T8Cf6wB789/n9Qbm3CjuZPx7VtTNHx2GOPgeM49Pb2QhAENDc3r9BIR2oX60TcSIGVY1ldYzOCJom0l836Qi4wlDYuz5R4mXElVfZ3CXF+vHPxt6grWofGkoUb7sULRdIotdTBpCtQaHbZzUzfFeX6ETG0DlwWRiBziUFnJ3whFxpKroj5oPZe1+/Bixx2t3w2TbPLo9eYcM2qP0tq25jqoaOjA3/7t38Lm80GURQxOTmJ73//+9i4cfmVaKbCTKQjSdEhxG8MFt0npYefze1yMJoi0142K4oCAuwUrIYSRdfojVobagrWpJzseTmewPi0SZwyc72i9iZFxlkO2A3laCnbBrNOuRA8Q2kRCM/vi5RHOcY8vZj0DaGp9MqYnyUIEoGwB6IkLOprlEdZ+iZP4/TgW1HDS5k7tv5jzG1jio5HH30UTz75ZFRktLe341vf+hZ+85vfJDnd5UmppR5WfXHS4fioGykVv+hgaB0IloAoCor5DGQjjUVmXOicgjvIKtohdzH8rAcSJEWXVgBAxxiTtgZeCo9ClSsrEauhWDFjMBmG0kKCBEHkFGkil2c+vrAbWtoQV4TKqLXBHRhDIDyliLtwnth82PsSdrd8NqlraEzREQgE5kQ1Nm3ahHB45Rnj6DWmlLqHamkDSiy1CRk7XVn3sRWxTinndXRPerGlWv2ksCDrAwFy2Vyg3NNOpBa9MjfPS54euPyjaCrdquiyw0phdlJ5XnQoDy9wCHE+FBor4vq8nMDoD7uXzTm93LHoClFqqUvq/hRTdFitVhw8eBA33HADAODgwYOw2VZueVKyCUvF5moUm6sT2mYlCA5gpvFbV5pER4mlBjeu+wIkSVJ87N6JU5j0DWFL3b6UQ72j7i50j7dj0HkWGlqPSe8AymO41MaDwzeMQWcnqgvW5LzoODf6HrwhJzbX7gNFKpOLJn9nnBCGHurZ3q9U5CVlY5wCQi6X9YXdiL9wM3eRJBHvdb8Al38UJEFhZ/PtsOiLou93jryHrvE2EIgYJFYXrAbLh3D4/K/ACyxIgsbu1juXrMBbW7kbr57+Ecqs9XPuU5tqbog5v5h3tW9961t4+umnsX37dmzfvh3/+Z//iYcffjjmwLlGkPXitY4fo2M4fU6sYT4Ih28k58sb5bLZ7jRWsJAEpdhNaDbekAMO31DKXh2j7i6cHDwETzBSnilJEk4OHsKouyvlOUbLZnM8SRmIRImcvhFF1/pri9ZjR9On4y4RzJMYMz2R4vt+Z0c68gADjrMQRA63bvwbbKm7GR/2vhR9L8T5cf7Se7h1w5exb93/hfe6/gBJktA1dhx2Qxlu3vAl1BdvwJmhw0vu4+TgIZh1BepEOurq6vD8888jEAhAFMU5be5XEjSlgSgJSftm9E92gBVCaCrZEnfy2ZinF2dHjmJD9fUx+7AsZ9JtEDbhHYRBY1blpjG7gsWgsSQ9Ts9ExPWXAAG9xhz1d+mZaE852jHjAZP7okNeAlEy4TOVv2ueOCAIGLW2uCus9Boz6os25n1Tphmb6kOlvRVAJKrr8A1H39MxRtx2xd+BJCj4wi5oaB0IgoDdWBbNHWP5EIgYOYSiJGJXy2eSml9M0XHq1Ck888wzcLlcc8LRP/vZz5La4XKFJjUgQCTd6GnYdQF+1o3mOLKxZaKeCjl+c6i1G0GRRFoMwiRJxIm+A7AainBV46cUH18fNQhLrbeJLxTxESFJak7nWrkJViqsJCt0Xggr6tEhI0kSJIj5agkViKfZ5WxIgkRr+XYVZ7S84IRQ9MECiNgvzK7sIQkKnSPvon3gIFZX7AAQyTkccV/E79v+BWE+gFs2fGnJfVTYmtA58i4q7S0giRkZsVhzvtnEFB1f/epXcc8996CpqWlFl4cRBAGG0iYd6QjzwYQ9IeQDJ9c9ATQ0hVq7MS0GYZwUhAQRRoUrV2TkZOMgm5qAMunsmAo6AMy13o7npI6FdgWJDk5goWOUjc5OeodwvO9ltJRuQ0PJJkXHzpMnVRhKN+fhWJKkeeJ4dcUOtJRtw8Ez+zHq7kbn6LtYV3kNWsu3w+kfxZudv8AnN/8/i+6jd+IkAODM8JFZrxLKlMzqdDrcfffdMQdaCTBJmnVJkgRWCMKqKYr94VlEn0hXgHtkU5EFr50fgTfEwaxTpjnXQnBSJD9G6XJZGf106D3IpRbpaCjehGM9f0KQ9cGss0erJBqKU7/JaWg9aFID5dNoswtB5CFKQrQdvVLIS1359vbKI0oCesbbYTeWJdSRe9Tdjb7JU1hVfvWKX2YpsdRi0NmJ+uINGJ8amPN9eAITaOt/FdetugckQYEkaRAEAS2tBzMdWdcxppj3nDu2fjXp+S0qOkZGRgAAq1evxk9+8hPs3bsXFDWjlioq4itnyiU0lA7+sDvhChZeZCFJYkIeHQCiB0E29l9RmqYiM147H0km3VSpnlsmK0USPJWIGCyEjjHCoi9O+em6zNoIg8aCMB8ARTIw6wrQULxJkeoVk86etm7JmUSURJSYa2FWuDeHhp4pmc2jLIHwFLrG21Bpb01IdIiSAE9wAt6Qc8WLjtrCtRhxd+Glkz8AAOxsvgNnho/ArCtETeEaFBjL8dKpH4AAgUp7C8qsDbDoivBO129xfvQ9iJKInU23Lzj2R/2v44raG3H0wvMLvh9PnseiouOee+6J/vv999+fk8NBEATeeOONmIPnGlUFq1BsqYUEKSFvSFk0JL68op3ePvfD4LOTSdUVHQGQUC/SQRIkdjT9j5THcQfGIUoCWsu2x1WGlmc+DKXB5rp9io9LU3nRoRZyBUqiSd75CpYZiAWuQbONBTfV3DDvmmLQWnDj2i/EHLtoWgjG6v67FIuKjkOHDiU9aK5SaW9JajtB5KChdHE3e5MhCBLbG25TtD9IttI4yyBMTTjJDwOpUXydX2mGXOcAAFX2VaqM7/KPgRNCKLHUqjJ+LjPbpyOPsvjCiZXLyuRFR3qoLlwDAAiwU9hQfd2c99r6Xo1rjEVFx9e+9rUlN3zsscfi2kEewKIvwvVrPp/UtislVDjbIExNyplNWNWgblK00z8Kh28Y1QWroWOMCW/PCyxG3d3Qa8wJhZgT4ezIEQRZL26I4+lmueINOXHJ04MSc62iVugRjxcmLzpUwJdk92eG0kBLG6KiJY86HO97BSHWh0FnJ6aCk9HXJUnEhHcQW+o+FnOMRUXHtm3blJllDjE+1Y/+ydNoLNmMAlP6clpEUQBBkDldPVRfaAJBqG8QRhHMHHc+NXD4htE9fgJ2Y1lSomNsqg+ixKPKvkq1v7mGNsAbckIQeVVM0rIBb9CB7vET0DFGxfuvNJdeqUop7krHH3KDJKikWk4YtTY4/SM5fUxnmrrCdXAHxjHq6Z6zxEIQJDbW7I1rjEX/Mrt27UJxcXE0oTRPpIrE4R9BOduMRLIOvCEHQlwAdkNpwr0aTg68gVFPN/auvhcMnbsXOS1NocZmxEWVvDpG3V24ONaGS+FBBC/0orFEmaTMhUi1xX2FrRk6xqhqH4nZZbP6JeyOlzNctK298udNXdF6xcdc6UiSBE4Mw6i1JtlqogZ6jTkvpTO+jwAAIABJREFUOlSkyFyNInM1agrXRn2kEmXRv8wDDzyAp59+Gvfccw8IgphjDLZSE0mjvhkJlrAOOs9hwHEGO5tvh5lKLJNeLs9jhWBOiw4gkkz6xsVL8Ic5GLXKlc3KluJB1oeQ6IHTPwLfoBMAVBEeBiY10UEQhGrLKjLyBSO8AkRHvinb8oAgCFzT+mcQRD6p7euLNyg8ozyLkazgAJYQHU8//TSAfELpbBjZPjpB0cEm0dZeRjOrbNaY25oDjdOio8fpw/py5Z7yZUtxQeQAIPoUpISl+EKkEukYn+qHSWdX3WpbQxsA5HY5NiewANSJdPSMt2Pc24/NtftSugCnwqi7Cz0T7fCFXDDp7IqVVGeafJQit4kZwzp16hT2798PlmXxxS9+EVdddRXefjt9Tc+yCfnikqhDqNxYi0ni4qRZQV4dzUWRG63SyaSypbggciBAgJq27VXCUnwhtIwRBIiEDcJ4kcOpwUP4oOdFVTrgzmYlWOzzonrLKwF2Cu7AWMbOSzl65w05IUGCN+RUrCFgppgKOuDwDYOffjhIFEHkcW70PfRMu2XmUY+usbZ5r3WOvBfXtjEl5aOPPor7778fBw4cgFarxe9+9zvcf//92LNnT+IzXebIFy82wax1lg+CobQgk1inTDa6shxRq2xWthQXRB4ESMgmK2oZhJEECZ3GBFEUEtrukqcHvMihtmi96knDpeY62JpLoEsiYW+5IEkSCBBgSOVFB5Nhg7CeiXaIkohA2AO9xqRoQ8BMMeA4gyHXucgytC5xQzeSoDDg6IRJa0ND8UYVZpjnzPBRcEII5y8di1YaAREjvt6JdqyuuDrmGDHvgqIoYvfu3Xjrrbewb98+VFRUQBASu5jmCgylgd1QlnANOZtE3xWZlfBEKqNWt9mG4k3RdeLZPQiUsBRfjN0tn8WO5k8ntM2QU/bmaFVjSnNgaC1MOjtoUj3L+Uyzofo63LTur1TJ6WAybBDmC7kQ4vxg+SD8Yc+s15dvyWjUGEyT3MMAQRAwaq3whd2qRwpXKtHKv8u+Xoqksas5vq6zMSMder0ezzzzDI4dO4YHH3wQP/vZz2A0Jl4GmAsQBIntjbcltI0oieCEcNIOmCZtAZpLt6IwjSW6maKhMPLUrXSko9zWhJrCQXSOvAtJJBW1FF+MRLuPekNOuANjKDJVpS2xUxB5CCKfsZyEdKBWxEgWHXyGREekPHR0eg4sOIEFQ2lUi96lA1/YDYPGAjJGW/WlMGlt09WC/qTKbvMsTXXBKlQXrEJd0YY5LqeJEFN0fO9738Pzzz+Pf/u3f4PVasXY2BieeOKJpHa2EiFAYE/LXZCSbK+l15jQWHKFwrPKTvQMjSqrQZVus63lV6HUWo++C8PY3rxT8fEvJ8wHMRWchElrj+viN+Q8DyBitZ8ORFHA62eeQaGpElvrb03LPtON0zcChtYmFaqPRaYjHVZDKQYcZ0BTGvAiB0HkwFAaVaN3asLyIXBCCHZDaUrjzHYmzYsO5Tl45ie4Ye1f4OCZ/cACzUAU6TJbWlqK++67L/rzP/zDPyQ2yxxj1N0NX9iFppItcT1FEQQBg1bdSoRcoqnIjMM9YwhxAnRM8k88l6Ol9Si11GGIcCg25lJMegdxeugtrK3cjeqC1TE/L0oCdIwpbbbkJEmBJpmcXrZr6z8Ao8aa8DJXPOgZMwpNVdAyBsXHjofV5VdBEFl4AhMIsFOw6AuXdfWKnB+QaM+Vy5ktOorMVSnPK89cGkoiovbaVZ9LupVEvjYpQUbdXRj39qOuaH1cWfGCyEOSRFAkk3So90T/AdCkZp7XfS7SWGTGW91j6HF4saZMmVCxJEkQJSGtpXiJls2urdwFURISXpZJBQ1tyFnRIUoCBJFTzaPDaijG1vpbVBk7HmhKM69p13LOY4jmc6S4PGTS2WDUJGculic2H/W/jrqi9Xi36/e47Yq/TWoM1a7Coijim9/8Js6fPw+NRoNHH30UtbVzn+KcTifuuusu/OlPf4JWq4UgCHjsscfQ0dEBlmVx//3347rrsutGy0TLZsNxiY6xqT6cGjyENRW7UDPdLCdRIh73uWuBPpvZyaRKiY4Q58Ph88+itnAtgPTkL+iTMAhLp+AAIknKnsBUpMojxyz2Zzw6cs8YzOkfhVVfPEdEu/xjODf6HlZX7Eh6rT2TVNpbUWCsiBowJotZV4jdrXcqNKs8l1NqqcPP33kAEoCfHp3pzyYhcoe6d1fsnmyqiY6DBw+CZVk899xzaG9vx+OPP46nnnoq+v6RI0fwxBNPYHJypmnMCy+8AJ7n8etf/xpjY2N45ZVX1Jpe0syUzYZgQOxlE/lJMtEOs7PRUHr42eWblZ4ITdNeHUr2YPGGIu6jGtqAAETFxl0KHWMAQZAIckv/Hv6wB2eHj6Ch5ArVXUgvR0vrIUECJ4RzLpmUV9EYDJBLBE9CS+vTlocDRHIfjve+DLOuAFfPal8uSQI8wXF0j5+Iq+lWtkESJIxaa6ankScGu1o+g10tn8EbZ3+KvWvuTWoM1WJQbW1t2L17NwBg06ZN6OjomLtjksT+/fths808zR49ehRlZWX467/+azzwwAO4/vrr1Zpe0kQNwuL0zWBTMAaTYWhdtNIg11GjbFbuhqhGQuFiEAQJPWOKGekYcp2Dwz8SNZBLJ5pZ/VdyjRkLdHVEBwECXWNtGHKdV2X8xRhynYMoCfNyNwpMFbAbyjDhHYAnMJHWOSmB3HxQCVz+S+ge/ygqPPMoT7KCA1Ax0uHz+WAyzSSaUBQFnudB05Fd7tw5v4LA5XKhv78fTz/9ND788EN87Wtfwy9/+cuY+7pw4YJyE4/BlDAKL+/Fmc7TMFOxT+4Jrgte0YvzZ7ugIZNrnjfJOeATvTh+4hhoYmHx0tY23yFuORLkI5GIj3qGFfudxrgO+EUvei8MgSa0afuuprgAgqILHx7/YMGlE0kSMcB+AAkihrtcGCXS+zcMiix0UgXOdpwHRSzs17Fcj6uA6ISX82IoMAL/iDq/gz8cQNA3gjb3zPhqfl+SJGGQex+CxGE8HICjf+6+ONEAL+fF0VMvoYzJ/oZ08nclSgL62CPQk3aUM6mbek3yXZgShnCp3w0dmRtJ/Mv1PFwI1USHyWSC3++P/iyKYlRwLIbNZsO1114LgiCwbds29PX1xbWvlpYWmM3p8TYY8/ShfWAUDRV1qClcG/PzJ/omAa8PV67elnTDts6REPodQaxual2wLXtbWxu2bNmS1NjZSPmr/RhnCcV+p7fPX4RWKMK21VfjxIkTafuu/OGmiDspY1owZ2LM04vJAS1qC9dhdcXWtMwpEZbzccUJLPzhDdDSRtVKJwPnu8AJYWxZE/mO1P6+xjx9mBzQoLpgI9ZWbp/3viRJONbjhzswhuam2gWvFdnC7O/KE5yAo6sdNQUtWFOZ+vc34NDj7IgHdVWVqLS3pDxeprn8uPJ6vWl90FYa1ZZXNm/eHO3R0t7ejpaW2H/8LVu24PDhwwCAc+fOoby8XK3pJU2JpQY3rfuruAQHEOkOS4BMKYverC9CsblmxWRkNxWZ0e/yg+VTd77lBTZSUqgrTHuypFFrhV5jXnS/g7IDaRpzAlYKDKWBzVCqqlcDQ2vBCWzaqkYGHJEl6sWuPQRBoKkkcnPqmzydljkpgT8kV64o0+RxdtlsnuxDtUjHjTfeiHfeeQd33XUXJEnCd77zHezfvx81NTXYu3fvgtt89rOfxUMPPYTPfvazkCQJDz/8sFrTS5pEb/wNxVcgxPlTuuFV2VvTYo2dLTQWmnGkZxy9Th9aS1JMLiMIrK+6VrWEwlhwAgtI0rwoV5D1YtI3CJuhFGZdQUbm5g05cGrwLVTYmlCfY70qREkEAUJVoUlTWkgQIYg8aEpdO3lB5MGJHAqM5UseL4WmSqyvuhal1npV56MkskdHou0lFsOUFx1ZjWqigyRJPPLII3Nea2xsnPe5Q4cORf+t0Wjw2GOxS24yiSSJ8AQnQRJkXOHLdJk95RKzk0lTFR00yWQsxOryj+FYzwuoL96I1rK54XCa0mJV+dVpszxfCAIUvCEH/OHijM1BLS5c+gD9jg5c1fhJWPXq/H4MpQVFMuBFVnXRQZE0rm78VMwOrARBLLslhahHh0KiQ0PrQZMMfHnRkZXkzcESRALwfvcfUGAsx7aGT6RlnyHOjyHnOVj0RStCxKjVbTbd6DWRHkULVbAwlAZ1RZlN9pPLuFkh96pXeCEMSRJBk+r5dGyoui6tS3YEQcTtOyKIPIZd51FgrIBJoWULtfCF3aBJDbS0Mu6ukcZvNoQ4f0560Cx38qIjQUiCBE1qwPKxS2aDrA/H+15Gha05pf4pnBBG13gbqgtWrwjR0Tzt1aFE2ewHPS9CkkRsa/hE2i8+WtoY8eq4THSEOD80lC6lxlZKQFMaECDjOpaXGzMls+qJjnQdT0POc/AEJ9BYshk6Jr5mmw7fCM6OvINyayM21iy8nJ0trK+6FiwfVPT73NbwibQ6EOeJn5WRmagwGloXV6OnMB+AP+yO29Nj8f3Jfgq5d3NYiMaiSPJfqo3fJEmEOzAOXmQz8rRDEMSCXh1nR47izXO/QJgLpH1OsyEIAhpahzCf2XmoQTocScN8EE7fCEKcP/aHk0SSJPROnMSQ63xC+WTF5mqYdYUY9XRnfbt7m6FE8YepvODIXvKiIwkYSgtWCMXMWpdFgiYFN1J5f5Hxci8MvhAWnQYlJl3KyysBdgqixKfVFOxy9BozWCEUNT4KcX5MTA1Az5gz1ixsNhpan5NilhPCoEhGVWv5Se8gPuh9ERPeAdX24fAPw896UG5tTMjVOFLJshkA0DPxkVrTSxlRFFSp/hFEHk7fCKaC6WnwmCd+8qIjCRhaB0kSIcRI6pLXyjVUaqKDJMio0FkpNBWZ0ef0gROSty2XLzgWfQZFx2U9WIZdFyBBQlVBdlQjlVhqUW5thCSlxx4+XfACq3rFEkPJ7sTqOV8OOM4AWLxMdilKLHUw6wow4r4If9ij9NQUoXfyFA6e/Qmc/lFFxw2yXnzQ+2K0zDhP9pAXHUkgNyViYyyxKBXpACIXOC4Hn0gXo7HIDF6UMOBKPnTtDUVERyYjHZX2Vmys3gstbYAkSRhynQNJ0Ci3NWdsTrNpLr0S66r25JwHTH3xBtQXbVB1H7KoSXX5dDECrBfjU/2w6ouTauJGEAQap6Md3ePZGe3wh90QRA46Or5clXgxaCwgQOQrWLKQ/MJXEjQUb0JN4dqY4U55OUSJZlpaWj+dkb8ysrFnl83K1SyJMpUFosNuLAVQCgCY9A0hyHpRaW/Jye6n2UQykYFEiYoOPnZ+VzIMOs4CSO13KbXUo8zagNIsTUD3hV0gCUpxEzeSpKDXWPJeHVlIXnQkQbwlaGZdAcqsDXFnnC9FJqovMkljoSw6prAPFUmNUWZtgFFjzZoOqhNTgwCAKvvqDM9kBpf/EkbcXaiyt8JqyD2/DjVh6IhwjCepPBkq7E0QJQFl1oakxyAIAptqblBwVsohSRL8YTeMWqsqkTaT1oZxbz9YPpQ114A8edGRNJIkQYIEcomTpdLeophRz0oSHIAy3War7K1Ahi0KOD6M93tegFVfjPVV16Lc1qiaWVUyBNgpDDrPwqovzhnREWR9ODX0Jsos9agtWqfafuRIB6+S6DDrCrG6Yodi40U6GUuK+WGkSojzQxB5xUzBLseotQHefvjDbmjoMlX2kSdxcmshN02MT/XjtY4fR8Of6SDE+THpHcpIC/RMoEaL+0xAUxoEWS/8YTcIgoDNUJJVAlJ+AsylslmWD8LlH0WQU/fYIQkK2xo+gTWVuxUdV5IkxasupoKTOHzuWVy8dFzRcVPBH7U/V+fJQBYz+byO7CIvOpKAImlIEGNWk5wbfQ+9E6cU2eeouxvH+16G2z+myHjZjt2gRaFBm3TZ7KDzHD7sfTmaTJopLnm64Q050DtxEm+d+yVG3V0Znc/l5KIHDCdOG4Op6EYqU2Ash1GbYn+gy3AHxvFu129x/tIxxcY06QqgY4wYdl9AkPUpNm4qGLRWtJZdhSJztSrjl1hqsbvlTlTasyNpO0+EvOhIAvnpcKlqEkmS0O84g0ueHkX3udLKZnucPghi4uWcLv8oHL4hkETmVhBH3V04OXgoGp2a8A7i5OChrBIecjk3m0ORDjmxM11N/kQp9W7Is5HLPItMyt2MSYJEY8kVkCQRPRPtio2bCgaNBfXFG5KqzIkHDa2DUWtV1aslT+LkRUcSMFRsAcCLLCRJTMjQZymiomOFLK8AkbJZThAx6E78hugNOUCRNAwZbKgmX9zlm5+8lp4tF30gN48rXpTdSNUXHcf7XsFrHT9WTHiEOD8ueXph0tpRYCxXZEyZclsT9BozhlznVHVRzSZEUchXsGQZedGRBJqoKdDiCWQzHh3KZE3PGBGtrEgHkHhehygK8IXcMOsKMuo/4QtF1qyNWhvMusLosZBNttQRYWYBlUMlvDN9V9QXHdT0U7RSBmFDznOQIKKmcK3iuT8kQaKxeDMkSURvFgjfdy/+Dh1Dh1Xdx4e9L+HIhf+GKCobjcqTPHnRkQQkSYEimSXXwWc8OpSKdOTe2nssGpMUHb6wCxLEjPpzADOl