Claims reserving using Python.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a0445f0d",
   "metadata": {},
   "source": [
    "## 1. Business Problem\n",
    "\n",
    "* Calculate  claims reserve for multiple lines of business.Motor Insurance, Fire Insurance and Micro Insurance"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5d5606ff",
   "metadata": {},
   "source": [
    "## 2. Data Pre-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "0b9fdbc1",
   "metadata": {},
   "outputs": [],
   "source": [
    "#2.1. Load required libraries\n",
    "import chainladder as cl\n",
    "import numpy as np \n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "#Ignore warnings\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\") \n",
    "\n",
    "#Global change of number formats by adding a comma and keeping it to two d.p.\n",
    "pd.options.display.float_format = '{:,.0f}'.format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "add55a71",
   "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>Loss Year</th>\n",
       "      <th>Calendar Year</th>\n",
       "      <th>Paid Amount</th>\n",
       "      <th>Line of Business</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2018</td>\n",
       "      <td>2018</td>\n",
       "      <td>42977</td>\n",
       "      <td>Motor Insurance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2018</td>\n",
       "      <td>2019</td>\n",
       "      <td>18689</td>\n",
       "      <td>Motor Insurance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2018</td>\n",
       "      <td>2020</td>\n",
       "      <td>5005</td>\n",
       "      <td>Motor Insurance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2018</td>\n",
       "      <td>2021</td>\n",
       "      <td>924</td>\n",
       "      <td>Motor Insurance</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019</td>\n",
       "      <td>2019</td>\n",
       "      <td>53141</td>\n",
       "      <td>Motor Insurance</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Loss Year  Calendar Year  Paid Amount Line of Business\n",
       "0       2018           2018        42977  Motor Insurance\n",
       "1       2018           2019        18689  Motor Insurance\n",
       "2       2018           2020         5005  Motor Insurance\n",
       "3       2018           2021          924  Motor Insurance\n",
       "4       2019           2019        53141  Motor Insurance"
      ]
     },
     "execution_count": 150,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#2.2. Load and clean data for analysis\n",
    "df = pd.read_excel('claims_data.xlsx')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "45aed6b3",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 30 entries, 0 to 29\n",
      "Data columns (total 4 columns):\n",
      " #   Column            Non-Null Count  Dtype \n",
      "---  ------            --------------  ----- \n",
      " 0   Loss Year         30 non-null     int64 \n",
      " 1   Calendar Year     30 non-null     int64 \n",
      " 2   Paid Amount       30 non-null     int64 \n",
      " 3   Line of Business  30 non-null     object\n",
      "dtypes: int64(3), object(1)\n",
      "memory usage: 1.1+ KB\n"
     ]
    }
   ],
   "source": [
    "#2.3 browse through the data\n",
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "fe664083",
   "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>Loss Year</th>\n",
       "      <th>Calendar Year</th>\n",
       "      <th>Paid Amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2,019</td>\n",
       "      <td>2,020</td>\n",
       "      <td>31,800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>24,635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2,018</td>\n",
       "      <td>2,018</td>\n",
       "      <td>924</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2,018</td>\n",
       "      <td>2,019</td>\n",
       "      <td>10,150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2,019</td>\n",
       "      <td>2,020</td>\n",
       "      <td>30,283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2,020</td>\n",
       "      <td>2,021</td>\n",
       "      <td>52,398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2,021</td>\n",
       "      <td>2,021</td>\n",
       "      <td>79,464</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Loss Year  Calendar Year  Paid Amount\n",
       "count         30             30           30\n",
       "mean       2,019          2,020       31,800\n",
       "std            1              1       24,635\n",
       "min        2,018          2,018          924\n",
       "25%        2,018          2,019       10,150\n",
       "50%        2,019          2,020       30,283\n",
       "75%        2,020          2,021       52,398\n",
       "max        2,021          2,021       79,464"
      ]
     },
     "execution_count": 152,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5763f378",
   "metadata": {},
   "source": [
    "## 3. Feature Engineering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "a8f53b79",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Triangle Summary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Valuation:</th>\n",
       "      <td>2021-12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Grain:</th>\n",
       "      <td>OYDY</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Shape:</th>\n",
       "      <td>(3, 1, 4, 4)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Index:</th>\n",
       "      <td>[Line of Business]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Columns:</th>\n",
       "      <td>[Paid Amount]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "              Triangle Summary\n",
       "Valuation:             2021-12\n",
       "Grain:                    OYDY\n",
       "Shape:            (3, 1, 4, 4)\n",
       "Index:      [Line of Business]\n",
       "Columns:         [Paid Amount]"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#3.1. Create triangle object that stores incremental triangles by line of business\n",
    "triangle_df = cl.Triangle(\n",
    "    df,\n",
    "    origin = 'Loss Year',\n",
    "    development = 'Calendar Year',\n",
    "    columns = ['Paid Amount'],\n",
    "    index = ['Line of Business']\n",
    ")\n",
    "\n",
    "triangle_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "13652816",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Motor Insurance incremental paid loss triangle:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>42,977</td>\n",
       "      <td>18,689</td>\n",
       "      <td>5,005</td>\n",
       "      <td>924</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>53,141</td>\n",
       "      <td>27,368</td>\n",
       "      <td>1,661</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>66,462</td>\n",
       "      <td>24,640</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>77,671</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "         12     24    36  48\n",
       "2018 42,977 18,689 5,005 924\n",
       "2019 53,141 27,368 1,661 NaN\n",
       "2020 66,462 24,640   NaN NaN\n",
       "2021 77,671    NaN   NaN NaN"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fire Insurance incremental paid loss triangle:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>51,227</td>\n",
       "      <td>37,840</td>\n",
       "      <td>10,241</td>\n",
       "      <td>6,292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>66,979</td>\n",
       "      <td>58,025</td>\n",
       "      <td>15,191</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>61,853</td>\n",
       "      <td>52,789</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>79,464</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "         12     24     36    48\n",
       "2018 51,227 37,840 10,241 6,292\n",
       "2019 66,979 58,025 15,191   NaN\n",
       "2020 61,853 52,789    NaN   NaN\n",
       "2021 79,464    NaN    NaN   NaN"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro Insurance incremental paid loss triangle:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>33,198</td>\n",
       "      <td>10,120</td>\n",
       "      <td>3,014</td>\n",
       "      <td>2,222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>33,924</td>\n",
       "      <td>14,168</td>\n",
       "      <td>17,578</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>38,104</td>\n",
       "      <td>4,224</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>39,006</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "         12     24     36    48\n",
       "2018 33,198 10,120  3,014 2,222\n",
       "2019 33,924 14,168 17,578   NaN\n",
       "2020 38,104  4,224    NaN   NaN\n",
       "2021 39,006    NaN    NaN   NaN"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#3.2. Visualize the incremntal triangles \n",
    "\n",
    "#We shall use LoB and keys throught the rest of these tutorial.\n",
    "LoB = list(df['Line of Business'].unique())\n",
    "keys = LoB\n",
    "\n",
    "incremental_triangle_df = {}\n",
    "\n",
    "for lob in LoB:\n",
    "    incremental_triangle_df[lob] = triangle_df[triangle_df['Line of Business'].isin([lob])]\n",
    "    print( lob +' incremental paid loss triangle:')\n",
