{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Practical introduction\n",
    "\n",
    "## Predicting forest covertypes\n",
    "\n",
    "![trees](images/trees.jpg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Before we dive into the manifold machine learning approaches in the context of remote sensing, let's first take a look at a classic application: Predicting forest covertypes based on cartographic data. The samples in this dataset correspond to 30×30m patches of forest in the US (Colorado). The study area is depicted below:\n",
    "\n",
    "![](images/tree_cov_aoi.jpg) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "Image source: {cite}`Blackard1999`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "The data set is described in detail here: https://archive.ics.uci.edu/ml/datasets/Covertype\n",
    "\n",
    "It was also used in a kaggle competition: https://www.kaggle.com/c/forest-cover-type-prediction/overview"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "The cover types are encoded as:\n",
    "\n",
    "- 1: Spruce/Fir\n",
    "- 2: Lodgepole Pine\n",
    "- 3: Ponderosa Pine\n",
    "- 4: Cottonwood/Willow\n",
    "- 5: Aspen\n",
    "- 6: Douglas-fir\n",
    "- 7: Krummholz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Our goal is to build a model that can predict these cover types using all or part of the 54 predictors given in the data set.\n",
    "\n",
    "The original paper that was published on this data set achieved an overall accuracy of **70.58%.**\n",
    "\n",
    "Let's see if we can beat that."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "# general imports...\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Load the data set\n",
    "\n",
    "We can easily load the data set using the scikit-learn library:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    },
    "tags": [
     "pyflyby-cell"
    ]
   },
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_covtype\n",
    "dataset = fetch_covtype(as_frame=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "This gives us a dictionary with all the information we need:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['data', 'target', 'frame', 'target_names', 'feature_names', 'DESCR'])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "df            = dataset[\"frame\"]\n",
    "target_names  = dataset[\"target_names\"]\n",
    "feature_names = dataset[\"feature_names\"]\n",
    "class_names   = [\"Spruce/Fir\",\"Lodgepole Pine\",\"Ponderosa Pine\",\"Cottonwood/Willow\",\"Aspen\",\"Douglas-fir\",\"Krummholz\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Let's first take a look at the data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "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>Elevation</th>\n",
       "      <th>Aspect</th>\n",
       "      <th>Slope</th>\n",
       "      <th>Horizontal_Distance_To_Hydrology</th>\n",
       "      <th>Vertical_Distance_To_Hydrology</th>\n",
       "      <th>Horizontal_Distance_To_Roadways</th>\n",
       "      <th>Hillshade_9am</th>\n",
       "      <th>Hillshade_Noon</th>\n",
       "      <th>Hillshade_3pm</th>\n",
       "      <th>Horizontal_Distance_To_Fire_Points</th>\n",
       "      <th>...</th>\n",
       "      <th>Soil_Type_31</th>\n",
       "      <th>Soil_Type_32</th>\n",
       "      <th>Soil_Type_33</th>\n",
       "      <th>Soil_Type_34</th>\n",
       "      <th>Soil_Type_35</th>\n",
       "      <th>Soil_Type_36</th>\n",
       "      <th>Soil_Type_37</th>\n",
       "      <th>Soil_Type_38</th>\n",
       "      <th>Soil_Type_39</th>\n",
       "      <th>Cover_Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2596.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>258.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>510.0</td>\n",
       "      <td>221.0</td>\n",
       "      <td>232.0</td>\n",
       "      <td>148.0</td>\n",
       "      <td>6279.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2590.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>212.0</td>\n",
       "      <td>-6.0</td>\n",
       "      <td>390.0</td>\n",
       "      <td>220.0</td>\n",
       "      <td>235.0</td>\n",
       "      <td>151.0</td>\n",
       "      <td>6225.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2804.0</td>\n",
       "      <td>139.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>268.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>3180.0</td>\n",
       "      <td>234.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>135.0</td>\n",
       "      <td>6121.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2785.0</td>\n",
       "      <td>155.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>242.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>3090.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>122.0</td>\n",
       "      <td>6211.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2595.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>153.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>391.0</td>\n",
       "      <td>220.0</td>\n",
       "      <td>234.0</td>\n",
       "      <td>150.0</td>\n",
       "      <td>6172.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581007</th>\n",
       "      <td>2396.0</td>\n",
       "      <td>153.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>108.0</td>\n",
       "      <td>240.0</td>\n",
       "      <td>237.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>837.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581008</th>\n",
       "      <td>2391.0</td>\n",
       "      <td>152.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>240.0</td>\n",
       "      <td>237.0</td>\n",
       "      <td>119.0</td>\n",
       "      <td>845.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581009</th>\n",
       "      <td>2386.0</td>\n",
       "      <td>159.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>236.0</td>\n",
       "      <td>241.0</td>\n",
       "      <td>130.0</td>\n",
       "      <td>854.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581010</th>\n",
       "      <td>2384.0</td>\n",
       "      <td>170.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>230.0</td>\n",
       "      <td>245.0</td>\n",
       "      <td>143.0</td>\n",
       "      <td>864.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581011</th>\n",
       "      <td>2383.0</td>\n",
       "      <td>165.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>231.0</td>\n",
       "      <td>244.0</td>\n",
       "      <td>141.0</td>\n",
       "      <td>875.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>581012 rows × 55 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Elevation  Aspect  Slope  Horizontal_Distance_To_Hydrology  \\\n",
       "0          2596.0    51.0    3.0                             258.0   \n",
       "1          2590.0    56.0    2.0                             212.0   \n",
       "2          2804.0   139.0    9.0                             268.0   \n",
       "3          2785.0   155.0   18.0                             242.0   \n",
       "4          2595.0    45.0    2.0                             153.0   \n",
       "...           ...     ...    ...                               ...   \n",
       "581007     2396.0   153.0   20.0                              85.0   \n",
       "581008     2391.0   152.0   19.0                              67.0   \n",
       "581009     2386.0   159.0   17.0                              60.0   \n",
       "581010     2384.0   170.0   15.0                              60.0   \n",
       "581011     2383.0   165.0   13.0                              60.0   \n",
       "\n",
       "        Vertical_Distance_To_Hydrology  Horizontal_Distance_To_Roadways  \\\n",
       "0                                  0.0                            510.0   \n",
       "1                                 -6.0                            390.0   \n",
       "2                                 65.0                           3180.0   \n",
       "3                                118.0                           3090.0   \n",