1SRBgqG1s15XJ1s/Wfa03oXNtTdlehqKoaONKDBWKFKqHgv5YYBXQHSIkoBB51nQpAYVKpnHVdgj0Y4wH1TFfjxeBInDVGhS9cR4OZnUz2anI+tKJF8ymySNJZujEY+FUFx0xLGkk2vMRDqmEtpuxhSsQPE5JUJD8SacHDy04Ot51KPC3oyKNCUPKulK6gu5IIgCKu2toCkm5fEWgiQo7Gi6HZPeAbzb9Vv4Qi6YdHY0FG9Cua1JlX0uBCtFlnfUqlyRkQV+IOzJ+PUgT4S86EiShuKNS74vSiK0tFGxmniKpLG94ZPQZUGTsHTRVJhcpENHG1BiqYPNUKrGtOJGvoj3TLTDF3LDpLOl/eIeD1PBSfhCLpRY6lS72eUqM6Ijda8Oi74I166+W/WlgEnvwIwYlgBvyBn9OV3HJidF8rTU8uiQMWhmymYzezXII5MXHSpRbmtEua1RsfEIgpi21F45FBq1sOqYhMtmi8zVqpXhJUq5rSnrRMblDDo7MejsxM7mO2Cmlv/T4JDzPDghhLqiDap7osh29kq5ktIkA5DqCr+eiXaIkgB/2AOSIKM3/p6J9rSLDrWXGk3y8ko+mTRryOd0JEnf5Ckc6/5jWrPAJUkCH6OzbS5BEASaiy3odnghiplbf851ZvKFcqOCZcB5FhfHjqfFhM1qKEVz6daU84e6x09gwHFG8fLbhfCFXCBAQhQFhLlAVDClM8GZjUY61F1e0WtMIAgyLzqyiLzoSJIg64UrcGnRC7XDN4zxqX5FLyIn+g/g4Jn9EEResTGzncZCM8K8iGFPfGWzIc6P9oGDGPP0qjyz3EGbY6KDF8Jp8+iw6AvRWHJFSvkCHB9G9/hH6J08BQLqCyWTzg6CIGaSLMNuSJKU1gRnA2lPS+NDgiCxsfp6rK7Yqep+8sRPXnQkCROjbLZr/ARO9B8AFLyI0NNh1xVZNuuIb4llKjiJS54e+KYtlvPEJtcqozghDHqZlACPurvwRudPMekdRJD1KmYmuBRyIjNNMdBpTBBFAQF2Kq0JzlaqGuurrk3LvsqsDaoZkOVJnHxOR5IwUVOlhS/ULB8EQ2mXbAiXKJpZ+9QxyraCzlZml81e1xS7aZM3C9rZLzdk0ZELfX0kSQInhFVPUJRh+RBODh6CPYmk5VF3F04OvAF3cBwgCEiSmJaEztkJzgAgSSJoSpPT15R4GnTmSQ950ZEkmhilciwfVKxcNrrPHHsijQc50hFvMqncKCsvOuJnxgp9+YsOYTrnKV3LKyRBTtvtkwAS69LbNdYGX9gNURSgZQxRI7t0JHTOTnB2+cdwvO9lBNgp2I3qd2Od8A5inDsLT6AmLZ2NJ7wDODnwBlrKtqGmcK3q+8s0kiTive4X4PKPgiQo7Gy+HRZ9UfT9zpH30DXeBgLAxpq9qC5YDVES8WHPi3D4hiFIPDbV3IDqgtWqzC8vOpIkaoW+gAAQJRGcEIZJ4brweOzXcw1ZdFyM06vDG3KCobRpMYbKFQxaM/a03KW4SM4EvMiCIMi0iQ6KZECAACeEkWjNiSc4AZYPgqIY6JkZu/50O9bajaW4tvVzcwzs1MTlvwSfOB61q1cbDaUHL3IrJpl0wHEWgsjh1o1/g/GpAXzY+xL2rrkXQCTn7fyl93Dbpr+DIPL4/Yl/QdXWVegePwFREnHLxi/DH/agb/K0avPLi44k0TGmadfD+SFJuRGcUn1XZOJpNJdrlJh0MGnpuCIdvMAhwHpQYKzIqvbx2Q5JUDBoLZmehiLoGBNuWvuXkJCeaieCIEBT2rhFBy+w4EUOOsYIu7EMkiSBpjRzjtdMONbKgkMQeYT5AAwa9Y4HuaV9upbA5C7AvvDKcCUdm+pDpb0VAFBiqYHDNxx9T8cYcdsVfweSoOALu6ChdSAIAiOui7Aby3DwzH5IALY33Kba/PILXEli0tmwreHjqCponfceK0y7kS7hWJoMFn0RWsuuSksINFsgCAJNhWZ0TXpj2jZzQhgFxooV9f0oBS+wCLCJ+aFkKwRBpHXtnpkWHbFw+IZx9OJvcGrwTUiShIbiTWBo7TyBnCnHWkHk8W7X79DW96pqBmWj7i70TJxESPTgeO8raem4TFMaaGnjiol0cEJozr2HIIg5VZQkQaFz5F28dPIHqCtaDwAI8X5MBSexd81fYH3VNXjn4vOqzS8f6VABo9aOa1o/p/jTttwKeqXRVGxB+4gLo1NBVFgXd2TVa0zY1vDxNM4sdzjW8ycE2CncuPYLmZ5KSoS5AAKsF0atVbFmi7FgKC2C3OKimBc4nL90DIPOsyBAoNLeAglS1jnWUiSNAmMFBp1n0T1+As1lWxUdf9TdhfaBNxDmAiBAwhdOnxOqSWuFwz8CXuSiVYC5CkPp5ohgSZJATjcmlFldsQMtZdtw8Mx+jLq7oaUNqCpYBYIgUGZtgCc4qdr88qIjBfonOwAAtUXr5rxOEiT0mtzNBE83s7vNLiU68iSPhtbDG3JAEHlQ5PK9LEx4B9ExfBjrqq5BlX1+FFIN7MYyaGgdpNB80eH0j+L00FsIsl4YtTZsqLpuTvJktjnWtpZtw4R3AD0T7Si11s9JQEyVnon26cT7uTfBdCTOCiIPT2ACB07/CDZDSVa2I1CKEkstBp2dqC/egPGpgTmRX09gAm39r+K6VfeAJCiQJA2CIFBqqcOw6zzqitbD6RuJOrmqQX55JQV6J0+izzE/4Ybjw2D5kOJdHCVJwoc9L+K0yu2gs43GOHuw9Iy3Y9DZmY4p5RzaHCmblZ/w0pVICgCryq/ClrqPRatPZHiBxYm+AwiyXtQXb8SOpk+npVojFWhKg3WVeyBBwumhw4qaG/pCLtCkBiRJg4J21uvqLnuMursw4RuEhtaDABHtNZOOpZ1MUFu4FhTJ4KWTP8CHvS9ia/3HcWb4CAYcZ2E1FKPAWI6XTv0AL596CsXmapRZG9BStg2SJOHFk9/Hu12/x9VN/0O1+S3fR5osgKG0CC6wDt47eRI9E+3Y3nCbovkFBEFgKuRAiI/PnTNXiJbNLmEQJkkiusZPwKi1qFbqlcvMtkI3aMwxPp29ZEJ0XI4cLaIpDdZXXQMNbVhWfZOKzFWotLdi2HUevRMn0ViyOaXxBJFHkPXCpLPDG3LCpi+G1+eLvq924mzPRDsYSjvvmEhnr5l0QhAkdlwmGmabo22quQGbam6Y8z5F0tjV8pm0zC8vOlJAQ+ngFZ0QJWFOuFAuo1WjBJGhdCuqegWYu7yyGAHWC1Hi8/4cSZIr/VdmREf6HEkvjh1H19hxeENejJ38ABRJY++ae0FTGpRa69M2DyVZVX4VAmFPyp2aQ5wfH/W/hiDrRWPJFegcfQ9Ic+KsLzTjTiyIPAiCBEmQaS9NzhMhLzpSgJnjEDrjCyFfuNVIZNPQOgRYDyRJWjFloeUWPfQMha6Jxb06pqYTn/KiIzm0OWI8x6c50jHq7sKZ4SMIhD0QJBF8wA+SpDDgOIuGksxUoSgBQ2mxvTG1ssmpoAMn+g8gxPlQYWtBdcEaaGg9eiba4fX6YNYVpCW3Qo6wCCIHT3ASDKWFWVsAkz79pcl58jkdKbFY/xVWCIIgSNCk8k9bspDhhfQY62QDBEGgqciMLsfiFQJR+3P98m/NngnsxjJsqLoOBabyTE8lJbhpwyk6TaKjZ6I92qRNggQtY4BVX4JRT+7kCwRZX8INFMen+nGs5wWEOB9aSrdhfdU1IEkK5bYm7Gy+Aw3aa7Cz+Y60LG/IkRSKYECTGnB8CCHel7HS5JVOPtKRAhpaB5KgwAtz282H+SA0lF6VSETUsloIps1BMBtoLDLj9Kgb474QSs3zl628ISeAfKQjWQwai6qGUOliVdlVCBWuB0Wk59LmC7nAUFpoaD1IUYoaXuVK6F6SRHzY+yJCnB87dXdEjbaWYsh1Hh1Dh0ESNK6ouTHjS0yzS5MlSPCFXSAJGvocON6XI3nRkQJNJZvRXHrlvNdZPgSjJvbJmQxWQzE4IQRihQWpmosiF4iuSe+CokOCBL3GrLgLbJ7lhUlnh0lnT+v+vCFn5P+cd9bruRG6JwgSLWXb0D5wEB1Dh7Gt4RMxH6bshjKYtHasr74WVn12VOvMLk2e9A3heO/LODn4BnY03Z7W/J88+eWVlLi8RA6IlLWuq9yDxtLUMr4Xo7pgNa6ovSlnbKvjpTFGMumVdTdjT8td6ZxSTiFKIo5eeB7tAwczPZWUULpMPRaLhehzKXRfZm1AqaUersAlDDjOLPgZjg/DP20zbtRasbP5jqwRHJdTZKqGvxkcAAAgAElEQVRCQ/EVCLJenB99P9PTWXHkIx0pwAscvCEnNLQuGnYkCALltsYMzyz3iKfb7EpJrFUDkiAR4vzL+juUJAmvn3kGdmMZttbfmpZ9zg7dpzM5Mt2sqdwJp38E5y99gGJLzZylOH/Yg7a+VyFJInY0fXpBa/dso6l0C0RJiNqA50kfedGRAr6wC8d6XkBd0XqsKr86LfsMsF4MOc/BbixDsbk6LfvMBpqWMAhz+S8hyPlQbK7OqD/DckdD65e1OZgg8RAlIe1Lj3Lovm2qDVuat6R13+lCSxuwumIHTg2+ieM9L4Oi6Gg+S5DzgSJp1BdtBL1MlipIgsSq8qsyPY0VSX55JQU09Pz29pPeIbx9/tcYcV1UZZ8sH0TPxEdw+IZUGT9bqbQaoKXJBQ3CBp3ncGrwEMLc8r1hZgNaWg+WD6Z9iUIp5Iqu/Bq9OpRbm1BkqoYv7II35ESI8+PSVC88gQlU2JrRWr496yMcCzHpG8IHPX8CL3KxP5wnZfKiIwU0C5TMhjg/AuwURElUZ58LCJ2VAEkSaCw04+LE1Lybojc0CZKgYVxheS5KEz22hOV5bMnnYbrKZVcaBEEgzEd8SIKsD/6wGwQImPUF0ZL15cikdxBO/yg6R97J9FRWBPnllRSgSAYEQU43MYoQbWuvUhXFTMlsajeGSIvpdvhCLph09mWxDt1YZMbZMQ8c/jCKTJEbpCgJ8IXdsOgKF0zszRM/s11Jl1sV0Ki7C50j78LpG4EoCigwlmX98bwckd09GUoLlmRg0tlBkfSyLhFuKdsGp/8Shl0XUGCsQKW9JdNTymnyV+kUIAgCGko7J+qgphspEPHIJwkqJSv0UXcXTg4egjfkhARp2TRAitqhz1pi8YfdkCQx78+hAAXGClQXrAG1zFp/y8ezLxy58YX54LI4npcjcjkyTTGw6oujHYmXc4kwSVDYVL0XNMng7MhR+MPLV0AtB/KiI0UYSjdneUUWIGo9KRIEAQ2tS2l5pWeiHUDEadAXcgHS3NezlaZZXh0yU0HZiTQvOlKl3NaItZW7ll3DN/m4pUgaeo05mtOR7cfzcmROKTCxyOvLEIPWgrWVeyCIPNoHDkIQ+UxPKWdRTXSIoogHH3wQd955J/78z/8c/f398z7jdDpx0003IRyeayPe3d2NLVu2zHs9G1lfde2c8rwwr+7yChDJJE8lYcsXckGSJATZKbB8EJwYnn49uxW+HOn48fszSbpyl19LPtKxYvEEJyGKQlR0yBUU2X48L0fKbU3YWH09zLoCECBh1hVgY/X1ObGUVW5rRJV9FQDMWTLPoyyq5XQcPHgQLMviueeeQ3t7Ox5//HE89dRT0fePHDmCJ554ApOTk3O28/l8+O53vwuNZnlkoFsNcw1wikxV0DGGaNhRDa5q/FRKokN2UTRqbfCH3QhxfjCUNutDpC+djVTsvN0zjocPnMRD+zaiqXQLaovWgyKpGFvniUWQ9aJ7/CMUmMpRYWvO9HRiIkkSRj1d8IdckDAtPGedFtl+PC9XZrt75hqrK3YAgKrX75WOapGOtrY27N69GwCwadMmdHR0zN0xSWL//v2w2WYuDJIk4Z/+6Z/w93//99Drl08imySJkKarVeqLN2B91bWq7i/VsjQ5FKqlDaAoBhwfgigKWR0iffjASfzbkXPRnx957RQePnASQKREkiSyV3Q8fOBkdK7ZjCgJGHKdg9M3mumpxCTMB9E+8DpODb4Jvca8YA5VNh/PK5mHD5zED0+NZ3oaC0KRdFRwTAUnEWAXNyPMkxyqyTmfzweTyRT9maIo8DwPmo7scufOnfO2+Y//+A9cc801WLVqVUL7unDhQmqTTQEX3weX0IdyZhP0ZHqerDgpBE7yQ0tYQBGRpL+2trYEtg+AEuzwiw5IIgVR4iAGNBjp9mAE8Y+TLn54ahz/1TE57/XvvvERphzncGdLDSgi/shYIt9Vqsye+8jICP56Q0na9p0ogsTBy3rR7+9BeMwIIL3fVbz4hAk4+AsQwEFHWFHMbEJY9MId7gcrBaAhDLBRtRk5nrPx+8om5pzL+19J6/kgC5149smKAQxzx6EhjKhgrsCPTk/Gva0a5NJxpZroMJlM8Pv90Z9FUYwKjsX44x//iLKyMvz2t7/FxMQEvvjFL+KXv/xlzH21tLTAbM5M8lvfJINzow401zSi0FSBsyPvoMBYgaqCVtX22TXWhq7xM1hddwuKzFVoa2vDli3xOyF2DL0Nt8uF3fWfgM1QirfP/xolljqsrcw+N8WHD5xcUHAAQIUlDEY/gIMOGl/fd3tcY42MjODpL9yc1DwA4KF9GxPaZvbc/6tjEhUVFQmNkU4kSYLrTAesehu2NG5J+LhKBywfwtvnn4VJMqKlbBtqC9dljSFVNn5f2UQmz4fZ+45nn5Ik4dQQh1F3Fw6P9+O/OoJxb7vQvoHErh2zufy48nq9GX3QThXVRMfmzZvx5ptv4pZbbkF7eztaWmLXPr/++uvRf19//fV45pln1JqeYjBRg7AQwnwQI+6LIAlKVdGRiokTL3AY9XRDx5hQYCwHQZDY03rXslzDLDFGHCh/cnwSL557BftWVWDfqgpsrS4ERc5dOXz4wEk88topAEDFdD5IvMzeFojv4nH5NjLya9koPCIl4Nlphc4J4ekW8jpsqL4eRq012kY+T/aTyfPh8n3Hs0+CILC2YheOdF+AP3ga91/FQEtJGJgYxndfv4Sv3rgvqX1n43mfblS709x444145513cNddd0GSJHznO9/B/v37UVNTg71796q127Qj9/pghZDqHh0ycmVMMl4do55uCCKH+qINUTOtbBYc8km60AWr1MTCptegoagCh7ocODYwiUdeO4UCgwY3tkQEyL7WCjz93oWELzoyyVywvvlqO771+ulF389m4aGl9fCz2VP1wQkszo2+C6d/FDub7gBNMSix1GZ6WnkSYDHBIaPm+ZCK2PnW62dx8ByLz23gYNfzuOTToNjIwuU7hu++jpjCI5lrx0pAtbsNSZJ45JFH5rzW2Di/++qhQ4cW3H6x17MNWWBwfHiW6FA3CVa2X08m0jHk7AQAVNrnRmLcgXFcHDuO5tItsBlKU5/kIiTjhLqY8NjbbMSqYh2+uu/jmApxeOPiJRw4P4xXO0fwXHsfnmvvW3TMeC4AsS5Y9+1ahQsTU7g4OYWuCW/k/5NedIxmz007UWTzJ1ES0r7vy4+NQmMVLk11I8T5YdEVgRVCoKnlZVy20hn3BnGkZywj+45H7Pz+9ADqCkzwhTlMhTh4wzy8YQ6TvhA4UcJfXBGAK8jArudg0/FwBCLH34Wxj2D5mgOVViOsOgYWHQOLTgOrjoFVz+CjISfe7pmfLJsXHnkb9JRhZgmAdHh0AAAT7b+SWBjcG3LAE5xAkbkaeo1pznuCyMHhG4KG0sJWo47okJ0jZ+bjjP6cqPB48Ka1aCo8DrO+EARBwKrX4NMbavDpDTWQJAlnxzz4ygvH8fqFxSsxHnntFP7j6DnU2IzQMzT0DAW9hoKeoXFh3INTS4iHR147teAFTUdTaCm2gBNFnB+fWnDbm1dV4MGbNiz5+2aKDdXXZWS/s48NSZJwydOD/skOmHQFWFOxEw0lm7K6Qmk5kkquQaxtPxyYxH8cPY//bu8DK4hgKAKcsHAjQZIAfGEeniALqz69VgmnR904PX2e6xkKFh0Ds5aBZNTikjeEIgMHHxs57gLczPFXaOCgpSh4QiwGXH6E+PhF+koXHnnRkSJa2oCmki2wGorhCUwASN/ySqKupAHWC4bSodq+et57BcYKGLU2XJrqxSouAC1jUGSus5EdIlk+DF4MQ8+YQRAEeiba46r7n32S/t3ucrzffRwWXcG8zxEEgbVlNlxdV7yk6ACAMC+g2+FDkOPBi4l1V20uMuPja6vQVGRBc5EZLcUWVFoNIMlIYuNCT1p6hsIr50Zw9y+O4unPXAWzLv/kDsx1D/WFnOCEMCiShllXgKbSfHKm0qSSa7DYtmFewPMn+/GDo+dxbCCStNlSbMH/3NmKz29twJOHO+edD3dsrMHxQQf+5fBZ/LytG498bBP+cnvTvJysRBn2BBDkBGgoEqywcPPNL+1owVevXweLjoFJQ4Om5ueBDUwMo9jIRoWHTI29FBOP3hn9meUFeEKRaMk/v9mBH72ft+BfjLzoSBGaYqIXRW/IBT1jhpZW/oY9Gw2lxVWNn4KOMSa0XamlDsWrqzHHQWkagiBQW7gWZ0fewaCzU5ULvS/kgigJ8IedkCQJnBCGRVeYkHOkfIGTJAnXtH4urs8uFmJ98KYNcy62nCAiyPEIcgKCnIDvvXkGT727cJb45dvGs/8Hb9qAv9zehD/7+RE8196H9mEnnrt3D9aX25ccJ50EwlNw+kdhN5aldb9yIzEA0GvMoAQGesa84ropp4NUcg0W2nYqxMGkpfHD9y5i3BcCQQC3rqnEfbtW4Ybm8qgIX+h8eGjfRoQ4AU8ePovH3ujAl39zDP/57gU8+amtuKZxfsQ1VoSlY9SFJ946i2c/6gMniCg167Cm1Io3u+Yu8cR7/n739Utw+Y5FX9MzIkrNFnxy4/VzPquhKRSbKBSbdPjPz1yNcosh7uvOSoP65je/+c1MTyJZWJaFw+FAYWEhtNrMt7O2G8tQV7Re9Q6dBEFAxxijds+jo6OoqKiIc1ty0RJDo9aOAccZeENO1BatVbxr69hUL1yBMQgCB4pkQJIUtLQeZn0BagrXJDQWQRBgKE20z8ZiXNsUuXke7o590aFIAlqagknLwKbX4JY1VXFvG2v/1zSW4qF9G2HRafDnVzYgwAp48ewQfvphNyqtBmyqnB+xyQTj3n50DB+GRVcIv5uL+7hKlbGpXrBcECAAkqTAUFoQBAGzzp7wsZEpEjkPM8VieQ7yMS4fr4ls+37/JN7uGQdDkfibna342ed24m92rkJjkXnetUYev1kv4Ht37AIA0BSJ3Q2luHdrIxz+MF47P4qfftiNM5fc2FpTBNv0kou8/8vnKkkS3uy6hP/5uw/wv//YhpMjLjQXmfGdW6/Aj+/cgS9ub57zOyZy/u5qbELbUAiDrgmUm8MoNxO4unEHGks2L7ldItedWFx+XGXbfS9R8pEOBTgzfBRB1osr6xP3f0gWSZLAi2y0eiYWp4cOw6yzo7Zw/aKig6YYVNpb0O84gzFPH8pt8xN/U6HYXIv+yQ5QJAOLvijyIkGgoXgT/GEPjFpr3GN5Q07oGXNciYWLPWHFQyrbXj6GDEOR+OfbtmBnfTG++Ot38cVfv4ujPeP4t09vhZ7J7CkpL92lu2y2zNqIIed5GLXWqJgG8q6iShJPYuWvP+pFU5EFgiRBECVI0//vcXgx4A4sOf6Xrm7Gt29d+mYMRM6Htrb5DdUqrAbs/7Od+PLOVvyvP3yI354awItnh/D316yBJEl4/NCZOXMVRQmry6x44q2zODHkBADsbijBV65dg1tXV0UjLPI+F/p3PHz1xn14+EAZRCKIKvupOVG5WL+nPFcgH+GQyYsOBfCGHHAHxjDi7gJDaVBsrlF9nx/0/AnuwBhuWvdXMT/rD7sx7DqPQmMF6oqWTmCsKVwHLWNCoalSqalGCfMBGHV2mLUF4IQwTDrb9E2FwJELz6GlbPt0Ke/SZk+SJOH97j9Az5ixq+Uzce1bPtlHRkZSEg1KXjQ+tb4G68vtuPNnb+OZD7rQNuTAc5/fg+Zii2L7SBRtNF8oCCA9T1GiJGJ8qg86jQk6xgRB5KPHRq72+MhWLkx4cWEiOetvDa1Mou+2miIcvf9j+NWJXnztpY/w2BsdC37u0YORsnSCAD69oQb/+9o12F5bvOBngdTOXXnbtj43JrwDcAfGYTPEdidV69qxnMmLDgWQE0dPDR6CUWtLi+hgaC0kSOAFNuZnB52RniVVBbHt5Y1