    "    display(incremental_triangle_df[lob])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "0d53c9b7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Motor Insurance cumulative paid loss triangle:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>42,977</td>\n",
       "      <td>61,666</td>\n",
       "      <td>66,671</td>\n",
       "      <td>67,595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>53,141</td>\n",
       "      <td>80,509</td>\n",
       "      <td>82,170</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>66,462</td>\n",
       "      <td>91,102</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>77,671</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "            12        24        36        48\n",
       "2018 42,977.00 61,666.00 66,671.00 67,595.00\n",
       "2019 53,141.00 80,509.00 82,170.00       NaN\n",
       "2020 66,462.00 91,102.00       NaN       NaN\n",
       "2021 77,671.00       NaN       NaN       NaN"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fire Insurance cumulative paid loss triangle:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>51,227</td>\n",
       "      <td>89,067</td>\n",
       "      <td>99,308</td>\n",
       "      <td>105,600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>66,979</td>\n",
       "      <td>125,004</td>\n",
       "      <td>140,195</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>61,853</td>\n",
       "      <td>114,642</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>79,464</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "            12         24         36         48\n",
       "2018 51,227.00  89,067.00  99,308.00 105,600.00\n",
       "2019 66,979.00 125,004.00 140,195.00        NaN\n",
       "2020 61,853.00 114,642.00        NaN        NaN\n",
       "2021 79,464.00        NaN        NaN        NaN"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro Insurance cumulative paid loss triangle:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>33,198</td>\n",
       "      <td>43,318</td>\n",
       "      <td>46,332</td>\n",
       "      <td>48,554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>33,924</td>\n",
       "      <td>48,092</td>\n",
       "      <td>65,670</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>38,104</td>\n",
       "      <td>42,328</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>39,006</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "            12        24        36        48\n",
       "2018 33,198.00 43,318.00 46,332.00 48,554.00\n",
       "2019 33,924.00 48,092.00 65,670.00       NaN\n",
       "2020 38,104.00 42,328.00       NaN       NaN\n",
       "2021 39,006.00       NaN       NaN       NaN"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#3.3. Construct the cumulative development triangles using incr_to_cum()\n",
    "\n",
    "cumulative_triangle_df = {}\n",
    "keys = LoB\n",
    "for lob in LoB:\n",
    "    cumulative_triangle_df[lob] = incremental_triangle_df[lob].incr_to_cum()\n",
    "    print( lob +' cumulative paid loss triangle:')\n",
    "    display(cumulative_triangle_df[lob])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "1a0d0d5a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1oAAAFiCAYAAADvFKl3AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAAB9z0lEQVR4nO3dd3xUVfrH8c9JB5LQAiGELr13EBuI2FbF3hUVdVfdXXf96dp7w7aCvSCCrq66qyvuWhCRKKL0bgLSSwgBQkuA1Dm/P+4NJJCElCmZyff9euWVmXPP3PscmHkyzy3nGmstIiIiIiIi4j1hgQ5AREREREQk1KjQEhERERER8TIVWiIiIiIiIl6mQktERERERMTLVGiJiIiIiIh4mQotERERERERL1OhJSIitY4xJscY0yHQcYhI6DPGvGGMeTDQcUjoUaFVRxljNhhj8o0xCUe0LzbGWGNMu0qs4zpjzE9ejusRY8w/vLlOEam93Fx00C2sin9aWmtjrbXrvLSNycaYJ7yxLhEJHpX9rmOt/YO19nE/xXOar7cjtYcKrbptPXBF8RNjTC+gvr82boyJ8Ne2KlJb4hCpw851C6vin60VdTbGhPsrsOpSXhGpNXzyXcc4avX36GCIMdTpH79uex+4tsTzMcB7JTsYYxoaY94zxuwwxmw0xjxgjAkzxnQD3gCOd/dA76mov7vsOmPMbGPMi8aYLOCRYwXo7nH6gzFmtTFmjzHmVWOMcZd1NMb8YIzZa4zZaYz52G1v574uosR6UowxN5YXhzHmOGPM98aYLHddHxhjGpV4/QZjzJ3GmGXu9j42xsSUWD7aGLPEGLPPGLPWGHNmiX+Pd4wxGcaYdGPME8HwJVEk0NzPcEf38WRjzOvGmK+MMfuBEcaYlsaYT91cs94Y8+dKrrc4P4wxxmxyP+/3l1g+2BizwP0sZxpj/u62DzfGbDliXYf2ThvnaPy/jTH/MMbsA65z1/WLm7syjDGvGGOijhhjmfnNXX6TMSbNGJNtjEk1xvR326s1dpE6qjLfdUod9a7gb3qKMeZJY8xs4ADQwRgzzBgz3/1uMN8YM6wyQbnfRX4yxjxvjNntfpbPOmL5Ovfzv94Yc5XbXurMnyO/85QT4/Ulcsk6Y8zvS7x+uDFmizHm/4wx291cdX2J5fWMMS8Y5zvdXjfmeu6yocaYn938tdQYM7wyY69LVGjVbXOAeGNMN/fL/+XAkaftvQw0BDoAp+Akq+uttWnAH4Bf3D3QjSrqX2J9Q4B1QCLwZCXjPAcYBPQGLgXOcNsfB74FGgOt3G1X1pFxGOBpoCXQDWjN0YXgpcCZQHs3luvA+WKGk7TvAhoBJwMb3NdMBgqBjkA/4HTgxirEKSKOK3E+q3HAz8B/gaVAMjAS+Isx5ozyX36UE4Eu7msfMs7OI4AJwARrbTxwHPBJFdY5Gvg3Th74ACgC/gokAMe727r1iNeUmd+MMZfg5KBrgXjgPCDLODuuajp2kbqkMt91DjnG33SAa4CbcXJRNvAl8BLQFPg78KUxpmklYxsCrMLJEc8C7xhHA3edZ1lr44BhwJJKrvPIGDcC23FyTTzOd7IXi3fcuFrgfHdLBsYCrxpjGrvLngcGuDE0Af4GeIwxye7Yn3Db7wQ+NcY0q0KcIU+FlhTv6RkFpAHpxQtKJKR7rbXZ1toNwAs4H+CjVLL/Vmvty9baQmvtwUrGOM5au8dauwmYCfR12wuAtkBLa22utbYq14uVisNau8ZaO91am2et3YGTLE854jUvWWu3Wmt34XzRKY5jLDDJfb3HWpturV1pjEkEzgb+Yq3db63dDryI828kIod97u4R3WOM+bycPlOttbOttR6gF9DMWvuYtTbfvZbrbar22XrU/ewvxSla+rjtBUBHY0yCtTbHWjunCuv8xVr7uZsHDlprF1pr57h5ZgPwJkfnlfLy243As9ba+daxxlq7Eacoq+nYReqacr/rlKHMv+kllk+21v5qrS3E2Xm62lr7vvs5/yewEji3knFttNa+ba0tAqYASTg7gAE8QE9jTD1rbYa19tdKj7ZEjNbaAmvtl9batW4u+QFnJ/VJJfoXAI+5fb8CcoAu7o6dG4Db3X+HImvtz9baPOBq4Ctr7Vfuv9N0YAHO9x5xqdCS93H2FF/HEYfScfawROLsDSm2EWePR1kq039zNWLcVuLxASDWffw3nCNR84wxvxpjbqjCOkvFYYxJNMZ8ZJzT+/bh7O1KOOI15cXRGlhbxjba4vx7ZBR/icT5otW8CnGK1AXnW2sbuT/nl9On5Ge2LdCyRHG2B7iPw19QKqO8z/NYoDOw0j0N6JwqrPPIvNLZGPM/Y8w2N688hXfySk3HLlLXVPRd50jlffaKlfyct6T0dx6o+HvSkQ59/q21B9yHsdba/cBlOGcOZRhjvjTGdK3kOo+MEWPMWcaYOcaYXW7OOJvSuSjLLRyLFeeiBCCG8nPRJUfkohNxikVxqdCq49w9pOtxPnSfHbF4J4ePGhVrw+E9QbaK/ct6TbVZa7dZa2+y1rYEfg+8ZpxrOva7XUpe7NriyJcf8fwpt62Xe8rQ1ThFXGVsxjnFqKz2PCChxJfIeGttj0quV0QOK/mZ3QysL/G5amStjbPW1nhPqrV2tbX2CpwdIs8A/3ZP49lPiZziHsE/8hSZI/PK6zh7tzu5eeU+vJNXfDJ2kVB1jO86Ryrvs3dodSUeb6X0dx44+ntPtVhrp1lrR+EULitxjlzDEbmIo7/flIrRGBMNfIpzCmCie6nHV1QuF+0Ecik/F71/RC5qYK0dV4n11hkqtAScPbinuntQDnEPZX8CPGmMiTPGtAXu4PC5zZlAK+Ne3F2J/l5ljLnEGNPKfbobJ7F43FP/0oGrjTHh7pGuipImOOcx5wB73fOO76pCKO8A1xtjRhpnopBkY0xXa20GzuH5F4wx8e6y44wxR546JCJVMw/INsbc7V6oHW6M6WmMGVTTFRtjrjbGNHNPUdzjNnuA34AYY8zvjDGRwANA9DFWFwfsA3LcvdG3VCGUicCdxpgB7jUbHd2c6rOxi4S4Mr/rlKHMv+nl9P0K6GyMudIYE2GMuQzoDvyvJoG6Z9mMdnfy5OF8P/G4i5cAJxtj2hhjGgL3HmN1UTi5agdQaJwJN06vTBxuHpwE/N04k/CEG2OOd4u3fwDnGmPOcNtjjDOxRquK11q3qNAS3PN2F5Sz+E84e0/WAT8BH+J86AC+B34Fthljdlaiv7cNAuYaY3KAL3DOIS6+785NOMVSFtAD5+L5ijwK9Af24lzceaw9XodYa+fhXlzqvv4HDu/huhYnyaXiFIP/RofVRWrE3alzDs71TOtx9rpOxLmYu6bOBH5188oE4HL3equ9OBNZTMTZkbMf2FL+agDn4vArcS6Yfxv4uLJBWGv/hTP5x4fu6z8Hmvh47CIh6xjfdUr2q+hv+pF9s3A+j/+H833jb8A51tqdZfWvgjCcHdVbgV0413be4m5zOk4uWQYs5BhFnbU2G/gzzo7w3Tg56YsqxHInsByY78byDBBmrd2MMwHQfThF3Gac712qLUow1nrtTC4RERERERFBVaeIiIiIiIjXqdASERERERHxMhVaIiIiIiIiXqZCS0RERERExMtUaImIiIiIiHhZRKADqC0SEhJsu3btAh2GV+3fv58GDRoEOgyfCNWxheq4Fi5cuNNae+TNXes05ZzgEqpjC9VxKeeUFor5BkL3/Ruq44LQHFtF+UaFlqtdu3YsWHDM2ysElZSUFIYPHx7oMHwiVMcWquMyxmwMdAy1jXJOcAnVsYXquJRzSgvFfAOh+/4N1XFBaI6tonyjUwdFRERERES8TIWWiIiIiIiIl6nQEhERERER8TJdoyXiIwUFBWzZsoXc3NxKv6Zhw4akpaX5MCrfiomJoVWrVkRGRgY6FJE6RzlHRPypruWc6uQbFVoiPrJlyxbi4uJo164dxphKvSY7O5u4uDgfR+Yb1lqysrLYsmUL7du3D3Q4InWOco6I+FNdyjnVzTc6dVDER3Jzc2natGmlk0+wM8bQtGnTKu3ZEhHvUc4REX+qSzmnuvlGhZaID9WF5FNSXRuvSG1T1z6DdW28IrVNXfoMVmesKrREQtjmzZsZMWIE3bt3p0ePHkyYMAGAXbt2MWrUKDp16sSoUaPYvXs3ACtXruT4448nOjqa559/vtS6XnzxRXr06EHPnj254oortBdZRI6inCMi/lTbc44KLZEQFhERwQsvvEBqaipz5szh1VdfJTU1lXHjxjFy5EhWr17NyJEjGTduHABNmjThpZde4s477yy1nvT0dF566SUWLFjAihUrKCoq4qOPPgrEkESkFlPOERF/qu05R4WWSC2yZMs+Xp25hoUbd3tlfUlJSfTv3x+AuLg4unXrRnp6OlOnTmXMmDEAjBkzhs8//xyA5s2bM2jQoDJn1CksLOTgwYMUFhZy4MABWrZs6ZUYRSRwlHNExJ/qWs7RrIMifvDof38ldeu+Cvtk5xaQti0bayHMQNcWccTFlD+FaPeW8Tx8bo9Kx7BhwwYWL17MkCFDyMzMJCkpCYAWLVqQmZlZ4WuTk5O58847adOmDfXq1eP000/n9NNPr/S2RcS/lHNExJ+Uc8qmI1oitcS+3EKsdR57rPPcW3JycrjooosYP3488fHxpZYZY455gefu3buZOnUq69evZ+vWrezfv59//OMfXotPRPxPOUdE/Kku5hwd0RLxg8rskVm4cTdXvT2HgiIPkRFhTLi8HwPaNq7xtgsKCrjooou46qqruPDCCwFITEwkIyODpKQkMjIyaN68eYXr+O6772jfvj3NmjUD4MILL+Tnn3/m6quvrnF8IuJ9yjki4k/KOWXTES2RWmJA28a8fVVv7ji9Cx/cONQrycday9ixY+nWrRt33HHHofbzzjuPKVOmADBlyhRGjx5d4XratGnDnDlzOHDgANZaZsyYQbdu3Wocn4gEjnKOiPhTncw51lqf/ACTgO3AijKW/R9ggQT3uQFeAtYAy4D+JfqOAVa7P2NKtA8AlruveQkwbnsTYLrbfzrQuDLxDhgwwIaamTNnBjoEnwmGsaWmplb5Nfv27fNqDLNmzbKA7dWrl+3Tp4/t06eP/fLLL+3OnTvtqaeeajt27GhHjhxps7KyrLXWZmRk2OTkZBsXF2cbNmxok5OT7d69e6211j700EO2S5cutkePHvbqq6+2ubm5ZW6zrHEDC6yPck3xj3JO4AXD57K6gmFsyjmHKeeEfr6xNjg+l9URLOOqizmnqvnGlwnoZKD/kQkIaA1MAzaWSEBnA1+7iWgoMNceTibr3N+N3ceN3WXz3L7Gfe1ZbvuzwD3u43uAZyoTbygmoWD5oFZHMIytNiSgQAjglx7lnAALhs9ldQV0bJvmWvvD887vCijnHKacE/r5xtrQzTnBMq66mHOqmm98duqgtfZHYFcZi14E/oazp6fYaOA9N945QCNjTBJwBjDdWrvLWrsbZ8/Nme6yeGvtHHeA7wHnl1jXFPfxlBLtIhLClHMkJK36Bt49C75/DKacC5vnBToicSnniMix+HUyDGPMaCDdWrv0iNk/koHNJZ5vcdsqat9SRjtAorU2w328DUisIJ6bgZvBuWguJSWliiOq3XJyckJuTMWCYWwNGzYkOzu7Sq8pKiqq8mtqm9zc3Frzf6Oc41/B8LmsLn+Prf7+zbTaMpUWGTMweDCApzCfDd+/x6a2B8p8jXJO4NWmnBPq+QZCN+cEy7jqYs6par7xW6FljKkP3Af47UYY1lprjLEVLH8LeAtg4MCBdvjw4f4KzS9SUlIItTEVC4axpaWlERcXV6XXZGdnV/k1tU1MTAz9+vULdBjKOQEQDJ/L6vLL2KyFjbPh55fht28goh50+x2s/haKCgkLj6LDqdfSofXgMl+unBNYtS3nhHq+gdDNOcEyrrqYc6qab/x5ROs4oD1QvJenFbDIGDMYSMc5p7lYK7ctHRh+RHuK296qjP4AmcaYJGtthnvofbvXRyIiwUA5R4JDUSGkTXUKrK2LoX4CjLgfBo6FBk2d0wU3zIJ2J0E5RZbUCso5IlKK36Z3t9Yut9Y2t9a2s9a2wzkM3t9auw34ArjWOIYCe93D4tOA040xjY0xjXH2Ek1zl+0zxgw1Tja7FpjqbuoLnBl8cH9PRUTqHOUcqfXycmDO6/BSP/j3DZCXDeeMh7+ugFP+5hRZ4BRXJ/2fiqxaTjlHRI7ksyNaxph/4uylSTDGbAEetta+U073r3Bm5FkDHACuB7DW7jLGPA7Md/s9Zq0tvvD0VmAyUA9nNp6v3fZxwCfGmLE4M/5c6sVhiUgtpZwjQWNfBsx7ExZMgty90GYYnPUMdD4TwnR7y2ChnCMix+LLWQevsNYmWWsjrbWtjkw+7h6fne5ja629zVp7nLW2l7V2QYl+k6y1Hd2fd0u0L7DW9nRf80d3Vh6stVnW2pHW2k7W2tNKJCyROmfz5s2MGDGC7t2706NHDyZMmADArl27GDVqFJ06dWLUqFHs3r0bgJUrV3L88ccTHR3N888/X2pdEyZMoGfPnvTo0YPx48f7eyjHpJwjtd72NPj8VhjfC2ZPgA4j4MYZcMPX0PXskCiylHNKLVfOEfGx2p5zgj+ri0i5IiIieOGFF0hNTWXOnDm8+uqrpKamMm7cOEaOHMnq1asZOXIk48aNA6BJkya89NJL3HnnnaXWs2LFCt5++23mzZvH0qVL+d///seaNWsCMSSR4GItrPsB/nExvDYUfv0PDLwe/rQILp0CrQYGOkKvUs4REX+q7TlHhZZILRK2dSHMesFr98pJSkqif//+AMTFxdGtWzfS09OZOnUqY8Y4p/iPGTOGzz//HIDmzZszaNAgIiMjS60nLS2NIUOGUL9+fSIiIjjllFP47LPPvBKjSEgqKoBl/4I3T4b3zoOMpXDqA/DXX+Hs56BJ+0BHCCjniIh/1bWc49f7aInUWV/fA9uWV9wnbx/1M1eA9YAJg8SeEB1ffv8WveCscZUOYcOGDSxevJghQ4aQmZlJUlKSs5oWLcjMzKzwtT179uT+++8nKyuLevXq8dVXXzFwYGjtiRfxitx9sOg9Z5KLfVsgoTOc9zL0uhQiY/wXh3KOiPiTck6ZVGiJ1Ba5e53kA87v3L0VJ6AqyMnJ4aKLLmL8+PHEx5depzGGI26seZRu3bpx9913c/rpp9OgQQP69u1LeHi4V2ITCQn7tjrF1cLJkLcP2p4I5/wdOo6qvddeKeeIiD/VwZyjQkvEHyqzR2bzPJhyrnPKUXgUXDTRK9M5FxQUcNFFF3HVVVdx4YUXApCYmEhGRgZJSUlkZGTQvHnzY65n7NixjB07FoD77ruPVq1aHeMVInXAthXwyyuw/F/OF4fu58OwP0LygMDGpZwjIv6knFOmWrqbTaQOaj2YA5d8DKfeD2O+8ErysdYyduxYunXrxh133HGo/bzzzmPKlCkATJkyhdGjRx9zXdu3O/fE3LRpE5999hlXXnlljeMTCUrWwtrv4f0L4I0TIPULGHQT/HkJXPJu4IusylLOERF/qoM5R0e0RGoRT8sB0GW419Y3e/Zs3n//fXr16kXfvn0BeOqpp7jnnnu49NJLeeedd2jbti2ffPIJANu2bWPgwIHs27ePsLAwxo8fT2pqKvHx8Vx00UVkZWURGRnJq6++SqNGjbwWp0hQKMyHXz+Dn1+GzBUQ2wJGPuzMIlivcaCjqxblHBHxp7qWc1RoiYSwE088EffWK0eZMWPGUW0tWrRgy5YtZfafNWuWV2MTCRq5e2m96TOYcAtkb4Vm3WD0a9DrYoiIDnR0tYpyjoj4U23POSq0REREyrJnM8x9AxZO4bj8bGh/Mpz3EnQ8DY5xYbWIiIgKLRERkZIyljqnB65w76HS80IWRB3PwHPHBjYuEREJKiq0RERErIU1M+DnCbD+R4iKhaG3wJA/QKPW5KSkBDpCEREJMiq0RESk7irMg+X/dqZo354KcS1h1GPQfwzUaxTo6EREJIip0BIRkbrn4G5Y8C7MfRNytkHzHnD+G9DzIoiICnR0IiISAlRoiYhI3bF7I8x5HRa9BwX7ocMIOP81OO5UTXAhIiJepRsWi4SwzZs3M2LECLp3706PHj2YMGECALt27WLUqFF06tSJUaNGsXv3bgA++OADevfuTa9evRg2bBhLly49tK5vvvmGLl260LFjR8aNq8Qd4EVqk/RF8K/r4aW+MP9t6HYu/OEnuPZz6DhSRZaXKOeIiD/V9pyjQkskhEVERPDCCy+QmprKnDlzePXVV0lNTWXcuHGMHDmS1atXM3LkyEMJpX379vzwww8sX76cBx98kJtvvhmAoqIibrvtNr7++mtSU1P55z//SWpqaiCHJnJsHg/8Ng0mnwNvj4A138Hxf4Tbl8GFb0KLXoGOMOQo54iIP9X2nKNCS6QWWZ61nInLJ7Jk+xKvrC8pKYn+/fsDEBcXR7du3UhPT2fq1KmMGTMGgDFjxvD5558DMGzYMBo3bgzA0KFDD93Ub968eXTs2JEOHToQFRXF5ZdfztSpU70So4jXFeQ6pwa+NhQ+vBR2rYPTn4C/roDTH4eGyYGOsNZQzhERf6prOUfXaIn4wTPznmHlrpUV9snJz2HV7lVYLAZDl8ZdiI2KLbd/1yZduXvw3ZWOYcOGDSxevJghQ4aQmZlJUlIS4NwlPTMz86j+77zzDmeddRYA6enptG7d+tCyVq1aMXfu3EpvW8QvDuyCBe/A3Ldg/3bniNWFb0OPCyA8MtDR+ZVyjoj4k3JO2VRoidQS2QXZWCwAFkt2QXaFCagqcnJyuOiiixg/fjzx8fGllhljMEdcnzJz5kzeeecdfvrpJ69sX8Sndq2HOa/B4n9AwQHoeBoM+xO0P0XXXlVAOUdE/Kku5hwVWiJ+UJk9Mku2L+HGb2+k0FNIZFgk404aR9/mfWu87YKCAi666CKuuuoqLrzwQgASExPJyMggKSmJjIwMmjdvfqj/smXLuPHGG/n6669p2rQpAMnJyWzevPlQny1btpCcrNOvJMC2LICfX4a0L8CEQ+9L4fjbILFHoCMLOOUcEfEn5Zyy6RotkVqib/O+vHTiS/yx3x95+/S3vZJ8rLWMHTuWbt26cccddxxqP++885gyZQoAU6ZMYfTo0QBs2rSJCy+8kPfff5/OnTsf6j9o0CBWr17N+vXryc/P56OPPuK8886rcXwiVebxwMqvYNJZMHEkrJ0Jw/4Mf1nmTNOuIqvSlHNExJ/qYs7RES2RWqRX014MazfMa+ubPXs277//Pr169aJv374APPXUU9xzzz1ceumlvPPOO7Rt25ZPPvkEgMcee4ysrCxuvfVWwJnNZ8GCBURERPDKK69wxhlnUFRUxA033ECPHvpCK35UcBCWfgS/vAJZa6Bhazjjaeh/DUTHBTq6oKWcIyL+VNdyjgotkRB24oknYq0tc9mMGTOOaps4cSITJ04ss//ZZ5/N2Wef7dX4RI7pt+lOcbV1MeTthaQ+cNE70P18CNefsNpGOUdE/Km25xz9lRIRkdppbQp8eAlgwYTBWc/B4Js0wYWIiAQFXaMlIiK106zngeI9lQbys1VkiYhI0NARLRERqX12b4RNc5zZBAHCo6DdSYGNSUREpApUaImISO0z/UHnJsMXToSdK50iq/XgQEclIiJSaSq0RESkdtnwE6ROheH3QbffAb8LdEQiIiJVpmu0RESk9vAUwdf3ONO3D/tToKMRERGpNhVaIiFs8+bNjBgxgu7du9OjRw8mTJgAwK5duxg1ahSdOnVi1KhR7N69G4APPviA3r1706tXL4YNG8bSpUsPreuGG26gefPm9OzZMyBjkTpi0XuQuRxGPQZR9QMdjVSRco6I+FNtzzk+K7SMMZOMMduNMStKtD1njFlpjFlmjPmPMaZRiWX3GmPWGGNWGWPOKNF+ptu2xhhzT4n29saYuW77x8aYKLc92n2+xl3ezldjFKntIiIieOGFF0hNTWXOnDm8+uqrpKamMm7cOEaOHMnq1asZOXIk48aNA6B9+/b88MMPLF++nAcffJCbb7750Lquu+46vvnmm0AN5ZiUc0LAwT3w/ePQZhj0uCDQ0Ug1KOco54j4U23POb48ojUZOPOItulAT2ttb+A34F4AY0x34HKgh/ua14wx4caYcOBV4CygO3CF2xfgGeBFa21HYDcw1m0fC+x22190+4kEhbxly9j55lscWLzYK+tLSkqif//+AMTFxdGtWzfS09OZOnUqY8aMAWDMmDF8/vnnAAwbNozGjRsDMHToULZs2XJoXSeffDJNmjTxSlw+MhnlnOD2w7NwYBecNU7TuPuJck6NTEY5R6RK6lrO8dlkGNbaH4/cy2Kt/bbE0znAxe7j0cBH1to8YL0xZg1QPL3UGmvtOgBjzEfAaGNMGnAqcKXbZwrwCPC6u65H3PZ/A68YY4wt77bRIn6w7amnyEtbWWGfopwc8lauBGvBGKK7diU8Nrbc/tHdutLivvsqHcOGDRtYvHgxQ4YMITMzk6SkJABatGhBZmbmUf3feecdzjrrrEqvP9CUc4Lcjt9g3pvQ/xpI6hPoaIKeco7vKeeIHKacU7ZAzjp4A/Cx+zgZJyEV2+K2AWw+on0I0BTYY60tLKN/cvFrrLWFxpi9bv+dRwZgjLkZuBkgMTGRlJSUmo2olsnJyQm5MRULhrE1bNiQ7OxsAAryCygsKqqwf9HevU7yAbCWwr17sfXqlds/LL/g0PqPJScnhwsuuICnn34a4x4pKPlaY0yp5z/++CNvv/0206ZNK9Wek5ODx+OpcLu5ubm19f9GOcfHavK57LXsMRqaKObGjKSgFv67KOco51RDQHNOqOcbCI7PZXUEy7jqYs6par4JSKFljLkfKAQ+CMT2i1lr3wLeAhg4cKAdPnx4IMPxupSUFEJtTMWCYWxpaWnExcUBEPfIw8fsf2DxYjZddz22sBATGUmrF56nfr9+NY6joKCAiy++mGuuuYarrroKcP7o5uTkkJSUREZGBs2bNz8U67Jly/jzn//M119/Tbt27UqtKzY2lrCwsEN9yxITE0M/L8TtTco5/lHtz+Vv30LKQjj9CU4Ydr63w/IK5ZzKU86pHTkn1PMNBMfnsjqCZVx1MedUNd/4fdZBY8x1wDnAVSUOc6cDrUt0a+W2ldeeBTQyxkQc0V5qXe7yhm5/kVqtfr9+JLz2Ks3+/GfavDvJK8nHWsvYsWPp1q0bd9xxx6H28847jylTpgAwZcoURo8eDcCmTZu48MILef/99+ncuXONt18bKOfUcoX5MO0+aHIcDP59oKOpU5RzfEM5R6RsdTLnWGt99gO0A1aUeH4mkAo0O6JfD2ApEA20B9YB4ThH3Na5bVFunx7ua/4FXO4+fgO41X18G/CG+/hy4JPKxDpgwAAbambOnBnoEHwmGMaWmppa5dfs27fPqzHMmjXLArZXr162T58+tk+fPvbLL7+0O3futKeeeqrt2LGjHTlypM3KyrLWWjt27FjbqFGjQ31Lfi4uv/xy26JFCxsREWGTk5PtxIkTy9xmWeMGFlgf5priH+WcwKrW5/LnV6x9ON7aVd94PR5vUs6pHOWc2plzQjHfWBscn8vqCJZx1cWcU9V848vk808gAyjAObd4LLAG57ziJe7PGyX63w+sBVYBZ5VoPxtn5p61wP0l2jsA89x1/guIdttj3Odr3OUdKhNvKCahYPmgVkcwjK02JKBACNSXHuWcwKvy5zJnh7VPtbb2vQus9Xh8EpO3KOfUXso5x845oZhvrA2Oz2V1BMu46mLOqWq+8eWsg1eU0fxOBf2fBJ4so/0r4Ksy2tdxeMaeku25wCVVClZEgp5yThD6/gnIz4Ezn9Z07hJ0lHNE5Fj8fo2WiIgIGctg4WQYfDM06xLoaERERLxOhZaIiPiXtfDNvVCvMQy/O9DRiIiI+IQKLREfck7drTvq2nilmlKnwsaf4NT7nWJLvKaufQbr2nhFapu69BmszlhVaIn4SExMDFlZWXUmCVlrycrKIiYmJtChSG1WcBCmPwjNe0D/647Zff62+UxcPpEl25f4PLRgp5wjIv5Ul3JOdfNNQG5YLFIXtGrVii1btrBjx45KvyY3NzeovzTExMTQqlWrQIchtdkvr8CeTXDtFxBe8Z+gRZmLGDttLADR4dG8ffrb9G3e1w9BBiflHBHxp7qWc6qTb1RoifhIZGQk7du3r9JrUlJSqnTHcZGgsm8rzPo7dDsXOpxyzO7vLH8Hi7OntMBTwILMBSq0KqCcIyL+pJxzbCq0RETEP757BDxFMOrxY3bNyMlgzrY5hJkwDIbIsEgGJg70fYwiIiJeokJLRER8b/N8WPYxnHgHNKl4D6i1lifmPkG4CeeZU55hw74NDEwcqKNZIiISVFRoiYiIb3k88M3dENsCTrrjmN2nbZzGj1t+5K6Bd3Fa29P8EKCIiIj3qdASERHfWvYxpC+E89+A6LgKu+7N28u4uePo0bQHV3W7yk8BioiIeJ8KLRER8Z28bOfarOQB0PuyY3Z/ceGL7Mnbwxuj3iA8LNz38YmIiPiICi0REfGdWX+HnG1w2T8grOJbN87fNp9PV3/K9T2vp2uTrn4KUERExDd0w2IREfGNXevhl1edI1mtB1XYNa8oj8d+eYxWsa24pc8tfgpQRETEd3RES0REfGP6gxAWDqc9csyuby97mw37NvDmqDepF1HP97GJiIj4mI5oiYiI9637AdL+68wyGN+ywq5rdq/hnRXvcG6HcxnWcpifAhQREfEtFVoiIuJdRYXwzb3QqA0c/8cKu3qsh0d+eYTYyFjuHHSnnwIUERHxPRVaIiLiXYsmw/ZfYdTjEFnxaYCfrPqEpTuW8rdBf6NJTBP/xCciIuIHKrRERMR7Du6G75+EtidC99EVds3cn8n4ReM5Pul4zulwjp8CFBER8Q8VWiIi4j0pz0DuHjjzaTCmwq5Pz3uaIk8RDx7/IOYYfUVERIKNCi0REfGK+vs3wby3oP8YSOpdYd8ZG2cwY9MMbul7C63jWvspQhEREf9RoSUiIjVnLR3XTIKoWDj1gQq7Zudn89Tcp+japCvXdr/WTwGKiIj4l+6jJSIiNbf6W5rsXgxnPAUNEirsOmHRBHbm7mTCqROICNOfITm2vMIivliylbU7chjVvQUD2jYOdEgiIsekv3AiIlIzhfnwzb0cqJdM/UE3Vdh18fbFfLzqY67udjU9E3r6KUAJJrv355OWsY/UjH2kbnV+r87MpsiCASb/vIEPbhyqYktEaj0VWiIiUjPz3oRda1nT6yF6R0SV2y2/KJ9Hf36UpAZJ/Knfn/wYoNRG1lq27D7Ir1v3kbp176HCauve3EN9WsTH0L1lPI3qRzJ33S4sUFDoYc66LBVaIlLrqdASEZHqy9kOPzwLnU5nV9MBFXadtGISa/eu5dWRr1I/sr6fApTaIK+wiNWZOaWOUqVt3Ud2XiEAYQaOaxbLoPZN6J4UT4+WDemWFEfT2GgAFm7czVUT51BQ6CEyIoyhHZoGcjgiIpWiQktERKrv+8eh4IBzbdaK9HK7rdu7jreWvcVZ7c7i5FYn+zFA8be9BwqcgipjH79u3Uvq1n2s2Z5DoccCUD8qnG5J8ZzfL5nuLePpnhRPlxZxxESGl7vOAW0b88GNQ5mzLouhHZrqaJaIBAUVWiIiUj0ZS2HR+zD0VkjoBJRdaHmsh0d/fpR6EfX42+C/+TdG8ZniU/9KHqVK3bqP9D0HD/VpHhdN95bxnNq1Od1bOkeq2japT1hY1e+bNqBtYxVYIhJUVGiJiEjVWQtf3wP1m8ApFRdPn63+jEXbF/HYsMdIqFfxjIRSO+UXelizPafUUarUjH1k5x4+9a9Ds1gGtG3MNce3pXtSPN2S4mkWFx3gyEVEAkeFloiIVN2v/4FNP8M546Feo3K77Tiwg78v+DuDWgzi/I7n+ys6qYG9BwucWf9KHKVavT2bgiLn1L96keF0TYrjvD4t6dGyId1bxtMlMY56UeWf+iciUhep0BIRkarJPwDTH4LEXtC/4hsOPzP/GfKK8nho6EMYU/XTxcR3rLVkHfQwPTWz1FGqLbsPn/rXLC6a7knxnNKlGd2T4uneMp52TRsQXo1T/0RE6hqfFVrGmEnAOcB2a21Pt60J8DHQDtgAXGqt3W2cv74TgLOBA8B11tpF7mvGAA+4q33CWjvFbR8ATAbqAV8Bt1trbXnb8NU4RaR2UM7xo59fhr2b4YI3IKz8oxg/bP6BaRum8ad+f6Jdw3b+i0+OUlDknvq3tfT9qfYeLAAWYAy0T2hA39aNuHJIm0Oz/jWPiwl06LWWco6IHMsxCy1jzHHAFmttnjFmONAbeM9au+cYL50MvAK8V6LtHmCGtXacMeYe9/ndwFlAJ/dnCPA6MMRNJg8DAwELLDTGfOEmlNeBm4C5OAnoTODrCrYhIkFAOaeW27sFfnoRuo+GdieW221/wX6emPsEHRt15Poe1/sxQNmXW8DKjOxSR6lWZ+aQX+QBICYyjK4t4vld7yQisrdx/vCBdG0RR/2ounmSi3KOiPhKZbLqp8BAY0xH4C1gKvAhzl6ZcllrfzTGtDuieTQw3H08BUjBSQ6jcZKaBeYYYxoZY5LcvtOttbsAjDHTgTONMSlAvLV2jtv+HnA+TgIqbxsiEhyUc2qz7x4B64FRj1fY7eXFL5O5P5PnznqOyPBI/8RWx1hrydibe9RRqk27Dhzq07RBFN1bxnPDie0PTaXePuHwqX8pKVn0b1PnZ/JTzhERn6hMoeWx1hYaYy4AXrbWvmyMWVzN7SVaazPcx9uARPdxMrC5RL8tbltF7VvKaK9oG0cxxtwM3AyQmJhISkpKFYdTu+Xk5ITcmIqF6thCdVxVpJxTS8XvTaP/8n+xoe2lbFi6Hlhfannx+3dD3gY+3PYhJ8adyJ7UPaSkpgQkXm8K5Gdzze4iUrOKSKhnsMCmbA+b9nnYlO1hf4HTxwCJ9Q1t4sMY3CmSNvFhtIkLo2G0wZiDwEHYvY0tu0u/kZVzgBDNOcGebyojVN+/oTouCO2xlaUyhVaBMeYKYAxwrttW492T7nnGtqbrqck2rLVv4ey9YuDAgXb48OG+DMfvUlJSCLUxFQvVsYXquKpIOac28njg7YchriXtrhpPu6gGR3VJSUnhhJNP4KX/vUSz+s149pxniY2KDUCw3heIz+aW3QeYPHsDk+atx1PiXRUdEUbXFnGc1zGe7i0b0j0pnq4t4mgQXfVT/5RzgBDNOUGdbyopVN+/oTouCO2xlaUyWfl64A/Ak9ba9caY9sD71dxepjEmyVqb4R4y3+62pwOtS/Rr5balc/jweHF7itveqoz+FW1DRIKDck5ttPSfkLEELngLyiiyik35dQqrd69mwogJIVNk+Yu1lhXp+5ielsl3qZmkZuwrtTzMwHXD2nHf2d2ICA8LUJQhSTlHRHzimJnaWptqrf2ztfafxpjGQJy19plqbu8LnD1GuL+nlmi/1jiGAnvdw+LTgNONMY3dbZ8OTHOX7TPGDHVn8rn2iHWVtQ0RCQLKObVQ7j7n2qxWg6DXJeV221GwgzeWvsGotqM4tc2p/osviOUVFpGyajsPfL6c45/+nnNf+YlXvl9NbHQE953dlZev6EtMZBjhBqIiwvhd75YqsrxMOUdEfKUysw6mAOe5fRcC240xs621dxzjdf/E2UuTYIzZgjOrzjjgE2PMWGAjcKnb/Suci07X4Ex7ej2AtXaXMeZxYL7b77HiC0aBWzk87enX7g8VbENEgoByTi006wXYvx2u+AjCyv6Sb63lo10fERkWyT2D7/FzgMFl9/58Zq7azndpmfywagf784uoHxXOyZ2acVr3RE7t2pwmDaIO9W/ZqD5z1mUxtENTBrSt8xNXeJ1yjoj4SmVOHWxord1njLkRZ8ach40xy471ImvtFeUsGllGXwvcVs56JgGTymhfAPQsoz2rrG2ISNBQzqlNstbCnNegz5XQakC53b5Y+wW/5f7Gg0MfpHn95n4MMDhs2Lmf79Iy+TY1kwUbduGx0DwumtH9khnVPZHjOzQlJrLse5INaNtYBZZvKeeIiE9UptCKcM8BvhS438fxiIgo59Qm3z4IYZEw8qFyu2QdzOK5Bc/RIboDF3e+2I/B1V5FHsuSzXuYnprJd2mZrNmeA0DXFnHcNqIjo7on0rNlQ8LcadYloJRzRMQnKlNoPYZzDvFsa+18Y0wHYLVvwxKROkw5p7ZYOxNWfekUWfFJ5XZ7bsFz7C/Yz22JtxFm6u71Qwfzi5i1egffpWXy/crt7MzJJyLMMLRDU64e0oaR3RJp3aR+oMOUoynniIhPHLPQstb+C/hXiefrgIt8GZSI1F3KObVEUSF8cy80agtDyzzjCYDZ6bP5ct2X3NLnFpL2lF+Mhaod2XnMSHOOWs1avZO8Qg9xMRGM6NKc07onckrnZjSspxs212bKOSLiK5WZDKMV8DJwgts0C7jdWrul/FeJiFSPck4tsfBd2JEGl/0DImPK7HKg4ACPz3mc9g3bc2OvG/l51s9+DtL/rLWs3p5z6JTAJZv3YC0kN6rHFYPbMKp7IoPaNSEqou4e2Qs2yjki4iuVOXXwXeBDoHhO36vdtlG+CkpE6jTlnEA7sAtmPgntT4au55Tb7bUlr5Gek87kMycTFR5Vbr9gV1jkIS2riFn/S+W7tEw2Zh0AoE+rhtxxWmdO655I1xZxOLNwSxBSzhERn6hModXMWvtuieeTjTF/8VE8IiLKOYGWMg5y98KZ46Cc4iE1K5X3097n4s4XMyCx/NkIg1V2bgE//rbz0PVWew8WEBWxkROOa8rNJ3fgtG6JJMaXfaRPgo5yjoj4RGUKrSxjzNXAP93nVwBZvgtJROo45ZxA2p4G8yfCgOshsUeZXQo9hTzy8yM0iWnCXwf81c8B+s7WPQeZkZbJ9LTt/LJ2JwVFlsb1IzmtWyJJnh3ccsFwGkRX5s+mBBnlHBHxicr8xbgB59zlFwEL/Axc58OYRKRuU84JFGudCTCiY2FE+bNcf5D2AWm70njhlBeIj4r3Y4DeZa3l1637+C4tk+mpmfy6dR8A7RMacP0J7RnVPZH+bRoTHmZISUlRkRW6lHNExCcqM+vgRpw7ph9ijHkeuNNXQYlI3aWcE0CrvoZ1M+HMZ6BB0zK7bMnewqtLXmV4q+GMaht8l7DkFRYxd92uQ5NZZOzNxRgY0KYx957VldO6J3Jcs9hAhyl+pJwjIr5S3d1zl6IEJCL+o5zja4V5MO0+SOgCg8aW2cVayxNznsBguH/o/UEz+cOeA/nMXLWd71K388NvO8jJK6ReZDgndUrgjlGdObVrc5rGRgc6TKldlHNEpMaqW2gFx19X8bol25fw89afGdZyGH2b9w10OFJ3KOf42tw3YPd6uPpTCC/7vk9frf+K2Vtnc8/ge2jRoIWfA6yajVn7Dx21mr9hN0UeS7O4aM7t05JR3Zsz7LgEYiLDAx2m1F7KOSJSY+UWWsaYJuUtQgmoTpq/bT43fXsTRbaISSsmMfH0iSq2xGuUcwIoOxN+eA46nwkdTyuzy57cPTw7/1l6J/Tm8i6X+znAY/N4LEu27OE7t7j6LTMHgK4t4rjllOM4rXsivZMbEhamt5I4lHNExNcqOqK1EOei0LKSTb5vwpHaav3e9fzth79RZIsAZ9axBZkLVGiJNynnBMr3j0FhLpz+ZLldnl/wPPvy9vHQqIcID6sdR4IO5hcxe40zBft3advZmZNHeJhhSPsmXD6oDad1S6RN0/qBDlNqL+UcEfGpcgsta217fwYitdfUNVN5cu6ThJtwIsMi8VgPkWGRDEwcGOjQJIQo5wRI+iJY/AEM+yMkdCyzy5yMOUxdO5Ube91IlyZd/BxgaTuy85i5cjvfpmby05od5BZ4iIuO4JQuzRjVPZHhnZvTsH7Zpz6KlKScIyK+prlqpVz7C/bz5Jwn+e+6/zIwcSDjThpHxv4MFmQuYGDiQB3NEgl2xdO5128KJ99VZpfcwlwe++Ux2sS14fe9f+/nAGHhhl18uSKD/AJLasZeFm/eg7WQ3Kgelw1szajuLRjcvglREWF+j01ERKQiKrSkTGlZadz1411szt7MrX1v5eZeNxMeFk5ig0QVWCKhYsWnsHkOnPsSxDQss8sbS99gc/ZmJp4+kZiIGL+GN2v1Dq6dNA9rnecdEhrwl5GdGdU9kW5JcUEz66GIiNRNKrSkFGstH678kBcWvEDjmMZMPH0ig1oMCnRYIuJt+Qdg+kPQojf0u7rMLqt2rWLyr5M5v+P5DEka4t/wCj088J8Vh4qsMAMXDWjFbSPKPr1RRESktqnOrIMAWGt3eT8cCaQ9uXt48OcHSdmcwimtTuHxEx6ncUzjQIcldYRyjp/NngD70uGiiVDG5BZFniIe/eVRGkY35P8G/J9fQ7PWcs9ny9i46wCR4QaPxxIZEcbQDmXfRFmkOpRzRMTXKjvrYBtgt/u4EbAJ0EWkIWRh5kLu/vFusnKzuHvQ3VzV7SqdliP+ppzjL3s2w+zx0ONCaDuszC4frfqI5TuX88xJz9AoppFfw3vh29/4bFE6d4zqzAkdE5izLouhHZoyoK12/IhXKeeIiE8dc9ZBY8zbwH+stV+5z88CzvdLdOJzRZ4iJi6fyGtLXyM5Npl/nP0PejTtEeiwpA5SzvGj7x52fo96rMzFGTkZTFg0gROST+Cs9mf5MTD4YO5GXpm5hisGt+ZPp3bEGKMCS3xCOUdEfK0y0zQNLU4+ANbar4Gyd4FKUNl+YDs3T7+ZV5a8wpntzuSTcz5RkSW1gXKOL2382ZkE44TboVHroxZba3lyrnM/rQeHPujXI9vTUzN58PMVnNq1OY+P7qmj6uIvyjki4hOVmQxjqzHmAeAf7vOrgK2+C0n8YdaWWTww+wEOFh7ksWGPcX7H8/WlRmoL5Rxf8RTB13dDfLJTaJXh243f8sOWH7hz4J0kxyb7LbTFm3bzp38uomdyQ165sh8R4ZquXfxGOUdEfKIyf8muAJoB/3F/mrttEoQKigp4YcEL3DrjVprWa8pHv/uICzpdoCJLahPlHF9Z8gFsW+acMhjV4KjFe/P28vTcp+netDtXdbvKb2Gt37mfsVMW0DwuhknXDaJ+lCbEFb9SzhERnzjmXzN31p2yd31KUNmcvZm7f7yb5TuXc1mXy7hz4J1+vy+OyLEo5/hI7j6Y8Ri0HgI9Lyqzy4sLX2RP3h5eP+11IsL8U+zszMnjunfnYa1lyg2DSYiN9st2RYop54iIr1Q0vft4a+1fjDH/xZmVpxRr7Xk+jUy86psN3/Doz49iMPx9+N8Z1XZUoEMSKUU5x8d+fA7274ArP4EyjmDP3zafT1d/ynU9rqNb025+CelAfiFjJ88nc18uH940lPYJRx9lE/EV5RwR8bWKdlm+7/5+3h+BiG8cLDzIM/Oe4dPVn9K7WW+ePflZv153IVIFyjm+krUW5rwOfa+G5P5HLc4ryuOxXx4jOTaZW/rc4peQCos8/PHDxSxP38ub1wykfxvNLCh+p5wjIj5V0fTuC93fP/gvHPGmjPwMrvzyStbsWcPYnmO5rd9tRIZFBjoskTIp5/jQtPshIhpGPlTm4onLJ7Jh3wbeOO0N6kfW93k41loenLqC71du54nzezKqe6LPtylyJOUcEfG1Y56Eb4zpBDwNdAcOXdBjre3gw7ikBqy1fLr6U57b9hxx0XG8edqbDEvWTLVetXkebJgF7U6C1oMDHU1IUc7xsjUz4Lev4bRHIe7ogmbN7jVMXD6R33X4HSckn+CXkF75fg3/nLeZW4cfx9VD2/plmyLlUc4REV+pzNXO7wIPAy8CI4DrqdxshRIA2fnZPPbLY3yz4Ru6xHThjfPeIKFeQqDDCi6eIsjPgfz9kJcD+dklHufAtuUw5zWwHgiPhjFfqNjyLuUcbykqgGn3QeP2MPToUwI91sOjvzxKg8gG/G3Q3/wS0r8WbOaF6b9xYb9k7jqji1+2KXIMyjki4hOVKbTqWWtnGGOMtXYj8IgxZiFQ9jkoEjDLdyznrh/vYtv+bdze/3Y67OxQN4qsokKnGMpzi6P8HMjLLuexWyyV+3g/FByowrbznSNbKrS8STnHWxZMgh0r4fIPnVMHj/CvVf9iyY4lPHHCEzSJaeLzcH74bQf3fracEzsmMO6i3rqthNQWyjki4hOVKbTyjDFhwGpjzB+BdCC2Jhs1xvwVuBFnlp/lOHuPkoCPgKbAQuAaa22+MSYaeA8YAGQBl1lrN7jruRcYCxQBf7bWTnPbzwQmAOHARGvtuJrEW9t5rIf3U99n/MLxNKvfjMlnTqZv876kpKQEOrSyFeZXohg64ijSEcXQkL07YF6h06cwt5IbNhAVC9Gxzu+oBhAdB/Gt3MfF7bFlPy7+vXM1/Pt652hBeJRz+qB4k3KON+zPgplPQofh0OXsoxZn7s9k/KLxDEkawnnH+X5ytRXpe7n1HwvplBjH61f3JypCBwyk1lDOERGfqEyhdTtQH/gz8DjOYfUx1d2gMSbZXVd3a+1BY8wnwOXA2cCL1tqPjDFv4CSW193fu621HY0xlwPPAJcZY7q7r+sBtAS+M8Z0djfzKjAK2ALMN8Z8Ya1NrW7MtVnWwSwemP0AP6X/xMg2I3l02KM0jG7ouw2unwUrPoWmnaBhy/JPrytZGB060uQedfIUVG5bJgyi4o4ojGKhfgL7bFPqtTmu/GKorMeR9cuc1rrKEjrBmP/qGi3fUc7xhpSnnM/bGU+X+b5/et7TFHgKeGjoQz4/srR51wGunzyfhvUimXz9IOJiNCmP1CrKOSLiExXdR6s5cB/QEWdvzNPW2uu9uN16xpgCnOSWAZwKXOkunwI8gpOARruPAf4NvGKcbwWjgY+stXnAemPMGqD4G+8aa+06dxwfuX1DLgHNzZjLvbPuZW/eXu4fcj+XdbnMt1+YUqfCJ2Mo43YjjrAIt6iJK10YxSa6bQ2OXQxFNThcXEXElFsYpaWkkDh8uM+GekytB6vA8jLlHC/K/NU5bXDgWEjsftTiGRtnMGPTDP7S/y+0iW/j01D2HMjnunfnkVdQxIe3DCMxXjdJl9pBOUdEfK2iI1rv4Rzafhk4B3gJuK6mG7TWphtjngc2AQeBb93t7LHWFrrdtgDFN3tKBja7ry00xuzFOeyeDMwpseqSr9l8RPuQmsZdmxR6Cnl96eu8vext2sa35fXTXqdLEx9fVL7+R/js9xwqskwYDL4Zht56uEgq4xoQkSpQzinhwOLFHJg3n/qDB1G/X7/Kv9Ba+OYeiI6HEfcdtTg7P5un5j5F58adubbHtTUNs0K5BUXcOGUBm3cd5P2xg+mUGOfT7YlUkXKOiPhURYVWkrX2fvfxNGPMIm9s0BjTGGfPS3tgD/Av4ExvrLsasdwM3AyQmJhYe69pKmF34W4m75zMurx1DGkwhEsaXkLGsgwyyDiqb05OTs3HZC0tt35Fp9UTyY1uSpSnEOMpwpoIlua3Z9/S9TVbfzV5ZWy1UKiOq5KUc1yR69bR+MXxUFgIERHs/utfKOhQuZmmE3bMoef6H1nd8WbS5y07avknWZ+w4+AOrml4DbN/nF3doZSp5PvXYy2vLsljYWYRt/aJ5uCm5aRs8urm/CpUP5uhOq5KCumcE4zfcaoqVN+/oTouCO2xlaXCa7TcZFF87lZ4yefW2l3V3OZpwHpr7Q53G58BJwCNjDER7t6eVjgXo+L+bg1sMcZEAA1xLhYtbi9W8jXltZdirX0LeAtg4MCBdnggT0WrhO83fc/zs5+n0FPIUyc+xbnHnVth/5SUFGo0psJ8+OpOWD0FOp9JvQvfdmYwc69N6h/AU+dqPLZaKlTHVVnKOY6dq35jR1GRc3TK46FLQSEJlXlfFOTCa7dDs250uuJpOoWXTvFLti/hp69/4qpuV3Hd4OuOvb4qKn7/Wmt59L+pLMzcwAO/68aNJwX/7YhC9bMZquOqrFDOOcH2Hac6QvX9G6rjgtAeW1kqKrQa4hzqLnmRTPHeHgtU9y/nJmCoMaY+ziH1kcACYCZwMc6MPGOAqW7/L9znv7jLv7fWWmPMF8CHxpi/41wk2gmY58bbyRjTHifxXM7hc6JrpNqn8tRQflE+Lyx4gQ9Xfki3Jt147pTnaBvv45t85uyAT66BTb/AiXfAqQ9AWLiuTRJfUs5x1R88CBMVhS0owERGUn/woMq9cM5rsHsDXPM5HFFkFRQV8Ogvj5LYIJE/9vtjTcI7pomz1jP55w2MPbF9SBRZErKUc0TEp8ottKy17XyxQWvtXGPMv3GSWSGwGGePy5fAR8aYJ9y2d9yXvAO8714EugsnoWCt/dWdySfVXc9t1toiAHd61mk4055Ostb+WtO4DyxezKbrb8Dm52Oiomjz7iS/FFsb9m7gbz/+jbRdaVzd7Wr+OuCvRIVH+XajGUvhn1fCgZ1w0TvQ62Lfbk8E5ZyS6vfrR5t3J1Vtx072NvjxeWcq9+NGHLV40opJrNmzhldHvkqDyAY1Ca9CXyzdypNfpfG7Xkncf3Y3n21HpKaUc0TE1yozvbvXWWsfxrkLe0nrODybTsm+ucAl5aznSeDJMtq/Ar6qeaSHHZg3H5ufDx4PtqDA+QLk40Lrv2v/y+NzHicqPIqXRrzEiDZHf3nyuhWfwee3Qv0mcMM30NJ/R+5EfCUYc079fv2qlmNmPObcQPv0J45atH7vet5c9iZntDuDk1ud7MUoS0vLKuLF6UsZ3K4JL1zah7Aw3ZBY6qZgzDki4n0BKbSCUbVP5amGAwUHeHLuk3yx9gv6N+/PMyc/Q4sGLXy2PQA8HufmprOeh9ZD4LJ/QGxz325TRLwjfSEs+QBOuB2aHldqkcd6ePSXR4mJiOGewff4LIRV27J5aXEubZrG8ta1A4iJDPfZtkRERIKBCq1KqtapPNWwctdK7vrhLjbu28gf+vyB3/f+PRFhPv5vyt0H//k9rPoK+l0Dv3tBU7WLBAtr4eu7oUFzOOnOoxb/Z/V/WJi5kEeOf4SEegk+CSFj70Gue3ce0eGGKTcMplF9H5/eLCIiEgQq9Q3eGHMi0Mla+64xphkQa60NzLzeAVTlU3mqwFrLR6s+4vn5z9MwuiETT5/I4CQ/TDqxax388wrYuRrOeta5N5Yvb3osUgnKOVWw/F+wZT6MfhVi4kst2nlwJy8sfIEBiQO4oNMFPtn8vtwCrn93Ptm5hdw1IJrkRvV8sh0RX1LOERFfOGahZYx5GBgIdAHeBSKBf+BMVSpesDdvLw/NfojvN3/Pickn8uSJT9IkponvN7wuBT4Z4xRW13wGHYb7fpsix6CcUwX5+2H6w5DUF/ocPenYuHnjyC3M5eHjHybMhHl/84Ue/vD+QtZsz+Hd6wdRlK7r8SX4KOeIiK9U5ojWBUA/3ClPrbVbjTFxPo2qDlm8fTF3/3g3Ow7u4M6Bd3JN92t88oWoFGth7psw7T5I6AxXfAhNNAWz1BrKOZX103jI3gqXvAthpfPGD5t/YNqGadzW9zbaN2zv9U17PJa7/r2Un9dm8fdL+3BSp2aklHn3MJFaTzlHRHyiMoVWvns/BwtgjPHdvMB1SJGniEkrJvHqkldJapDE+2e9T8+Enr7fcGEefHkHLP4HdPkdXPgmROvvidQqyjmVsXsj/PwS9LwY2gwttWh/wX6emPsExzU8jrE9x/pk889OW8XUJVu564wuXNi/lU+2IeInyjki4hOVKbQ+Mca8iXNH85uAG4C3fRtWaNt5cCf3zLqHuRlzObPdmTx0/EPERfmh2MnOdG5CvHkunHwXDL/vqL3gIrWAck5lTH8IMDDq0aMWvbL4FTL3Z/LeWe8RGR7p9U2/98sG3vhhLVcNacOtw4879gtEajflHBHxiWMWWtba540xo4B9OOcvP2Stne7zyELU7PTZ3PfTfRwoOMAjxz/ChZ0uxPhj8omti+Gjq+DALrhkMvTwzYXxIjWlnFMJG36C1M9h+L3QsPTRpOU7lvNB2gdc2uVS+jbv6/VNf7NiGw9/8SundUvksdE9/ZO/RHxIOUdEfKUyk2HcAXyspFMzBZ4CXl78Mu+ueJeOjTryzunv0LFxR/9sfPm/Yept0KAZjP0Wknr7Z7si1aCccwyeIvjmHohvBcP+XGpRgaeAR355hGb1mnF7/9u9vumFG3dx+0eL6dOqES9f0Y9w3ZBYQoByjoj4SmVOHYwDvjXG7AI+Bv5lrc30bVihZUv2Fu7+8W6W7VzGxZ0v5m+D/ka9CD9MgWyL4LtH4KcXoc0wuPQ9iG3m++2K1IxyTkUWvw/blsPFkyCqfqlF7/36Hr/t/o3xw8d7/XTktTtyGDtlAUkNY3hnzEDqRemGxBIylHNExCeOeYGOtfZRa20P4DYgCfjBGPOdzyMLEd9u+JZL/3sp6/au47lTnuPh4x/2T5GVu5dey59yiqwB18G1U1VkSVBQzqnAwT0w43Foczz0uLDUok37NvH60tcZ2WYkI9uO9Opmt2fnMmbSPMKNc0PiprG6obmEDuUcEfGVSt2w2LUd2AZkAc19E07oyC3M5bn5z/HJb5/QK6EXz578LK3i/DQzV9Za+OflNN69Fs5+HgbdqJsQSzBSzjnSj8/BgSw489NSn2lrLY/NeYzIsEjuHXyvVze5P6+QsZMXkJWTz0c3D6VtU03IJiFLOUdEvOqYR7SMMbcaY1KAGUBT4CZrrS7yqcDaPWu54ssr+OS3T7i+x/VMOXOK/4qsNTPg7RGwfyfLej8Kg29SkSVBRTmnHDtXw9w3oN/V0LJvqUVfrP2CuRlz+Uv/v5DYINFrmywo8nDrB4v4deteXr2qH31aN/LaukVqC+UcEfGVyhzRag38xVq7xMexBD1rLZ+v+Zyn5j5F/cj6vH7a65yYfKK/Ng5zXoNvH4Bm3eCKD9mzdIN/ti3iXco5ZZl2P0TUg5EPlWrelbuL5xY8R99mfbmkyyVe25y1lvv/s5wfftvB0xf24tSu3ivgRGoZ5RwR8YlyCy1jTLy1dh/wnPu8Scnl1tpdPo4tqOTk5/DYnMf4ev3XDGkxhKdOeorm9f105kFBLvzvr7D0Q+h6DlzwJkTHAhv8s30RL1DOqcDq6bB6Gox6HGJL55Vn5z/L/oL9PHz8w4QZ790Xb/x3q/lkwRb+fGpHrhjcxmvrFaktlHNExNcqOqL1IXAOsBCwQMnzzyzQwYdxBZVfd/7KXT/eRXpOOn/q9yfG9hxLeJifZuTK3ubcHyt9gXNPnZP/ppsQS7BSzilLUQF8cy80OQ6G/KHUotnps/ly3Zf8vvfvvXq7iI/nb2LCjNVcPKAVfx3V2WvrFalllHNExKfKLbSstee4v9v7L5zgYq3l/dT3eXHRizSNacq7Z7xL/8T+/gsgfaFTZOXudaZu7z7af9sW8TLlnHLMexuyVsMVH0NE1KHmAwUHeHzO47SLb8dNvW/y2uZmrtzOff9Zwcmdm/H0hb10Q2IJWco5IuJrlZkMY0Zl2uqa3bm7+eP3f+S5Bc9xUvJJfHrep/4tspZ+DJPOgrBI5ybEKrLEzxZu3M2rM1ezcONur65XOaeE/TshZRwcdyp0PqPUoteXvk56TjoPHf8Q0eHemW592ZY93PrBIrq2iOO1q/oTGa6j4xL6lHNExFcqukYrBqgPJBhjGnP4kHo8kOyH2Gqt+dvmc8+P97A7bzf3DL6HK7te6b+9vh73JsQ/vwRtT4RLp0CDBP9sWwLO47HkF3nIK/CQV1REfqHH+SnyHH5c6CHviOfFy/MKi47Zp/hxXsn1FRaVWn4wv4jcQg8AMZFr+ODGoQxo27hGY1POOcLmefDNPZCXDWc8XWr20NSsVN5LfY+LOl3EoBaDvLK5TVkHuGHyfJo0iOLd6wYRG12Vu3+IBB/lHBHxtYr+kv4e+AvQEuf85eIEtA94xbdh1U6LMhfx2pLXmLdtHm3i2/DKyFfo1rSb/wI4uAc+vRHWTHfujXXmOAiP9N/2BYDcgiK+/XUb8zfspkuLONonNDhUwOSVU/gcLnRKLztU+LhtO3cfZNySH48qeoofF3qs18YRFRFGdHgYURElfsJLP46LiSC6xPPoiHCiIsJIy9jHwo27sUBBoYc567JqXGihnHPY5nkw+RwoyoOwcMjbd2hRoaeQR35+hMbRjfnrgL96ZXO79ucz5t15FBRZPrp5MM3jY7yyXpFaTjlHRHyqomu0JgATjDF/sta+7MeYaqXZ6bO55btbsFjCTBgPDHnAv0XWztXwzytg93o450UYeIP/tl0HFXksW/ccZN3O/azfkcP6nftZt3M/63bsJ33PwSqvzxicgiU8jKiI8FLFS8nCJjbS0KJJ/aOLoPAwoiPDiAoPL1UYHdnnyKIp+sjnkeFEhYcRGW5qdBR24cbdXDVxDgWFHiIjwhjaoWm111VMOaeEDbPAU+A8tu7z1oMB+CDtA9J2pfHcKc/RMLphjTd1ML+IsVPmk77nIB/eOISOzWNrvE6RYKCcIyK+dsxzQ6y1LxtjegLdgZgS7e/5MrDaZsXOFVicowkGw4qsFQxtOdQ/G189Hf49FsIj4NovoN0J/tluiLPWsmt//qEiav3O/axzi6oNWQfId0+NA4iLjqBDswYMateYdtkN+HntTiwQZuCyQW24bFDrQ4VOWUVURFjlCpuUlBSGDx/ow1F7x4C2jfngxqHMWZfF0A5NvXE06xDlHKDdSRAeDUX5EB7lPAe2ZG/h1SWvckqrUzij7RnHWMmxFXkst3+0mCWb9/D6Vf0Z2K7JsV8kEmKUc0SOkPpfZwdfr4sP7eST6jlmoWWMeRgYjpOAvgLOAn4C6lQCGpI0hInLJ1LgKSAyLJKBiX74Mmwt/PwyfPcwNO8BV3wIjXQ/m6o6kF/Ihp0HShVSxYXV3oMFh/pFhhvaNm1A+4QGjOjSnA7NGtA+IZb2CQ1IiI06VCgt3LibhRN3HTqac/GAVvRt3ShAowucAW0be7XAKqacg/OHbcwXzh+6didB68FYa3li7hMYDPcPub/G14Vaa3nki1/5NjWTR87tzpk9k7wUvEhwUc4RceXsgK/ugtT/OM8Xvef8LVKxVW2Vudr5YqAPsNhae70xJhH4h2/Dqn36Nu/L26e/zYLMBQxMHEjf5n19u8GCg/Df22HZx86Mgue/DlENfLvNIFZY5CF9z0HW7SguotyCasd+MvbmlurbsmEM7Zs14Nw+SXRIiKV9swZ0SGhAcqN6RFRiljVfHs0RQDnH0XpwqT9uX63/itnps7ln8D0kxda8KHrjh3W8P2cjN5/cgetO0OzWUqcp50jdlrvX2bH/y2tQcOBwe1F+qVPXpeoqU2gdtNZ6jDGFxph4YDvQ2sdx1Up9m/f1fYEFsG+rc3+srYtgxP1w8l2lZhyrq6y17MjJY/2O/fywuYBfvkpj7Q6nqNq06wAFRYcnioiPiaBDs1iOP64pHRIOH5lqn9CAelE1v5m0r47mCKCcc5Q9uXt4dv6z9Gzak8u7XF7j9X2+OJ1nvlnJuX1acs+ZXb0QoUhQU86Ruin/AMx7C356EXL3QI8LoOu5MPW2o05dl+qpTKG1wBjTCHgbZ1aeHOAXXwZVp21Z4BRZ+Tlw2QfQ7ZxAR+R3OXmFbDg0+YRzZGr9zv2s37Gf7LzCQ/2iIjbQvmkDOjWP4/QeLWif0IDj3NP9GteP1I1Wg5dyzhFeWPgCe/P28taotwgPq9mOgtlrdnLXv5cytEMTnr+kN2Fh+pxInaecI3VLYT4smgI/Pgc5mdDpdDj1AUjq4yxv1LrUqetSfZWZDONW9+EbxphvgHhr7TLfhlVHLfnQOV0wLgmu+Q8kdg90RD5TUORh864DrNuxv8Q1Uzms27Gf7dl5h/oZA8mN6tE+oQEX9k+mQzPnyFTmmuVceOYIwvUlMeQo55Q2N2Mun6/5nLE9x9KlSZcarSstYx9/eH8hHRJiefOagURH1PzorkiwU86ROsNTBMv/BTOfgj0boc0wuGQKtD2+dL8jTl2X6qvohsX9K1pmrV3km5DqoKJCZ8KLX16B9ic7b/r6wT/7l7WW7dl5rC0+KrXj8CQUm3YdoKjEPaGaNIiifUIDTunc7NA1U+0TYmnbtD4xkUd/GUzZGqYiK8Qo5xwttzCXx355jNZxrflDnz/UaF3pew5y3bvzaBAdwbvXD6JhPd2DT+o25RypM6yFlV/C90/AjjRo0Ruu+hQ6jtSlKT5W0RGtFypYZoFTvRxL3XRwN/z7Blj7PQz+PZzxZK2+CfHCjbuPmgRiX24B60scmSp5ut+B/KJDr42JDKNd0wZ0T4rnd72SnGum3KKqUf2oQA1Jag/lnBKWbF/ChEUT2JS9ibdPf5uYiOrfRHjvgQKumzSPA3lF/OuW42nZqJ4XIxUJWso5EvrWpcCMxyB9ITTtBJdMhm6jIezYk39JzVV0w+IR/gykTtqxyrkJ8Z5NcO5LMGBMoCOq0MKNu7ny7TnkF3oIM4bOibHsyMlnZ87hU/3CDLRuUp/2CQ0Y3L4JHRIaHDrdr0V8jK4HkXIp5xy2ZPsSxk4bS74nnzATRkx49YusvMIibn5/ARuy9jPl+sF0bRHvxUhFgpdyjoS0zfPh+8dg/Y8Q3wrOewX6XOHck1X8pjL30bq2rPaa3MjPveh0ItATZ6/RDcAq4GOgHbABuNRau9s4MxpMAM4GDgDXFR/ON8aMAR5wV/uEtXaK2z4AmAzUw7knxu3W2sPnqdUGv02DT2+EiGgY89+jz4+tZXbtz+fJL1PJc2/iW2Qt2bmFjOxafL+pBnRo1oDWTerrug+pEeUcWJC5gEJPYann1Znx1OOx/N8nS5m7fhcTLu/LsI4J1Q1JJGQp50hIyfwVvn8SVn0J9RPgzGdg4PXO903xu8qUtYNKPI4BRgKLqNmN/CYA31hrLzbGRAH1gfuAGdbaccaYe4B7gLtxbhzYyf0ZArwODDHGNAEeBgbiJLGFxpgvrLW73T43AXNxEtCZwNc1iNd7rIXZ4+G7R6FFL7j8Q2d2l1oqt6CIKT9v4JWZa8jJLXSui7KWyIgwJlzRT1Ociy/U+ZwzMHEgUeFRNb5B+tNfp/G/ZRncc1ZXRvdNrm44IqGuzuccCQG71sHMp53JLqLjnVkEh9wC0bGBjqxOq8ysg38q+dzdS/NRdTdojGkInAxc564/H8g3xozGuTM7wBQgBScBjQbec/fUzDHGNDLGJLl9p1trd7nrnQ6caYxJwZkxaI7b/h5wPrUhARUchC/+5HwIelwIo1+FqPqBjqpM1lr+tyyDZ75ZyZbdBxnRpRn3nd2NfbmFulGv+JRyjndukD7pp/W8PWs9Y45vy+9P7lDdUERCnnKOBLV9W51p2he9B2GRcMLtzk8ITKoWCqpzouZ+oH0Nttke2AG8a4zpg3PPituBRGtthttnG5DoPk4GNpd4/Ra3raL2LWW0B9bedPjoSshYCiMfghPvqLUzvSzcuIsnvkxj8aY9dEuK5x9je3Nip8OnHKnAEj+rkzmnJjdI/2p5Bo9/mcoZPRJ56NweuqecSNXUyZwjQebALvjp7zDvbWfa9gHXw8l3QlyLQEcmJVTmGq3/4hyyBggDugOf1HCb/YE/WWvnGmMm4Bw+P8Raa40xPj/X2BhzM3AzQGJiIikpKT7ZTvzeNHquGEeYJ4+0nveRVTQAfvjBJ9sqKScnp0pj2n7Aw79+y2f+tiIaRRvG9ozihORCCtNXkJLuuziro6pjCxahOq6qUM6pmVW7inhuQS7HNQzjopbZzPrR97mmWCi/f0N1bKE6rqoI1Zzjr+84gRSq79+KxhVeeIBWW/5L682fE150kMzE4WxodwW59RJh4UpgpV9jrapQ/T8rT2WOaD1f4nEhsNFau6W8zpWwBdhirZ3rPv83TgLKNMYkWWsz3EPm293l6UDJi5hauW3pHD4EX9ye4ra3KqP/Uay1bwFvAQwcONAOHz68rG41s+h9mPUQxCfDFR/Rq3lX72+jHCkpKVRmTHsPFPDy96uZ8ssGIsLC+Mtpnbj55A7Uj6q9M9NUdmzBJlTHVUXKOdW0Zns2t7/+C22aNOBftwyjcQP/3jYhlN+/oTq2UB1XFYVkzvHLd5wAC9X3b5njKsiFBe/AvBfgQBZ0PQdOfYAWzbsRTMewQvX/rDzHnETfWvuDtfYHYDGQBhxwL9CsFmvtNmCzMaaL2zQSSAW+AIrnNx8DTHUffwFcaxxDgb3uofdpwOnGmMbGmMbA6cA0d9k+Y8xQdyafa0usy3+KCuHre+CLP0LbYXDT9+DHIqsy8gs9TPppPac8P5N3Zq/ngn7JpNw1nL+c1rlWF1kS2pRzqidzXy5jJs0nMjyMKTcM9nuRJRKslHOkVisqhIVT4OX+MO0+52bDN30Pl38AzbsFOjo5hsqcOngz8BiQC3gAg3OIvSZXV/8J+MCdiWcdcD1O0feJMWYssBG41O37Fc6Up2twpj29HsBau8sY8zgw3+33WPEFo8CtHJ729Gv8fYHogV3w7+udm8QNvRVGPV6r7ltgrWXar9sY9/VKNmQd4KROCdx3dje6Jen+OhJ4yjlVl51bwPXvzmf3gXw+vvl4WjepnZPsiNRGyjlSK3k88OtnMPMp2LUWWg2CC96A9icHOjKpgsp8+78L6Gmt3emtjVprl+BMV3qkkWX0tcBt5axnEjCpjPYFOPeu8L/tac5NiPelO7MK9rs6IGGUZ+nmPTz5ZRrzNuyiU/NY3r1+EMM7N9PF8lKbKOdUQUGRh1s/WMSqzGwmjhlIr1YNAxGGSDBTzpHaw1qaZC2ANx+AzOXQvAdc8RF0PrPWTqIm5atMobUWZw+LHMvKr+CzmyCyPlz3JbQeHOiIDtmy+wDPTVvF1CVbSYiN4skLenLZwNZEhB/z7FERf1POqSRrLXd/uoxZq3fy7EW9GdGleaBDEglGyjlSO2yYDTMeo/fmOdC4PVw4EXpeBGH6rhasKlNo3Qv8bIyZC+QVN1pr/+yzqIKNtTDreedO3El9nJsQN6wdM63uyy3gtZlrmTR7PQa4bcRx/OGU44iLiQx0aCLlUc6ppBe+/Y3PFqXzl9M6cemg2nvjc5FaTjlHAmvrYpjxOKydAXFJrOp8C10uexzC9V0t2FWm0HoT+B5YjnPuspSUfwCm3uacR9vrEjjvZYisF+ioKCzyMGNTAf83K4Ws/flc2C+ZO8/oQstGgY9N5BiUcyrhg7kbeWXmGi4f1JrbR3YKdDgiwUw5RwJjxyqY+SSkToV6jZ1r+gffRMbsuXRRkRUSKlNoRVpr7/B5JMFoz2bnJsTblsNpjzp34g7w+bPWWr5fuZ2nvkpj7Y58hrRvwuTfddd1GxJMlHOO4bvUTB78fAUjujTjifN76hpLkZpRzhH/2rMJUsbB0n86l5uccg8cfxvEaFKyUFOZQutrd0ae/1L6kPqu8l9SB2z8BT6+Gory4cpPoPPpgY6IFel7eeqrNH5em0WHhAb8uV80f710qL6ESbBRzqnAks17+OM/F9GjZUNeubK/rrMUqTnlHPGPnO3w4/OwYBKYMGdm6hP/Cg0SAh2Z+EhlCq0r3N/3lmir6bSnwW3hZPjyTmjUxpkJplnngIazbW8uz01bxWeLt9CoXiSPnteDK4e0YfasH1VkSTBSzinHhp37uWHyfJrFRTPpukE0iK49t40QCWLKOeJbB/fAzy/BnNehMA/6XwMn/63WXM8vvnPMv9LW2vb+CCQobPwZpj8EW+bDcSPh4necc2oDZH9eIW/+sJa3Zq3D44GbT+rArSM60rCezuuV4KWcU7adOXmMeXce1lqmXD+YZnHRgQ5JJCQo54jP5O+HuW/C7PGQuxd6Xgwj7oOmxwU6MvGTytyw+Nqy2q2173k/nFpszQz44GKwHgiLgJPvCliRVeSx/GvBZp7/9jd25uRxTu8k7j6zq25SKiFBOedoB/ILGTt5Ptv25vLhTUPp0Cw20CGJhAzlHPG6wnzn7Kcfn4P92517YJ36ALToFejIxM8qc97JoBKPY3ButrcIqFsJKH2hU2SBM537pp+h7fF+D+OH33bw1JdprMrMZkDbxrx17QD6twncUTURH1DOKaGwyMOfPlzM8vS9vHH1AAa01eddxMuUc8Q7PEWw7GNIedqZ8KLtiXDZP6DNkEBHJgFSmVMH/1TyuTGmEfCRrwKqtToMh1l/dya/CI+Cdif5dfOrtmXz5Fdp/PjbDto0qc9rV/XnrJ4tdA2WhBzlnMOstTw49VdmrNzO46N7cHqPFoEOSSTkKOdIjVkLaf+F75+AnasgqS+cMx6OOzXgs1FLYFXnSur9QN07n7n1YBjzBWyY5RRZrQf7ZbPbs3N5cfpvfDx/M7HRETzwu25cc3xboiPC/bJ9kVqgTuachRt3M376b8xas5Nbhh/HNce3C3RIInVFncw5Ug3WwrqZMOMx56bDCZ3h0veg23kqsASo3DVa/8WZfQcgDOgOfOLLoGqt1oP9VmAdzC/i7VnreOOHtRQUebhuWHv+PLIjjepH+WX7IoGinOMUWZe/9QsFRZYwAyO7Ng90SCIhSzlHqmXzPKfA2jALGraB81+H3pdBmHaEy2GVOaL1fInHhcBGa+0WH8VT53k8ls8Wp/P8tFVs25fLmT1acM9ZXWmX0CDQoYn4S53POXPWZVHkcb73GWDu+l0MbNcksEGJhK46n3OkCratcE4R/O1raNAcznoOBoyBCM0EK0crt9AyxnQEEq21PxzRfoIxJtpau9bn0dUxP6/dyZNfpvHr1n30adWQl67ox+D2+nIldYNyzmFDOzQlKiKMgkIPkRFhDO3QNNAhiYQc5Rypkqy1MPMpWPEpxMTDyIdgyB8gSjvCpXwVHdEaT+mb9xXb5y471wfx1Elrtucw7us0vkvbTnKjeky4vC/n9m5JWJjO75U6ZTzKOQAMaNuYD24cypx1WQzt0FQzDYr4xniUc+RY9qbDj8/Covedo1Yn3QHD/hTQ+6hK8Kio0Eq01i4/stFau9wY0853IdUdWTl5jP9uNR/O20T9yHDuPrMr15/QjphInd8rdZJyTgkD2jZWgSXiW8o5Ur79WfDT32He287tfQbdCCf9H8QlBjoy8YOFG3d7ZWdnRYVWowqW1av2FoXcgiLenb2B12au4UBBEVcObsNfTutE01id3yt1WqMKlinniIi3NapgmXJOXbV2Jvz0ImyeD0W50OcKOOVuaNw20JFJDRV5LPsOFrC3nJ/iZRuy9jNv/S48FqIjwvjwpqHVLrYqKrQWGGNusta+XbLRGHMjsLBaW6vjPB7Lf5dt5dlvVpG+5yCndWvOPWd1pWPzuECHJlIbKOeIiD8p58hhe7c4k1ws/afz3ITBhW9Br0sCG5eUUljkObpAyi0sVSjtPVB2EZWdV1jhuqMiwmhYL5Iij8Wdj4qCIg9z1mX5pND6C/AfY8xVHE44A4Eo4IJqba0Om79hF0/8L5WlW/bSPSme5y7uzbCOCYEOS6Q2+QvKOSLiP39BOadus9aZnn3eW7DyS+f5IQb2bApYaKGsoKxiqYIiqeTy/flFFa47JtIplop/WjaKoWtSXKm2sn7i60UeunRn4cbdXDVxjlcmpCq30LLWZgLDjDEjgJ5u85fW2u+rvbU6aMPO/Yz7eiXf/LqNFvExPH9JHy7sl6yJLkSOoJwjIv6knFOH5eXAso+d6692pEG9JjDsz5A8AD67GYryITwK2p0U6EgDojLXJ+UXesoulI5xWt6unFzyvvm6wu3XiwwvVQS1alyfhi1LFkYRNKx/dKHUsF4k0RE1n+fAmxNSHfM+WtbamcDMam+hjtpzIJ+XZqzh/TkbiAwP445RnbnppA7Ui9JEFyIVUc4REX9SzqlDstbC/Imw+API2wtJfWD0a9DzQoh0L8uLa+Ec5Wp3ErQeHNh4/cBaS3ZeITuz89iZk8+89VmM/241RR5LWJjhpI4JRISHHVVIHSyo+MhS/ajwUkVQ6yb16Vkvkuyd2+jRuUOZRZLzOMIrxVJNeWtCqsrcsFiqIK+wiPd/2chLM1aTk1fIZYNa89fTOtM8PibQoYmIiIjULR4PrJnunB645jsIi4Qe58Pgm6HVIDBHnGHUenDQF1jWWvYdLGRHTh47svPYmVPiJzu/xHPncV6hp8z1FHksCzfuJrlxPRrWi6Rt0/qlT7mrf3Sh1LBeJPExkURFhJW5zpSU3Qwf3smXw69VVGh5ibWWr1dsY9zXK9m06wAnd27GfWd3pWuL+ECHJiIiIlK3HNztHLma/zbs3gCxLWD4fTDguqCcot3jsew5WOAWS3nsKFEo7cwuXThl5eSTX3R08RQeZmjSIIqE2GgSYqM4rlksCXHOY6ctmh3Zedz3n+UUFjnXJ02+YbBuNVIDKrS8YPGm3Tz5ZRoLNu6mS2IcU24YzCmdmwU6LBEREZG6ZdsKp7ha9gkUHIA2w2Dkw9DtXAiPDHR