       "4                                 -1.0                            391.0   \n",
       "...                                ...                              ...   \n",
       "581007                            17.0                            108.0   \n",
       "581008                            12.0                             95.0   \n",
       "581009                             7.0                             90.0   \n",
       "581010                             5.0                             90.0   \n",
       "581011                             4.0                             67.0   \n",
       "\n",
       "        Hillshade_9am  Hillshade_Noon  Hillshade_3pm  \\\n",
       "0               221.0           232.0          148.0   \n",
       "1               220.0           235.0          151.0   \n",
       "2               234.0           238.0          135.0   \n",
       "3               238.0           238.0          122.0   \n",
       "4               220.0           234.0          150.0   \n",
       "...               ...             ...            ...   \n",
       "581007          240.0           237.0          118.0   \n",
       "581008          240.0           237.0          119.0   \n",
       "581009          236.0           241.0          130.0   \n",
       "581010          230.0           245.0          143.0   \n",
       "581011          231.0           244.0          141.0   \n",
       "\n",
       "        Horizontal_Distance_To_Fire_Points  ...  Soil_Type_31  Soil_Type_32  \\\n",
       "0                                   6279.0  ...           0.0           0.0   \n",
       "1                                   6225.0  ...           0.0           0.0   \n",
       "2                                   6121.0  ...           0.0           0.0   \n",
       "3                                   6211.0  ...           0.0           0.0   \n",
       "4                                   6172.0  ...           0.0           0.0   \n",
       "...                                    ...  ...           ...           ...   \n",
       "581007                               837.0  ...           0.0           0.0   \n",
       "581008                               845.0  ...           0.0           0.0   \n",
       "581009                               854.0  ...           0.0           0.0   \n",
       "581010                               864.0  ...           0.0           0.0   \n",
       "581011                               875.0  ...           0.0           0.0   \n",
       "\n",
       "        Soil_Type_33  Soil_Type_34  Soil_Type_35  Soil_Type_36  Soil_Type_37  \\\n",
       "0                0.0           0.0           0.0           0.0           0.0   \n",
       "1                0.0           0.0           0.0           0.0           0.0   \n",
       "2                0.0           0.0           0.0           0.0           0.0   \n",
       "3                0.0           0.0           0.0           0.0           0.0   \n",
       "4                0.0           0.0           0.0           0.0           0.0   \n",
       "...              ...           ...           ...           ...           ...   \n",
       "581007           0.0           0.0           0.0           0.0           0.0   \n",
       "581008           0.0           0.0           0.0           0.0           0.0   \n",
       "581009           0.0           0.0           0.0           0.0           0.0   \n",
       "581010           0.0           0.0           0.0           0.0           0.0   \n",
       "581011           0.0           0.0           0.0           0.0           0.0   \n",
       "\n",
       "        Soil_Type_38  Soil_Type_39  Cover_Type  \n",
       "0                0.0           0.0           5  \n",
       "1                0.0           0.0           5  \n",
       "2                0.0           0.0           2  \n",
       "3                0.0           0.0           2  \n",
       "4                0.0           0.0           5  \n",
       "...              ...           ...         ...  \n",
       "581007           0.0           0.0           3  \n",
       "581008           0.0           0.0           3  \n",
       "581009           0.0           0.0           3  \n",
       "581010           0.0           0.0           3  \n",
       "581011           0.0           0.0           3  \n",
       "\n",
       "[581012 rows x 55 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "source": [
    "We can see that the data is already structured the way we need it to be able to feed it into the machine learning process. However, we have to split the data into different subsets in order to be able to assess the performances of the models we are going to build."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Split the data set\n",
    "\n",
    "In the context of ML, we normally split a data set into the following subsets:\n",
    "\n",
    "- `train` (used for training the model)\n",
    "- `test` (used for testing the final model)\n",
    "\n",
    "The `train` set can further be split into `train` and `validation` subsets. The `validation` subset is used within the feature selection and hyperparameter tuning processes in order to find the best possible model. The `test` set is used to evaluate the performance of the final model. The exact procedure of these splits can differ and is dependent on the data structure and the problem we want to solve.\n",
    "\n",
    "We can use the built-in `train_test_split` function to make a random split into `80% train` and `20% test` data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(df[feature_names], df[target_names], test_size=0.2, random_state=42)\n",
    "y_train = y_train.values.ravel() # convert to 1d array\n",
    "y_test  = y_test.values.ravel()  # convert to 1d array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Samples in train data: 464809\n",
      "Samples in test data:  116203\n"
     ]
    }
   ],
   "source": [
    "print(\"Samples in train data: {0:5d}\\nSamples in test data:  {1:5d}\".format(y_train.size,y_test.size))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Baseline model\n",
    "For comparison purposes, let's go ahead and build a simple Decision Tree classification model, fit it to the data and check its performance..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier, plot_tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(random_state=42)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt_model = DecisionTreeClassifier(random_state=42)\n",
    "dt_model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9389516621773965"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dt_model.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "OK, that looks already pretty good. With an average accuracy of ~93.8% the simple Decision Tree model already achieves good results on the validation data set."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "We can plot the fitted model to get an idea of what the model learned and how it uses the data to predict forest cover types:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1728x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(24,10))\n",