aKxqK1Tk51lXuQYWtaZ6gmQpOQksbcOHSMfjDbqyt3LVkpUKAnYIg8gm3s1/sCSvebdWgsciMo/d/DP/rhQ/xw/cuYuuTL+PHd12N2zdEvChSdTNMFM2cyqj0iI6+ydPwBCdQV7gOG2tyx8Mn23jwpg14t3ccBy9eWvD9f7huDb62dz1IggBFRv4jCQIUQYAkif+/vTuPjqpOEz7+vVWVSmVfCdnZQliC7AIKbiyCOIqO+KIiqGgrtr4N40K3tLjBiEu34+ior3oc2m1GEW0bRQRBbOgGAkTCEpYQyUpCyJ5Utkqq7vtHTJFAJakktYbnc47nkErVrV9+3qr73N/yPJ2OlDj6Ll5RFBZOGMyxs5XtRjhseWTqMP7z1kkOe+/ODIhIoaQmj7yyY3YFHSDBxoUk6HCA1m2z4Pztstb30f56R9pFrg6LaqawIhMfrS/9gwfZfXyLaqa6vszuD5Y9FEWxOYIS7BfJFUm38nPOFs5UnKTeVM24xOvx0dm+6NU0lAEQ1EfK2Rt8tLwzfwrTBkWxdMNe/s+HO/ndVcMJ8tXx79vO3+W54svLel4116PB+Rk/axsrySo+gF5rYETsxfWYhGNUN5hY8tketp06S6Beh9HUPvjuycLo7ry2p+wZPQn3d926hojAePz1IZyt+oVhMZOdvn6vL3JaldlLib5N0OGqk/D8HWnnQUdxVQ4mcwNxYcloulH+fd/pb9l3+luamht71U6A44V7OFG0F7Ol41EGg08Ak4bcRP/ggZTXFrHnl79SbzLafG5NQ8v8bXAfCTpaLZwwmNRlcxnRP4Q3dp1oF3C0rarrTBqNFp1G75I1Hapq4UjB37GoZkbGTXP6VvNL1dGiCib9x3f89Uge1wzpz8mnbmmXJ6a7a5x6+tqeuPD9LuTqdRKtu/wMPgHUSwXaHpGRDgfwafNl2TYAcep7WpOSdX5xKKg4CUB8WPcq9/YPHkhlXTEFFScZ1K/niayq60vJLTtCgD6E5OjLO32uTuPD2MRZnCreT0XtWXx9bAdwNfW/jnT4ecaOD0caGR3KvJR4jhdXXfQ7VyUVmpZ8Oz5aX9IPOjfIaWiqw9RcT3TIYKJDBjv1vS5Vn6adZumGvdSZzDxx7Uj+fe44dFpNr9YauHqdgjtGWDqTEDGCxIgUjyk06G0k6HCAQN9QFBRUVNdNr1jTr3c+0pEQPpwgQ5g1vbW94sOGc6r4AHnlGQyMHNWj7bOqqnK8cDcAI2Kn2pVRUlEUkqMnYVHN1ucbGyoJNJwf6q9uKMNX5+f0fCju8PyWQ53OYbsi8Ohu/pee8tMHMnXofCwW16dc7+tMzWYe35jG2/88SZCvD1/cM41/Hd1+rZkjFla6iiftBJHMuL0jQYcDRAUP4LKE6zic/6PLgo5Q/yiGx1xBiH8UxWR3+Lye3kX66HyJDU2ioOIkJTUFRAV3f3FsUVUWFXVn6R88kMig+G69tvWDXWosIC17MwMiRzEsejKKomHCwNkeWQm1r2gym2hsqkVVbWdy7C1VVWlsrsXgE4hWo/PogoPeKL+ilgUf7SQ1r5RR0aF8ce81JLtxR5SjeNJOEFVVOVW8nzpTNWMTZ7q1Ld5GPu0OEh0ymIiAWJd9gQb4hrYp7X1x0GFRLTSZG3u1xiQxIoWCipPklWV0O+hoNps4WZSKRtEyLOaKHrfB3ycYf98QckqPcK4qF0XRUGeqItAQRlMf3FLZVRZVBdBrNZgtll6niu5I5tl95JcfI0Tt3pScvfLKMsgs3s+4xJlEBiU45T0uFRfubtqWWcTCT3ZRWtvIXeMH8f/mTybAt++k2nd3sNFKURQq685RXltoLVAo7CNBhwM0mU3klWXgrw92eEKtniqtyedg3g+kxE6za6usLcF+kYT5R9PQZMRsae5WQFVSk09jcx1JURPw1wf16P0B/H2DmTJkHv84tYGCihNoNFqCDJHdKhbnbToKPBaOH8SOrLM8vTmd746f4cO7pjI4oud925HWqTszXW/H7q46UzWZxfvQKNpub3kW7bXdwqqqKnqdlme+T0en0fBft01i6RXJsu7AiQZEpFBeW0huWQYpcdPc3RyvIUGHAxRWZnEg+zsURSE2NJkhUc6/A7dYzOzP3vTraMfFc/D55cdRVcv5zJ89NG7A9daU1N0REzoEP30QQTYKsnWXj9YXH40eX58AGptqqaorJsQ/Cq1GZ3exOG/T0Rx2eV0jv92QyheHchn35295bd5ElkxKcujFpXV0zKw2OeyY0HJhPFqwE7OlmZSEq/rkmhxXuTBnxuofWhJlxYf4s/6eqztNkiUco1/wAAw+gRRWZpIcPanLkgyihQQdvVRUmcWxM7swW1q+oGsaSl1yB64oGqrqS7CoZvQXBB0NTUZKavII9uvX66Cju9sYVVX9tX2KQ3N81DZWEqAPQavxwWwxoVVaTt3uFIvzNrbmsMP9ffnfRVdx86gEHv0ylQfX7+WbjALeu30KUUGOWU/Uui7J0SMdBeXHKa8tJCpoADEhnhcoujoRW091lqRrwbiBEnC4iEbRkBA+glPF+ymsyGRA5Ch3N8krSJ6OXjpdkt7uLlP5tUvblup2BkVR0OsMNvN0nKloqYya0MNplQs1Ntdxomiv9bidOVedy4GczdQ1VjvkvVsFGsJAAYOPf8vojtL6uPMTWLnTs7PHXHQRVBSFu8YP4tATNzE9KZpvMgoY/adv2Hg0v93znt9yqEe5PaxBh+q4oKPeVMOJs6noNHpGxk3zuGH/1gu5q/Kh9FRX9VP+/NMxj25/XxMfPhxF0ZBblmG94RKdk6Cjly4q/qO0Pu78O3AfreGijKSqqlJQfhKtRkdMiGPWl6iqSm7pUbJLDnX6wTJbmjletJvy2kJUHLvzoaPCX5dyQbCEsAC2PDST1+ZNpLqhiVvX/cSD6/dQ09DUq4uoM6ZXFEVDWEA0w2OvcNmWXHvZKtcuF25hD1+dHyNirmBU/NXuborXkOmVXgo0hFHTUE5oQH9Q2z7u/Dtwvc5ATUMZljZbG6vqS6hvqiE+bFi7ap29YfAJIDpkEEVVv1BeW0REoO3y3adL0mloMjIockybnTWO0TpVdbok3Zq3QwqCgUajsOzqEcxMjmHxp//gg9QsvjycS2X9+YChu7k9DD6BjEmYQc6pAoe10+ATwIQBcxx2PFt6Mj3S0ciBqxKxddezs8dQYmzgnd22Rx2lkqnrJUakuLsJXkWCjl4a3G8sh/J/bMkrobR/3Nlas59aOH+BCfWP4uphd9CuMQ6QGDGKoqpfyCvLsBl01JmqyS45hK/OnyFR4xz63q1iQpMu+SCjIynRoexZdgMz3/mBf+aUXPT77lxEtRod76UaKSys44rOk8h2qaHJiLGxksjAeKdOqVwYPHT0d6qqSna5kQP5Zbyx8zh7cks7PKYnBh655UY2ZtgOBiXgcK/axiq0Gp3HjeR5Ggk6esmdd+ChAdGoqDSea/+4v97xiYBC/aMINkRyrjqHepMRP31gu9+fKNqLRTUzLGaKw0ZYRPes3X7UZsDR6oWth7GoKs/P6TwgbnsBj91yqMcXMlVVOXpmF6U1+UwafBPhATE9Ok5XbE2PQMtFOK+ilrSCctIKyjiQX0ZafhkV9Y7fCuwKJcYG5ry3nTNVdbzyL+Mxmpo9IkOngNKaAg7kfMegyDEMi5ns7uZ4NAk6HMBdd+ADIlIYEJFCWkka0LKIU6NoiQiMc/hdpaIoJEakcPTM38kvP0Zy9PlS0g1NRsqNZwjzj3bYOhLhHC9tP8q2zCLGxIYzJi6McXHhjIoOxV/f8lXQegG/ZUQxMUEmXtja8rqeXNAKK09RWpNPRGAcYf7RjvwzrDqbHnl1Rwb1Te1TrCdFBnH9sFgmJkQwPj6crScLebmDtPM3jozzmAt5TUMTN76/ncySalZcl8Lj17Uf0veUdl6qwgNi0GsNFFScIKn/BMmy2wnpmT5CVVVOFO2locnIdcPv7rAsfG/EhA6hsq6Y/sGD2j1u8AnkqmELMJubPW5XwqWkq2ymo2NCURSFA/ll7G0zraBRFIZFBaNV4OjZKoZHGhkbU0O4XxP3jS9g/c/GdsfvTFFlFqdL0qmqL8XYUIG/PoiUuKudcl50tZOjvsnMiKhgFl8+hAnxLUFG2AVl0K9NisZXp73oOAadhk3HzrD0i738562X42tHiXVnaWgy86/rfiKtoJwlk5J48cbz05cSbHgGjUZLfPgITpccpKgyq8cJGR1BVS3s+eVvVNQWoVG0TB16W7vUCccL95B1Lg0FGJM4g4TwEdbfVdadY9Oht1gw+Wl0GudkspWgw4vVNlZRUH6CeksNFbVF1JmqiA0d6pSAA1rm+i9cpa2qKoqitCR6krPJ7eypyNnYbObY2SoOninnUGE5hworSM0txWS2MDzSyE3DS/DVtSxOjg40cdPwEtb/vLvd8W0pqsxqyVGjtuRVaTY30mwxUFVXbHdWWnsWgzaZLWw/VcTXR/K6PN7tYweyYnrn+RNsJWJbPHEw8//yd97fe4ojRRV8cc81xIa4PplZs9nCwk938WPWWW65LIF35k+WwN5DJUaMILskndyyDOLChrnt/1Ne2THMliZuHPNbzlXnsT97EzNG3gNAQ1MtJ8/u4eaxyzBbmvnrz68Rf/lwFEXB1NzAgexNTh+lkcuEF2toqiW79BBaSzj5FScAXBZhty6aSsvZTHL0JPoFdb8gnHCOripy+uq0jIsPZ1z8+Wyxz32fzuofjjAloQoAi6XlC7NfgAmzRbE+Xmos4GjBTqA1EVzLf6qqotG0jAZYsNDU3IBOq8egC7A7a2xni0HNFgu7Tp/j8/QcvjyUR1ldIwDBvjqqG5ttHq876xxsJWLb9X/n8NAXe/mfn7O5/D++Y/09VzN1kOMS3nVFVVUe3pDK10fyuS6pP58uvAqdVrIceCqDTyBRwQMprs6msq6YsADnTCl2pbg6h7iwYQBEBSdSZjzTpo0B3E8HhHMAAA4qSURBVDxuGRpFi7GxAr3OgKIoqKrK7qyvGD9gNtuPf+TU9knQ4cWs5e3VeoqrsgnQhzht7ryt/ac3WZOiqapKoCFCgg4P092KnM/NGYuiKNTV5gDQ0KzBX6+gUUBVIDbY0pJunZb8M4qioFFaL4AKiqJQU1+GRqNFQUGn9bUmcbMnZ42txaCqqjJnRByfH8zhi0O5FFW3VBbuH2Tg0WnDWDB2IFMG9GP1D4c7Hdmx14XP99fr+OiuqUxMiODJb9KY/vZWXr/lcpZe6ZqaJis3HeS/92UxPj6cr+67FoOPlFT3dAMiUiipyaOmodxtQUeTucG6sxFaPqsW1Wyt3K1RtBwv3E163jZGxF4JQHreNhLChxPeQToER5Kgw4u1Bh21lnMEqUG/Zsdz7pdhUWUWZ6tPW9O+a7U+FFWcIiooQbazepieXHT/a0cqBZVnaWjWUFjdMk3nr9dRWK1h/Gvfsu7OqfzLyLtsvv6fpzZQ01COoigEtynm1lXOmo7WZqz+4Yi1pkiYn54HpiSxYOxArhnSv12F3a5GdnpDUVryoIyJDWPBRzt59Kt9HMgv463bJjs1CPjzjgxe2ZFBcr9gNj0wnWCD7AjzBmEBMVw7fGG3y0c4ko/WQJO50fqzqqrWgKPViNgrSY6exLaMdRRV/sLpknT89cFknt1PvcnID0c/4IbRS53SPhmr82KlNQVU1ZVQb6nE2FiJxkkLf9o6XZKOj8YXza/zfgH6EFCcn/ZduMZt46YzMPz8duiB4YFMSohk8uAp1JqamffBDlZ8k0aT+eKMsz3JGtvVYlCAO8cNpPC5+bx7+xVMHxrTLuBo9ezsMTxz/WinbR29Nima/f92IxPiw/nL/l+49q0t5FfUtvs73jt8rpMj2G/dvixWfPszcSH+fP/gDIfV1BHO11qewp2iggdQ8Ot0+7nqvHYjLlV1Jfx4/GNrIKLR6FAUhdsmPskNox/ihtEP4acPZNao+53WPhnp8FJFlVkcKfipZdgMLYH6UI4X/gO9Vu/UEQdjQwUoEGQIx2IxW3Ny9OXCa5eSmNAk5o25mS8P/khzUzmXxSYyuN9Y5oQmMXlgOXd8tIs//3SMf2af43/uvooBbQKUnuSssadexdB+wejt2D3i7J0ciWEB/P3R2fx2QyofHTjN5a9v4vPF1/BT1tle5TVpu3j2b0fzeXD9XsL99Xz/4Ix2/Su8R3F1DmcqMhmbOOOiUQZnGxCRQmFlFpsOvQ3A1KHzyTiziyBDBIkRIwkPiGHT4bdRUIgLSyY6ZLBL2ydBh5dqHVnQKFrMWKwJSJ1d6r017btWo2u3yrmvF167lMSEJvHodUmkpaUxYegE6+NjYsPZt3wuD2/Yy/8ezGHCa5v44I4rmTcqod1r7Tn/VFXlh8witp862+nzPC3plZ+Pjv++40ouT4jk3/62nxlvb21b/aDbWUzbjvTklBv5PD0HPx8t3z4wnZHR8pnyVhW1RZyrzqG4Ktvl086KouHKpFvbPda24vfYxJmMTZzZ4etvv/wPTmsbyPSK12otNBfi3w9fTVCbx5074iCF1y5tQQYfPl44jXdvn0L9r/kjHv/bAUzN55NwdVbdVlVVvj9xhmlvfs8N721nd04JN6XE88Dki7+YPS3gaKUoCr+dNoyFEwZha5zG3oJxF04tfXTgNE1mCxvuvUbK03u5hPCRAOSW2U48dymTkQ4v1TricPHjzr07ksJrQlEUHpgylMkDIrnjo528vvM4u3Napls+OnDa5rZXVVXZfKKQ1VsPsS+vDIB5oxJYNWu0detubIi/16T1fn7LIT7cf7rD37+w9TDv7s5kYmIECaEBJIT6Ex8aYP33un1ZvLjt6EWvs6iwJ6eE64c5fxeBcJ4A3xAigxIorcmnur60XXKuS50EHV6qtdCcrcedTQqvCYDLYsJIXT6XR77cxydpp0l5eSONbRaYtm57nZAQwZofjnAgvyXYuPWyRFZdfxljYsPbHa+723w9XXldI5uOnen6iRfwxEJzovsGRKRwpiKTnSc/w0frS6AhTG7QkKDDa7UdcaipMRJkCJcTWrhcoK8Pf7nzSqoaTHxjo/pp65ZXRYH5Ywbw9KzLuCwmrMPjecuFtquU8607aSrrTeRX1pFfWUt+ZR0FlbVsOlbA4SJZeN3XNTWbqDfVYFHNhPn3p6ah3HqjeCl/T0vQ4cVaRxzSqtsv+BPClV7YethmwNHWQ1ck89Ztfav6pj0p58P8fQnz92V07PlAa83ccZ1uFfb0qSVhn+zSdPz0QaiqBetKf5y/2N/TSdAhhHC6qED35i5wlp4mJrMnYBHezdhQga/u4hwrl3p6AQk6hBC9Ys9UQ1++kLb+bYWFhd36O52ZSVW4n7sW+3s6CTqEEL12qd+5Pzt7DGlptgvPdfU6W/8W3s+di/09mQQdQgiHkDv3npE+6pskvYBtEnQIIRxG7tyFOE/SC1xMgg4hhENJsCGE6IikQRdCCCGES0jQIYQQQgiXcFrQYbFYeOaZZ1iwYAGLFi0iNzf3oueUl5dz/fXX09jYCEBNTQ1Lly7l7rvvZsGCBRw8eNBZzRNCCCGEizkt6Ni2bRsmk4nPP/+cxx9/nJdeeqnd73ft2sWSJUsoLS21PrZu3TqmTJnCJ598wtq1a3nhhRec1TwhhBBCuJjTFpKmpaVx1VVXATB27FiOHm1fUVGj0bBu3Tpuu+0262P33nsver0eALPZjK+vr7OaJ4QQQggXc1rQYTQaCQwMtP6s1Wppbm5Gp2t5y6lTp170muDgYABKSkp48sknWblypV3vlZmZ6YAWe7e0tDR3N8FrSF/ZT/qqe6S/7Cd9Zb++1FdOCzoCAwOpra21/myxWKwBR2dOnjzJY489xooVK5g0aZJd75WcnExQUFCP2+rt0tLSmDBBCr7ZQ/rKftJX3SP9ZT/pK/td2Fc1NTVefaPttKBj/Pjx7Nixg7lz55Kenk5ycnKXr8nKymLZsmW8/vrrDB8+vMvnWywWAOrq6nrdXm9XU1Pj7iZ4Dekr+0lfdY/0l/2kr+zXtq9ar3et1z9vo6iqqjrjwBaLheeee47MzExUVeXFF19k586dJCYmMmPGDOvzpk+fzubNm/H19eXhhx/m5MmTxMXFAS2jJe+8806H71FcXExBQecltYUQQoi+Jj4+nv79+7u7Gd3mtKDDFZqbmykrK8NgMKDRSMoRIYQQfZvFYqGhoYGIiAi7lix4Gq8OOoQQQgjhPWR4QAghhBAuIUGHEEIIIVxCgg4hhBBCuIQEHUIIIYRwCe9b+iraueWWW6yJ0eLj41m7dq2bW+R5Dh06xJ/+9Cc+/vhjcnNz+cMf/oCiKAwdOpRnn31Wdj610bavMjIyWLp0KQMHDgTgzjvvZO7cue5toAdoampi5cqVnDlzBpPJxMMPP0xSUpKcVzbY6qvo6Gg5rzpgNpt5+umnyc7ORqvVsnbtWlRV7VPnlgQdXqy1Ou/HH3/s5pZ4rvfff5+NGzfi5+cHwNq1a1m+fDmTJ0/mmWeeYfv27cyaNcvNrfQMF/bVsWPHuO+++1iyZImbW+ZZNm7cSGhoKK+++ioVFRXceuutDB8+XM4rG2z11SOPPCLnVQd27NgBwGeffUZqaqo16OhL55b3hkuCEydOUF9fz5IlS1i8eDHp6enubpLHSUxM5M0337T+nJGRYU2vf/XVV7N79253Nc3jXNhXR48e5aeffmLhwoWsXLkSo9HoxtZ5jjlz5rBs2TLrz1qtVs6rDtjqKzmvOjZz5kxWr14NQGFhIZGRkX3u3JKgw4sZDAbuv/9+PvjgA55//nmeeOIJmpub3d0sjzJ79ux2CXRUVUVRFAACAgIkFXMbF/bV6NGjWbFiBZ9++ikJCQm89dZbbmyd5wgICCAwMBCj0cjvfvc7li9fLudVB2z1lZxXndPpdPz+979n9erVzJ49u8+dWxJ0eLFBgwZx8803oygKgwYNIjQ0lJKSEnc3y6O1nQutra21VjYWF5s1axajRo2y/vvYsWNubpHnKCoqYvHixcybN4+bbrpJzqtOXNhXcl517eWXX2bLli2sWrXKOo0OfePckqDDi23YsIGXXnoJaKlDYzQa6devn5tb5dlGjhxJamoqADt37mTixIlubpHnuv/++zl8+DAAe/bsISUlxc0t8gylpaUsWbKEJ598kvnz5wNyXnXEVl/JedWxr7/+mnfffRcAPz8/FEVh1KhRferckjToXsxkMvHUU09RWFiIoig88cQTjB8/3t3N8jgFBQU89thjrF+/nuzsbFatWkVTUxODBw9mzZo1aLVadzfRY7Ttq4yMDFavXo2Pjw+RkZGsXr2awMBAdzfR7dasWcPmzZsZPHiw9bE//vGPrFmzRs6rC9jqq+XLl/Pqq6/KeWVDXV0dTz31FKWlpTQ3N/Ob3/yGIUOG9KnvLAk6hBBCCOESMr0ihBBCCJeQoEMIIYQQLiFBhxBCCCFcQoIOIYQQQriEBB1CCCGEcAkJOoQQ3ZKamsqiRYvc3QwhhBeSoEMIIYQQLiFBhxCixz788EMWLVpEfX29u5sihPACUtpeCNEjX331FVu3buW9997Dz8/P3c0RQngBGekQQnRbZmYmq1atYvHixQQEBLi7OUIILyFBhxCi2wICAnjzzTd55ZVXqKurc3dzhBBeQoIOIUS3xcXFMX36dCZNmsQbb7zh7uYIIbyEBB1CiB5bsWIF33zzDRkZGe5uihDCC0iVWSGEEEK4hIx0CCGEEMIlJOgQQgghhEtI0CGEEEIIl5CgQwghhBAuIUGHEEIIIVxCgg4hhBBCuIQEHUIIIYRwCQk6hBBCCOES/x/WbtIvR9KenQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of clusters based on silhouette score: [3]\n"
     ]
    }
   ],
   "source": [
    "# Silhouette Score for Hierarchical Clustering\n",
    "\n",
    "# Import ElbowVisualizer\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "model = AgglomerativeClustering()\n",
    "# k is range of number of clusters.\n",