0pRR5LLv25x91xGnRqny+2L7EKZzcImrX/nwKPfaodUSEGadIinOKpS4t4g4VUs3iog8VUAmxUTSuH1WpOQLaJTTwyvVJokKrRjbvOsAz36zkf8sySIiN5ukLe3HpwNaEa6ILEREREf8oKoCV/3Mmt9g4GyLqQe9LYNBNkNTbr6EUFnnYtT+/xGl7ZR91Ki6eyqidiDDQfFcWCXHRtGgYQ8/k+MMFk3sEqpn7vGG9SK9PsOat65NEhVa17D1YwGsz1/Du7A2EhcGfT+3IzaccR2y0/jlFRERE/CEyfw/88BwsmATZW6FRWxj1OPS7Guo3qdY6y5pxL7/QQ9b+w9c37SjzeieniNp9IL/0LPGumMiwQ8VSq8b16demUYmjTe7pe+4RqEVzfmLEiBE1+JeR2kKVQRXMW5/Fmz+uY966LHLyi7iwXyvuPKMzSQ11A3kRERERn7MW0hfC3Dc5fsVnYAvhuFPhnBeh0ygIq96MdbkFzmRm475Z6cy4ZyCpYQw5eUXsPVhQ5mvqR4WTEBtNs7ho2ic0YFC7JoeOOjUrcd1TQlw0DaLCMUdOvFGOyvaT2k+FViXNXJnJDZMXYIEwA89e1JtLBrYOdFgiIiIioa8gF379zJk9cOtiiIpja8szaXXBI5BQvVnsdmTnMXPldqanZfLT6p0cLCgirN5Gouqvo/BABxpE9+DUromlroFKiI12TtuLi6J+lL5GS8X0Dqmk1Ix9FB8JNsD27LxAhiMiIiIS+vZsdk4NXDQFDmRBQhc4+3noczlrfllIqyoUWdZaVmVmMyNtO9NTM1m6JQsi9tCs0X4G9SokN2INqfucW6pF2QjGDpjAZb17HmOtIuVToVVJQzskEBO5hoJCZ7rLoR2aBjokERERkdBjrXPT4LlvwqqvnLYuZzv3vmp/8tH3vipHoaeQLfsymLFmJbPW/8aK7RvIKdyOidxNTP29xHXdg8XDQWDxQTAYjHF2q4eZIrLNKuAk34xR6oSAFVrGmHBgAZBurT3HGNMe+AhoCiwErrHW5htjooH3gAFAFnCZtXaDu457gbFAEfBna+00t/1MYAIQDky01o6rabwD2jbmgxuHarpLkSAVbDlHRIKX8k015eXAso+c2QN3rIR6TeCE22HgDdCozVHdPdbDtv3b2JK9ha37t5Kek87WnK1s3LuZ9Xu2sKdgB1DiflKxhiYRTWnXsDXtGvakZWxLkmOTSY5NpmVsS7Yd2MYt02+hwFNAZFgkAxMH+m/sEpICeUTrdiANKL6j7zPAi9baj4wxb+Akl9fd37uttR2NMZe7/S4zxnQHLgd6AC2B74wxnd11vQqMArYA840xX1hrU2sasKa7FAlqQZdzRCRoKd9Uxc41zr2vlnwIefsgqQ+Mfg1PjwvYUZDN1v1b2bJ2KVtzth4qqNKznaLKs6n0jXkjbCPyDjbEU5BEPdOb7s3ackK7zozq1I12jZKJrOBeWq3iWvH26W+zIHMBAxMH0rd5Xx8PXEJdQAotY0wr4HfAk8Adxple5VTgSrfLFOARnCQ02n0M8G/gFbf/aOAja20esN4YswYY7PZbY61d527rI7dvcCchEak25RwR8Rflm0ryFOH5bRpZ894gPf0XtkZGk96uF+kJHdhqC9i64Z9sXfECBZ7SM/4l1EugZWxLejTtSfO8roRFtee39Ei27qyPLWxI9xZNOa1bc0Z2S6RXcsMq32Oqb/O+KrDEawJ1RGs88Dcgzn3eFNhjrS10n28Bkt3HycBmAGttoTFmr9s/GZhTYp0lX7P5iPYhXo5fRILLeJRzRMQ/xqN8AziTT2TlZjlHonLcI1F71rF12yLSs7eQYTzkhYVByxbOC/K30GTXAZJjk+napCuntjmVVrGtaBnbkpaxLYkNa8acddnMSMtk+vwd7D1YQFR4GMcf15Q/DG7Oqd0SSW6kW+5I7eH3QssYcw6w3Vq70Bgz3N/bPyKWm4GbARITE0lJSQlkOF6Xk5MTcmMqFqpjC9VxBZJyjv+E8vs3VMcWquMKlGDON5Hr1hH5228UdO5MQYcOldqGtZb9nv1kFWaxq3BX6d9Fzu8CW/qIVKMiD8mFBbQz9elZvwv1YnvSNKoZTcKb0CSiCdFh0e7KgWzYnunhq+0HWbx9Fb/tTqPIQlwk9GkeQdc4y4BWMdSLOAB5G1i9ZAOrq/oPVQuF8ucylMdWlkAc0ToBOM8YczYQg3P+8gSgkTEmwt3j0wpId/unA62BLcaYCKAhzgWjxe3FSr6mvPZSrLVvAW8BDBw40A4fPrzGg6tNUlJSCLUxFQvVsYXquAJMOcdPQvn9G6pjC9VxBVBQ5psDixezYcIEKCiAyEjaTZ5M/X79sNayL3+fcyTKnWii5O/0nHQOFh4sta74qHiSY5PpGduTlg1a0DIni1Yb5tEyYwUtiaRBr4ud2QNb9CozliKPZcnm3XyXtp3vUjNZvd1Zf6fmsdx8SiKndWtO39aNCQ8zIfv+DdVxQWiPrSx+L7SstfcC9wK4e3vutNZeZYz5F3Axzqw8Y4Cp7ku+cJ//4i7/3lprjTFfAB8aY/6Oc6FoJ2Aezm2uOrkz/KTjXExafF60iNQxyjki4i/Bmm/WzpyKyc8n3EJRfj5T/nEX321uSHpOOvsL9pfqGxsZS3JsMq3jWjM0aeihGfuKf8dFxUF2pnPfq1mTIDsDGrWFEQ9D36ugfpOjtr8/r5BZq3fwXdp2Zq7cTtb+fCLCDIPbN+GKwW04rVsibZrWr+kwRfyuNt1H627gI2PME8Bi4B23/R3gffdC0F04SQVr7a/GmE9wLgAtBG6z1hYBGGP+CEzDmfp0krX2V7+ORESCgXKOiPhLrc43qW3C6BIOFEFhOMxLOkDLBt0YmDiw1PTnyXHJxEfFl70Sa2HLApj3Fvz6H/AUwHEj4Zzx0GkUhIWX6r51z0FmpGXyXdp2flmbRX6Rh/iYCEZ0bc5p3RI5uXMzGtYrf4ZAkWAQ0ELLWpsCpLiP13F4Rp2SfXKBS8p5/ZM4s/oc2f4V8JUXQxWREKCcIyL+Ekz5psvJ5zHu6v/QeUMBq9tFcfe1r1V+5r2CXFjxqVNgZSyB6HgYdKPzk9DxUDePx7Ji616+S3WKq9SMfQC0T2jAtce3ZWS3RAa2a0xkeJg3hyYSULXpiJaIiIiI+Fnf5n2554Z3WZC5gAsre/+oPZtgwSRYOAUO7oJmXeF3L0DvyyDamXDxYH4Rs9fsZMbKTGakbWd7dh5hBga2bcK9Z3XltO6JHNcs1reDEwkgFVoiIiIidVyl7h9lLaz/0Tl6tco9qNblbGdyi/YngzFs35fLjKWbmJGWyU9rdpJb4CE2OoJTOjdjZLfmjOjSnMYNonw+HpHaQIWWiIiIiJQvLxuWfgTz3oadq6BeEzjhdhg4FtuwFWkZ2cz4fg3fpWWydMteAFo1rsflg9owsltzhrRvSlSETgmUukeFloiIiIgcbedqmD8RlnwIefsgqS+c/zp5XUczZ9MBZvyQyYy01aTvOYgx0KdVI+46owsjuzWnS2IcxphAj0AkoFRoiYiIiIjDUwSrp8O8N2Ht9xAWCT0uYG+v65i+rw0zVmznx89msT+/iHqR4ZzYKYHbR3ZiRNfmNIuLDnT0IrWKCi0RERGRum71dzD3ddi2AnK2YeOSyBp8F/+LOJ3/rS1k4bu7sXYPifHRjO6XzKhuiRx/XFNiIsOPvW6ROkqFloiIiEhdtnke9sNLMNaDxfB9y5t5cs/prPsxH9hBz+R4/nxqJ0Z1T6RHy3idEihSSSq0REREROqw9CXf0sJjCTdQZA2LN++lbYeGjD0lkZFdE2nRMCbQIYoEJRVaIiIiInXYL0Xd+R2RRNpCCoigdb/TufPCo+6vLCJVpEJLREREpA5r328E1y98gAH2VxaaHtw1YGSgQxIJCSq0REREROqwAW0bc9eN1zJnXRZ3dWjKgLaNAx2SSEhQoSUiIiJSxw1o21gFloiX6TbdIiIiIiIiXqZCS0RERERExMtUaImIiIiIiHiZCi0REREREREvU6ElIiIiIiLiZSq0REREREREvEyFloiIiIiIiJep0BIREREREfEyFVoiIiIiIiJepkJLRERERETEy1RoiYiIiIiIeJkKLRERERERES9ToSUiIiIiIuJlKrRERERERES8TIWWiIiIiIiIl6nQEhERERER8TIVWiIiIiIiIl6mQktERERERMTL/F5oGWNaG2NmGmNSjTG/GmNud9ubGGOmG2NWu78bu+3GGPOSMWaNMWaZMaZ/iXWNcfuvNsaMKdE+wBiz3H3NS8YY4+9xikjtoJwjIv6knCMixQJxRKsQ+D9rbXdgKHCbMaY7cA8ww1rbCZjhPgc4C+jk/twMvA5OwgIeBoYAg4GHi5OW2+emEq870w/jEpHaSTlHRPxJOUdEgAAUWtbaDGvtIvdxNpAGJAOjgSlutynA+e7j0cB71jEHaGSMSQLOAKZba3dZa3cD04Ez3WXx1to51loLvFdiXSJSxyjniIg/KeeISLGAXqNljGkH9APmAonW2gx30TYg0X2cDGwu8bItbltF7VvKaBeROk45R0T8STlHpG6LCNSGjTGxwKfAX6y1+0qeXmyttcYY64cYbsY5TE9iYiIpKSm+3qRf5eTkhNyYioXq2EJ1XLWBco7vhfL7N1THFqrjqg0CnXNCPd9A6L5/Q3VcENpjK0tACi1jTCRO8vnAWvuZ25xpjEmy1ma4h8W3u+3pQOsSL2/ltqUDw49oT3HbW5XR/yjW2reAtwAGDhxohw8fXla3oJWSkkKojalYqI4tVMcVaMo5/hHK799QHVuojivQakPOCfV8A6H7/g3VcUFoj60sgZh10ADvAGnW2r+XWPQFUDyjzhhgaon2a91ZeYYCe91D79OA040xjd2LQ08HprnL9hljhrrburbEukSkjlHOERF/Us4RkWKBOKJ1AnANsNwYs8Rtuw8YB3xijBkLbAQudZd9BZwNrAEOANcDWGt3GWMeB+a7/R6z1u5yH98KTAbqAV+7PyJSNynniIg/KeeICBCAQsta+xNQ3v0eRpbR3wK3lbOuScCkMtoXAD1rEKaIhAjlHBHxJ+UcESkW0FkHRUREREREQpEKLRERERERES9ToSUiIiIiIuJlKrRERERERES8TIWWiIiIiIiIl6nQEhERERER8TIVWiIiIiIiIl6mQktERERERMTLVGiJiIiIiIh4mQotERERERERL1OhJSIiIiIi4mUqtERERERERLxMhZaIiIiIiIiXqdASERERERHxMhVaIiIiIiIiXqZCS0RERERExMtUaImIiIiIiHiZCi0REREREREvU6ElIiIiIiLiZSq0REREREREvEyFloiIiIiIiJep0BIREREREfEyFVoiIiIiIiJepkJLRERERETEy1RoiYiIiIiIeJkKLRERERERES9ToSUiIiIiIuJlKrRERERERES8TIWWiIiIiIiIl6nQEhERERER8bKQLbSMMWcaY1YZY9YYY+4JdDwiEtqUc0TEn5RzRGq/kCy0jDHhwKvAWUB34ApjTPfARiUioUo5R0T8STlHJDiEZKEFDAbWWGvXWWvzgY+A0QGOSURCl3KOiPiTco5IEAjVQisZ2Fzi+Ra3TUTEF5RzRMSflHNEgkBEoAMIJGPMzcDN7tMcY8yqQMbjAwnAzkAH4SOhOrZQHVfbQAdQGyjnBLVQHVuojqvO55w6kG8gdN+/oTouCM2xlZtvQrXQSgdal3jeym0rxVr7FvCWv4LyN2PMAmvtwEDH4QuhOrZQHVcdoJxDaL9/Q3VsoTquOuCYOSfU8w2E7vs3VMcFoT22soTqqYPzgU7GmPbGmCjgcuCLAMckIqFLOUdE/Ek5RyQIhOQRLWttoTHmj8A0IByYZK39NcBhiUiIUs4REX9SzhEJDiFZaAFYa78Cvgp0HAEWyqcMhOrYQnVcIU85Bwjt92+oji1UxxXylHOA0H3/huq4ILTHdhRjrQ10DCIiIiIiIiElVK/REhERERERCRgVWiHCGNPaGDPTGJNqjPnVGHO7297EGDPdGLPa/d040LFWhTEmxhgzzxiz1B3Xo257e2PMXGPMGmPMx+7FwEHJGBNujFlsjPmf+zxkxiahKVTzDYR+zlG+kWCknBO8n826nnNUaIWOQuD/rLXdgaHAbcaY7sA9wAxrbSdghvs8mOQBp1pr+wB9gTONMUOBZ4AXrbUdgd3A2MCFWGO3A2klnofS2CQ0hWq+gdDPOco3EoyUc4L3s1mnc44KrRBhrc2w1i5yH2fjvKmTgdHAFLfbFOD8gARYTdaR4z6NdH8scCrwb7c96MZVzBjTCvgdMNF9bgiRsUnoCtV8A6Gdc5RvJFgp5wTn2JRzVGiFJGNMO6AfMBdItNZmuIu2AYmBiqu63MPOS4DtwHRgLbDHWlvodtmCk3CD0Xjgb4DHfd6U0Bmb1AGhlm8gpHPOeJRvJMgp5wSV8dTxnKNCK8QYY2KBT4G/WGv3lVxmnSkmg26aSWttkbW2L86d7wcDXQMbkXcYY84BtltrFwY6FpHqCMV8A6GZc5RvJBQo5wQP5RxHyN5Hqy4yxkTiJKAPrLWfuc2Zxpgka22GMSYJZ29JULLW7jHGzASOBxoZYyLcvSKtgPTARlctJwDnGWPOBmKAeGACoTE2CXGhnm8g5HKO8o0ENeWcoPtsKuegI1ohwz3v9R0gzVr79xKLvgDGuI/HAFP9HVtNGGOaGWMauY/rAaNwzs2eCVzsdgu6cQFYa++11ray1rYDLge+t9ZeRQiMTUJbqOYbCN2co3wjwUw5J/jGppzj0A2LQ4Qx5kRgFrCcw+fC3odzDvMnQBtgI3CptXZXQIKsBmNMb5yLJcNxdgx8Yq19zBjTAfgIaAIsBq621uYFLtKaMcYMB+601p4TamOT0BOq+QbqRs5RvpFgo5wT3J/NupxzVGiJiIiIiIh4mU4dFBERERER8TIVWiIiIiIiIl6mQktERERERMTLVGiJiIiIiIh4mQotERERERERL1OhVccYY4qMMUuMMb8aY5YaY/7PGOPV94Expp0xZoU31+kNxpj7Kli2wRiz3BizzBjzrTGmRRXX/XMV+082xlx87J4iwU05p9xlyjkiPqCcU+4y5ZwAUKFV9xy01va11vbAuSneWcDDAY7JX8pNQK4R1trewIJK9AXAGBMBYK0dVsPYREKVck75lHNEvE85p3zKOX6mQqsOs9ZuB24G/mgc4caY54wx8909Hr8HMMZ8ZIz5XfHrivdSlNe/JGNMjDHmXXcvymJjzAi3/TpjzFRjTIoxZrUx5mG3vZ0xZqW7jd+MMR8YY04zxsx2+w12+zUwxkwyxsxz1zu6xHo/M8Z84/Z/1m0fB9Rz93J9cIx/mh+BjhX8eww3xswyxnwBpLptOe5v475mhTvmy0q0v2KMWWWM+Q5oXs3/NpGgpZxTLuUcER9QzimXco6/WGv1U4d+gJwy2vYAiTjJ6AG3LRpnj0d74AJgitseBWwG6lXQvx2wwm3/P2CS+7grsAmIAa4DMoCm7rpWAAPd1xYCvXB2BCwEJgEGGA187q7rKZy7iQM0An4DGrjrXQc0dLezEWhd3thL/BtsABLcx68Az1QwvuHAfqD9kf+uwEXAdJw7vCe6400CLizR3tL9N7840O8H/ejH1z/KOeX+uyjn6Ec/PvhRzin330U5JwA/EYgcdjrQ2xw+p7Yh0An4GphgjIkGzgR+tNYeNMaU1/+3Eus8EXgZwFq70hizEejsLpturc0CMMZ85vb9HFhvrV3utv8KzLDWWmPMcpwEVRzrecaYO93nMUAb9/EMa+1e9/WpQFucpHksM40xRcAy4AFgYjnjywfmWWvXl7GOE4F/WmuLgExjzA/AIODkEu1bjTHfVyIekVCnnKOcI+JPyjnKOX6lQquOM8Z0AIqA7Th7U/5krZ1WRr8U4AzgMuCj4uay+htj2lVy87ac53kl2jwlnns4/J41wEXW2lVHbHvIEa8vovLv8xHW2p0l1lXe+Ibj7OkRkSpSzilFOUfEx5RzSlHO8TNdo1WHGWOaAW8Ar1jnePA04BZjTKS7vLMxpoHb/WPgeuAk4Bu3raL+xWYBVxUvx9kbU5w0Rhljmhhj6gHnA7OrEP404E9uksAY068SrykojrUK2zjW+I40C7jMPe+5Gc4ennk450MXtycBI6oQh0hIUM6p1DaUc0S8RDmnUttQzvEhHdGqe+oZY5YAkTjnCL8P/N1dNhHnkPUi94O9AycxAHzr9p1qrc2vRP9irwGvu4fDC4HrrLV5bt6YB3wKtAL+Ya1dUIW9RI8D44Flxpm2dT1wzjFe85bbf5G19qpKbKMy4zvSf4DjgaU4e67+Zq3dZoz5D3AqzkWlm4BfKrF9kVCgnKOcI+JPyjnKObWGcQp8Ef8yxlwHDLTW/jHQsYhI6FPOERF/Us4R0KmDIiIiIiIiXqcjWiIiIiIiIl6mI1oiIiIiIiJepkJLRERERETEy1RoiYiIiIiIeJkKLRERERERES9ToSUiIiIiIuJlKrRERERERES87P8BkCb2sYt+R0EAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#3.4. Plot cumulative paid claims triangles\n",
    "\n",
    "fig, (ax1,ax2,ax3) = plt.subplots(1, 3,figsize=(12, 5))\n",
    "\n",
    "\n",
    "ax1.plot(cumulative_triangle_df['Motor Insurance'].T,marker='.')\n",
    "ax1.set_title('Motor Insurance')\n",
    "ax1.set_ylabel('Cumulative Paid Loss')\n",
    "ax1.set_xlabel('Development Period')\n",
    "ax1.set_ylim(0, 150000)\n",
    "ax1.legend(cumulative_triangle_df['Motor Insurance'].T.columns)\n",
    "ax1.grid()\n",
    "\n",
    "\n",
    "ax2.set_title('Fire Insurance')\n",
    "ax2.plot(cumulative_triangle_df['Fire Insurance'].T,marker='.')\n",
    "ax2.set_ylabel('Cumulative Paid Loss')\n",
    "ax2.set_xlabel('Development Period')\n",
    "ax2.set_ylim(0, 150000)\n",
    "ax2.legend(cumulative_triangle_df['Fire Insurance'].T.columns)\n",
    "ax2.grid()\n",
    "\n",
    "\n",
    "ax3.set_title('Micro Insurance')\n",
    "ax3.plot(cumulative_triangle_df['Micro Insurance'].T,marker='.')\n",
    "ax3.set_ylabel('Cumulative Paid Loss')\n",
    "ax3.set_xlabel('Development Period')\n",
    "ax3.set_ylim(0, 150000)\n",
    "ax3.legend(cumulative_triangle_df['Micro Insurance'].T.columns)\n",
    "ax3.grid()\n",
    "\n",
    "\n",
    "plt.tight_layout(pad=1.0, w_pad=1.0, h_pad=1.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "6fff0da6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Motor Insurance age-to-age factors:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "#T_3561e_row0_col0 {\n",
       "  background-color: #fb7c5c;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_3561e_row0_col1, #T_3561e_row1_col0 {\n",
       "  background-color: #67000d;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_3561e_row0_col2, #T_3561e_row1_col1, #T_3561e_row1_col2, #T_3561e_row2_col0, #T_3561e_row2_col2 {\n",
       "  background-color: #fff5f0;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_3561e_row2_col1 {\n",
       "  background-color: #000000;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "</style>\n",
       "<table id=\"T_3561e_\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th class=\"col_heading level0 col0\" >12-24</th>\n",
       "      <th class=\"col_heading level0 col1\" >24-36</th>\n",
       "      <th class=\"col_heading level0 col2\" >36-48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_3561e_level0_row0\" class=\"row_heading level0 row0\" >2018</th>\n",
       "      <td id=\"T_3561e_row0_col0\" class=\"data row0 col0\" >1.4349</td>\n",
       "      <td id=\"T_3561e_row0_col1\" class=\"data row0 col1\" >1.0812</td>\n",
       "      <td id=\"T_3561e_row0_col2\" class=\"data row0 col2\" >1.0139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_3561e_level0_row1\" class=\"row_heading level0 row1\" >2019</th>\n",
       "      <td id=\"T_3561e_row1_col0\" class=\"data row1 col0\" >1.5150</td>\n",
       "      <td id=\"T_3561e_row1_col1\" class=\"data row1 col1\" >1.0206</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_3561e_level0_row2\" class=\"row_heading level0 row2\" >2020</th>\n",
       "      <td id=\"T_3561e_row2_col0\" class=\"data row2 col0\" >1.3707</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fire Insurance age-to-age factors:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "#T_525be_row0_col0, #T_525be_row0_col1, #T_525be_row0_col2, #T_525be_row1_col2, #T_525be_row2_col2 {\n",
       "  background-color: #fff5f0;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_525be_row1_col0, #T_525be_row1_col1 {\n",
       "  background-color: #67000d;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_525be_row2_col0 {\n",
       "  background-color: #980c13;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_525be_row2_col1 {\n",
       "  background-color: #000000;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "</style>\n",
       "<table id=\"T_525be_\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th class=\"col_heading level0 col0\" >12-24</th>\n",
       "      <th class=\"col_heading level0 col1\" >24-36</th>\n",
       "      <th class=\"col_heading level0 col2\" >36-48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_525be_level0_row0\" class=\"row_heading level0 row0\" >2018</th>\n",
       "      <td id=\"T_525be_row0_col0\" class=\"data row0 col0\" >1.7387</td>\n",
       "      <td id=\"T_525be_row0_col1\" class=\"data row0 col1\" >1.1150</td>\n",
       "      <td id=\"T_525be_row0_col2\" class=\"data row0 col2\" >1.0634</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_525be_level0_row1\" class=\"row_heading level0 row1\" >2019</th>\n",
       "      <td id=\"T_525be_row1_col0\" class=\"data row1 col0\" >1.8663</td>\n",
       "      <td id=\"T_525be_row1_col1\" class=\"data row1 col1\" >1.1215</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_525be_level0_row2\" class=\"row_heading level0 row2\" >2020</th>\n",
       "      <td id=\"T_525be_row2_col0\" class=\"data row2 col0\" >1.8535</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro Insurance age-to-age factors:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style type=\"text/css\">\n",
       "#T_687e9_row0_col0 {\n",
       "  background-color: #ed392b;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_687e9_row0_col1, #T_687e9_row0_col2, #T_687e9_row1_col2, #T_687e9_row2_col0, #T_687e9_row2_col2 {\n",
       "  background-color: #fff5f0;\n",
       "  color: #000000;\n",
       "}\n",
       "#T_687e9_row1_col0, #T_687e9_row1_col1 {\n",
       "  background-color: #67000d;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "#T_687e9_row2_col1 {\n",
       "  background-color: #000000;\n",
       "  color: #f1f1f1;\n",
       "}\n",
       "</style>\n",
       "<table id=\"T_687e9_\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th class=\"blank level0\" >&nbsp;</th>\n",
       "      <th class=\"col_heading level0 col0\" >12-24</th>\n",
       "      <th class=\"col_heading level0 col1\" >24-36</th>\n",
       "      <th class=\"col_heading level0 col2\" >36-48</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th id=\"T_687e9_level0_row0\" class=\"row_heading level0 row0\" >2018</th>\n",
       "      <td id=\"T_687e9_row0_col0\" class=\"data row0 col0\" >1.3048</td>\n",
       "      <td id=\"T_687e9_row0_col1\" class=\"data row0 col1\" >1.0696</td>\n",
       "      <td id=\"T_687e9_row0_col2\" class=\"data row0 col2\" >1.0480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_687e9_level0_row1\" class=\"row_heading level0 row1\" >2019</th>\n",
       "      <td id=\"T_687e9_row1_col0\" class=\"data row1 col0\" >1.4176</td>\n",
       "      <td id=\"T_687e9_row1_col1\" class=\"data row1 col1\" >1.3655</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th id=\"T_687e9_level0_row2\" class=\"row_heading level0 row2\" >2020</th>\n",
       "      <td id=\"T_687e9_row2_col0\" class=\"data row2 col0\" >1.1109</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#3.5 Explore the Link ratios before adjusment of development factors\n",
    "for lob in LoB:\n",
    "    print( lob +' age-to-age factors:')\n",
    "    display(cumulative_triangle_df[lob].link_ratio.heatmap(cmap='Reds'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bf7f554b",
   "metadata": {},
   "source": [
    "## 4. Actuarial Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "98605913",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Motor Insurance calendar year effect test:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>2019</th>\n",
       "      <th>2020</th>\n",
       "      <th>2021</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "       2019   2020   2021\n",
       "2018  False  False  False"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fire Insurance calendar year effect test:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>2019</th>\n",
       "      <th>2020</th>\n",
       "      <th>2021</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "       2019   2020   2021\n",
       "2018  False  False  False"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro Insurance calendar year effect test:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>2019</th>\n",
       "      <th>2020</th>\n",
       "      <th>2021</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018</th>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "       2019   2020   2021\n",
       "2018  False  False  False"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Motor Insurance development factors test:\n"
     ]
    },
    {
     "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>Paid Amount</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LoB</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Motor Insurance</th>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 Paid Amount\n",
       "LoB                         \n",
       "Motor Insurance        False"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fire Insurance development factors test:\n"
     ]
    },
    {
     "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>Paid Amount</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LoB</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Fire Insurance</th>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                Paid Amount\n",
       "LoB                        \n",
       "Fire Insurance        False"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro Insurance development factors test:\n"
     ]
    },
    {
     "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>Paid Amount</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LoB</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Micro Insurance</th>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 Paid Amount\n",
       "LoB                         \n",
       "Micro Insurance        False"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#4.1. Peform an indepenence test as a suitable check for using the chainladder method\n",
    "# The multiplicative chainladder method is based on the strong assumptions of independence across origin years and across valuation years. Mack developed tests to verify if these assumptions hol\n",