    "plot_tree(dt_model,feature_names=feature_names,filled=True,class_names=class_names,proportion=True,precision=0,max_depth=2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "Looking at the scores of the different forest cover types, we can see, that, due to the different class counts, some classes are recognized worse by the model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxIAAADSCAYAAADexPbiAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAnqElEQVR4nO3deZhlVX3v//fHFkFsbGTQtAi2Iogo2gLiBAaHnwo44IiIMlwj8SZOV9EQR7jJjai5SpQoFxMZHQhXQSMOeEEmQbCBphsVCEITRRQRaUaJtN/fH3uVHstT1XWgqk6d4v16nnrOPmuvvfd37dq1zvmetfapVBWSJEmSNIj7DTsASZIkSaPHREKSJEnSwEwkJEmSJA3MREKSJEnSwEwkJEmSJA3MREKSJEnSwEwkJEmS5rgkhyQ5YdhxSL1MJKQRk+TMJL9Osu6wY5Gk+S7JzknOS7I6yU1JvpvkKcOOa65LsirJ84Ydh2aWiYQ0QpIsAXYBCnjJLB73/rN1LEmaK5I8GPga8ElgI2Az4FDgrmHGJc0VJhLSaNkX+B5wDLDfWGGSzZN8Ockvk/wqyRE9696Y5EdJbk3ywyTbt/JK8pieesck+fu2vGuSnyb5myQ/B45O8pAkX2vH+HVbfkTP9hslOTrJz9r6U1r5ZUle3FNvnSQ3Jlk6Q+dIkqbL1gBV9YWqWlNVd1bVaVW1AiDJlknOaP3ujUk+l2TDsY3bp/LvSrIiye1J/jXJw5J8o/XJ/y/JQ1rdJa1fPrD1o9cneedEgSV5WhspuTnJpUl2naRu39eIJPdL8r4k1ya5IclxSRa1dbsm+em4/fx+lKFNtfq3ts2tSX6QZMe27nhgC+Dfk9yW5N334NxrBJhISKNlX+Bz7ecF7QVpAd0nZtcCS+g+MfsiQJJXAYe07R5MN4rxqyke68/oPoF7JHAgXX9xdHu+BXAncERP/eOB9YHHAw8FPt7KjwNe11Nvd+D6qlo+xTgkaViuBNYkOTbJbmNv+nsE+BDwcOBxwOZ0fW6vVwD/H11S8mLgG8B7gE3o+tW3jqv/bGAr4PnAwf2mByXZDDgV+Hu6fvog4EtJNu1Td8LXCGD/9vNs4NHAQv64X1+bl7R9bQh8dWzbqno98J/Ai6tqYVV9ZIB9aoSYSEgjIsnOdG/i/62qLgJ+DLwW2InuRexdVXV7Vf2mqs5tm/0F8JGq+n51rqqqa6d4yN8BH6yqu9qncL+qqi9V1R1VdSvwv4A/b7EtBnYD3lRVv66q31bVWW0/JwC7tykCAK+nSzokaU6rqluAnemmk34G+GWSryZ5WFt/VVV9u/WTvwQ+RusXe3yyqn5RVdcB5wAXVNUlVXUXcDLw5HH1D219+Uq6D2/27hPa64CvV9XXq+p3VfVtYBndBzXjTfYasQ/wsaq6uqpuA/4WeM0A01nPbTGsoevXnzTF7TRPmEhIo2M/4LSqurE9/3wr2xy4tqru7rPN5nQJxz3xy6r6zdiTJOsn+T9tCPwW4Gxgw/Zp1+bATVX16/E7qaqfAd8FXtGG/HejG1GRpDmvqn5UVftX1SOAJ9C9KT8cIMlDk3wxyXWtXzyBbqSh1y96lu/s83zhuPo/6Vm+th1vvEcCr2rTmm5OcjNdwrO4T93JXiMe3o7Re7z7Aw/rU7efn/cs3wGs5z119y3+sqURkOSBwKuBBe2eBYB16YaTfwFskeT+fV4ofgJsOcFu76CbijTmz4De+bA1rv47gccCT62qn7d7HC6hG9r/CbBRkg2r6uY+xzqWbnTk/sD57ZM5SRopVXV5kmOAv2xFH6LrK59YVb9KsieDTQ3qZ3Pg8ra8BfCzPnV+AhxfVW+cwv5+wsSvET+jS0rGbAHcTfe68nB6XiPah0Z/MnVqEuNfQzQPOSIhjYY9gTXAtsDS9vM4umHyPYHrgcOSPCjJekme2bb7F+CgJDuk85gkYy8ay4HXJlmQ5IX86XD8eBvQfXp2c5KNgA+Oraiq6+nm/X6q3ZS9TpJn9Wx7CrA98Da6eyYkac5Lsk2Sd459sUSSzemmGn2vVdkAuI2uX9wMeNc0HPb9bQT48cABwIl96pwAvDjJC1ofvl67OfoRfepeyMSvEV8A/keSRyVZCPwDcGJLOK6kG2HYI8k6wPvoPsCaql/Q3XehecxEQhoN+wFHV9V/VtXPx37oPvnam+4GvsfQ3dz2U2AvgKo6ie5ehs8Dt9K9od+o7fNtbbub6ebJnrKWGA4HHgjcSPci+s1x618P/Jbuk7QbgLePraiqO4EvAY8Cvjz1ZkvSUN0KPBW4IMntdH3fZXQjtNB9Fez2wGq6m5+no387C7gKOB34x6o6bXyFqvoJ8FK6m7Z/STfq8C76vK9r9y/0fY0APkt3b8PZwDXAb4C3tO1WA39F94HUdcDt/PGo9dp8CHhfm3p10ADbaYSkypEnSTMvyQeAravqdWutLEn3Men+T9A1wDoT3M8gzTneIyFpxrWpUG+gG7WQJEnzgFObJM2oJG+kG3b/RlWdPex4JEnS9HBqkyRJkqSBOSIhSZIkaWAmEpIkSZIG5s3WI2iTTTapJUuWDDsMSbpHLrroohurapB/bDXy7LcljarJ+mwTiRG0ZMkSli1bNuwwJOkeSXLtsGOYbfbbkkbVZH22U5skSZIkDcxEQpIkSdLATCQkSZIkDcxEQpIkSdLATCQkSZIkDcxvbRpBK69bzZKDTx12GJL6WHXYHsMOQXPQdPfbXmeS5gJHJCRJkiQNzERCkiRJ0sBMJCRJkiQNzERCkiRJ0sBGOpFIclvP8u5J/iPJFkOMZ1WSTQaov3+SI2YyJkmSJGkmjHQiMSbJc4FPAi+sqv8ct85vppIkSZKm2ci/yU6yC/AZYPeq+nErOwa4CXgycHGSW4Hbquof2/rLgBe1XXwTOBd4GnApcDRwKPBQYJ+qujDJIcCjgMXA1sA7Wv3dgOuAF1fVb9v+3pLkxcA6wKuq6vIkGwGfBR4N3AEcWFUrxrVjec/Tx9IlRWfd6xMkSZIkzYBRH5FYF/gKsGdVXT5u3dbA86rqnWvZx2OAfwKeCGwDvBbYGTgIeE9PvS2BPYCXAicA36mq7YA7W/mYG6tqe+DTbR/QJSaXVNUT2z6PGx9EVS2tqqXA+4FlwHlriVuSJEkamlFPJH5L94b7DX3WnVRVa6awj2uqamVV/Q74AXB6VRWwEljSU+8bbdRhJbCAbiSDPvW+3B4v6infGTgeoKrOADZOsmh8IEm2Aj4K7NUzwjG27sAky5IsW3PH6ik0S5I0TPbbkua7UU8kfge8GnhKkveMW3d7z/Ld/HFb1+tZvmvc/u7qWb7/+Hot4fhtSzYmrAes6SlPn9ir90mSBwH/Bryxqn72J5WrjqqqHatqxwXr/0kOIkmaY+y3Jc13o55IUFV30N3vsE+SfiMTAKuA7QGSbE93v8NsOhvYpx1/V7rpT7eMq3M0cHRVnTO7oUmSJEmDG/mbrQGq6qYkLwTOTnJjnypfAvZtNzR/H7hyNuMDDgGOTrKC7mbr/XpXJnkk8Epg6yT/rRX/RVUtm9UoJUmSpCka6USiqhb2LP+EP4w0fGVcvTuB50+wmyf01Nu/Z3nV2LqqOmSS4x7Ss7ykZ3kZsGtbvonuJu3x8R8DHNOejvzokCRJku47fPMqSZIkaWAmEpIkSZIGZiIhSZIkaWAmEpIkSZIGNtI3W99XbbfZIpYdtsfaK0qS5gT7bUnzkSMSkiRJkgZmIiFJkiRpYCYSkiRJkgbmPRIjaOV1q1ly8KnDDkNzxCrnXUtz3nT12/69S5pLHJGQJEmSNDATCUmSJEkDM5GQJEmSNDATCUmSJEkDG/lEIsmaJMuT/CDJpUnekWTa25XkkCQHTcN+tmnxXpJkyyTnTUd8kiRJ0myaD9/adGdVLQVI8lDg88Ai4IPDDGoSewJfqaqx+J4xvkKSBVW1ZlajkiRJkgYw8iMSvarqBuBA4M3prJfk6CQr2wjAswGS7J/kiLHtknwtya5t+Q1JrkxyZpLP9Nbrqf/GJN9vIyBfSrJ+K39Vksta+dl9ttsdeDvwF0m+08pua4+7JvlOks8DK6f3zEiSJEnTaz6MSPyRqrq6TW16KPC6VrZdkm2A05JsPdG2SR4OvB/YHrgVOAO4tE/VL1fVZ9o2fw+8Afgk8AHgBVV1XZIN+8T29SRHArdV1T/22e9OwBOq6popN1iSJEkagnk1ItEj7XFn4HiAqrocuBaYMJGgeyN/VlXdVFW/BU6aoN4TkpyTZCWwD/D4Vv5d4JgkbwQW3IO4L5woiUhyYJJlSZatuWP1Pdi1JGk22W9Lmu/mXSKR5NHAGuAG/pBQjHc3f9z29cY2n+JhjgHeXFXbAYeObV9VbwLeB2wOLE+ycZtatTzJ16ew39snWlFVR1XVjlW144L1F00xTEnSsNhvS5rv5lUikWRT4EjgiKoq4Gy6EQPalKYtgCuAVcDSJPdLsjndSATAhcCfJ3lIkvsDr5jgUBsA1ydZZ2z/7RhbVtUFVfUB4EZg86o6oKqWVtXu091eSZIkaVjmwz0SD0yyHFiHbqTheOBjbd2ngCPbFKS7gf2r6q4k3wWuobup+TLgYoB2b8M/ABcAPwN+CPQbj35/q3Nt28cGrfyjSbaiG9k4nf73V0iSJEkjb+QTiaqa8F6EqvoNsH+f8qJnJGGcz1fVUW1E4mTgtLbNIT3bfxr4dJ/9vnwK8R4y7vnC9ngmcObatpckSZLmgnk1tWmaHNJGOC6jG7U4ZajRSJIkSXPQyI9ITLequtf/vVqSJEma7xyRkCRJkjQwRyRG0HabLWLZYXsMOwxJ0hTZb0uajxyRkCRJkjQwEwlJkiRJAzORkCRJkjQwEwlJkiRJA/Nm6xG08rrVLDn41GGHMbBV3mgo6T5qqv22/aSkUeKIhCRJkqSBmUhIkiRJGpiJhCRJkqSBmUhIkiRJGpiJxBQleVmSSrLNsGORJEmShs1EYur2Bs4FXjPsQCRJkqRhM5GYgiQLgWcCb6AlEkkWJzk7yfIklyXZpZXfluR/J7k4yelJNm3lWyb5ZpKLkpwzNrKR5Jgkn0hyXpKrk7xySM2UJEmSpsxEYmr2BL5ZVVcCNyXZHngt8K2qWgo8CVje6j4IuLiqtgfOAj7Yyo8C3lJVOwAHAZ/q2f9iYGfgRcBh/QJIcmCSZUmWrblj9TQ2TZI0E+y3Jc13/kO6qdkbOLwtf7E9/3fgs0nWAU6pquVt/e+AE9vyCcCX24jGM4CTkoztc92e/Z9SVb8DfpjkYf0CqKqj6JIR1l28VU1DmyRJM8h+W9J8ZyKxFkk2Bp4DPCFJAQuAAt4NPAvYAzg+yUer6rg+uyi6kZ+b2+hFP3f1HnK6YpckSZJmilOb1u6VwHFV9ciqWlJVmwPX0CURN1TVZ4B/BbZv9e/XtoFu+tO5VXULcE2SVwGk86RZbYUkSZI0jRyRWLu9+dP7Fr4EHAPcnuS3wG3Avm3d7cDjk1wErAb2auX7AJ9O8j5gHbopUpfObOiSJEnSzDCRWIuq2rVP2SeAT0yyzfuB948ruwZ4YZ+6+497vvAehipJkiTNGqc2SZIkSRqYicQ0c0RBkiRJ9wUmEpIkSZIG5j0SI2i7zRax7LA9hh2GJGmK7LclzUeOSEiSJEkamImEJEmSpIGZSEiSJEkamImEJEmSpIF5s/UIWnndapYcfOq93s8qb/yTpFkx1m/b70qaTxyRkCRJkjQwEwlJkiRJAzORkCRJkjQwEwlJkiRJA5tSIpHkz5J8McmPk/wwydeTbD1B3SVJXtvzfGmS3acr4OmWZP8kR/Q8X5zktCQnJ9mzp/yKJO/ref6lJC9P8qYk+7ayY5K8si2fmWTHWWyKJEmSNGvWmkgkCXAycGZVbVlV2wLvAR42wSZLgNf2PF8KzNlEoo8XAt8CzgOeAZBkY+A24Ok99Z4OnFdVR1bVcbMepSRJkjREUxmReDbw26o6cqygqpYD5yb5aJLLkqxMsldbfRiwS5LlSf4G+J/AXu35Xkk2SnJKkhVJvpfkiQBJDkny2fZJ/tVJ3trK392z/PEkZ7Tl5yY5oS3v3WK4LMmHx+KcpPyAJFcmOQt45rj2vhD4BvBdWiLRHr8GbJrOo4A7q+rnLe6DJjuB/eJI8uokH2vLb0tydVveMsm5a/+1SJIkScMzlf8j8QTgoj7lL6cbbXgSsAnw/SRnAwcDB1XViwCS/ALYsare3J5/ErikqvZM8hzguLYfgG3oEpcNgCuSfBo4G3gn8AlgR2DdJOsAOwPnJHk48GFgB+DXwGltStKFE5RfABzaylcD3wEuabEtAB5bVT9Msi7whCQPoEskzgIeDTwOeDJdorFWk8R3NvCuVm0X4FdJNhtrV5/9HAgcCLDgwZtO5dCSpCGy35Y0392bm613Br5QVWuq6hd0b7SfMsXtjgeoqjOAjZMsautOraq7qupG4Aa66VMXATsk2QC4CzifLqHYhe4N91Popl39sqruBj4HPGuS8qf2lP8XcGJPbE+lSzSoqruAHwDbA09r5efTJRXPoJv6NBV946iqnwMLW7s2Bz7f4htr1x+pqqOqaseq2nHB+ovGr5YkzTH225Lmu6kkEj+g+zR9vNzDY/bbrtrjXT1la4D7V9VvgVXAAXRv3s+hG7XYEvjRJHFMFl9NUL4b8M2e5+fRvbnfoKp+DXyPPyQSUxqRWEsc59O16wq6du1Cd+/FVPctSZIkDcVUEokz6KYTvXGsIMlT6Kbp7JVkQZJN6d5wXwjcSjc1acz452cD+7T97ArcWFW3rCWGs4GD2uM5wJuA5VVVdCMFf55kkzY1aW+60ZHJyndNsnGbIvWqnuM8Fzi95/l3gb8ELm3PV9CNTmxBl2BNxURxjG/XJXQJ0l1VtXqK+5YkSZKGYq33SFRVJXkZcHiSg4Hf0I0QvB1YSPcmu4B3t5uPfwXcneRS4BjgWODgJMuBDwGHAEcnWQHcAew3hTjPAd4LnF9Vtyf5TSujqq5P8rd09zoE+HpVfQVgkvJD6EYDrgcuBsaSod+MS2rOo7sv4kPtWHcnuQH4SVX9bgpxTxpfa8PmwNlVtSbJT4DLp7JfSZIkaZjSfaivJK8DHlFVhw07lrVZd/FWtXi/w+/1flYdtse9D0aSBpTkoqq6T/2fnbF+235X0qiZrM+eyrc23SdU1QnDjkGSJEkaFffmW5skSZIk3UeZSEiSJEkamFObRtB2my1imfNsJWlk2G9Lmo8ckZAkSZI0MBMJSZIkSQMzkZAkSZI0MO+RGEErr1vNkoNPHXYY0kjwe/s1F8y1ftu/C0nTwREJSZIkSQMzkZAkSZI0MBMJSZIkSQMzkZAkSZI0MBMJSZIkSQOb1UQiyZoky5NcluSkJOtPwz4PSXLQdMR3L2JYkuTO1rYfJjkyyf2SvCTJwcOMTZIkSZoJsz0icWdVLa2qJwD/Bbxplo9Pkpn6ytsfV9VS4InAtsCeVfXVqjpsho4nSZIkDc0wpzadAzwmyUZJTkmyIsn3kjwRfj/S8NkkZya5OslbxzZM8t4kVyT5f8Bje8q3TPLNJBclOSfJNq38mCQfS/Id4MNJlrZjrUhycpKHtHpvbSMKK5J8sZXtlOS8JJe0x8cyiaq6GzivtW3/JEf0xPCJto+rk7yyJ+53Jfl+O+6h03R+JUmSpBkzlESijQrsBqwEDgUuqaonAu8Bjuupug3wAmAn4INJ1kmyA/Aa4MnAy4Gn9NQ/CnhLVe0AHAR8qmfd1sDzquqd7Rh/0465Evhgq3Mw8ORWPjZacjnwrKp6MvAB4B/W0rb1gee2/Y63GNgZeBFwWKv/fGCr1salwA5JntVnvwcmWZZk2Zo7Vk8WgiRpDrDfljTfzfZ/tn5gkuVt+RzgX4ELgFcAVNUZSTZOsqjVObWq7gLuSnID8DBgF+DkqroDIMlX2+NC4BnASUnGjrduz7FPqqo1bd8bVtVZrfxY4KS2vAL4XJJTgFNa2SLg2CRbAQWsM0HbtmxtK+ArVfWNJPuPq3NKVf0O+GGSh7Wy57efS9rzhXSJxdm9G1bVUXSJEusu3qomiEGSNEfYb0ua72Y7kbiz3Ufwe+l5199jrMO9q6dsDX+It1+HfD/g5vH773H7FOLbA3gW8BLg/UkeD/wd8J2qelmSJcCZE2z740mOPaa3Pel5/FBV/Z8pxCdJkiTNCXPh61/PBvYBSLIrcGNV3bKW+i9L8sAkGwAvBmjbXJPkVW1fSfKk8RtX1Wrg10l2aUWvB85Kcj9g86r6DvBuYEO60YFFwHWt7v73vJkT+hbw39qICkk2S/LQGTiOJEmSNG1me0Sin0OAo5OsAO4A9pusclVdnOREYDlwLd0UqTH7AJ9O8j66KUhfBC7ts5v9gCPb/QxXAwcAC4AT2tSnAB+vqpuTfIRuatM7gDPucSsnbs9pSR4HnN8GZ24DXgfcMN3HkiRJkqZLqpy2OWrWXbxVLd7v8GGHIY2EVYftMewQNE6Si6pqx2HHMZvmWr/t34WkqZqsz54LU5skSZIkjRgTCUmSJEkDM5GQJEmSNLC5cLO1BrTdZotY5vxWSRoZ9tuS5iNHJCRJkiQNzERCkiRJ0sBMJCRJkiQNzHskRtDK61az5OBThx2GdJ/ld/BrUPbbkoZtJl67HJGQJEmSNDATCUmSJEkDM5GQJEmSNDATCUmSJEkDm/VEIslt93C7JUkum+54JjneIUkOGqD+/kl+mWR5kh8meWMr/59JnjdzkUqSJEmzz29tml4nVtWbkzwU+EGSr1bVB4YdlCRJkjTd5sTUpiRLk3wvyYokJyd5SCvfIcmlSc4H/rqn/vpJ/q3VPzHJBUl2bOuen+T8JBcnOSnJwla+KsmHk1zYfh7Tyh+Z5PS2r9OTbNEnvi2TfDPJRUnOSbLNZO2pqhuAHwOPTHJMklf2xHBoi23l2H6