    "visualizer = KElbowVisualizer(model, k=(2,30),metric='silhouette', timings= True, locate_elbow=False)\n",
    "visualizer.fit(cluster_df)        # Fit the data to the visualizer\n",
    "visualizer.show()        # Finalize and render the figure\n",
    "\n",
    "df3 = pd.DataFrame(visualizer.k_values_,columns=['centers'])\n",
    "df3['scores'] = visualizer.k_scores_\n",
    "df4 = df3[df3.scores == df3.scores.max()]\n",
    "print('Optimal number of clusters based on silhouette score:', df4['centers'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of clusters based on silhouette score: [2]\n"
     ]
    }
   ],
   "source": [
    "# Calinski Harabasz Score for Hierarchical Clustering\n",
    "\n",
    "# Import ElbowVisualizer\n",
    "from yellowbrick.cluster import KElbowVisualizer\n",
    "model = AgglomerativeClustering()\n",
    "# k is range of number of clusters.\n",
    "visualizer = KElbowVisualizer(model, k=(2,30),metric='calinski_harabasz', timings= True, locate_elbow=False)\n",
    "visualizer.fit(cluster_df)        # Fit the data to the visualizer\n",
    "visualizer.show()        # Finalize and render the figure\n",
    "\n",
    "\n",
    "df3 = pd.DataFrame(visualizer.k_values_,columns=['centers'])\n",
    "df3['scores'] = visualizer.k_scores_\n",
    "df4 = df3[df3.scores == df3.scores.max()]\n",
    "print('Optimal number of clusters based on silhouette score:', df4['centers'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of clusters based on DB score: [8]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFlCAYAAADyLnFSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3XdcU9f7B/DPTSDsvUQZgooIqLhFEQdaR9040LpqtbU/W1tRq7baoV9Xv3Y4Wq1+Oyzu0bra4sBRV1EpKkNwIQiyZM8Ekvv7A4kiCQmQ5Cbheb9efdXcJPc+OSJPzrnnPIdhWZYFIYQQQnQej+sACCGEEKIalNQJIYQQPUFJnRBCCNETlNQJIYQQPUFJnRBCCNETlNQJIYQQPUFJnZB6pKWloUOHDhgzZgzGjBmDUaNGITQ0FH/++WeTzjtmzBgUFRWpJb6JEyciOjq6SecdNGgQYmNj6xyPiIjA9OnTAQCbNm3C0aNHm3QdXfbbb7/hnXfekT5mWRZr1qzBiBEj8PTpUw4jI82ZAdcBEKLtjI2NcezYMenj9PR0zJo1C3w+H0OHDm3UOV8+X1O9Gt+ff/6J5cuX4/Tp0yq7hiwffPCBWs+vS8RiMT7++GOkpKRg7969sLa25jok0kxRT52QBmrVqhUWLFiAH3/8EQCQnJyMN998E5MmTcLAgQPx7rvvQigU4sCBA5g3b570fQ8fPkS/fv0gFovRvn175OXlAQAOHTqE8ePHY+zYsZg1axYePnwIALh58yYmTJiA8ePHY/z48Th16pRS8RUUFMDBwUH6+MCBAxg5ciRGjx6N2bNnIzk5GQCwbNky6WeQ9bjGpk2bMHjwYEyYMAFnzpyR+fqOHTtiy5YtCA0NxaBBg7B3716ZsW3evBmjRo3C+PHj8dZbbyE7OxsAcPv2bUycOBEjR47EuHHjcO3aNWkbTJo0Sfqev//+G0B1L3nq1KkYN26cdORAXju+bNGiRfjpp5+kj/fu3YsPP/wQpaWlWLBgAcaMGYNx48ZhxYoVkEgkSrQ2IBKJsGDBAhQXF+OXX36hhE64xRJC5Hry5Anr7+9f5/i9e/fYzp07syzLsuvXr2ePHj3KsizLikQiduTIkWxERARbXFzMdu/enc3OzmZZlmW//PJL9uuvv2ZZlmW9vLzY3NxcNioqip06dSpbVlbGsizLXrp0iR02bBjLsiw7Y8YM9uTJkyzLsuzdu3fZzz//XGZ83t7e7OjRo9nRo0ezAwYMYH19fdkLFy6wLMuyV69eZQcPHszm5uayLMuyR44cYYcPH85KJBJ26dKl7P/+9z/puV5+PHDgQPbOnTvsmTNn2BEjRrDFxcVsZWUl+/bbb7PTpk2r83ovLy82PDycZVmWjY2NZf38/NiKiopasT59+pTt2rUrKxQKWZZl2R9//JE9c+YMKxKJ2L59+7Lnz5+Xvn/kyJFsXl4eGxAQwN66dUva5j179mRTU1PZI0eOsD169GCLi4tZlmXrbceXXbt2jR05cqT08YQJE9grV66wv//+Ozt79myWZVm2qqqK/eSTT9jHjx/Xef/Ljhw5wk6bNo2dMWMG6+3tzSYmJtb7ekI0gYbfCWkEhmFgbGwMAFiyZAmuXLmCnTt34vHjx8jOzkZZWRnMzc0xZMgQHD9+HLNmzcKJEyewZ8+eWue5cOECUlJSEBoaKj1WVFSEgoICDB8+HKtWrcK5c+fQp08fhIWFyYzl1eH3q1evYv78+Th+/DguXbqEESNGwNbWFgAwfvx4rFmzBmlpaUp9zmvXrmHIkCEwNzcHAISEhCA8PFzma4ODgwEAvr6+EIlEKCsrg5GRkfR5JycneHt7Y9y4cQgKCkJQUBACAgIQHx8PHo+HAQMGAAD8/Pxw4sQJXLx4EW5ubujcuTMAoF27dujatSuuX78OhmHQvn17aVz1tePLPedevXpBKBQiNjYWJiYmyMvLQ0BAANLS0vDNN99g+vTp6NOnD2bOnAl3d3eF7XP9+nX83//9HwICAvDBBx/g8OHD0pgI4QIldUIaITY2Fl5eXgCAsLAwiMViDB8+HAMGDEBGRgbY51sqTJo0CStXrkSbNm3Qpk0buLq61jqPRCLBmDFjsGTJEunj7OxsWFlZITQ0FAMHDsSVK1dw6dIlbN26FREREbUSpSx9+vSBm5sbYmNjZQ4hsyyLqqoqMAwjjRMAKisrZZ7v5dfw+Xy5162Ji2GYOu8DAB6Ph927dyM2NhbXrl3D2rVr0a9fP4wePVr6nhr37t2DWCyuc7wmdkNDQ5iamkqP19eOL2MYBhMmTMCxY8dgaGiICRMmgGEYuLq64syZM4iKisI///yDN998E6tWrcKgQYPkfl6guq0/+OADsCyLmzdv4qOPPsJ3331XJ25CNIXuqRPSQMnJyfj+++8xe/ZsAMDly5cxf/58jBgxAkD1/WGxWAwA8Pf3BwB89913mDhxYp1zBQYG4o8//pDeW963bx9mzpwJAAgNDcXdu3cxfvx4rF69GkVFRcjJyVEqvvT0dHTo0AH9+vXDn3/+Kb1/f+TIEVhbW8Pd3R02NjaIi4sDAGRlZeH69et1zhUUFISIiAgUFRVBIpE0aYJfYmIiRo4ciTZt2uCdd97BrFmzEBsbC09PTzAMgytXrgAA4uPjMXPmTHTu3BmPHj3CnTt3AAD379/HjRs30LNnzzrnrq8dXzVu3DicO3cOp06dwvjx4wFU31tfvnw5AgMDsWTJEgQGBiIhIUHhZxIIBACqvyz897//RXx8PLZt29bwxiFERainTogCFRUVGDNmDIDq3qaRkRHCwsKkw8ULFy7E/PnzYWpqCnNzc/To0QOpqanS90+cOBHff/89Bg8eXOfcgYGBmDt3LmbPng2GYWBubo6tW7eCYRgsXrwYa9euxbfffguGYfDee+/BxcWl3viA6l7qqlWr4OHhAQ8PD8yaNQszZ86ERCKBra0tfvjhB/B4PEyfPh2LFy/G0KFD4eLigt69e9c5d//+/ZGUlISQkBBYWlrC29sb+fn5jWpHb29vDB8+HCEhITA1NYWxsTFWrFgBgUCALVu2YO3atfjyyy9haGiILVu2wM7ODps2bcLq1atRUVEBhmGwbt06eHh4ICYmRul2fJWDgwN8fHxQVVUFJycnAMDYsWNx/fp1jBgxAiYmJnB2dpZOwJs7dy5CQ0OltxfksbGxwTfffIOZM2fCz88PQUFBjWonQpqCYV8dIyOEEEKITqLhd0IIIURPUFInhBBC9AQldUIIIURPUFInhBBC9IROz36XSCQoLS2FoaEhrQslhBCi91iWRWVlJczMzMDj1e2X63RSLy0txb1797gOgxBCCNEoLy8vWFhY1Dmu00nd0NAQQPWHqykC0RzFxcXBz8+P6zB0ArWV8qitGobaS3nUVsp7ta1EIhHu3bsnzX+v0umkXjPkLhAIFJbO1HfN/fM3BLWV8qitGobaS3nUVsqT1VbybjnTRDlCCCFET1BSJ4QQQvQEJXVCCCFET1BSJ4QQQvQEJXVCCCFET1BSJ4QQQvQEJXVCCCFET+j0OnVCCCFEkf0xyVgfGYeErEL4OFlhWbAfQrt4cB2WWlBSJ4QQorf2xyTjjd2XpY9jMwqkj/UxsdPwOyGEEL21PjJO5vENkfEajkQzKKkTQgjRWwlZhXKOF2g4Es2gpE4IIURv+ThZyTlureFINIOSOiGEEL21LFj2bnBLg301HIlmUFInhBCilx7lFiMhsxA/TQ5AJ2cbGPAYMADMBQYI8nTiOjy1oKROCCFELy06dhNrzsbCRGCAmMUjIfzvNGwe1xMVVWJcTs7mOjy1oCVthBBC9M6pxKc4Hp+Gfp6OmNjZXXr83b5eGNSuBbzl3GvXddRTJ4QQoldEVWJ8ePQGeAyDTeN6gGEY6XMMw0gTeqVYgif5pVyFqRaU1AkhhOiVTX8n4l5OEeb18ULnlrYyX1NRKUbQ1giM2BmJ8soqDUeoPpTUCSGE6I2CchH+c/YO7EyN8MWwznJfZ2zIRw9XeyRkFWLpiX81GKF60T11QgghesPaRICTcwbhWakQtqZG9b52w6iuOP8gE99dScKwDq0wokMrDUWpPtRTJ4QQolf6eTphXEc3ha8zMTTAnmn9IODz8Nb+q8gqLtdAdOpFSZ0QQojOE0skCDt2Aw+fFTfofZ1a2mD9yK7ILqnAWweugWVZNUWoGWodfr99+zY2btyI8PBw6bGcnByEhYVJH9+9exeLFi1CaGgogoKC0Lp1awCAv78/Fi1apM7wCCGE6Ikd1+5j09+JKK6ows7JAQ167/uB3rj5JBeTu7SuNVNeF6ktqe/cuRPHjx+HiYlJreMODg7SJB8TE4NvvvkGkyZNQmpqKnx9fbF9+3Z1hVSv5rTfLiGE6JNnJRVY+dctWBob4j8j/Bv8fh6PQfgbgWqITPPUNvzu5uaGLVu2yH2eZVmsXr0an3/+Ofh8PuLj45GVlYXp06dj7ty5ePTokbpCq6Nmv93YjAKIJax0v939Mckai4EQQkjjrIy4hfxyET57rROcLEwUv6EeOSUVWHTsJioqxSqKTrMYVo03ENLS0hAWFoaDBw/WeS4yMhKnT5/Ghg0bAAA3btzAs2fPMHz4cNy8eRPr1q3DkSNH6j2/UChEXJzsvXIbYuqfD/GgQFjneFtrI+wd0abJ5yeEEKIed/PKMSsiGa2tjLBnuCcMeE0bPt8ck4Xdd3Mxpb0tFnZroaIoVc/Pzw9GRnVn93O2pO348eOYMWOG9LGfnx/4fD4AoHv37sjKygLLskrd35D34ZSVvP+uzOOPi0To1q1bo8+rKdHR0ToRpzagtlIetVXDUHspT5Vt9fEPZ8EC+CG0H3p5OTf5fN/7VeLGN39iX1IeZvTvgtfat2x6kE3walsp6sxyNvs9Pj4eXbt2lT7eunUrdu3aBQBITExEy5YtNTZhobntt0sIIfrip9A+2DS2B4JVkNABwMzIELunBcKQz8Ob+67iWUmFSs6rKRpL6idOnMCBAwcAAHl5eTAzM6uVtN9++23cuHED06ZNw7p167Bu3TpNhdbs9tslhBB90crKFO/181bpObu62GH1MH9kFpdj7kHdWuam1uF3FxcX6f30UaNGSY/b2tri2LFjtV5rZWWFHTt2qDMcuWpmub935Dryy0Vo72CJT4d2otnvhBCipb65mID2jlZqqwK3aIAPTiWl46/EdHTYcAyPckt0YmUUFZ95LrSLB94OaAcA+N/kAK3+SyOEkObsblYhlp38Fwt+uw5RlXpmqfN4DMZ3dEOlmMX9nGKdWRlFSf0l9mbGAICc0roz4QkhhHCPZVl8ePQGqiQsvhrTHQIDvtquteOf+zKPb4iMV9s1m4qS+ksGtm2Br0Z3g28L2RPnCCGEcOto3BOcvZeB19q3xGhfF7VeKyGrUM7xArVetylol7aXdHGxRRcX2XvvEkII4c7+mGSsPRuL+MzqRPual7PaV0j5OFkhNqNuAvd21N6OH/XUCSGEaLWaqp81CR0AFp+IVvu9bXkro4wM+KgSS9R67caipP6SnJIKDPr+NFb8GcN1KIQQQp5bHym72Iq6722HdvHAnmmB6ORsAwMeA78W1vB1skJ0Wi7ePRyllUvdaPj9JSaGfFx8mAVTATULIYRoCy7vbYd28ai1Gqq4ohKDt5/Bnn8f4YMgb/g526g9hoag7PUSM4EBjAx4OldBiBBC9Fl7B0uZiZ2Lqp8WxoY4OWcQkrKLtC6hAzT8XgvDMHAwM0ZOKSV1QgjRFl3lTGDmquqng7kxAj0dAQClwkpEJKZzEocslNRfYW9mhGe0Tp0QQrRCpViCvx9lQ8Bn4ONkBQMeg07ONtgzLVArioRN2X0Jo/53HkdjU7kOBQANv9dhb26MW0/zUVEphrGh+ooaEEIIUezw7RSk5pdift/22Dy+J9fh1PHJ4I648CALU3dfwp9zgzGgLbfbtVJP/RVBno4I6eSGCjWVHiSE6Jf9Mcnw33gCAfsS4L/xhFaXENU1LMviqwsJ4DEMPuzfgetwZOrl7oAjs/pDwgJjf7qAf9NyOY2HkvorPhnSCQdn9oe1iYDrUAghWq5m/XRsRgHELHSiNrguOf8gEzHpeRjfyQ2edhZchyPXkPYtEf5GIEpElRixMxL3coo4i4WSOiGENBJX66ebCwGfj97u9gjT0l76yyZ2dsfW8b2QVyZC9BPueut0T/0Vt5/m4eCtFIz2dUEvdweuwyGEaDFdrA2uSwI9HXFlwXCuw1DavD5eGNSuBbwcLDmLgZL6K+7nFGN9ZBxaWBhTUieE1KuFhQnSC8vqHOdi/bS+KRVWwszIkOswGqwmoYuqxJi25zISswuRmF2ksb3Yafj9FfZmRgBAy9oIIfUqrqhEibBS5nNcrZ/WFxlFZWj1xRGsPn2H61Aabc7BazhyJxXxmYUa3YudkvorHMyf76leQkmdECLfl+fjUFhRiZBObujkbAM+AxjyeWAAuFqbcR2eTtt6OQnFwko4WRhzHUqj3U7Pl3lc3fMtKKm/wkHaU6eqcoQQ2Z7kl+LrC3fR0tIEP4f2Qczikbg2xQdn5g0GwzCYsfcyCstFXIepk0qEldh+9R4czI0wvbsn1+E02t1sbuZbUFJ/ha0pDb8TQuq34q9bqKgSY/XwLrXu+/bzdMLyYD88zivFe79dV2sMNevjBUt269X6+J+vP0BBuQjz+3rDxFB3p335OMnec13d8y10t8XUxIDPg7OlCXgM15EQQrTVmz3bAACmd6876Wnla50QlfoMQ71bqu36Nevja9TcrwWgFaVTG6tKLMG3f9+FsQEf7/bx4jqcJlkW7Ffr76iGuudbUFKX4cmnIWAYyuqEENkGtG0htxyoIZ+HiLeD1fo7pL718bqc1K+l5CAlvxTvBHjB3lx376cDL75cbYiMR0JWAXycrLE02Fftfz+U1GWghE4IkSUqJQcO5sYKq5vV/A4pqhBhy6VELB3kBwO+au52siyL+EzZ92V1fX18P08nxH80GmYC/UhNr+7Frgn60XIqlphViPvPijDYy1mn7+kQQlRHVCXGzL1X8LSoHI9XjpfOv6nPpxG3seVSIiRs9bB8U4klEoQduwkJK/t5b0fZ93FVaX9MMtZHxiEhq1Ata6/ba+Az6DOaKCfD1xcTMPanC3hSULeoBCGkefrh2j3cf1aMmT3aKJXQAeCz1zrB1doUq8/cwbXHOU26fqmwEuN/voitl5PgYmUq8zVllVXILi5v0nXqU6vWvYrXXq87G4tLj7LAsnK+sRClUFKXoaYATU4JLWsjhAD5ZUKsOn0HlsaG+LQBPW4bUyPsmhoICcti+p7LKKpo3DK3gnIRBn5/GicT0jDYyxl3lozCnmmB6ORsAwMeg47O1gjydMSj3BL03RKhtg1F/nM6Vubxpq69TswqxIq/bmH5yRi6/dlENLYsQ00BGlrWRggBgLVn45BXJsL617tKfz8oq38bJywb5Id1kXFY8PsN/DKlb4Ovb2lkiDb2Fujc0hbfT+gFQz6vzv1almWx6vQdrDp9B4GbI3B09gD08XBs8LVkkUhY/C/qvtrWXn99MQEAsGigT5POQ6inLpMd9dQJIc89yi3G1suJcLcxw/v9vBt1js+GdkZ3Vzvs+zcZiXI2gZHlcV4JAIDHY/Dr1EDsmNQbhnIm3DEMg8+GdsaOSb1RWCFC8vP3NlVCZgEGfHcK7x6OkrvU19POvNHnzywqR/jNR2hrb4HRvi6NPg+pRj11GRzManrqlNQJae6MDPiY0NkdI31cYGzIb9Q5DPk87H4jEAXlInjLKUryql+uP8S8w/9g+4TemNWzjdxk/qq3erVDcDtntLatTrRVYkmjZt5XVIqxPjIO68/FoVIswfhObhjUtoXMojpPCspw4UGm3GV+9fn+ShJEYgk+7N8BfB71M5uKkroMtKkLIaRGKytThL8R2OTztHtpO85KsQQ8BjKTGMuy+PzUbfznTCxsTASN6gXXJHSxRILQ8EvwsDXHhpFdwWtAVa1SURW2XU2Ck7kxtozvidF+rgAAG1NBrbXXgZ6O2PnPfQzfEYlfpvTF5C6tlb+GsBLbribBztQIM7u3adBnJLKpNanfvn0bGzduRHh4uPRYTk4OwsLCpI/v3r2LRYsWYdy4cViyZAlyc3NhZmaGDRs2wNbWVp3hyeXbwhrXPxwBNxvalIGQ5oplWVx9nIO+KrovXSM1vxSTdl3EaD9XfDy4Y63nhFVizD14DXuik+Fha44/5g5q0hKvvDIR7mYV4PfYVDwpKMUvU/rWO9qQVybE/Zwi9HJ3gJ2ZEU7MGYQOjlawMH5RClfW2uuxfq4I+eUi3jn0D4LbtVC6cIyYZfFun/awNhHAVE/WpnNNbWMdO3fuxIoVKyAU1u7tOjg4IDw8HOHh4QgLC4OPjw8mTZqEffv2wcvLC3v37sXYsWPx/fffqys0hUwFBujmatfgCTGEEP1x4NZjBG09hQ1yqrc1lrmRAZ4WlePzU7dxPfWZ9HhRhQgjdkRiT3QyernZ4+qCYU1es+1gboxL7w9DP09HHLqdgmE7ziKvrO4IJMuy2PtvMnw2HMO4ny+g4PlmND3d7GsldHmCvZxx8b3XcHhW/wZVgrM0FmDVcH+EDaAJcqqitq9Gbm5u2LJlCz766COZz7Msi9WrV2Pjxo3g8/mIjo7GnDlzAABBQUGcJnWgetiqRFgFKxMBp3EQQjSvolKMj/+IgYDPw0R/d5We29bUCLum9sWQ7Wcw5sfzcDA3QmJ2ETo4WoFhGIzr6IbwN/qqrPCVrakRIt4ejFn7ruDQ7RT02xKBeX3a48eo+0jIKkRbOwsIDPiIzciHiSEfnw/t3KiKbp1bvhhZLSgXYUNkHD4f1hlGBrJHBp4WlsHJwpjuo6uY2pL60KFDkZaWJvf5c+fOoV27dvD0rN5ar6SkBBYW1aUXzczMUFxcrPS14uJU+00aAEYfvQ+GAY6Naafyc6tDdHQ01yHoDGor5TXXtvo14RlS8kvxhrct8h/fQ/Rj5d6nbHtZAghsaY5L6SXIfr7KJu556ddJHiZIuHO7EVHXb5GPKQRCW+xJzMOHR29Ijyc9X9Pe1toI/w1yRSvzCty5FdOka337byb2JuYhMj4ZXwa5wkJQO7GzLIth3/2BIpEEe4Z7wtiAEnt9GvLvkLObGMePH8eMGTOkj83NzVFaWgoAKC0thaWlpby31uHn5wcjI+UqPCmrxYUMPMgtRrdu3VR6XnWIjo7WiTi1AbWV8pprW+WUVODX347C1lSATW8Mho2S1eMa2l75557KPH4ouRwfhwQrfZ6G+LU7cG3t73iUW3e5m6mJCUb3D1DJdXZ2EqNi72X8dicVH1zOwh9zB8HF+sUcpR//uoT43AqM9nVB3149VHJNffXqz5VQKKy3I8vZ16P4+Hh07dpV+rhr1664ePEiAODvv//m/JeJvbkxSoRVqKgUcxoHIUSzvjh1G0UVlfj0tU5KJ/TGUFchF0VS8kvlXFf59fOKGBvysX96P7wX2B5xmQXouzkCcRn50uf3JOYCABYNUO82pM2RxpL6iRMncODAAQBAXl4ezMzMapUDnDJlCu7fv48pU6bgwIEDeO+99zQVmkwvlrXRWnVCmpPWtubo7mqHeX3aq/U6PnLWq/s4WevFdfk8Hr4d2wMbRnZFWmEZhmw/i13XH6DD+mO4mFYME0M+0gpkf8EgjafW4XcXFxccPHgQADBq1CjpcVtbWxw7dqzWa01MTLB582Z1htMgNTPfc0qEtYaNCCH6bfFAX4T192nQmu7GWBbshzd2X65zfGmwenuvmrwuwzBYPNAXLa1MEZWSg9kHrkmfK68U4409lwEGOr0HvLah2QlyOFBPnZBmY39M9XIuwZLd8N94AgdvP1b7NUO7eNTalKWTsw32TAtUe4Lj4rpTu3rg4sMsmc81dTMYUhut9pfjtfYtYW0igJeD8hP2CCG6p2Y70Ro124kC6u9ByirkoglcXFfePXt1zyFobqinLkcPN3vMD/SGu23jNyoghGgniYRFbEY+tly6i3cPRcl8DfUgVYurOQTNDSV1QojeY1kWCZkFKK6oBACUiCrR9as/8OHRmygSVsp8D/UgVWtZsJ/M4+qeQ9Dc0PC7HE/ySzEl/BKGtHfGZ0M7cx0OIXpvf0wy1kfGISGrED5OVlgW7Kf0EHGd9w7yQ1dXO5x/kIkLDzJx8WEWsoorcGBGECZ0doelsQAfD/aDu405Np6PlxZgeRn1IFWr5u9yQ2Q84jPz4dvCBkuDfWmSnIpRUpfDyICHayk5cLYy4ToUQvReY+5rl1dWoaBchL3Ryfjo5L+137un9uzuFhYmCO3SGs6WL/49fzHMHwBgKuBzMgu9Oaq5l99cCxtpAiV1OWxNjcAwQC5tv0qI2q2Xs2nKgt+u42R8Gn6Z0hcGfB4SswoxaNtpFJSLIKyS1HtOK2NDrH29Kwa2dYKXg2Wtuhgve7kHWbOdKPUgia6ipC6HAZ8HGxMBckpoSRsh6iZvZnRumQj7Yh5j07iesDMzgrmRASyNDOFuYwYrYwFsTAU4dCsFrIz3loqqMK+Pl1LX52oWOiGqRkm9HvZmxsihdeqEqJ2PkxViM+pOTPNysMCZeUNgbVK9/aeLtRkSl4+t9Zq7WSdkvpfuiZPmiGa/18PBzAi5pSJIJLL6AYQQVZE3M/qzoZ3hYm1W7/acNKuakBeop16PwV7OcLMxg1AshgmPmooQdWnKfW26J07IC5Sp6vEpLWUjRCNupefB2kSA6LDXG1Vzne6JE1KNht8JIZxbfeYOXt95DjeePOM6FEJ0GiX1elxPfYYlx6MR/SSX61AI0VuP80pwPC4N3Vxs0dPNnutwCNFplNTrkZhdiK8vJiAmPY/rUAjRW99dToKEZfF+vw5y15ITQpRD99Tr4WBWvac6bb/avDWlfCmpX4mwEj9G3YeThTEm+btzHQ4hOo+Sej3spXuqU1W55orLbTmbg/Cbj1BYUYkPgzrAyIDPdTiE6Dwafq9HTVLPKaGk3lytOyu7fClty6kaAgMePGzN8Y6Sld8IIfWjnno9HMxp+L05O/8gE3GZsrffpG05VeNOdpxAAAAgAElEQVStXu3wZo+2jVrGRgipi5J6PcwEBrA3M4IhnwY0mpPErEIsPfkvTiakyX0NlSBtOomEBY/HUEInRIUoW9WDYRhkrZqEo7MHch0K0aDj8U9wMiENQZ6OWD3cX+ZrqARp0yRlF6LN2t+x799krkMhRK9QT500exWVYvzvn/uYG9AORgZ8LOjXAR2crDDSxwUMw8DTzhwbIuMRl1kACcvijW5Uvayptl5OQmp+KQxoFIwQlaKkrkBsRj7u5RRhlI8LBDQ7V69IJCz233qMT/6MQWp+KSQsiwVBHWBsyMcoX1fp62pKkGYXl6PVF0fw6FkJh1HrvoJyEXbdeAhXa1OM83NV/AZCiNIoqSuwITIO+2Ie4/GK8XC1MeM6HNJI0rXmmQXwOf8UYzu6IuLuU9x4kgsBn4fFA3wwo0ebes/haGGC/m0ccf5BFtIKSuFiTT8PjfHz9QcoFVVhxZCO1FMnRMUoqSvwYga8kJK6jpK11rxm/+3J/q2x9vUuaG1rrtS5Qjq54/yDLPwem4r3+3VQS7z6TCyRYOvlRJgY8jGndzuuwyFE79DXZAWka9VpWZvOWh8pe615W3sL7J3eT+mEDgBjO7qCYYAjd1JVFV6zciopA4/zSjGtmydsTY24DocQvUM9dQXszF701IluSsgqlHn8cV7D7407W5ri3T7t0dbeAizLUq3yBhrWviVOzBmEdvYWXIdCiF6ipK6Ag/nzUrEl1FPXVT5OVtLh9trHG7fWfMv4nk0Nqdni8RiM6NCK6zAI0VsKh98LCwuxYsUKzJgxAwUFBVi+fDkKC2X3fPSR/fOeOg2/664hXs4yjzd1rblEwjbp/c3NkTspSCso5ToMQvSawqS+cuVKdOzYEQUFBTA1NYWjoyOWLFmiidi0QjcXW9xaPBKLBlCxEV1UJqrCkTup4DGAl4Ml+AzQydkGe6YFNmmt+cbz8Wj1xWFkFZerMFr9lVNSgel7LmPoD2fBsvRliBB1UZjU09LSMHnyZPB4PAgEAixcuBCZmZmaiE0rmBsZoqOzDaxNBFyHQhph7dlYpOSXYvEAX9xdNgbXpvggZvHIJhePMeAxyC6pwNG4JyqKVL/t/Oc+hFUSvBPgRfMQCFEjhUmdz+ejuLhY+g/x8ePH4PGUmzR/+/ZtTJ8+vc7xO3fuYOrUqZgyZQoWLFgAobB6EtrYsWMxffp0TJ8+HcuXL2/I51CrSrEEeWU0UU7XJGYVYuOFBLham2LFkI4qPff4TtV7fx+5naLS8+qjSrEE264kwcLIELN61l8LgBDSNAonyi1YsADTp09HRkYG/u///g+3bt3C2rVrFZ54586dOH78OExMTGodZ1kWK1euxObNm+Hu7o5Dhw4hPT0drVpVT54JDw9v5EdRn1afH4aThTFiPxrNdSikASolEnR0tsbHgzvCzMhQped2szFDTzc7XHiYhWclFbB/Xs+A1HX4dgqeFpXj/X7esDSmES9C1Elhl9vBwQE//fQTNmzYgJCQEBw/fhwDBgxQeGI3Nzds2bKlzvHk5GRYW1tj165dmDZtGgoKCuDp6YnExESUl5dj9uzZmDFjBm7dutWoD6QOdmZGNFFOB3V0tsE/HwzHWDWVIg3p5A6xhMWxeBqCr8+WS4lgGOC9wPZch0KI3lPYU1+4cCH++usvpRL5y4YOHYq0tLpbV+bn5yMmJgYrV66Eu7s75s2bBz8/P9ja2uKtt97CxIkT8fjxY8ydOxcREREwMFC86i4uTnZxEVUxQSUelApx4+ZN8LT0fmB0dDTXIWiNEpEYz8qr0NpKdnETVbWVF08EAPj5Uiz8DYpUck5t09S2qqiSwBwiBLWyQGHKfUTr+d0K+neoPGor5TWkrRRmzLZt22Lr1q3o3LkzjI1fDDH26NGjUcFZW1vD3d0dbdu2BQD069cPcXFxmDlzJtzd3cEwDDw8PGBtbY2cnBw4O8tejvQyPz8/GBmprzqV++0i3M5Jg2eHjrAz074qWNHR0ejWrRvXYWiND4/ewParyfjr7WAMbNui1nOqbKtuAL6UWKKvpyO6uTuo5JzaRFVtdbrXi73T9Rn9O1QetZXyXm0roVBYb0dWYVIvKChAVFQUoqKipMcYhsGvv/7aqABdXV1RWlqKlJQUuLu74+bNm5gwYQIOHz6Me/fu4fPPP0dWVhZKSkrg4KAdvyhf1H+v0MqkTl64lZ6H7y4noY2dOfq0Vv/Pz6KBtNRRnpcr7ul7QidEWyhM6jUT10pKSiCRSGBpadmoC504cQJlZWWYPHky1qxZg0WLFoFlWXTp0gUDBgyASCTC8uXLMWXKFDAMg7Vr1yo19K4J0vrvJUK0d+Q4GCKXRMJi/pEoSFgWW8b3hJEGt8pNLyxDKytTjV1PF3wWcRs3nuTih4m94UabIRGiEQqz5pMnT7Bw4UI8efIELMuiZcuW+Pbbb9G6dWuFJ3dxccHBgwcBAKNGjZIeDwgIwOHDh2u9ViAQ4Kuvvmpg+Joxxs8VrW3N4Wmn/MYfRPN+uv4A/6Q8wyR/dwxp31Jj15206yJOxKch84uJsNKDegavblO7LNivwev6KyrF+OHaPUhYVvqlmBCifgpnv3/66aeYM2cOoqKicP36dbz99ttYuXKlJmLTGr3cHfB2gBdaUk9Maz0rqcDyP/6FhZEhvhrdXaPX7tzSBiKxBCcT6k4M1TU129TGZhRAzFZvU/vG7svYH5Os9Pv9N56AxfK9eFYqRN/WjjAVaMeIGyHNgcKknp+fj2HDhkkfjxgxAgUFdTfHIIRLDMNgpI8LVg3rrPEvXyE1hWj0YDtWedvUboiMBwDMPXANC367jnVnY/Hz9QeISEzH7ad5KCwX1fpCUFMW/0RCmtJfCAghTafwK7RAIEB8fDx8fasnBMXFxdUpKKPv7ucUYdqeyxjr54rlg1VbmYyohp2ZEX6e0peTuuLeTlbwbWGFiMR0FFdUwsJYtYVuNKVKLEF8puwv7AlZBRBLJPjlxkNIZLTxiiEdcUxOydwNkfFNLstLCFGOwqT+8ccf4/3334e1tTVYlkVhYSG++eYbTcSmNQR8Hm4+yYW3oxXXoZBXVIklOHs/A0PbtwTDMJzVFQ/p5I5Vp+/gj7tpOpvAJu66CHkbz/k4WYPHMEj7LAQZReXS/zKLq//fp7Uj1snp5Sdk0cgeIZqiMKn7+/vj1KlTePz4MSQSCVq1agVz8+Y1Yaxmos8zqiqndb67koSwYzfxzZjuWBDUgbM4Qjq5YdXpOzgW90Rnk/qc3u2QXy7CpUfZdZ5bGuwLhmHgZGECJwsT+MvYEl3V+9YTQhpO4T31P//8E+PHj0e7du1gYmKC119/HWfPntVEbFrDVGAAYwM+npXSpi7aJL2wDJ9G3IKtqQBTu3KbSH1bWOPYWwPxv0kBnMbREFeSs9FvSwQyi6q3j33dxwUX5g/FnmmB6ORs0+BtapcF+8k83tR96wkhylPYU9+2bRt+/vlnANX13H/77TfMnj0bgwcPVntw2oJhGDiYG1FPXcssPn4TJcIq/DCxN+cbqtRM1NMFJcJKfPJnDL67kgQAiEh8Wmv3tNAuHgjt4tHgql81iX9DZDwSsgrg42SNpcG+OjtyQYguUpjUKysrYW9vL31sZ2fHyWQkrtmbGeNejn7W99ZFZ5Ke4uCtFPR2t8fsnm25Dkcqt1SIOxn5dcrTaotTiU8x7/A/SM0vhbejJXZMCkBfD9VVVKr5QkAI4YbCpN6tWzeEhYVh1KhRYBgGf/zxB/z9/TURm1YZ0aEVOjpbQyyRgK/kfvJEPYRVYrz/23XwGAbfhfTSqhKkQVsjkFZYhswvJsLEkLv12dICMlmF8HGywrJgPzwtLMeSE9Ew4DH4eLAfPhncCcaGmqu6RwhRP4W/dT777DOEh4fjwIEDMDAwQI8ePTBlyhRNxKZVVg1vfl9ktM3LicrR3BijfFrBv5Ut12HVMtrXFV+ej8fppAyMUdOWr4rUrBevUVNA5usx3RDo4YhN43poXbsRQlRDYZdTIBBg1KhR2L59O2bOnAkDAwNIJBJNxEaIVK1KZxIWGUXlOBavfYVNQjrXFKLhbo9ReQVkfrn+CBffG0oJnRA9pjCpf/bZZ/j222/x4MEDLFmyBPHx8VixYoUmYtMqlx5l4YPfr+PO03yuQ2mWFFU60xbdXGzhbmOGE/FpEFaJOYkhIatQznFaL06IvlOY1GNjY7FmzRr89ddfCAkJwdq1a5GcrF29I02IzyzE1stJiM2gpM4FXUlUDMMgpJM7iioqcfZeBicx+DjJLpJE68UJ0X8Kk7pYLIZEIkFkZCSCgoJQXl6O8vJyTcSmVRzMqwvQ5NJadU7oUqIK6ewGAPg3LY+T69N6cUKaL4VJfezYsQgMDESrVq3QuXNnhISEYPLkyZqITavYm1Wvg86hteqc0KVE1cvNHo9XjMfK1zpxcv2RPi4wM+SDz2NgwGMaVECGEKLbFM5+f/PNNzFz5kzwni/j2r17N2xtm99EGwdpqVjqqXMhtIsHKsUSzNp3FQyAjs42WlvYhGEYuNqYcXb9Pf8mo7RSjM9e64RPh3bmLA5CiOYptZCW99K67OaY0IEX9d9zSiipcyXI0wkAMKWrB8LfCOQ4mvqJJRKcvZeJzOJyzOzRRvEbVOjNHm1gYsjHcG8ZBdoJIXqNu+oYOsbOzAh2pkYwNqDCM1zJLK6ey9HCQvu3/mXA4K0DVyGsEmNqVw8Y8jX3cyMw4GNGd81+kSCEaAfKUEri83jIXj0Ju6f14zqUZqutvSX2zwjClK6tuQ5FIR6PwbiObsgrE+HiwyyNXfeX6w9RUC7S2PUIIdpFYU89ISEB27dvR2FhYa2a77/++qtaAyPkVXZmRpj4vLiLLgjp5IbvryThyJ0UDPZyVvv1riZn460DV/HnXTccnNlf7dcjhGgfhUl96dKlmDx5Mtq1aweG0Z4a21y4+SQXD54VYVLn1lpVb5xop36ejnAwN8LvsanYOr6n2vcM+PriXQDA/EBvtV6HEKK9FCZ1Y2NjTJs2TROxaL21Z2NxLO4Jhni1hN3ziXNEcz46EY2jsU/w59uD0NbekutwFOLzeOjobINz9zNh/NFe+Lao3lhFHTP2HzwrwtG4VHR3tUOQp+p2XSOE6BaFST0wMBDh4eEIDAyEkdGLRNayZUu1BqaNagrQPCutoKTOgeS8EjzMLYalkSHXoShlf0wyzt3PBABIWFa6sQoAlSf2TX8ngmWBhf07NPsRNUKaM4VJ/dixYwCAn3/+WXqMYRhERkaqLyotJS1AUyJEe+oMaVxWUTl4DKMzX6jqq1evyqSeVybELzcewM3GDBM66c6cA0KI6ilM6ufOndNEHDrhRQEaqirHhaySCjiaG+vMfvaaqlefWypEbzcHvO7TCgYaXDpHCNE+cpP6li1b8P7772P58uUyn1+3bp3agtJWdtJSsVSAhguZxeVoY2fBdRhK83GyQmxG3QSu6nr17RwscebdIbVWpxBCmie5Sd3Xt7qmds+ePTUWjLZ7sakL9dQ1rVRYiRJhFZx0oPBMjWXBftJ76C/7IEh1s9MrxRJpYRu6l04IkZvUvb298fTpU/Tq1UuT8Wi1QA9HJCwdjZaWplyH0uyIWRYfBnVAOwftn/Veo+a++YbIeCRkFcDGRICcUiEi72diZo82TU7CLMsiYNNf6OhsjZ9C+1BSJ4TIT+rTpk0DwzAQCoXIzc2Fq6sreDweUlNT4ebmhoiICE3GqRXMjQzR3lH2FqBEvSyNBfhqTHeuw2iw0C4e0uReKZag/9ZT2PtvMoLbOWNWz6aVcj2dlIGY9Dx4O1pSQieEAKinTOy5c+cQGRmJHj16IDw8HKdPn0ZERAT2798PLy8vTcaoVcpEVcgpoeF30nCGfB72TAuElbEh3v89CnflTKRT1tcXEwAAYQN8VBEeIUQPKJwq+/DhQ3Tv/qKH1KlTJyQnJyt18tu3b2P69Ol1jt+5cwdTp07FlClTsGDBAgiFQkgkEnz66aeYPHkypk+fjpSUlAZ8DM1xWHkQr+9sfsv5uPZHQhpm7L2MmLQ8rkNpEg87C+yYFIB+nk6wMxU0+jy3n+bh7L0MDGzrhK4udiqMkBCiyxQuaWvRogU2bdqEESNGgGVZHDt2DK1bt1Z44p07d+L48eMwMak9sYllWaxcuRKbN2+Gu7s7Dh06hPT0dDx48AAikQgHDhzArVu3sH79emzbtq3RH0xdHMyNaE91DkQ/ycWe6GTM1IPdxyZ0dkdIJ7cmDZl/87wk7ML+1EsnhLygsKf+3//+F0VFRQgLC8OiRYtQVVWl1HI2Nzc3bNmypc7x5ORkWFtbY9euXZg2bRoKCgrg6emJ6Oho9OtXvQOav78/4uJkF+7gmr2ZMXJo9rvGZRZXt3kLS92Z/V6fmoT+1910HI970qD3FpaLcPh2CrwdLWnPdEJILQp76lZWVli5cmWDTzx06FCkpaXVOZ6fn4+YmBisXLkS7u7umDdvHvz8/FBSUgJzc3Pp6/h8PqqqqmBgoHjLd01+ARCIhSgTiXEl6oZW7a0eHR3NdQhqlfQkAwCQ+TAJFemKfybqoy1tVVBRhYnH74PPMNg93BMtzZUfjt833APPyqsQE/OvGiPUnrbSFdReyqO2Ul5D2qreJW0Mw4Bl2VrDhDWP796926jgrK2t4e7ujrZt2wIA+vXrh7i4OJibm6O0tFT6OolEolRCBwA/P79adenVyfNuGaIyS+Hq5QM3GzONXFOR6OhodOvWjesw1Ep4JQcGvBIMDOjZpB3ytK2tNhvYYe7Ba1h/uwAX5g+VrjlXRBOfQNvaSttReymP2kp5r7aVUCistyMrN2smJiaqNrLnXF1dUVpaipSUFLi7u+PmzZuYMGEC3NzccP78eYwYMQK3bt3S2hn2DubVVeWelVZoTVJvDjKLy+Fobqx3W96+2bMNzt3PwL6Yx/gs4hbWvt613tcfj3sCY0M+hng50zI2QkgdCrvCW7dulXn8vffea9CFTpw4gbKyMkyePBlr1qzBokWLwLIsunTpggEDBkAikeDKlSsIDQ0Fy7JYu3Ztg86vKZP8W6NTSxu4WlNC1yQPW3OYCpo27K6NGIbB9xN64XpqLjaci8eAti3wWnvZOyCKJRKEHbuJjKJypH4aojMb2xBCNKdBvyUrKytx6dIldO7cWanXu7i44ODBgwCAUaNGSY8HBATg8OHDtV7L4/GwatWqhoTDiYDWDgho7cB1GM3OmXeHcB2C2lgaC7Bvej/03RKBLZcS5Sb132OfIDmvBO8EeFFCJ4TIpDCpv9ojnz9/PmbPnq22gAhpjrq52uH0O4PR291e5vMsy+LrCwlgGODD/h00HB0hRFc0ePp2aWkpnj59qo5YdMKdp/nosvEkNp6P5zqUZuNxXgm2XUlCXEY+16GoVVAbJwgM+ABQp2rh1cc5iEp9hpE+LvDSofr3hBDNUthTHzRokHRCDsuyKCwsxJw5c9QemLYy4DG4k5GP3rmye1RE9W4+ycV7v13HN2O6w8/Zhutw1O7HqPv44PcbOP3OYPTxcATwUklYKjZDCKmHwqQeHh4u/TPDMLC0tKy1nry5qZn9nlNCVeU0Jft54Rld2na1KdrZW0JYJcEbey4jOux12JgI0NPVHlViFv08HbkOjxCixRQm9ZYtW2Lfvn34559/UFVVhd69e2PatGng8bSn8Iom2ZoKwDC0p7omZRaXAwCcLIw5jkQzgto44dPXOuHzU7fRZs3vKBVVwcfJCsuC/WgZGyGkXgqT+pdffomUlBSEhISAZVn89ttvSE1NxYoVKzQRn9bh83iwNTFCDtV/15iapN6imfTUAaCNffVoWFFFJQAgNqMAb+y+DODFPu2EEPIqhUn9ypUrOHr0qLRnPmDAgFrL05ojezMjPKOeusZkSYffm0dPHQC+PCd7IuaGyHhK6oQQuRQmdbFYjKqqKggEAuljPp+v9sC02fhObsgrE3EdRrORU1IBAZ8Ha5PGb1WqaxLk7LWekFWg4UgIIbpEYVIfNWoUZsyYgddffx0A8Mcff2DkyJFqD0yb/WdEF65DaFYuvT8UeWWiZnU/2cfJCrEZdRO4j5M1B9EQQnSFwqQ+b948+Pj44Nq1a2BZFvPmzcOAAQM0EBoh1fg8nnTVQXOxLNhPeg/9ZUuDfTmIhhCiK5QqExsUFIQ2bdogPj4enp6e6o5J6529l4Hf7qTivcD28GlBPSd1ElWJEZ9ZCFdrU9g3o8Rec998Q2Q8ErIK4ONkjaXBvnQ/nRBSL7lJPSkpCStWrIC9vT2mTZuGBQsWwM3NDenp6Vi6dClCQkI0GadWicvIxw/X7mFIe2dK6mqWnFeC7t/8gbd6tcWOSQFch6NRoV08KIkTQhpEblL/9NNP8c4776C4uBjvvvsudu3ahS5duiA9PR3z5s1r1kndzqymAA3NgFe3zOcz35vTcjZCCGksuUm9vLwcgwcPBgBs374dXbpUTw5r1aoVDA0NNROdlnIwr94hK5fWqqtdVjMrPEMIIU0htyzcy8vWTE1Naz3Hsqz6ItIB9jU9dVqrrnYvkjr11AkhRBG5PfWCggIcPXoULMtK/wy82NSlOXN4vpf1M+qpq11mMyw8QwghjSU3qffu3RtRUVF1/gwAvXr1Un9kWszezAhOFsYwNVRq8QBpgqxmWCKWEEIaS25WWrdunSbj0ClmRoZ4+vlErsNoFhYP8MUoX1e4WptxHQohhGg96moSrebtZAVvJyuuwyCEEJ3QPPdPVYGrydnYE/2I6zD0XnOflEkIIQ3RoKReUlKC+/fvqysWnbLq9B3M2HsFZaIqrkPRWyzLwuaTAxj303muQyGEEJ2gMKkfOnQIy5YtQ15eHkaMGIEFCxZg+/btmohNq9nTDHi1yy8XoVhYyXUYhBCiMxQm9X379iEsLAwnT55EcHAwTpw4gdOnT2siNq1Ws8EI7auuPplFz2e+W9LMd0IIUYZSw++Ojo64ePEiBgwYAAMDAwiF1Dt9kdSpLdQl63kZXidzSuqEEKIMhUm9bdu2eOedd5CWloaAgAB8+OGH6NSpkyZi02p2z4ffqf67+tT01J0sqfAMIYQoQ+GStrVr1yImJgZeXl4QCAQYPXo0goKCNBGbVnN4XiqW6r+rj7RELPXUCSFEKQp76hKJBDdv3sTatWtRUlKChIQESCQSTcSm1YZ4OeP+x2MxN6Ad16HorW6udlgy0Bd+zrS9LSGEKENhUl+1ahXKy8sRHx8PPp+P1NRUfPzxx5qITatZGBvC084CJlQqVm36eTph/ciu8HKw5DoUQgjRCQqTenx8PMLCwmBgYAATExNs2LABiYmJmohNq7Esi4JykfS+LyGEEMI1hUmdYRiIRCIwDAMAyM/Pl/65OZOwLOxXHsCU8L+5DkVvvX3wGuYcuMp1GIQQojMUjh3PmDEDb775JnJycrBmzRqcPXsW8+fPV+rkt2/fxsaNGxEeHl7r+M8//4zDhw/D1tYWAPDFF1/Aw8MDQUFBaN26NQDA398fixYtauDH0Rw+jwdbEyPk0EQ5tfkjIR1mArq9QQghylL4G3Ps2LHw8/NDVFQUxGIxtm3bBm9vb4Un3rlzJ44fPw4Tk7ozl+Pj47Fhwwb4+flJj6WkpMDX11enqtU5mBvRkjY1EUskyCmtQBs7e65DIYQQnSF3+P38+ep620ePHkVcXBzMzMxgaWmJxMREHD16VOGJ3dzcsGXLFpnPxcfHY8eOHZgyZQp++OEH6bGsrCxMnz4dc+fOxaNH2r9Zir2ZMfLKRBDTagCVyy0VQixh4Uj7qBNCiNLk9tRjY2MxcOBAREVFyXx+7Nix9Z546NChSEtLk/nc66+/jqlTp8Lc3Bzvvfcezp8/DwcHB7z99tsYPnw4bt68iSVLluDIkSNKfYi4uDilXqdqBpXlkLAsLly7AWtjboeJo6OjOb2+qj0oqB4B4VUUq/yz6VtbqRO1VcNQeymP2kp5DWkruZlowYIFAIDg4GBpeVhVYFkWM2fOhIWFBQCgf//+SEhIwOzZs8Hn8wEA3bt3R1ZWFliWVWpSnp+fH4yMjFQSX0O0fSTChbRitGrrzeme39HR0ejWrRtn11eHvKSnAB6ho6cbunVTXQVDfWwrdaG2ahhqL+VRWynv1bYSCoX1dmQVzn4/fvw4Bg0ahM8++0wl36xKSkowcuRIlJaWgmVZREVFwc/PD1u3bsWuXbsAAImJiWjZsqXWz7Kf2b0Nfp3aF04W3JQx3R+TDP+NJxCwLwH+G09gf0wyJ3Gog5EBH0GejmjvSGvUCSFEWQq735s3b0ZJSQnOnj2LHTt2IDU1FcOGDcMHH3zQoAudOHECZWVlmDx5MhYuXIgZM2ZAIBAgICAA/fv3h7+/P5YsWYKLFy+Cz+dj3bp1jf5QmtLHwxF9PBw5ufb+mGS8sfuy9HFsRoH0cWgXD05iUqWgNk44P38o12EQQohOUWpM3dzcHN26dUNmZiYyMjIQExOj1MldXFxw8OBBAMCoUaOkx8eOHVvnnryVlRV27NihbNzN3vpI2cMvGyLj9SKpE0IIaTiFSf3nn3/GyZMnIRKJMHr0aOzYsQMtWrTQRGxaLyolB28duIZ3Atrh/X4dNHrthKxCOccLNBqHuhy89RiZReWY07sdTGmtOiGEKEXhb8usrCz85z//QYcOmk1auoDP4+FuViGS80o0fm0fJyvEZtRN4NYmAlSJJTDgK5wuodV+jHqAs/cyaMMcQghpAIW/+RcuXIi0tDQcPXoUR48exZEjR7Bp0yZNxKb1HJ7vqf6Mg6pyy4L9ZB5/VirEsB1ndb4oTlZxOSyNDWnDHEIIaQCFvzEXLVqEwsJCpKamonv37oiKikLXrl01EZvWczCvnvXORVIP7eKBiw+zsOPaffAYwK+FDRYEeeNEfBqOxT3BkhPR+GVKX+j8k3IAAB4YSURBVI3HpSpZxRVoQYVnCCGkQRT21JOSkvDrr79iyJAhmDNnDvbt24f09HRNxKb1TAUGMDHk4xlHvWIrYwEAYFuwO2IWj8SbPdvi8Mz++HpMd2wcpbtrQKvE1SViW3C0VJAQQnSVwqRuZ2cHhmHg4eGBpKQkuLq6orKyUhOx6QQHc2NOeuoAcPf5ZLnWli8K7/B4DD4I6gD756MIEYnpWHj0BirFulPK9lmpECwLKhFLCCENpHD4vV27dli9ejWmTJmCxYsXIzs7GyzLaiI2nRDq3xqVHNV+T8ouhJ2pEWzklKhlWRbrzsbhcnI2bqXnYf+MIDjpQKJ8VloBQz6PeuqEENJACpP6559/jpiYGLRt2xbvv/8+rl27hq+++koTsemEdSO5mV8grBLjUV4JervJ38WMYRicnDMIsw9cxW93UtHzmz9xeFZ/9KjnPdrAz9kG5RumQqRDowuEEKINFA6/P3z4EM+ePcOFCxfg5eWFFStWwMvLSxOxkXoUlIsQ5OmIgNYO9b7OwtgQB2cEYc0If6QXlaH/d6fwU9QDDUXZeAzDwMiAz3UYhBCiU+T21HNzc7FgwQLcv38f7u7uYBgGycnJ6NKlC7766ivphizN3cmENByNTcWyYD+0tddcnXInCxOcffc1AIp38GEYBsuCO8K/lS3e2H0Z268mwciAh/+ej0dCViF8nKywLNhPayrRPXhWhNxSIfxaWMPMyJDrcAghRGfI7al/9dVX6NatG65cuYJDhw7h4MGDuHLlCtq3b481a9ZoMkatdudpPn6+/hAPn2m+AE1DDfNuhesfjsDMHm0wY+8VxGYUQCxhpXXjtWVDmG1X7qHP5gi5VfMIIYTIJjepx8TEICwsDIaGL3pKAoEAYWFhSEhI0EhwusBeWoBGs8vawm8+wneXE1FRKW7Q+9rYW2DnP/dlPrchMl4VoTVZVnE5ANA6dUIIaSC5SV3e/uQMw4DH0+0SpKpkb8ZNAZrvLidi8fFoGPAavj2ttteNzyqu/oLkSLPfCSGkQeRm5/r2Mtf2fc41ycG8+suPJsuysiyLxOwitHOwaFSNdx8nKznHrZsamkpklZTDxkRAE+UIIaSB5E6Uu3//PoKDg+scZ1kWOTk5ag1Kl3DRU88oKkexsBLtHWUnZ0WWBfvV2ou9xtJg36aGphKZRRVoYUlD74QQ0lByk/qpU6c0GYfOcjAzgouVKcyNNLfxSGJ29fB5h0Ym9ZpZ7hsi4xGXWQAJy2JKl9ZaMfu9UixBbpkQHZ21Y9SAEEJ0idxM1KpVK03GobPszY2R8mmIRq+ZlF0EAGjv2PgldKFdPBDaxQN5ZUK4fHEYMel5YFmW81srPAa4/uEI8OgWDyGENBjta6mDCspFMDLgwbuRPfWX2Zoa4ZPBHdHKygwSlgWf42TK5/HQzdWO0xgIIURXUVJXgfMPMpFeWIZp3Tw1cr3lgzvio0G+YKCaBPzJkE4qOY8qVIol4DMMeI2Y1U8IIc0drU1Tgc8jbmPWvisQa3BjFz6Pp/LEV1QhavC6d1ULv/kIxkv3aE0hHEII0SWU1FXA3twILAvkl4nUfq0yURX2/ZuM+zlFKj3v77GpaPXFYeyPeazS8zZUVnE5xBJWulc8IYQQ5VFSV4EXVeXUv6ztblYhpu25jC2XElV63i6tbFEmEuOXG9xu9pL1fL2/ExWeIYSQBqOkrgIOz9eq52igVGzNcjZVTJJ7WWtbcwS3a4FLj7JxT8WjAA2RWUQlYgkhpLEoqauAg7nmCtCoYjmbPLN6tgUA/HKdu9569vOeek2bEkIIUR4ldRWwM9Ncqdi7NYVn5JR6bYpxHV1hZWyIX28+QpVYc5P+XpZZVA57MyMYNqL8LSGENHe0pE0Fxvi6ImXleDhqoHeZlF0ICyNDOKuhjKqJoQGmdvXAtqv3cPVxDoLaOKn8GoqEDfBBmahK49clhBB9QEldBSyMDWFhbKj4hU0klkhwP6cYnVvaqK3y2+KBvvi/vu3h04KbMq1zerfj5LqEEKIPKKmrAMuyeFYqRKVYgpZWpmq7Dp/HQ9aqichV47371rbmajs3IYQQ9aIblyoQfvMRWnx2CG6rjsB/4wm1Fk6xNBbAw85CbecHqr+kXE3Oxpmkp2q9zqvuPM1Hvy0R+OX6Q41elxBC9AUl9SbaH5OMN/dfBQCwAGIzCvDG7stqSewpeSVIzS+FRMKq/Nwvyy8XIXjbGSw6fhMsq95rvSwlvwRXH+fgmQaWBhJCiD6ipN5E6yPjZB7fEBmv8mutORsLj//8Jl2rri62/9/evUY1daZ7AP+HcJWAVLloBRRUtJjxhnePek4dB8crtVpFB2q12pvT0dapl2o7rU7V1p41q2rXsa4ezxp6QcfptDIdtaNja73UaVOkAkWsICMCARUhCSGEZJ8PXEYhaC57Z4fw/32SneTN4+NePrx77/d5uwVgjjoGeRU1+PbaTUm/604VuqZiHsnGM0RETpG0qOfk5CAtLa3d8f3792PmzJlIS0tDWloaioqKUF9fj1//+tdYvHgxVqxYgVu3bkkZmmjytbYLbL72tujfVaCtgdJHgQHh0l5+B4AnxvQHAOx345r1Sh0bzxARuUKyor5v3z5s2rQJJlP7h7ry8vKwY8cOZGRkICMjA/Hx8fj444+RkJCAjz76CCkpKXj33XelCk1UiR2sF0+MEv/p8YLKWvTvGQJ/X6XoY7c1LaE3+nTvhszsq25bYtYyU2dRJyJyjmRFPTY2Frt27bL5Wl5eHt577z2kpqZi7969AACNRoNJkyYBACZPnoxz585JFZqo1k9V2zy+buoQUb/nhr4eN+tMknSSs0Xp44PHR8ejtt6MTy7+yy3fqW2eqbPvOxGRcyRb0pacnIzS0lKbr82cOROLFy+GSqXCqlWrcPLkSej1eoSENF1WDg4Ohk6ns/u7cnNt39d2h4EAtk7og/15N1Bca0JsiD+eVEdgoPUWNBrxbiFcqKwDAIRZjdBoNO1et3XMVUlBDQgLUCL/8hVoUC36+G2FC0aMjgpGSUEeSiXcT12KXHkr5soxzJf9mCv7OZIrt69TFwQBjz/+eGsBnzJlCvLz86FSqWAwGAAABoMBoaH2z0jVajUCAgIkidceSUnAhkel/Y7sby4DuIopQxOQlDTgrtc0Gg2SkpJE/84kALMmjYWvm1q2SvBXaEeqXHkj5soxzJf9mCv7tc2VyWS650TW7U+/6/V6zJo1CwaDAYIg4Pz581Cr1Rg5ciS++uorAMCpU6c67T+40dyIGqP4+6rPVcfg8xUP4xeDHhR97HtxV0EnIiLXue1/7KysLBw4cAAhISFYs2YN0tPTsXjxYgwYMABTpkxBamoqLl++jNTUVBw4cACrVq1yV2ii+cflcjzw8gHsOXNJ9LEjVIGYPrgP+kjYsa4je88VYvre47BYpdvkpd5swUtZGvz5hxLJvoOIyNtJevk9OjoaBw8eBADMnj279XhKSgpSUlLuem9QUBDeeecdKcOR3JBeYTBbrDhTXCn62HUNjejmL09X3+zSW/h7YTmOF1YgebA0VwoqdEa8/WU+fpUUj0eH9pXkO4iIvB2vrYooKiQI/XuG4NzVKlG7vtWbLei+MRPz/+8r0cZ0hDvWrFe0rlHnk+9ERM5iURfZhLgI1NSbkSdi85nLN2phFQSEB8vzMOCY2HAkRnXHZ7nXJNtMpqK2ZTkb16gTETmLRV1kE+MiAQBniqtEG7OgshYAMNhNa9TbUigUeGLMADRYrPjo+yJJvkOrb2o8wzXqRETOY1EX2cR+EQCAs1fFu69+qbnX+6BI293r3GFJUhx8fRT43/NXJNnkRVvLFrFERK7ifuoiGxzZHf+zYBwmx0eKNuaPzf3lH+qgJa07RIUEYfMvhmJgeCgEAVCI3Bum0SogyE/JmToRkQtY1EXm46PAinEDRR3zUmUtAn2ViA0LFnVcR22aNlSysV//5XC8Nn2YZOMTEXUFLOoSsVitqGuwICTQz+Wx1k1VQ1trhI+ErVMdcdNggirAFwEibyyjEHv6T0TUxfCeugRyy6vRc9NBvHrsgijjLRjWF6smDRZlLFd9qClC9GuH8BeRN3k58uN1fOfGvduJiLwRi7oE+oeHoL7RIuoT8J5iVExPNFis2P/PK6KO+9gfv8LTf/pG1DGJiLoaFnUJBPn5Iim6B7Kv34LBZHZprPfOFSLp7b/irARd6pwxKLI7BkWE4nhhOfzXfoDhO7OQmV3s0ph6kxl1DRY+JEdE5CIWdYlM6BcJi1XAP128pJxTVo0LZdWytYhtKzO7GJeqmtbNWwQBF8tvY8kHp10q7BU6Np4hIhIDi7pEJsY1r1d3cYbdskY9IUKexjNtbT9he8u/HSfynB6zorap8QxbxBIRuYZFXSITmpvQnHbxvnpBZS36PhDsMTP1/OY18+2PO98WV6vnTJ2ISAyeUSm8UGRIEPY8OhYjo3s4PUaNsQHltUa376F+L4lR3XGxvH0BT4wKc3pMbS1bxBIRiYFFXUJPT0hw6fMt967l6vluy/qpaiz54HS74+umDnF6zF+NisPE+AjEyNxch4ios2NRl5ggCDCaLU5dPg8J8MPTExLw8MBeEkTmnEUj4gA03UPP195G7APBeGFKYutxZ4QG+mPYg85f0SAioia8py6hspo6xLz+Zzzl5Prrh6K6Y8+jYzF7SIzIkblm0Yg4ZK+dhb0LxqPoph51DY0ujXfb2IBGi1Wk6IiIui4WdQn1CgmCqdEi6o5tnmRmYh/4KBT4LPeaS+PMeO8EQjd+LMnub0REXQmLuoR8fBQY3y8CV28ZUFZT5/Dnl2WexcbPv5cgMnFEqAIxMS4CZ0uqoG1ea+6MCp0RkapA9n4nInIRi7rE/iOuaQvWM1cdW9pmtljxoaYIX13RShGWaOaqYyAIQFZeqVOfFwQBWl09l7MREYmARV1iE/o1FXVHm9AU3dSh0SpgUKR8e6jbY07z/X5nL8HX1ptR38gWsUREYmBRl9iomJ7wV/rgjINFvaDS85az2dI/PAQ/6x2GH7U1Tj3s1toiVsWZOhGRq7ikTWKBfkr8d8ooxPcIcehzBc2d2zx9pg4AR1ZORZQqyKn93rW65haxoZypExG5ikXdDZ6ZMMjhzxQ093x/KMrzi3rv0G5Of7Z/eAjenT8WI/pwnToRkat4+d2NzA5cno4JC8aIPj0Q10MlYUTiuVRZgzf/kevQ3xEA+nTvhqfGJ2BMbLhEkRERdR0s6m5gNDdiyI7PMOf9k3Z/5vVfDsd3L8yEn7Jz/BPtOX0JGz7PxikPf1qfiMibdY6K0ckF+fnCKgDnrlbBYvXOzmlz1c49Bb/x8+8x7g9/c2odPxER3Y1F3U0m9IuAzmRGbsX9tyjNLa/G747mIKfslhsiE8fk/lEIC/LHZ7nXHOoMl1txG99eu4lgD9laloioM2NRd5MJcU37q5+1Y3/108VV2PL3H5BTVi11WKLxU/pgZmIflNbU4ftS+38Z0erqEeirRGign4TRERF1DSzqbjKxuQmNPevVLzU/+T64Eyxnu5Mzl+C1OiOiQtgilohIDJIW9ZycHKSlpXX4+ubNm7Fz587Wn1NSUpCWloa0tDRs2LBBytDcblBkKHp2C8BZO9rFtjSeGRTh2Y1n2koe9CB6dPOH0Wyx6/2CIKBCV49ebBFLRCQKyW5k7tu3D4cPH0ZQkO3/sDMzM1FYWIjRo0cDAEwmEwAgIyNDqpBkpVAosGXGcHQP9IPVKtyzUculyhr0Dg1C9yB/N0boOlWAH8p+t8DuJ/arjQ0wW6yIZItYIiJRSDZTj42Nxa5du2y+lp2djZycHCxcuLD1WEFBAYxGI5YtW4b09HRcuHBBqtBk89T4BCwaEXfPgl7X0IiSaoPHt4ftiCNL8CxWAemj4jF1YC8JIyIi6jokm6knJyejtLT9zl2VlZXYvXs3du/ejSNHjrQeDwwMxPLly7FgwQJcvXoVK1aswNGjR+Hre/8Qc3NzRY1dalZBgE8H95Cv6RoQEeSLHooGaDQau8d05L1S+2vRbZTUmvDc8Kj7vndVQiCAOrfG70m58nTMlWOYL/sxV/ZzJFduX0d09OhRVFdXY+XKlaiqqkJ9fT3i4+Mxa9Ys9O3bFwqFAnFxcQgLC0NVVRV69+593zHVajUCAgLcEL1rBEHAw+9+gQaLFWee/6XN9yQBSPnP8bBYrVD62Dfr1Wg0SEpKEjFS16w5exTnSm5hx8KHEaHyrEvrnpYrT8ZcOYb5sh9zZb+2uTKZTPecyLr96ff09HR88sknyMjIwMqVKzFr1izMmzcPhw4dwvbt2wEAWq0Wer0eERER7g5PUgqFAo1WAd9euwm9yXzP99pb0D3RXHUMrIKAv+bfe4/1v/14HWs+/bb1aX8iInKN2ypHVlYWDhw40OHr8+fPh06nQ2pqKtasWYM33njDrkvvnc3EuEhYrALOl9yw+fqRH6/ji0tlnbrz3Bw7l7Z9fUWLd74uwA2DyR1hERF5PUmrZnR0NA4ePAgAmD17drvX582b1/pnf39/vP3221KG4xEm9GtuQnO1ClMT2t9a2Ph5Nq7c1OH27xe5OzTRDIwIRWJUd/z9UjkMJjOCA2w3ltHqm7dd5ZI2IiJRdN5rvJ1US1G31YTGYrWisKoWgyJDndqb3JPMVcegvtGCLwrLO3xPhc4IAIjikjYiIlGwqLtZuCoQgyJC8U3JjXaX2P9VbUB9o6XTNZ2xJeVnsRjfNwL+91jiVqmrR7C/L1QdzOSJiMgx3nfTuhN48b8S0WgV0GgVcGfNa+kkNziqc7WHtWVUTE+cfn76Pd9T0dwiloiIxMGiLoPlYwfaPF7QSXu+O0MQBDwQ5I/osGC5QyEi8hos6jISBOGujUwKq5pn6p20m1xbNw0mbPj8e8SEBWPzL4be9ZpCocDFl+bIFBkRkXfiPXWZLM74GsN2Zt11bM+8sbjy8iMY5CUz9ZAAX/wppwT7//mTQ3usExGRc1jUZWI0NyKvogbXqg2tx3x8FOjXQ+VQ/3RP5u+rxIyH+qCk2tBub/h/VRuQmV2M4ps6maIjIvI+3lE9OqGJcc37q19tWtpW19AIjR2d5jqbjvZY/7pIiyUfnL7nkjciInIMi7pMWpvQFDftr37h+i2M+cPf8NqxH+QMS3TTBz8If6VPu6Ku1bHxDBGR2FjUZZIU0xMBvj44e7WpqLcsZxvkJQ/JtQgN9MfDA3shp6z6rkvtWjaeISISHZ9+l0mArxKjY8Jx9moVdPXm1k1NHvKCNeptLR0zAIMju9+1SU0FZ+pERKJjUZfRk+MGYvaQaFgE4d+NZ7zkyfc7LRjWFwuG9b3rGGfqRETiY1GXUdqo+NY/X6qsQXhwAHoGe/6+8K5oWZuv1dUjNNAPQX48BYmIxMJ76h7A1GhB0S29V87SW3xdpMWwt7Lwx++KAAAnnp2Gb9fMkDkqIiLvwmmSzNYe/g6Hc0vx9apkeHN7lkhVIHIrbuOz3Gt4fHR/9OgWgB7dvPuqBBGRu3GmLrMaoxlXburgr1RiXN8IucORzKDI7hgcGYovLpXhtrEBJbf0qDdb5A6LiMirsKjLzLe5e9yYP3yO4TuzkJldLHNE0pmrjoHRbMHu0wWI//1f8FKWRu6QiIi8Cou6jDKzi/HeuUIAgFUALpbfxpIPTnttYW/pLvfeucsA+OQ7EZHYWNRltP1Ers3jO07kuTkS9xgdE47eoUG4XlMHAIjiGnUiIlGxqMsoX1vTwfHbbo7EPXx8FJiVGI0gPyUA4I3jF732qgQRkRxY1GWU2EH3uMSoMDdH4h6Z2cXY981lGJsfkCupNnj17QYiIndjUZfR+qlqm8fXTR3i5kjco6vdbiAicjeuU5fRohFxAJqKWr72NhKjwrBu6pDW496mq91uICJyNxZ1mS0aEee1RbytxKjuuFjevoB76+0GIiJ34+V3cpuudruBiMjdOFMnt+lqtxuIiNyNRZ3cqivdbiAicjdeficiIvISLOpERERegkWdiIjIS0ha1HNycpCWltbh65s3b8bOnTsBAFarFa+88goWLlyItLQ0lJSUSBkaERGR15GsqO/btw+bNm2CyWSy+XpmZiYKCwtbfz5+/DgaGhpw4MABvPjii9i+fbtUoREREXklyYp6bGwsdu3aZfO17Oxs5OTkYOHCha3HNBoNJk2aBAAYPnw4cnNttxQlIiIi2yRb0pacnIzS0tJ2xysrK7F7927s3r0bR44caT2u1+uhUqlaf1YqlWhsbISv7/1D5C8ATb8UkX2YK/sxV45hvuzHXNnPkVy5fZ360aNHUV1djZUrV6Kqqgr19fWIj4+HSqWCwWBofZ/VarWroAOAWq1GQECAVCF7PI1Gg6SkJLnD6BSYK/sxV45hvuzHXNmvba5MJtM9J7JuL+rp6elIT08HAHzyyScoKirCvHnzcOzYMZw8eRIzZszAhQsXkJCQcN+xBEEAADQ0NEgac2fQ0bML1B5zZT/myjHMl/2YK/vdmauWetdS/9pyW1HPyspCXV3dXffR7zRt2jScOXMGixYtgiAIeOONN+47ptlsBoC7HrjrqngLwn7Mlf2YK8cwX/ZjruxnK1dmsxmBgYHtjiuEjsp9J2C1WmEwGODn5weFQiF3OERERJISBAFmsxnBwcHw8Wn/rHunLupERET0b+woR0RE5CVY1ImIiLwEizoREZGXYFEnIiLyEm5fp07iSklJQUhICAAgOjoa27Ztkzkiz5OTk4OdO3ciIyMDJSUlWL9+PRQKBQYOHIhXX33V5hOkXdWducrLy8PTTz+Nfv36AQBSU1MxY8YMeQP0AGazGRs3bsT169fR0NCAZ555BgMGDOB5ZYOtXPXq1YvnVQcsFgs2bdqE4uJiKJVKbNu2DYIgOHRusah3Yi0NCTIyMmSOxHPt27cPhw8fRlBQEABg27ZtWL16NcaOHYtXXnkFJ06cwLRp02SO0jO0zVV+fj6eeOIJLFu2TObIPMvhw4cRFhaGt956C9XV1XjkkUcwePBgnlc22MrVc889x/OqAydPngTQtOHZ+fPnW4u6I+cWf5XsxAoKCmA0GrFs2TKkp6fjwoULcofkcdpuLJSXl4cxY8YAACZPnoyzZ8/KFZrHaZur3NxcfPnll1iyZAk2btwIvV4vY3SeY/r06fjNb37T+rNSqeR51QFbueJ51bGf//zn2LJlCwCgrKwM4eHhDp9bLOqdWGBgIJYvX473338fr732GtauXYvGxka5w/IoycnJd+0hIAhCa6Oi4OBg6HQ6uULzOG1zNXToULz00kv48MMPERMTgz179sgYnecIDg6GSqWCXq/H888/j9WrV/O86oCtXPG8ujdfX1+sW7cOW7ZsQXJyssPnFot6JxYXF4c5c+ZAoVAgLi4OYWFhqKqqkjssj3bnvSiDwYDQ0FAZo/Fs06ZNg1qtbv1zfn6+zBF5jvLycqSnp2Pu3LmYPXs2z6t7aJsrnlf3t2PHDhw7dgybN2++q++7PecWi3ondujQIWzfvh0AoNVqodfrERERIXNUni0xMRHnz58HAJw6dQqjRo2SOSLPtXz5cvzwww8AgHPnzmHIkCEyR+QZbty4gWXLluG3v/0t5s+fD4DnVUds5YrnVcc+/fRT7N27FwAQFBQEhUIBtVrt0LnFNrGdWENDAzZs2ICysjIoFAqsXbsWI0eOlDssj1NaWooXXngBBw8eRHFxMTZv3gyz2Yz4+Hhs3boVSqVS7hA9xp25ysvLw5YtW+Dn54fw8HBs2bIFKpVK7hBlt3XrVhw5cgTx8fGtx15++WVs3bqV51UbtnK1evVqvPXWWzyvbKirq8OGDRtw48YNNDY2YsWKFejfv79D/2exqBMREXkJXn4nIiLyEizqREREXoJFnYiIyEuwqBMREXkJFnUiIiIvwaJORHY7f/480tLSWn/W6/V47LHHWvslEJG8uKELETnFYDDgySefxJgxY7B27Vq5wyEisKgTkRPq6uqwcuVKjBs3DqtXr5Y7HCJqxsvvROQQo9GIp556CoWFhVi6dKnc4RDRHVjUicghFy9exPjx4zFjxgxs2rRJ7nCI6A4s6kTkkBEjRuDZZ5/F+vXrcfnyZWRmZsodEhE1Y1EnIof4+fkBaNpF6s0338Sbb76Jn376SeaoiAhgUSciFwwbNgxLly7FmjVr7tr3mYjkwV3aiIiIvARn6kRERF6CRZ2IiMhLsKgTERF5CRZ1IiIiL8GiTkRE5CVY1ImIiLwEizoREZGXYFEnIiLyEv8PtwnWwO5fYy8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Davies Bouldin score for Hierarchical Clustering\n",
    "\n",
    "from sklearn.metrics import davies_bouldin_score \n",
    "\n",
    "def get_AG_clustering_score(data, center):\n",
    "    '''\n",
    "    returns the AG_clustering score regarding Davies Bouldin for points to centers\n",
    "    INPUT:\n",
    "        data - the dataset you want to fit AG_clustering to\n",
    "        center - the number of centers you want (the k value)\n",
    "    OUTPUT:\n",
    "        score - the Davies Bouldin score for the AG_clustering model fit to the data\n",
    "    '''\n",
    "    #instantiate AG_clustering\n",
    "    AG_clustering = AgglomerativeClustering(n_clusters= center)\n",
    "\n",
    "    # Then fit the model to your data using the fit method\n",
    "    model = AG_clustering.fit_predict(cluster_df)\n",
    "    \n",
    "    # Calculate Silhoutte Score\n",
    "\n",
    "    score = davies_bouldin_score(cluster_df, model)\n",
    "    \n",
    "    return score\n",
    "\n",
    "scores = []\n",
    "centers = list(range(2,30))\n",
    "\n",
    "for center in centers:\n",
    "    scores.append(get_AG_clustering_score(cluster_df, center))\n",
    "    \n",
    "plt.plot(centers, scores, linestyle='--', marker='o', color='b');\n",
    "plt.xlabel('K');\n",
    "plt.ylabel('Davies Bouldin score');\n",
    "plt.title('Davies Bouldin score vs. K');\n",
    "\n",
    "\n",
    "df3 = pd.DataFrame(centers,columns=['centers'])\n",
    "df3['scores'] = scores\n",
    "df4 = df3[df3.scores == df3.scores.min()]\n",
    "print('Optimal number of clusters based on DB score:', df4['centers'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Dendrogram for Hierarchical Clustering\n",
    "import scipy.cluster.hierarchy as shc\n",
    "from matplotlib import pyplot\n",
    "pyplot.figure(figsize=(10, 7))  \n",
    "pyplot.title(\"Dendrograms\")  \n",
    "dend = shc.dendrogram(shc.linkage(cluster_df, method='ward'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Silhouette Score: 0.2291\n",
      "Calinski Harabasz Score: 718.8155\n",
      "Davies Bouldin Score: 1.4191\n"
     ]
    }
   ],
   "source": [
    "# Agglomerative clustering\n",
    "from numpy import unique\n",
    "from numpy import where\n",
    "from sklearn.cluster import AgglomerativeClustering\n",
    "from matplotlib import pyplot\n",
    "\n",
    "# define the model\n",
    "model = AgglomerativeClustering(n_clusters=4)\n",
    "# fit model and predict clusters\n",
    "yhat = model.fit(cluster_df)\n",
    "yhat_2 = model.fit_predict(cluster_df)\n",
    "# retrieve unique clusters\n",
    "clusters = unique(yhat)\n",
    "\n",
    "\n",
    "# Calculate cluster validation metrics\n",
    "\n",
    "score_AGclustering_s = silhouette_score(cluster_df, yhat.labels_, metric='euclidean')\n",
    "\n",
    "score_AGclustering_c = calinski_harabasz_score(cluster_df, yhat.labels_)\n",
    "\n",
    "score_AGclustering_d = davies_bouldin_score(cluster_df, yhat_2)\n",
    "\n",
    "\n",
    "\n",
    "print('Silhouette Score: %.4f' % score_AGclustering_s)\n",
    "print('Calinski Harabasz Score: %.4f' % score_AGclustering_c)\n",
    "\n",
    "print('Davies Bouldin Score: %.4f' % score_AGclustering_d)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# DBSCAN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.9335816413107338\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 360x360 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# parameter tuning for eps\n",
    "from sklearn.neighbors import NearestNeighbors\n",
    "nearest_neighbors = NearestNeighbors(n_neighbors=11)\n",
    "neighbors = nearest_neighbors.fit(cluster_df)\n",
    "distances, indices = neighbors.kneighbors(cluster_df)\n",
    "distances = np.sort(distances[:,10], axis=0)\n",
    "\n",
    "from kneed import KneeLocator\n",
    "i = np.arange(len(distances))\n",
    "knee = KneeLocator(i, distances, S=1, curve='convex', direction='increasing', interp_method='polynomial')\n",
    "fig = plt.figure(figsize=(5, 5))\n",
    "knee.plot_knee()\n",
    "plt.xlabel(\"Points\")\n",
    "plt.ylabel(\"Distance\")\n",
    "\n",
    "print(distances[knee.knee])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of min_samples based on silhouette score: [18]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Silhouette Score for DBSCAN\n",
    "\n",
    "from sklearn.metrics import silhouette_score \n",
    "from sklearn.cluster import DBSCAN\n",
    "\n",
    "def get_dbscan_score(data, center):\n",
    "    '''\n",
    "    \n",
    "    INPUT:\n",
    "        data - the dataset you want to fit kmeans to\n",
    "        center - the number of centers you want (the k value)\n",
    "    OUTPUT:\n",
    "        score - the Silhouette Score for DBSCAN\n",
    "    '''\n",
    "    #instantiate kmeans\n",
    "    dbscan = DBSCAN(eps= 1.9335816413107338, min_samples=center)\n",
    "\n",
    "    # Then fit the model to your data using the fit method\n",
    "    model = dbscan.fit(cluster_df)\n",
    "    \n",
    "    # Calculate Silhoutte Score\n",
    "\n",
    "    score = silhouette_score(cluster_df, model.labels_, metric='euclidean')\n",
    "    \n",
    "    return score\n",
    "\n",
    "scores = []\n",
    "centers = list(range(2,30))\n",
    "\n",
    "for center in centers:\n",
    "    scores.append(get_dbscan_score(cluster_df, center))\n",
    "    \n",
    "plt.plot(centers, scores, linestyle='--', marker='o', color='b');\n",
    "plt.xlabel('min_samples');\n",
    "plt.ylabel('Silhouette Score');\n",
    "plt.title('Silhouette Score vs. min_samples');\n",
    "\n",
    "df3 = pd.DataFrame(centers,columns=['min_samples'])\n",
    "df3['scores'] = scores\n",
    "df4 = df3[df3.scores == df3.scores.max()]\n",
    "print('Optimal number of min_samples based on silhouette score:', df4['min_samples'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of min_sample based on Calinski Harabasz score: [29]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calinski Harabasz for DBSCAN\n",
    "\n",
    "from sklearn.metrics import calinski_harabasz_score  \n",
    "\n",
    "def get_dbscan_score(data, center):\n",
    "    '''\n",
    "    \n",
    "    INPUT:\n",
    "        data - the dataset you want to fit dbscan to\n",
    "        center - the number of centers you want (the k value)\n",
    "    OUTPUT:\n",
    "        score - the Calinski Harabasz score for the dbscan model fit to the data\n",
    "    '''\n",
    "    #instantiate dbscan\n",
    "    dbscan = DBSCAN(eps= 1.9335816413107338, min_samples=center)\n",
    "\n",
    "    # Then fit the model to your data using the fit method\n",
    "    model = dbscan.fit(cluster_df)\n",
    "    \n",
    "    # Calculate Silhoutte Score\n",
    "\n",
    "    score = calinski_harabasz_score(cluster_df, dbscan.labels_)\n",
    "    \n",
    "    return score\n",
    "\n",
    "scores = []\n",
    "centers = list(range(2,30))\n",
    "\n",
    "for center in centers:\n",
    "    scores.append(get_dbscan_score(cluster_df, center))\n",
    "    \n",
    "plt.plot(centers, scores, linestyle='--', marker='o', color='b');\n",