    "# False indicates that correlation between years is not sufficiently large.\n",
    "\n",
    "#Mack's test for the calendar year effect\n",
    "for lob in LoB:\n",
    "    print( lob +' calendar year effect test:')\n",
    "    display(cumulative_triangle_df[lob].valuation_correlation(0.05).z_critical)\n",
    "\n",
    "#Mack's test for correlations between subsequent development factors\n",
    "for lob in LoB:\n",
    "    print( lob +' development factors test:')\n",
    "    display(cumulative_triangle_df[lob].development_correlation(0.05).t_critical)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "ecaf26db",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Motor Insurance losses: \n"
     ]
    },
    {
     "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>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "      <th>Ultimate Losses</th>\n",
       "      <th>Claims Reserve</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Loss Year</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>2018</th>\n",
       "      <td>42,977</td>\n",
       "      <td>61,666</td>\n",
       "      <td>66,671</td>\n",
       "      <td>67,595</td>\n",
       "      <td>67,595</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>53,141</td>\n",
       "      <td>80,509</td>\n",
       "      <td>82,170</td>\n",
       "      <td></td>\n",
       "      <td>83,309</td>\n",
       "      <td>1,139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>66,462</td>\n",
       "      <td>91,102</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>96,695</td>\n",
       "      <td>5,593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>77,671</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>118,288</td>\n",
       "      <td>40,617</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Total</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>365,887</td>\n",
       "      <td>47,349</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              12     24     36     48  Ultimate Losses Claims Reserve\n",
       "Loss Year                                                            \n",
       "2018      42,977 61,666 66,671 67,595           67,595               \n",
       "2019      53,141 80,509 82,170                  83,309          1,139\n",
       "2020      66,462 91,102                         96,695          5,593\n",
       "2021      77,671                               118,288         40,617\n",
       "Total                                          365,887         47,349"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fire Insurance losses: \n"
     ]
    },
    {
     "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>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "      <th>Ultimate Losses</th>\n",
       "      <th>Claims Reserve</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Loss Year</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>2018</th>\n",
       "      <td>51,227</td>\n",
       "      <td>89,067</td>\n",
       "      <td>99,308</td>\n",
       "      <td>105,600</td>\n",
       "      <td>105,600</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>66,979</td>\n",
       "      <td>125,004</td>\n",
       "      <td>140,195</td>\n",
       "      <td></td>\n",
       "      <td>149,078</td>\n",
       "      <td>8,883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>61,853</td>\n",
       "      <td>114,642</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>136,388</td>\n",
       "      <td>21,746</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>79,464</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>172,586</td>\n",
       "      <td>93,122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Total</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>563,652</td>\n",
       "      <td>123,751</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              12      24      36      48  Ultimate Losses Claims Reserve\n",
       "Loss Year                                                               \n",
       "2018      51,227  89,067  99,308 105,600          105,600               \n",
       "2019      66,979 125,004 140,195                  149,078          8,883\n",
       "2020      61,853 114,642                          136,388         21,746\n",
       "2021      79,464                                  172,586         93,122\n",
       "Total                                             563,652        123,751"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Micro Insurance losses: \n"
     ]
    },
    {
     "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>12</th>\n",
       "      <th>24</th>\n",
       "      <th>36</th>\n",
       "      <th>48</th>\n",
       "      <th>Ultimate Losses</th>\n",
       "      <th>Claims Reserve</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Loss Year</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>2018</th>\n",
       "      <td>33,198</td>\n",
       "      <td>43,318</td>\n",
       "      <td>46,332</td>\n",
       "      <td>48,554</td>\n",
       "      <td>48,554</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019</th>\n",
       "      <td>33,924</td>\n",
       "      <td>48,092</td>\n",
       "      <td>65,670</td>\n",
       "      <td></td>\n",
       "      <td>68,819</td>\n",
       "      <td>3,149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020</th>\n",
       "      <td>38,104</td>\n",
       "      <td>42,328</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>54,351</td>\n",
       "      <td>12,023</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021</th>\n",
       "      <td>39,006</td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>63,656</td>\n",
       "      <td>24,650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Total</th>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td></td>\n",
       "      <td>235,380</td>\n",
       "      <td>39,822</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              12     24     36     48  Ultimate Losses Claims Reserve\n",
       "Loss Year                                                            \n",
       "2018      33,198 43,318 46,332 48,554           48,554               \n",
       "2019      33,924 48,092 65,670                  68,819          3,149\n",
       "2020      38,104 42,328                         54,351         12,023\n",
       "2021      39,006                                63,656         24,650\n",
       "Total                                          235,380         39,822"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#4.2. Apply the basic chain-ladder algorithim to predict claims reserves\n",
    "cum_tri_df= {}\n",
    "ult_df = {}\n",
    "ibnr_df = {}\n",
    "cum_ult_df = {}\n",
    "combined_df = {}\n",
    "\n",
    "for lob in LoB:\n",
    "    print(lob + ' losses: ')  \n",
    "    cum_tri_df[lob] = incremental_triangle_df[lob].incr_to_cum().to_frame()\n",
    "    ult_df[lob] = cl.Chainladder().fit(cumulative_triangle_df[lob]).ultimate_.to_frame()\n",
    "    ibnr_df[lob] = cl.Chainladder().fit(cumulative_triangle_df[lob]).ibnr_.to_frame()\n",
    "    cum_ult_df[lob] = pd.merge(cumulative_tri_df[lob],ult_df[lob], how= 'left',on=[cum_tri_df[lob].index, ult_df[lob].index])\n",
    "    cum_ult_df[lob].drop('key_1',axis=1,inplace=True)\n",
    "    cum_ult_df[lob].rename(columns={'key_0':'Loss Year','2261':'Ultimate Losses'}, inplace=True)\n",
    "    combined_df[lob] = pd.merge(cum_ult_df[lob],ibnr_df[lob],how='left', on=[cum_ult_df[lob]['Loss Year'],ibnr_df[lob].index])\n",
    "    combined_df[lob].drop('key_0',axis=1,inplace=True)\n",
    "    combined_df[lob].drop('key_1',axis=1,inplace=True)\n",
    "    combined_df[lob].rename(columns={'2261':'Claims Reserve'}, inplace=True)\n",
    "    combined_df[lob].set_index('Loss Year',inplace=True)\n",
    "    combined_df[lob].loc[\"Total\"] = combined_df[lob].loc[:,['Ultimate Losses','Claims Reserve']].sum()\n",
    "    display(combined_df[lob].replace(np.nan, \"\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "a50310fb",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Fire Insurance</th>\n",
       "      <th>Micro Insurance</th>\n",
       "      <th>Motor Insurance</th>\n",
       "      <th>Total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Basic Chain-Ladder Method</th>\n",
       "      <td>123,751</td>\n",
       "      <td>39,822</td>\n",
       "      <td>47,349</td>\n",
       "      <td>210,921</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Fire Insurance  Micro Insurance  Motor Insurance  \\\n",
       "Basic Chain-Ladder Method         123,751           39,822           47,349   \n",
       "\n",
       "                            Total  \n",
       "Basic Chain-Ladder Method 210,921  "
      ]
     },
     "execution_count": 162,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#4.3. Sumarize claims reserve totals by line of business\n",
    "claims_reserve = {}\n",
    "keys = LoB\n",
    "for lob in LoB:\n",
    "    claims_reserve[lob]=cl.Chainladder().fit(cumulative_triangle_df[lob]).ibnr_.sum()\n",
    "    \n",
    "results = pd.DataFrame.from_records(claims_reserve, index =['Basic Chain-Ladder Method'])\n",
    "results['Total']=results.sum(axis=1)\n",
    "results"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}