SPCjJZ5N8P8klSV46LSdWkiRJmiFzIpEAjgP+pqqeCKwEPtjKjwbeWlVPH1f/r4Bft/p/B+wAkGQT4H3A86pqe2AZ8I6e7W6pqp2AI4DDW9kRwHFtX58DPtEnvqOAt1TVDsBBwKcma0ySRwOPBq7qs/rGFtun274A3gucUVVPAZ4NfDTJgyY7hiRJkjRMQ5/alGQRsGFVndWKjgVO6lN+PLBbW94Z+CeAqrosyYpW/jRgW+C7SQAeAJzfc7gv9Dx+vC0/HXh5zzE+Mi6+hcAzWkxjxetO0Jy9kuwM3AX8ZVXd1LPNmC+3x4t6jvt84CU992SsB2wB/KgnjgOBAwEWPHjTCQ4vSZor7LclzXdDTyQmEaAmWTdR+berau8J1tcEyxPVgW7U5uaqWjpB/V4nVtWb11Lnrva4hj+c/wCvqKorJtqoqo6iGxlh3cVbTRS7JGmOsN+WNN8NfWpTVa0Gfp1kl1b0euCsqroZWN0+4QfYp2ezc4FXAyTZFtiulX8PeGbP/Q/rJ9m6Z7u9eh7HRirOA17Tc4xzx8V3C3BNkle1fSbJk+5hcyfyLeAtacMXSZ48zfuXJEmSptUwRiTWT/LTnucfA/YDjkyyPnA1cEBbdwDw2SR30L3ZHvMp4Ng2pekSYAWwuqp+mWR/4AtJxqYfvQ+4si2vm+QCugRqbNTire0Y7wJ+2XPsXvsAn07yPmAd4IvApfeo9f39Hd09GytaMrEKeNE07l+SJEmaVqkavdHWJAuAdarqN0m2BE4Htq6q/5pkm1XAjlV14yyFOWPWXbxVLd7v8GGHId1nrTpsj2GHMNKSXFRVOw47jtlkvy1p2O7pa9dkffZcvkdiMusD30myDt39Bf99siRCkiRJ0vQayUSiqm4FBvo0q6qWzEw0kiRJ0n3P0G+2liRJkjR6TCQkSZIkDWwkpzbd12232SKWebOnJI0M+21J85EjEpIkSZIGZiIhSZIkaWAmEpIkSZIG5j0SI2jldatZcvCpww5DGmn+UznNJvttaebZr88+RyQkSZIkDcxEQpIkSdLATCQkSZIkDcxEQpIkSdLA5m0ikeS9SX6QZEWS5UmeOqQ41klyUVte02IZ+1mS5LxhxCVJkiTdG/PyW5uSPB14EbB9Vd2VZBPgAVPc9v5Vdfc0hrMzMJYs3FlVS8etf0afGBZU1ZppjEGSJEmaVvMykQAWAzdW1V0AVXUjQJJVwInAs1u911bVVUmOAW4CngxcnORW4Laq+se23WXAi6pqVZJ9gYOAAlZU1euTbAocCWzR9vv2qvpuW34h8I2JAk1yW1UtTLIr8EHgemApsO29PQmSJEnSTJmvU5tOAzZPcmWSTyX58551t1TVTsARwOE95VsDz6uqd0600ySPB94LPKeqngS8ra36J+DjVfUU4BXAv/Rs9mzgzLb8wJ5pTSf3OcROwHuryiRCkiRJc9q8HJGoqtuS7ADsQvdG/sQkB7fVX+h5/HjPZidNYTrRc4D/OzbCUVU3tfLnAdsmGav34CQbABsAN1XVHa2839SmXhdW1TX9ViQ5EDgQYMGDN11LmJKkYbPfljTfzctEAqAlBWcCZyZZCew3tqq3Ws/y7T3Ld/PHozXrtceM22bM/YCnV9WdvYVJXg18a4Cwb59oRVUdBRwFsO7irfrFIEmaQ+y3Jc1383JqU5LHJtmqp2gpcG1b3qvn8fwJdrEK2L7ta3vgUa38dODVSTZu6zZq5acBb+45/tK2OOn9EZIkSdKomq8jEguBTybZkG504Sq64eUXAesmuYAuidp7gu2/BOybZDnwfeBKgKr6QZL/BZyVZA1wCbA/8Fbgn5OsoDunZyf5a2Crqrp8RlooSZIkDdG8TCSq6iL6f60qwD9X1aHj6u8/7vmdwPMn2PexwLHjym7kDyMdY8faGfjeuHoL++xvYXs8kz/clC1JkiTNafMykZgLqupc4NxhxyFJkiTNhPtUIlFVS4YdgyRJkjQfzMubrSVJkiTNrPvUiMR8sd1mi1h22B7DDkOSNEX225LmI0ckJEmSJA3MREKSJEnSwEwkJEmSJA3MREKSJEnSwEwkJEmSJA3MREKSJEnSwEwkJEmSJA0sVTXsGDSgJLcCVww7jj42AW4cdhATmKuxGddgjGswczWux1bVBsMOYjbN4X77npir19U9MZ/aAvOrPbZl7nhkVW3ab4X/kG40XVFVOw47iPGSLJuLccHcjc24BmNcg5nLcQ07hiGYk/32PTFXr6t7Yj61BeZXe2zLaHBqkyRJkqSBmUhIkiRJGpiJxGg6atgBTGCuxgVzNzbjGoxxDca45o751GbbMnfNp/bYlhHgzdaSJEmSBuaIhCRJkqSBmUjMAUlemOSKJFclObjP+kVJ/j3JpUl+kOSAtW2bZKMk307yH+3xIbMVV5LNk3wnyY9a+dt6tjkkyXVJlref3WcrrrZuVZKV7djLesqHeb4e23M+lie5Jcnb27rZOF8PSXJykhVJLkzyhLVtO0vnq29cc+D6mux8DfP6muh8zfT19dkkNyS5bIL1SfKJFveKJNuvrU3Tcb7mirX93oal3+9tsvOe5G9bG65I8oKe8h3aNX9V+z2nla+b5MRWfkGSJTPYlr59wii2J8l67e/30taWQ0e1LT1xLEhySZKvzYO2/EkfP8rtmRZV5c8Qf4AFwI+BRwMPAC4Fth1X5z3Ah9vypsBNre6E2wIfAQ5uywePbT9LcS0Gtm/lGwBX9sR1CHDQMM5Xe74K2KTPfod2vvrs5+d039k8W+fro8AH2/I2wOlr23aWztdEcQ37+uob1xy4viaMa6aur7aPZwHbA5dNsH534BtAgKcBF8z09TVXfqbyextibH/ye5vovAPbttjXBR7V2rSgrbsQeHr7/X4D2K2V/xVwZFt+DXDiDLalb58wiu1px13YltcBLmh/NyPXlp42vQP4PPC1Ub7O2jFWMa6PH+X2TMePIxLDtxNwVVVdXVX/BXwReOm4OgVs0DLWhXRvQO9ey7YvBY5ty8cCe85WXFV1fVVdDFBVtwI/AjYb8PjTHtda9ju08zWuznOBH1fVtQMe/97EtS1wOkBVXQ4sSfKwtWw7G+erb1xz4Pqa6HxNZmjna1yd6b6+qKqz6a7libwUOK463wM2TLKYmb2+5oqp/N6GYoLf20Tn/aXAF6vqrqq6BrgK2Kn9Hh9cVedX987nuHHbjO3r/wLPHfvUdQbaMlGfMHLtaX8nt7Wn67SfGsW2ACR5BLAH8C89xSPZlknMt/YMxERi+DYDftLz/Kf86ZuiI4DHAT8DVgJvq6rfrWXbh1XV9dB1ssBDZzGu32vDck+m+1RlzJvTTXH4bAafsnBv4yrgtCQXJTmwZ5s5cb7oPoH4wriymT5flwIvB0iyE/BI4BFr2XY2ztdEcf3ekK6vyeIa5vW11vPF9F9fUzFR7DN5fc0VU/m9zSUTnffJfoc/7VP+R9tU1d3AamDjGYu8GdcnjGR72lSg5cANwLeramTbAhwOvBvofa0b1bZA/z5+lNtzr5lIDF+/TLPGPX8BsBx4OLAUOCLJg6e47TDi6naQLAS+BLy9qm5pxZ8Gtmz1rwf+9yzH9cyq2h7YDfjrJM8a8PgzFRdJHgC8BDipZ5vZOF+HAQ9pL1xvAS6hGykZ9vU1UVzdDoZ3fU0W1zCvr7Wdr5m4vqZiothn8vqaK+ZLG+/J73DW2z5Bn9C3ap+yOdOeqlpTVUvpPgjYKT33YfUxZ9uS5EXADVV10VQ36VM2J9rSY5A+fhTac6+ZSAzfT4HNe54/gu4T614HAF9uQ55XAdfQzYGebNtftOEz2uMNsxgXSdah69A/V1VfHtugqn7ROsnfAZ+hG/qftbiq6mft8Qbg5J7jD/V8NbsBF1fVL8YKZuN8VdUtVXVAe+Hal+7+jWvWsu2Mn69J4hrq9TVZXMO8viaLq5mJ6+vexD6T19dcMZV+YS6Z6LxP9jt8RJ/yP9omyf2BRUw+Be5emaBPGNn2AFTVzcCZwAsZzbY8E3hJklV00/qek+SEEW0LMGEfP7LtmQ4mEsP3fWCrJI9qnxi+BvjquDr/STe3mTbn+bHA1WvZ9qvAfm15P+ArsxVXm8/3r8CPqupjvRuM/bE1LwP6ftPLDMX1oCQbtPIHAc/vOf7QzlfP+r0ZN+1kNs5Xkg3bOoC/AM5un+YN9fqaKK5hX1+TxDXU62uS3+OYmbi+puKrwL7pPA1Y3Yb/Z/L6mium0i/MJROd968Cr0n3jTKPArYCLmy/x1uTPK39Xe47bpuxfb0SOKPNB592k/QJI9eeJJsm2bAtPxB4HnD5KLalqv62qh5RVUvorv0zqup1o9gW6Pr1Cfr4kWzPtKk5cMf3ff2H7ltNrqS7o/+9rexNwJva8sOB0+jm1V8GvG6ybVv5xnQ3Xv5He9xotuICdqYbiltBN5VnObB7W3d8q7+C7g9m8SzG9Wi6eeSXAj+YK+errVsf+BWwaNw+Z+N8Pb21+3Lgy8BD5sj11TeuOXB9TRTXsK+vyX6PM3l9fYFuWtRv6T5Ne8O4uAL8c4t7JbDjbFxfc+VnojYO+2eC39uE5x14b2vDFbRvmGnlO9L1Zz+muw9s7B/drkc3je4qum+oefQMtqVvnzCK7QGeSDctcUWL4wOtfOTaMq5du/KHb20aybYwQR8/qu2Zrh//s7UkSZKkgTm1SZIkSdLATCQkSZIkDcxEQpIkSdLATCQkSZIkDcxEQpIkSdLATCQkSZIkDcxEQpIkSdLATCQkSZIkDez/B7H6HHuaHKKRAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 864x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_scores_dt = [dt_model.score(X_test[y_test==covertype],y_test[y_test==covertype]) for covertype in np.unique(y_test)]\n",
    "class_counts    = np.unique(y_test,return_counts=True)[1]\n",
    "\n",
    "fig, ax = plt.subplots(1,2,figsize=(12,3),sharey=True)\n",
    "ax[0].barh(range(len(class_scores_dt)),class_scores_dt)\n",
    "ax[0].set_xlim((0.8,1.0))\n",
    "ax[0].set_yticklabels(class_names)\n",
    "ax[0].set_yticks(range(7))\n",
    "ax[0].set_title(\"Accuracy\")\n",
    "ax[1].barh(range(len(class_counts)),class_counts)\n",
    "ax[1].set_title(\"Sample count\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "To account for imbalances in the sample counts of the classes, we will also calculate the [balanced accuracy score](https://scikit-learn.org/stable/modules/model_evaluation.html#balanced-accuracy-score):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import balanced_accuracy_score\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9389516621773965\n",
      "0.8832865638983036\n"
     ]
    }
   ],
   "source": [
    "y_pred = dt_model.predict(X_test)\n",
    "print(accuracy_score(y_test, y_pred))\n",
    "print(balanced_accuracy_score(y_test, y_pred, adjusted=True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "Still pretty high, but we can see that the balanced accuracy score only reaches around 88% due to some classes having lower scores."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Random Forest model\n",
    "\n",
    "Let's see if we can achieve better results with a more advanced model. For this purpose, we'll take a random forest model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 4min 17s, sys: 1.66 s, total: 4min 19s\n",
      "Wall time: 26.1 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "model = RandomForestClassifier(random_state=42, n_jobs= -1, oob_score=True, class_weight='balanced')\n",
    "model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "When working with random forests, we don't need an independent validation data set to assess model performance: We can just use the internally calculated [OOB score (out-of-bag)](https://www.stat.berkeley.edu/~breiman/RandomForests/cc_home.htm#ooberr): "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9537250784730933"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.oob_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "Let's compare that to the validation score:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9555088939184014"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "As we can see, both values are pretty close to each other, indicating that the OOB score really was a good estimate of the general model performance. However, we want to compare the results to those of the baseline model, so we also calculate the balanced accuracy score:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9555088939184014\n",
      "0.8928133354659079\n"
     ]
    }
   ],
   "source": [
    "y_pred = model.predict(X_test)\n",
    "print(accuracy_score(y_test, y_pred))\n",
    "print(balanced_accuracy_score(y_test, y_pred, adjusted=True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "We see, even without further adjustments, the random forest model performs better on the data set. But let's see if we can do even better..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Clean up the data set\n",