    "plt.xlabel('min_sample');\n",
    "plt.ylabel('Calinski Harabasz');\n",
    "plt.title('Calinski Harabasz vs. min_sample');\n",
    "\n",
    "df3 = pd.DataFrame(centers,columns=['min_sample'])\n",
    "df3['scores'] = scores\n",
    "df4 = df3[df3.scores == df3.scores.max()]\n",
    "print('Optimal number of min_sample based on Calinski Harabasz score:', df4['min_sample'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of min_samples based on DB score: [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Davies Bouldin score for DBSCAN\n",
    "\n",
    "from sklearn.metrics import davies_bouldin_score \n",
    "\n",
    "def get_dbscan_score(data, center):\n",
    "    '''\n",
    "    returns the dbscan score regarding Davies Bouldin for points to centers\n",
    "    INPUT:\n",
    "        data - the dataset you want to fit dbscan to\n",
    "        center - the number of centers you want (the k value)\n",
    "    OUTPUT:\n",
    "        score - the Davies Bouldin score for the dbscan model fit to the data\n",
    "    '''\n",
    "    #instantiate dbscan\n",
    "    dbscan = DBSCAN(eps= 1.9335816413107338, min_samples=center)\n",
    "\n",
    "    # Then fit the model to your data using the fit method\n",
    "    model = dbscan.fit_predict(cluster_df)\n",
    "    \n",
    "    # Calculate Silhoutte Score\n",
    "\n",
    "    score = davies_bouldin_score(cluster_df, model)\n",
    "    \n",
    "    return score\n",
    "\n",
    "scores = []\n",
    "centers = list(range(2,30))\n",
    "\n",
    "for center in centers:\n",
    "    scores.append(get_dbscan_score(cluster_df, center))\n",
    "    \n",
    "plt.plot(centers, scores, linestyle='--', marker='o', color='b');\n",
    "plt.xlabel('min_samples');\n",
    "plt.ylabel('Davies Bouldin score');\n",
    "plt.title('Davies Bouldin score vs. min_samples');\n",
    "\n",
    "\n",
    "df3 = pd.DataFrame(centers,columns=['min_samples'])\n",
    "df3['scores'] = scores\n",
    "df4 = df3[df3.scores == df3.scores.min()]\n",
    "print('Optimal number of min_samples based on DB score:', df4['min_samples'].tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Silhouetter Score: 0.3782\n",
      "Calinski Harabasz Score: 345.2103\n",
      "Davies Bouldin Score: 2.2108\n"
     ]
    }
   ],
   "source": [
    "# dbscan clustering\n",
    "from numpy import unique\n",
    "from numpy import where\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.cluster import DBSCAN\n",
    "from matplotlib import pyplot\n",
    "# define dataset\n",
    "# define the model\n",
    "model = DBSCAN(eps=1.9335816413107338, min_samples= 18)\n",
    "\n",
    "# rule of thumb for min_samples: 2*len(cluster_df.columns)\n",
    "\n",
    "# fit model and predict clusters\n",
    "yhat = model.fit_predict(cluster_df)\n",
    "# retrieve unique clusters\n",
    "clusters = unique(yhat)\n",
    "\n",
    "\n",
    "# Calculate cluster validation metrics\n",
    "\n",
    "score_dbsacn_s = silhouette_score(cluster_df, yhat, metric='euclidean')\n",
    "\n",
    "score_dbsacn_c = calinski_harabasz_score(cluster_df, yhat)\n",
    "\n",
    "score_dbsacn_d = davies_bouldin_score(cluster_df, yhat)\n",
    "\n",
    "\n",
    "\n",
    "print('Silhouette Score: %.4f' % score_dbsacn_s)\n",
    "print('Calinski Harabasz Score: %.4f' % score_dbsacn_c)\n",
    "print('Davies Bouldin Score: %.4f' % score_dbsacn_d)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# GMM Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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>Covariance_Type</th>\n",
       "      <th>N_Components</th>\n",
       "      <th>BIC_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>112</th>\n",
       "      <td>full</td>\n",
       "      <td>26</td>\n",
       "      <td>10171.280409</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Covariance_Type  N_Components     BIC_Score\n",
       "112            full            26  10171.280409"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# BIC for GMM\n",
    "\n",
    "from sklearn.mixture import GaussianMixture\n",
    "n_components = range(1, 30)\n",
    "covariance_type = ['spherical', 'tied', 'diag', 'full']\n",
    "score=[]\n",
    "for cov in covariance_type:\n",
    "    for n_comp in n_components:\n",
    "        gmm=GaussianMixture(n_components=n_comp,covariance_type=cov, random_state = 10)\n",
    "        gmm.fit(cluster_df)\n",
    "        score.append((cov,n_comp,gmm.bic(cluster_df)))\n",
    "score_1 = pd.DataFrame(score)\n",
    "score_1.columns = ['Covariance_Type', 'N_Components','BIC_Score']\n",
    "score_2 = score_1[score_1.BIC_Score == score_1.BIC_Score.min()]\n",
    "\n",
    "score_2.head(n=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "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>Covariance_Type</th>\n",
       "      <th>N_Components</th>\n",
       "      <th>Silhouette_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>tied</td>\n",
       "      <td>2</td>\n",
       "      <td>0.320604</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Covariance_Type  N_Components  Silhouette_Score\n",
       "28            tied             2          0.320604"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Silhouette Score for GMM\n",
    "\n",
    "\n",
    "n_components = range(2, 30)\n",
    "covariance_type = ['spherical', 'tied', 'diag', 'full']\n",
    "score=[]\n",
    "for cov in covariance_type:\n",
    "    for n_comp in n_components:\n",
    "        gmm=GaussianMixture(n_components=n_comp,covariance_type=cov,random_state = 10)\n",
    "        model = gmm.fit(cluster_df)\n",
    "        model_2 = model.predict(cluster_df)\n",
    "        score_s = silhouette_score(cluster_df, model_2, metric='euclidean')\n",
    "        score.append((cov,n_comp,score_s))\n",
    "score_1 = pd.DataFrame(score)\n",
    "score_1.columns = ['Covariance_Type', 'N_Components','Silhouette_Score']\n",
    "score_2 = score_1[score_1.Silhouette_Score == score_1.Silhouette_Score.max()]\n",
    "score_2.head(n=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "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>Covariance_Type</th>\n",
       "      <th>N_Components</th>\n",
       "      <th>Calinski_Harabasz_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>spherical</td>\n",
       "      <td>4</td>\n",
       "      <td>878.787589</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Covariance_Type  N_Components  Calinski_Harabasz_Score\n",
       "2       spherical             4               878.787589"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calinski Harabasz Score for GMM\n",
    "\n",
    "\n",
    "n_components = range(2, 30)\n",
    "covariance_type = ['spherical', 'tied', 'diag', 'full']\n",
    "score=[]\n",
    "for cov in covariance_type:\n",
    "    for n_comp in n_components:\n",
    "        gmm=GaussianMixture(n_components=n_comp,covariance_type=cov, random_state = 10)\n",
    "        model = gmm.fit(cluster_df)\n",
    "        model_2 = model.predict(cluster_df)\n",
    "        score_c = calinski_harabasz_score(cluster_df, model_2)\n",
    "        score.append((cov,n_comp,score_c))\n",
    "score_1 = pd.DataFrame(score)\n",
    "score_1.columns = ['Covariance_Type', 'N_Components','Calinski_Harabasz_Score']\n",
    "score_2 = score_1[score_1.Calinski_Harabasz_Score == score_1.Calinski_Harabasz_Score.max()]\n",
    "score_2.head(n=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "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>Covariance_Type</th>\n",
       "      <th>N_Components</th>\n",
       "      <th>Davies_Bouldin_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>90</th>\n",
       "      <td>full</td>\n",
       "      <td>8</td>\n",
       "      <td>3.84953</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Covariance_Type  N_Components  Davies_Bouldin_Score\n",
       "90            full             8               3.84953"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Davies Bouldin Score for GMM\n",
    "\n",
    "\n",
    "n_components = range(2, 30)\n",
    "covariance_type = ['spherical', 'tied', 'diag', 'full']\n",
    "score=[]\n",
    "for cov in covariance_type:\n",
    "    for n_comp in n_components:\n",
    "        gmm=GaussianMixture(n_components=n_comp,covariance_type=cov, random_state = 10)\n",
    "        model = gmm.fit(cluster_df)\n",
    "        model_2 = model.predict(cluster_df)\n",
    "        score_d = davies_bouldin_score(cluster_df, model_2)\n",
    "        score.append((cov,n_comp,score_d))\n",
    "score_1 = pd.DataFrame(score)\n",
    "score_1.columns = ['Covariance_Type', 'N_Components','Davies_Bouldin_Score']\n",
    "score_2 = score_1[score_1.Davies_Bouldin_Score == score_1.Davies_Bouldin_Score.max()]\n",
    "score_2.head(n=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Silhouetter Score: 0.0697\n",
      "Calinski Harabasz Score: 191.3066\n",
      "Davies Bouldin Score: 2.4507\n"
     ]
    }
   ],
   "source": [
    "# gaussian mixture clustering\n",
    "from numpy import unique\n",
    "from numpy import where\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.mixture import GaussianMixture\n",
    "from matplotlib import pyplot\n",
    "# define the model\n",
    "model = GaussianMixture(n_components= 26,covariance_type= \"full\", random_state = 10)\n",
    "# fit the model\n",
    "model.fit(cluster_df)\n",
    "# assign a cluster to each example\n",
    "yhat = model.predict(cluster_df)\n",
    "# retrieve unique clusters\n",
    "clusters = unique(yhat)\n",
    "\n",
    "\n",
    "\n",
    "# Calculate cluster validation score\n",
    "\n",
    "score_dbsacn_s = silhouette_score(cluster_df, yhat, metric='euclidean')\n",
    "\n",
    "score_dbsacn_c = calinski_harabasz_score(cluster_df, yhat)\n",
    "\n",
    "score_dbsacn_d = davies_bouldin_score(cluster_df, yhat)\n",
    "\n",
    "\n",
    "\n",
    "print('Silhouette Score: %.4f' % score_dbsacn_s)\n",
    "print('Calinski Harabasz Score: %.4f' % score_dbsacn_c)\n",
    "print('Davies Bouldin Score: %.4f' % score_dbsacn_d)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}