    "\n",
    "The most important thing you **always** have to do before jumping into the ML training phase is to **clean up your data!**\n",
    "\n",
    "So let's take a look at the data again:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "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>Elevation</th>\n",
       "      <th>Aspect</th>\n",
       "      <th>Slope</th>\n",
       "      <th>Horizontal_Distance_To_Hydrology</th>\n",
       "      <th>Vertical_Distance_To_Hydrology</th>\n",
       "      <th>Horizontal_Distance_To_Roadways</th>\n",
       "      <th>Hillshade_9am</th>\n",
       "      <th>Hillshade_Noon</th>\n",
       "      <th>Hillshade_3pm</th>\n",
       "      <th>Horizontal_Distance_To_Fire_Points</th>\n",
       "      <th>...</th>\n",
       "      <th>Soil_Type_31</th>\n",
       "      <th>Soil_Type_32</th>\n",
       "      <th>Soil_Type_33</th>\n",
       "      <th>Soil_Type_34</th>\n",
       "      <th>Soil_Type_35</th>\n",
       "      <th>Soil_Type_36</th>\n",
       "      <th>Soil_Type_37</th>\n",
       "      <th>Soil_Type_38</th>\n",
       "      <th>Soil_Type_39</th>\n",
       "      <th>Cover_Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "      <td>581012.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2959.365301</td>\n",
       "      <td>155.656807</td>\n",
       "      <td>14.103704</td>\n",
       "      <td>269.428217</td>\n",
       "      <td>46.418855</td>\n",
       "      <td>2350.146611</td>\n",
       "      <td>212.146049</td>\n",
       "      <td>223.318716</td>\n",
       "      <td>142.528263</td>\n",
       "      <td>1980.291226</td>\n",
       "      <td>...</td>\n",
       "      <td>0.090392</td>\n",
       "      <td>0.077716</td>\n",
       "      <td>0.002773</td>\n",
       "      <td>0.003255</td>\n",
       "      <td>0.000205</td>\n",
       "      <td>0.000513</td>\n",
       "      <td>0.026803</td>\n",
       "      <td>0.023762</td>\n",
       "      <td>0.015060</td>\n",
       "      <td>2.051471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>279.984734</td>\n",
       "      <td>111.913721</td>\n",
       "      <td>7.488242</td>\n",
       "      <td>212.549356</td>\n",
       "      <td>58.295232</td>\n",
       "      <td>1559.254870</td>\n",
       "      <td>26.769889</td>\n",
       "      <td>19.768697</td>\n",
       "      <td>38.274529</td>\n",
       "      <td>1324.195210</td>\n",
       "      <td>...</td>\n",
       "      <td>0.286743</td>\n",
       "      <td>0.267725</td>\n",
       "      <td>0.052584</td>\n",
       "      <td>0.056957</td>\n",
       "      <td>0.014310</td>\n",
       "      <td>0.022641</td>\n",
       "      <td>0.161508</td>\n",
       "      <td>0.152307</td>\n",
       "      <td>0.121791</td>\n",
       "      <td>1.396504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1859.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-173.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2809.000000</td>\n",
       "      <td>58.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>108.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>1106.000000</td>\n",
       "      <td>198.000000</td>\n",
       "      <td>213.000000</td>\n",
       "      <td>119.000000</td>\n",
       "      <td>1024.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2996.000000</td>\n",
       "      <td>127.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>218.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>1997.000000</td>\n",
       "      <td>218.000000</td>\n",
       "      <td>226.000000</td>\n",
       "      <td>143.000000</td>\n",
       "      <td>1710.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>3163.000000</td>\n",
       "      <td>260.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>384.000000</td>\n",
       "      <td>69.000000</td>\n",
       "      <td>3328.000000</td>\n",
       "      <td>231.000000</td>\n",
       "      <td>237.000000</td>\n",
       "      <td>168.000000</td>\n",
       "      <td>2550.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3858.000000</td>\n",
       "      <td>360.000000</td>\n",
       "      <td>66.000000</td>\n",
       "      <td>1397.000000</td>\n",
       "      <td>601.000000</td>\n",
       "      <td>7117.000000</td>\n",
       "      <td>254.000000</td>\n",
       "      <td>254.000000</td>\n",
       "      <td>254.000000</td>\n",
       "      <td>7173.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>7.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 55 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           Elevation         Aspect          Slope  \\\n",
       "count  581012.000000  581012.000000  581012.000000   \n",
       "mean     2959.365301     155.656807      14.103704   \n",
       "std       279.984734     111.913721       7.488242   \n",
       "min      1859.000000       0.000000       0.000000   \n",
       "25%      2809.000000      58.000000       9.000000   \n",
       "50%      2996.000000     127.000000      13.000000   \n",
       "75%      3163.000000     260.000000      18.000000   \n",
       "max      3858.000000     360.000000      66.000000   \n",
       "\n",
       "       Horizontal_Distance_To_Hydrology  Vertical_Distance_To_Hydrology  \\\n",
       "count                     581012.000000                   581012.000000   \n",
       "mean                         269.428217                       46.418855   \n",
       "std                          212.549356                       58.295232   \n",
       "min                            0.000000                     -173.000000   \n",
       "25%                          108.000000                        7.000000   \n",
       "50%                          218.000000                       30.000000   \n",
       "75%                          384.000000                       69.000000   \n",
       "max                         1397.000000                      601.000000   \n",
       "\n",
       "       Horizontal_Distance_To_Roadways  Hillshade_9am  Hillshade_Noon  \\\n",
       "count                    581012.000000  581012.000000   581012.000000   \n",
       "mean                       2350.146611     212.146049      223.318716   \n",
       "std                        1559.254870      26.769889       19.768697   \n",
       "min                           0.000000       0.000000        0.000000   \n",
       "25%                        1106.000000     198.000000      213.000000   \n",
       "50%                        1997.000000     218.000000      226.000000   \n",
       "75%                        3328.000000     231.000000      237.000000   \n",
       "max                        7117.000000     254.000000      254.000000   \n",
       "\n",
       "       Hillshade_3pm  Horizontal_Distance_To_Fire_Points  ...   Soil_Type_31  \\\n",
       "count  581012.000000                       581012.000000  ...  581012.000000   \n",
       "mean      142.528263                         1980.291226  ...       0.090392   \n",
       "std        38.274529                         1324.195210  ...       0.286743   \n",
       "min         0.000000                            0.000000  ...       0.000000   \n",
       "25%       119.000000                         1024.000000  ...       0.000000   \n",
       "50%       143.000000                         1710.000000  ...       0.000000   \n",
       "75%       168.000000                         2550.000000  ...       0.000000   \n",
       "max       254.000000                         7173.000000  ...       1.000000   \n",
       "\n",
       "        Soil_Type_32   Soil_Type_33   Soil_Type_34   Soil_Type_35  \\\n",
       "count  581012.000000  581012.000000  581012.000000  581012.000000   \n",
       "mean        0.077716       0.002773       0.003255       0.000205   \n",
       "std         0.267725       0.052584       0.056957       0.014310   \n",
       "min         0.000000       0.000000       0.000000       0.000000   \n",
       "25%         0.000000       0.000000       0.000000       0.000000   \n",
       "50%         0.000000       0.000000       0.000000       0.000000   \n",
       "75%         0.000000       0.000000       0.000000       0.000000   \n",
       "max         1.000000       1.000000       1.000000       1.000000   \n",
       "\n",
       "        Soil_Type_36   Soil_Type_37   Soil_Type_38   Soil_Type_39  \\\n",
       "count  581012.000000  581012.000000  581012.000000  581012.000000   \n",
       "mean        0.000513       0.026803       0.023762       0.015060   \n",
       "std         0.022641       0.161508       0.152307       0.121791   \n",
       "min         0.000000       0.000000       0.000000       0.000000   \n",
       "25%         0.000000       0.000000       0.000000       0.000000   \n",
       "50%         0.000000       0.000000       0.000000       0.000000   \n",
       "75%         0.000000       0.000000       0.000000       0.000000   \n",
       "max         1.000000       1.000000       1.000000       1.000000   \n",
       "\n",
       "          Cover_Type  \n",
       "count  581012.000000  \n",
       "mean        2.051471  \n",
       "std         1.396504  \n",
       "min         1.000000  \n",
       "25%         1.000000  \n",
       "50%         2.000000  \n",
       "75%         2.000000  \n",
       "max         7.000000  \n",
       "\n",
       "[8 rows x 55 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Elevation', 'Aspect', 'Slope', 'Horizontal_Distance_To_Hydrology',\n",
       "       'Vertical_Distance_To_Hydrology', 'Horizontal_Distance_To_Roadways',\n",
       "       'Hillshade_9am', 'Hillshade_Noon', 'Hillshade_3pm',\n",
       "       'Horizontal_Distance_To_Fire_Points', 'Wilderness_Area_0',\n",
       "       'Wilderness_Area_1', 'Wilderness_Area_2', 'Wilderness_Area_3',\n",
       "       'Soil_Type_0', 'Soil_Type_1', 'Soil_Type_2', 'Soil_Type_3',\n",
       "       'Soil_Type_4', 'Soil_Type_5', 'Soil_Type_6', 'Soil_Type_7',\n",
       "       'Soil_Type_8', 'Soil_Type_9', 'Soil_Type_10', 'Soil_Type_11',\n",
       "       'Soil_Type_12', 'Soil_Type_13', 'Soil_Type_14', 'Soil_Type_15',\n",
       "       'Soil_Type_16', 'Soil_Type_17', 'Soil_Type_18', 'Soil_Type_19',\n",
       "       'Soil_Type_20', 'Soil_Type_21', 'Soil_Type_22', 'Soil_Type_23',\n",
       "       'Soil_Type_24', 'Soil_Type_25', 'Soil_Type_26', 'Soil_Type_27',\n",
       "       'Soil_Type_28', 'Soil_Type_29', 'Soil_Type_30', 'Soil_Type_31',\n",
       "       'Soil_Type_32', 'Soil_Type_33', 'Soil_Type_34', 'Soil_Type_35',\n",
       "       'Soil_Type_36', 'Soil_Type_37', 'Soil_Type_38', 'Soil_Type_39',\n",
       "       'Cover_Type'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "Well, this looks ok. However, there are many 1-hot-features in the data set (wilderness areas and soil types). We can recode these into 2 categorical variables to reduce the dimensionality of the data set significantly:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "df_clean = df[df.keys()[:10]]                                            # clip data set to first 10 features (without wilderness area and soil types)\n",
    "df_clean[\"wilderness_area\"] = df[df.keys()[10:14]].values.argmax(axis=1) # add wilderness areas encoded as one variable with values 0-3\n",
    "df_clean[\"soil_type\"] = df[df.keys()[14:-1]].values.argmax(axis=1)       # add soil types encoded as one variable with vlaues 0-39\n",
    "df_clean[\"Cover_Type\"] = df[\"Cover_Type\"]                                # add target feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "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>Elevation</th>\n",
       "      <th>Aspect</th>\n",
       "      <th>Slope</th>\n",
       "      <th>Horizontal_Distance_To_Hydrology</th>\n",
       "      <th>Vertical_Distance_To_Hydrology</th>\n",
       "      <th>Horizontal_Distance_To_Roadways</th>\n",
       "      <th>Hillshade_9am</th>\n",
       "      <th>Hillshade_Noon</th>\n",
       "      <th>Hillshade_3pm</th>\n",
       "      <th>Horizontal_Distance_To_Fire_Points</th>\n",
       "      <th>wilderness_area</th>\n",
       "      <th>soil_type</th>\n",
       "      <th>Cover_Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2596.0</td>\n",
       "      <td>51.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>258.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>510.0</td>\n",
       "      <td>221.0</td>\n",
       "      <td>232.0</td>\n",
       "      <td>148.0</td>\n",
       "      <td>6279.0</td>\n",
       "      <td>0</td>\n",
       "      <td>28</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2590.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>212.0</td>\n",
       "      <td>-6.0</td>\n",
       "      <td>390.0</td>\n",
       "      <td>220.0</td>\n",
       "      <td>235.0</td>\n",
       "      <td>151.0</td>\n",
       "      <td>6225.0</td>\n",
       "      <td>0</td>\n",
       "      <td>28</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2804.0</td>\n",
       "      <td>139.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>268.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>3180.0</td>\n",
       "      <td>234.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>135.0</td>\n",
       "      <td>6121.0</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2785.0</td>\n",
       "      <td>155.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>242.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>3090.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>122.0</td>\n",
       "      <td>6211.0</td>\n",
       "      <td>0</td>\n",
       "      <td>29</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2595.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>153.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>391.0</td>\n",
       "      <td>220.0</td>\n",
       "      <td>234.0</td>\n",
       "      <td>150.0</td>\n",
       "      <td>6172.0</td>\n",
       "      <td>0</td>\n",
       "      <td>28</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581007</th>\n",
       "      <td>2396.0</td>\n",
       "      <td>153.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>108.0</td>\n",
       "      <td>240.0</td>\n",
       "      <td>237.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>837.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581008</th>\n",
       "      <td>2391.0</td>\n",
       "      <td>152.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>95.0</td>\n",
       "      <td>240.0</td>\n",
       "      <td>237.0</td>\n",
       "      <td>119.0</td>\n",
       "      <td>845.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581009</th>\n",
       "      <td>2386.0</td>\n",
       "      <td>159.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>236.0</td>\n",
       "      <td>241.0</td>\n",
       "      <td>130.0</td>\n",
       "      <td>854.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581010</th>\n",
       "      <td>2384.0</td>\n",
       "      <td>170.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>230.0</td>\n",
       "      <td>245.0</td>\n",
       "      <td>143.0</td>\n",
       "      <td>864.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>581011</th>\n",
       "      <td>2383.0</td>\n",
       "      <td>165.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>231.0</td>\n",
       "      <td>244.0</td>\n",
       "      <td>141.0</td>\n",
       "      <td>875.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>581012 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Elevation  Aspect  Slope  Horizontal_Distance_To_Hydrology  \\\n",
       "0          2596.0    51.0    3.0                             258.0   \n",
       "1          2590.0    56.0    2.0                             212.0   \n",
       "2          2804.0   139.0    9.0                             268.0   \n",
       "3          2785.0   155.0   18.0                             242.0   \n",
       "4          2595.0    45.0    2.0                             153.0   \n",
       "...           ...     ...    ...                               ...   \n",
       "581007     2396.0   153.0   20.0                              85.0   \n",
       "581008     2391.0   152.0   19.0                              67.0   \n",
       "581009     2386.0   159.0   17.0                              60.0   \n",
       "581010     2384.0   170.0   15.0                              60.0   \n",
       "581011     2383.0   165.0   13.0                              60.0   \n",
       "\n",
       "        Vertical_Distance_To_Hydrology  Horizontal_Distance_To_Roadways  \\\n",
       "0                                  0.0                            510.0   \n",
       "1                                 -6.0                            390.0   \n",
       "2                                 65.0                           3180.0   \n",
       "3                                118.0                           3090.0   \n",
       "4                                 -1.0                            391.0   \n",
       "...                                ...                              ...   \n",
       "581007                            17.0                            108.0   \n",
       "581008                            12.0                             95.0   \n",
       "581009                             7.0                             90.0   \n",
       "581010                             5.0                             90.0   \n",
       "581011                             4.0                             67.0   \n",
       "\n",
       "        Hillshade_9am  Hillshade_Noon  Hillshade_3pm  \\\n",
       "0               221.0           232.0          148.0   \n",
       "1               220.0           235.0          151.0   \n",
       "2               234.0           238.0          135.0   \n",
       "3               238.0           238.0          122.0   \n",
       "4               220.0           234.0          150.0   \n",
       "...               ...             ...            ...   \n",
       "581007          240.0           237.0          118.0   \n",
       "581008          240.0           237.0          119.0   \n",
       "581009          236.0           241.0          130.0   \n",
       "581010          230.0           245.0          143.0   \n",
       "581011          231.0           244.0          141.0   \n",
       "\n",
       "        Horizontal_Distance_To_Fire_Points  wilderness_area  soil_type  \\\n",
       "0                                   6279.0                0         28   \n",
       "1                                   6225.0                0         28   \n",
       "2                                   6121.0                0         11   \n",
       "3                                   6211.0                0         29   \n",
       "4                                   6172.0                0         28   \n",
       "...                                    ...              ...        ...   \n",
       "581007                               837.0                2          1   \n",
       "581008                               845.0                2          1   \n",
       "581009                               854.0                2          1   \n",
       "581010                               864.0                2          1   \n",
       "581011                               875.0                2          1   \n",
       "\n",
       "        Cover_Type  \n",
       "0                5  \n",
       "1                5  \n",
       "2                2  \n",
       "3                2  \n",
       "4                5  \n",
       "...            ...  \n",
       "581007           3  \n",
       "581008           3  \n",
       "581009           3  \n",
       "581010           3  \n",
       "581011           3  \n",
       "\n",
       "[581012 rows x 13 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_clean"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "We could now also further examine the data set for specific features. For example, it always makes sense to look at the correlation of the predictors among each other as well as with the target feature to get an understanding of the relationships between the variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAj4AAAIjCAYAAAADehHpAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAACLHUlEQVR4nOzdd5xcVfnH8c83hZIEEkIXgdA7BAi9F6kiICDtBwQURCmCoqB0FQWRIlURIXQBAQFBQIHQW4DQm1QVpCoQOsnz++M5Qy7L7mbL7GyS+b5fr3ntzL137nPOzOzcZ84591xFBGZmZmbNoE9vF8DMzMysUZz4mJmZWdNw4mNmZmZNw4mPmZmZNQ0nPmZmZtY0nPiYmZlZ03DiY2Y2GZA0UtLt3Xj+XyXtUs8y9RZJv5V0aG+Xw6ZOTnzMzApJO0gaI2mcpFdKMrF6b5erJUlHSDq/uiwiNo6Ic3qrTB3R0eQuIvaMiJ81okzWfJz4mJkBkr4PnAj8ApgdmAc4Ddi8C/vq15Fl9kWS+vZ2GWzq5sTHzJqepMHAT4G9IuLyiHgvIj6JiKsj4odlm2klnSjp5XI7UdK0Zd3akv4l6UBJ/wHOLq0yf5J0vqR3gJGSBkv6Q2lN+rekn7d1oJf0G0n/lPSOpPslrVGWbwT8BNi2tEw9VJaPlvStcr+PpEMkvSjpNUnnljoiaZikkLSLpJckvSHp4HZem1GSTiutX+Mk3SFpjlL//0p6UtKyle0PkvSspHclPS5py7J8MeC3wCplP/+r7P90SddKeg9Ypyz7eVl/oKS7a4mjpO9IekzSdF17t63ZOfExM4NVgOmAK9rZ5mBgZWA4sAywInBIZf0cwFBgXmCPsmxz4E/AEOAC4BzgU2BBYFlgA+BbbcS7r8QaClwIXCppuoi4jmyVujgiBkXEMq08d2S5rQPMDwwCTmmxzerAIsB6wGElMWnLN0pdZwE+Au4CHiiP/wQcX9n2WWANYDBwJHC+pDkj4glgT+CuUu4hlefsABwFzAC07Ao7FvgYOETSQqXu/xcRH7ZTXrM2OfExM4OZgTci4tN2ttkR+GlEvBYRr5MH9Z0q6ycAh0fERxHxQVl2V0T8OSImADMCGwP7lRal14ATgO1aCxYR50fEmxHxaUQcB0xLJiodsSNwfEQ8FxHjgB8D27XobjsyIj6IiIeAh8hkri1XRMT9Jdm4AvgwIs6NiPHAxWQSVyv3pRHxckRMiIiLgWfIJLE9V0bEHeU5n0toymu3M7AvcBXwq4h4sCMvgllr3OdsZgZvArNI6tdO8vMl4MXK4xfLsprXW2mF+Gfl/rxAf+AVSbVlfVps8xlJPyBbg74EBJk4zTLpqrRZ1n7k2KWa/1Tuv0+2CrXl1cr9D1p5/NlzJe0MfB8YVhYN6kC5W30NaiLiBUk3A5sAp05iX2btcouPmVl23XwIbNHONi+TyUvNPGVZTbTynOqyf5LdRLNExJBymzEilmj5pDKe50Cyi2mm0i30NlDLmFqLNamyfsrnE5a6kzQv8Htgb2DmUu5HmXS5262PpE3I7sgbya4vsy5z4mNmTS8i3gYOA06VtIWkAZL6S9pY0q/KZheR40xmlTRL2f78tvbZSoxXgBuA4yTNWAYgLyBprVY2n4FMVF4H+kk6jGzxqXkVGCapre/wi4D9Jc0naRATxwS115VXDwPJJOZ1AEm7AktW1r8KfFnSNB3dYXmt/0C2fu0CbFYSIbMuceJjZgZExPFkF80h5IH7n2TLxZ/LJj8HxgAPA4+Qg3t/3skwOwPTAI8D/yUHBs/ZynbXA38Fnia7qT7k891Bl5a/b0p6oJXnnwWcB9wKPF+ev08ny9ppEfE4cBzZgvYqsBRwR2WTm4DHgP9IeqODuz2DHAN0bUS8CXwTOFPSzPUruTUTRUyqxdTMzMxs6uAWHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfMzMzaxqeudmaxnRSzNDgXH+eOYY0NF6NBgxoeMzxr73Z8Jh9BnR4Opi6GT/uo4bHBOg7sPF1jU8nNDympumlw9Kn4xseUkOHNjzm28/9u+ExAQYPa23Whp7zwmtv8cY749TaOic+1jRmoA9b0diE4JTdNmpovBots3zDY4477dyGxxy49DwNj/nW3f9oeEyAmUYMa3jMT94c1/CY08w9W8NjAox/6+2Gx+y37fYNj/nXHQ5ueEyAjX/9o4bGW+mAX7W5zl1dZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNJz4mJmZWdNw4mNmZmZNw4mPmZmZNQ0nPmZmZtY0nPhYmySNlzS2cjuoLB8taUSdY+0naUDl8bWShtQzhpmZmS9ZYe35ICKGNyjWfsD5wPsAEbFJg+KamVkTcYuPdYukDSTdJekBSZdKGiRpY0mXVLZZW9LV5f7pksZIekzSkWXZvsCXgJsl3VyWvSBplnL/+5IeLbf9yrJhkp6Q9PuyrxskTd/g6puZ2RTGiY+1Z/oWXV3bVleWxOQQYP2IWA4YA3wf+BuwsqSBZdNtgYvL/YMjYgSwNLCWpKUj4iTgZWCdiFinRYzlgV2BlYCVgd0lLVtWLwScGhFLAP8DtmpZAUl7lERrzIdE914NMzOb4rmry9ozqa6ulYHFgTskAUwD3BURn0q6DthM0p+ATYHapXm/IWkP8rM3Z3n+w+3EWB24IiLeA5B0ObAGcBXwfESMLdvdDwxr+eSIOAM4A2BW9XXmY2bW5Jz4WHcI+FtEbN/KuouBvYC3gPsi4l1J8wEHACtExH8ljQKm60CMtnxUuT8ecFeXmZm1y11d1h13A6tJWhBA0gBJC5d1o4HlgN2Z2M01I/Ae8Lak2YGNK/t6F5ihlRi3AluUfQ8EtgRuq3dFzMysObjFx9ozvaSxlcfXRcRBtQcR8bqkkcBFkqYtiw8Bno6I8ZL+AowEdinbPyTpQeAx4Dngjsq+zwD+KumV6jifiHigtAzdWxadGREPShpWv2qamVmzcOJjbYqIvm0sX7ty/yZghTa22xvYu8WykW1sezJwcuXxsMr944HjW2z/ArBk5fGv26iGmZnZZ9zVZWZmZk3DiY+ZmZk1DSc+ZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNJz4mJmZWdNQhK/baM1h+Tlnjnt226ihMff+xVUNjVdzyvfWa3zQgQMbH/P99xsfc/reuSRc3+8c0vCY/9vuGw2POebptxoeE2CdQ7dreMwJDz/a8Jh95py94TEBmDChoeFWOus67n/lzVav9egWHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfMzMzaxpOfMzMzKxpOPExMzOzpuHEx8zMzJqGEx8zMzNrGk58zMzMrGk48TEzM7Om4cTH6krSlpJC0qINiDVE0nd7Oo6ZmU09nPhYvW0P3A404op/QwAnPmZm1mFOfKxuJA0CVgO+SUl8JM0p6VZJYyU9KmmNsnycpOMkPSDpRkmzluULSLpO0v2Sbqu1HEmaXdIVkh4qt1WBo4EFyr6P7ZVKm5nZFMWJj9XTFsB1EfE08Jak5YAdgOsjYjiwDDC2bDsQeCAilgNuAQ4vy88A9omI5YEDgNPK8pOAWyJiGWA54DHgIODZiBgeET9srUCS9pA0RtKYN97/sK6VNTOzKU+/3i6ATVW2B04s9/9YHl8NnCWpP/DniBhb1k8ALi73zwcuLy1GqwKXSqrtc9ryd11gZ4CIGA+8LWmmSRUoIs4gkymWn3Pm6GrFzMxs6uDEx+pC0sxkcrKkpAD6AgH8CFgT2BQ4T9KxEXFuK7sIsgXyf6V1yMzMrO7c1WX1sjVwbkTMGxHDImJu4Hky6XktIn4P/IHspoL87G1d7u8A3B4R7wDPS9oGQGmZss2NwHfK8r6SZgTeBWZoQN3MzGwq4cTH6mV74IoWyy4DRgFjJT0IbAX8pqx7D1hC0v1kS9FPy/IdgW9Keogcx7N5Wf49YB1JjwD3A0tExJvAHWXQtAc3m5nZJLmry+oiItZuZdlJ5KDktp5zKHBoi2XPAxu1su2rTEyCqst36EJxzcysSbnFx8zMzJqGEx/rFRExqLfLYGZmzceJj5mZmTUNJz5mZmbWNJz4mJmZWdNw4mNmZmZNw4mPmZmZNQ0nPmZmZtY0nPiYmZlZ0/DMzdY0NGAAWmb5hsY85XvvNTRezd6/ubHhMU85fKuGx/zgiX82POaAnbae9EY9YMIT9zY85uBzRjU85rq3XdPwmABIDQ/54YtvNjzmgKWWaHhMgPEPPdrYgJ980uYqt/iYmZlZ03DiY2ZmZk3DiY+ZmZk1DSc+ZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNJz4mJmZWdNw4mNmZmZNw4mPNYSkgyU9JulhSWMlrSRptKQRvV02MzNrHr5Wl/U4SasAXwWWi4iPJM0CTNPLxTIzsybkFh9rhDmBNyLiI4CIeCMiXq5uIGl7SY9IelTSMZXl4yQdJ+kBSTdKmrUsX0DSdZLul3SbpEUbWiMzM5siOfGxRrgBmFvS05JOk7RWdaWkLwHHAOsCw4EVJG1RVg8EHoiI5YBbgMPL8jOAfSJieeAA4LQer4WZmU3x3NVlPS4ixklaHlgDWAe4WNJBlU1WAEZHxOsAki4A1gT+DEwALi7bnQ9cLmkQsCpwqaTaPqZtLbakPYA9AOaZZaY61srMzKZETnysISJiPDAaGC3pEWCXymq1+qQ2dkW2VP4vIoZ3IO4ZZOsQIxaYOzoRx8zMpkLu6rIeJ2kRSQtVFg0HXqw8vgdYS9IskvoC25PdWpCf0a3L/R2A2yPiHeB5SduU/UvSMj1ZBzMzmzq4xccaYRBwsqQhwKfAP8jupz8BRMQrkn4M3Ey2/lwbEVeW574HLCHpfuBtYNuyfEfgdEmHAP2BPwIPNaY6ZmY2pXLiYz0uIu4nx+S0tHZlmwuBC9t4/qHAoS2WPQ9sVL9SmplZM3BXl5mZmTUNJz42WYuIQb1dBjMzm3o48TEzM7Om4cTHzMzMmoYTHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfMzMzaxqK8HUbrTksO8OAGL38wg2NOXC1JRoa7zP9+zc85N5HXtbwmDvMOrjhMVfdcYWGxwRg6NDGx3z33cbHbKZjUi/U9ZN/vd7wmAD9F1+gofFW+t1V3P/vN1q9ALZbfMzMzKxpOPExMzOzpuHEx8zMzJqGEx8zMzNrGk58zMzMrGk48TEzM7Om4cTHzMzMmoYTHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyaRocTH0njWjweKemUzgST9DVJB3XmOZPY3xBJ3+3gtuPaWTdM0geSHpT0hKR7Je1SWd9uuSUNl7RJ50pfP5I2lDS23MZJeqrcP7cT++jW+1tew0c7We5Of4bMzMy6o1+jAknqFxFXAVfVcbdDgO8Cp9VhX89GxLIAkuYHLpfUJyLO7kC5hwMjgGvrUI5Oi4jrgesBJI0GDoiIMb1RlpbK+/5pb5fDzMwM6tTVJWleSTdKerj8nacsHyXpeEk3A8dUf+FXWijGltaWtSQNlfTnsp+7JS1dtj1C0lmSRkt6TtK+JfTRwAJlH8dKGlTiPyDpEUmbd6U+EfEc8H1g3xK/Wu5tJD0q6SFJt0qaBvgpsG0px7aSVpR0Z2lBulPSIpX9XC7pOknPSPpV5TXcqJT7IUk3lmUDS73vK/vqVH0kfb+U9VFJ+3XltZA0g6TnJfUvj2eU9IKk/pKWL+W9C9ir8pyRki6VdDVwQ1vva4s4bX2GFijPuU/ST2stU5LOq74eki6Q9LWu1NHMzJpHZ1p8ppc0tvJ4KBNbQU4Bzo2IcyTtBpwEbFHWLQysHxHjJY2sPTkihgNI2gz4EXAncDzwYERsIWld4FyyNQVgUWAdYAbgKUmnAwcBS1b21Q/YMiLekTQLcLekqyIiOlHPmgdKzJYOAzaMiH9LGhIRH0s6DBgREXuXcswIrBkRn0paH/gFsFV5/nBgWeCjUo+TgQ+B35fnPC9paNn2YOCmiNhN0hDgXkl/j4j3JlV4ScsDuwIrAQLukXRLRDzYxlNafX8j4t3SirQp8GdgO+CyiPhE0tnAPhFxi6RjW+xvFWDpiHir1LGt97Wmrc/Qb4DfRMRFkvasbH8msD9wpaTBwKrALi32iaQ9gD0A5p62fxtVNzOzZtGZFp8PImJ47UYmADWrABeW++cBq1fWXRoR41vboaSFgGOBbSPik/K88wAi4iZg5nJQA7gmIj6KiDeA14DZW9sl8AtJDwN/B+ZqY7uOUBvL7wBGSdod6NvGNoOBS5VjXk4AlqisuzEi3o6ID4HHgXmBlYFbI+J5gIh4q2y7AXBQSUhGA9MB83Sw/KsDV0TEexExDrgcWKOd7dt7f88kkyjK37PL+zIkIm4py89rsb+/VerR3vta09ZnaBXg0nK/tp4Sd0FJswHbk8nYF7rUIuKMiBgRESNm7t+wnl0zM5tM9dSRoNrC0mrrhKSBwCXA7hHxcm1xO/v6qLJsPK2XfUdgVmD50iLxApksdMWywBNfKEzEnpJWIltAxkoa3spzfwbcHBFbShpGJi01rdVDfP41qxGwVUQ81YXyt5W4dVpE3KEcvLwW0DciHi0tUO21pFXf9/be1zbDdqBo55Hv+XbAbh3Y3szMmly9Tme/kzz4QB6Ibu/Ac84Gzo6I2yrLbi3PR9LawBsR8U47+3iX7PqqGQy8VpKedcjWlE4rycqvgZNbWbdARNwTEYcBbwBzt1GOf5f7IzsQ8i5gLUnzlRi1rq7rgX0kqSxfthPVuBXYQtKAkmRuCdw2iee051zgIvJ9IyL+B7wtqdYys+MkyjKp97Wtz9DdTOwm3K7Fc0YB+5XyPNbBepiZWROrV4vPvsBZkn4IvM7EbpFWSZoX2BpYuIznAPgWcATZjfIw8D6tjNmoiog3Jd1RupT+ChwDXC1pDDAWeLITdVhA0oNkC9G7wMkRcXYr2x1buugE3Ag8BLzExC6pXwK/As6R9H3gpkkFjojXy1iUyyX1IbvyvkK2HJ0IPFySnxeAr3akMhHxgKRRwL1l0ZntjO/piAuAn5PJT82u5Pv+PuWssjYcwaTf17Y+Q/sB50v6AXAN8HbtCRHxqqQnyLFHZmZmk6Sujfu1ZiNpa2DziNipwXEHkOOPQtJ2wPYRsXll3SPAchHxdnv7AVh2hgExevmFe7bALQxcbYlJb9QT+jd+IPfeR17W8Jg7zNpyqFjPW3XHFRoeE4ChQye9Tb29+27jYzbTMakX6vrJv15veEyA/osv0NB4K/3uKu7/9xutDvnwaE+bpHJW1sZAb0zSuDxwSmnx+h9lLE85W+4s4PiOJD1mZmbQZImPpKX44tlHH0XESr1Rnq6QtCHZpVf1fERs2YHnzkx2z7W0XkS82dbzImKfzpWyfsoYsGVaWf53On6Gm5mZGdBkiU9EPMIX54+ZolRnae7Cc99kCq+/mZlZd/gipWZmZtY0nPiYmZlZ03DiY2ZmZk3DiY+ZmZk1DSc+ZmZm1jSa6qwua259BkzDwKUbfAb8++83Nl7xwRP/bHjM3phM8MLXGz+F06qzzdbwmACatavXW+66GDyu4TF5+eVJb9MTZp654SE/ebDxV9rpP2cvTIQJ8NZbk96mnj79wjWrP+MWHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfMzMzaxpOfMzMzKxpOPExMzOzpuHEx8zMzJqGEx8zMzNrGk58zMzMrGk48TEzM7Om4cTHzMzMmkaPJj6SRkvasMWy/SSd1sHnj5T0pcrjMyUt3oVyrC3pL5OI87qkByU9I+l6SatW1v9U0vrtPH+LrpSrXiQdLGlsuY2v3N+3g8//wusjaZSkrTtRhpGSTulkuTsVw8zMrLt6usXnImC7Fsu2K8vbJakvMBL4LPGJiG9FxOP1LGDFxRGxbEQsBBwNXC5psRL3sIj4ezvP3QLotcQnIo6KiOERMRz4oHY/Ik7qrTLVSOrX22UwMzOr6enE50/AVyVNCyBpGJnIDJB0l6QHJF0qaVBZ/4KkwyTdDmwPjAAuKK0X05cWpBFl243K8x+SdGNZtqKkO0vLzZ2SFulKoSPiZuAMYI+y389aJiQdLelxSQ9L+nVpGfoacGwp5wKSdpd0XynbZZIGVPZzUinbc9XWDkk/kvRIec7RZdkCkq6TdL+k2yQt2tE6SJpO0tllnw9KWqcrr4Wk9SRdUXn8FUmXl/u7Snpa0i3AapVtRkk6XtLNwDGShku6u7xmV0iaqY04D5bynlX5zGwi6UlJt5fX7i+S+pSWuVnLNn0k/UPSLK3sdw9JYySNeeODj7vyEpiZ2VSkRxOfiHgTuBfYqCzaDrgROBhYPyKWA8YA36887cOIWD0izi/rdiytFx/UNigHvN8DW0XEMsA2ZdWTwJoRsSxwGPCLbhT/AeBziYakocCWwBIRsTTw84i4E7gK+GEp57PA5RGxQinbE8A3K7uZE1gd+CrZsoSkjclWo5XKc35Vtj0D2CcilgcOADrURVjsBRARS5FJ5DmSpmtn+zUqXWRjyWQO4CZgsVqSAewKnC1pTuBIMuH5Cl9s8VqYfI9/AJwLHFhes0eAw6sblnKNArYt5e0HfKcs/x2wcUSsDsxa6jQBOB/YsexifeChiHijZaUi4oyIGBERI2aZfpp2qm9mZs2gEYObq91d2wHPkwfJO8oBdhdg3sr2F3dgnysDt0bE8wAR8VZZPhi4VNKjwAnAEt0ot1pZ9g7wIXCmpK8D77fx3CVLC80j5MG5Wo4/R8SE0mU3e1m2PnB2RLxfq09pBVu11GcsmQDM2Ynyrw6cV/b3JPAimYy05bZKF9lwMpkjIqLs5/8kDQFWAf4KrASMjojXI+Jjvvi+XRoR4yUNBoZExC1l+TnAmi22XQR4PiKebrHNosBztfeZz3eRngXsXO7vBpzdTt3MzMyA/GXd0/4MHC9pOWB64EHgbxGxfRvbv9eBfQqIVpb/DLg5IrYs3WqjO13aiZYlW2s+ExGfSloRWI9M4vYG1m3luaOALSLiIUkjgbUr6z6q3Fflb8v69AH+V5KQrmgtceuqs4GryaTv0vI6QOvvQU1H3seatsraZh0i4p+SXpW0LpmE7djWtmZmZjU93uITEePIBOQs8hf73cBqkhYEkDRAUlstEe8CM7Sy/C5gLUnzlX0MLcsHA/8u90d2tcyS1iLH9/y+xfJBwOCIuBbYDxjeRjlnAF6R1J+OHZBvAHarjAUaGhHvAM9L2qYsk6RlOlGNW2uxy+s7D/BUJ57/mYh4GXgZOIRM6gDuAdaWNHOp5zZtPPdt4L+S1iiLdgJuabHZk8Cw2meiss2TwPwliQXYtsXzziS7vC6JiPFdqJqZmTWZRs3jcxGwDPDHiHidTEoukvQwmQi1NWh3FPDbMu5k+trCso89yDOvHmJiN8uvgF9KugPo28kyblviPA38hBw/9ESLbWYA/lLKfQuwf1n+R+CHZXDuAsChZGLwN/Lg3a6IuI7sWhpTurUOKKt2BL5Z6vgYsHkn6nMa0Ld0t10MjIyIjybxnPZcAPyzdlZdRLwCHEEmoX8nx0S1ZRdy8PfDZLL40+rKiPiQHDt0aSnvBOC3ZVzXd4HrlAPeXwXerjz1KmAQ7uYyM7MOUg7hMGufco6eByPiDw2OOygixin71k4FnomIE8q6EcAJEbFGuzsplp99SNyzbcvhRT2sb2fz7/r44Il/Njzmgw/8p+ExL3z97UlvVGen/GKHhscE0KyzT3qjOov3xzU8Ji+/3PiYADPP3PCQnzz4WMNj9p/9CyfVNkaDc42VLhrN/a/+r9XhEp652SZJ0v3A0mS3UqPtXlrBHiO7Mn9XynQQcBnw414ok5mZTaGaanI5SbsC32ux+I6I2Ks3ytMVkg7mi+NpLo2Iozrw3A2BY1osfj4itmzveeV0+l5RWndOaGX50ZTpAMzMzDqqqRKfiDibKXw8SElwJpnktPHc64Hr61siMzOzKYe7uszMzKxpOPExMzOzpuHEx8zMzJqGEx8zMzNrGk01uNma2/hxH/HW3f9oaMyh6w9vaLyaATtt3fCYqy52T+NjzjZbw2Pu/ZMLGx4T4JQDN214zE9eavzcTP1nGtjwmADjbm/8nDoznHRcw2N+eOThk96oB0y3aWtXd+o5mq7t7yO3+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNJz4mJmZWdNw4mNmZmZNw4mPmZmZNQ0nPmZmZtY0nPiYmZlZ03DiY2ZmZk3DiY+ZmZk1DSc+ZmZm1jQ6lfhIGtfi8UhJp3RyH1+TdFBnnjOJ/Q2R9N0ObjuunXXDJH0g6UFJT0i6V9IulfXtllvScEmbdK709SNpQ0ljy22cpKfK/XM7sY/x5TmPSrpa0pA6la3N193MzKyRGtriI6lfRFwVEUfXcbdDgA4lPh3wbEQsGxGLAdsB+0vaFaAD5R4O9FriExHXR8TwiBgOjAF2LI937sRuPijPWRJ4C9irJ8pqZmbWW+qW+EiaV9KNkh4uf+cpy0dJOl7SzcAx1VaiSgvF2NLaspakoZL+XPZzt6Sly7ZHSDpL0mhJz0nat4Q+Glig7ONYSYNK/AckPSJp867UJyKeA74P7FviV8u9TWkVeUjSrZKmAX4KbFvKsa2kFSXdWVqQ7pS0SGU/l0u6TtIzkn5VeQ03KuV+SNKNZdnAUu/7yr46VR9J3y9lfVTSfp146l3AXGUfw8t78bCkKyTNVJbvXsr1kKTLJA0oy+eTdFdZ97NKWU6T9LVy/wpJZ5X735T083L/z5Lul/SYpD0q60+o7Gf38pkaKOmaEv9RSdu2Uv89JI2RNObNTz/tzEtnZmZToc4mPtNXkxXyYF9zCnBuRCwNXACcVFm3MLB+RPygurNKC8WhZCvFncCRwINlPz8Bql01iwIbAisCh0vqDxxEttQMj4gfAh8CW0bEcsA6wHGS1Ml61jxQYrZ0GLBhRCwDfC0iPi7LLi7luBh4ElgzIpYt635Ref5wYFtgKTJZmlvSrMDvga3Kfrcp2x4M3BQRK5T6HCtpYEcKL2l5YFdgJWBlYHdJy3bgeX2B9YCryqJzgQPLe/IIcHhZfnlErFDK+wTwzbL8N8Dppcz/qez6VmCNcn8uYPFyf3XgtnJ/t4hYHhgB7CtpZuCPwNfK+02p09nARsDLEbFMaaW6rmVdIuKMiBgRESNm7tdvUlU3M7OpXGcTn1pXSC1hOayybhXgwnL/PPJgVnNpRIxvbYeSFgKOBbaNiE/K884DiIibgJklDS6bXxMRH0XEG8BrwOyt7RL4haSHgb+TB9jWtuuIthKmO4BRknYH+raxzWDgUkmPAicAS1TW3RgRb0fEh8DjwLxkYnJrRDwPEBFvlW03AA4qieZoYDpgng6Wf3Xgioh4LyLGAZczMfFozfQlzpvAUOBv5bUfEhG3lG3OAdYs95eUdJukR4AdK3VcDbio3D+vsv/bgDUkLV7q/aqkOcnPzp1lm30lPQTcDcwNLBQR7wE3AV+VtCjQPyIeIZOw9SUdI2mNiHi7g6+LmZk1qZ4c4xOV+++1tkFpubgE2D0iXq4tbmdfH1WWjQda+wm/IzArsHxJzl4lk4WuWJZsyfh8YSL2BA4hD8xjS6tESz8Dbi4tEZu1KENr9RCff81qRLYC1RLOeSLiC2VqQ2dbuj4or9m8wDRMeozPKGDviFiKbKmr1vELdYmIfwMzkS01t5KJ0DeAcRHxrqS1gfWBVUor0oOVfZ4JjGRiaw8R8TSwPJkA/VJSNRE3MzP7gnomPneSA4Ihk4/bO/Ccs4GzI+K2yrJby/MpB8I3IuKddvbxLjBD5fFg4LWI+ETSOuRBvNMkDQN+DZzcyroFIuKeiDgMeINMgForx7/L/ZEdCHkXsJak+UqMoWX59cA+te66jnRVVdwKbCFpQEkyt2Ril1KbSsvJvsABwPvAfyXVWop2AmqtPzMAr5QuqB0ru7iDz38Wqu4C9mNi4nNApUyDgf9GxPulZWflSpnuIV/nHSitSZK+BLwfEeeT79Vyk6qbmZk1t3oOetgXOEvSD4HXyV/mbZI0L7A1sLCk3cribwFHAGeXrqr3gV1a30OKiDcl3VG6lP4KHANcLWkMMJYca9NRC0iqtTK8C5wcEWe3st2xpYtOwI3AQ8BLTOyS+iXwK+AcSd8nu2naFRGvl8G8l0vqQ3blfYVsOToReLgkPy8AX+1IZSLiAUmjgHvLojMj4sEOPvfB0uW0Hfke/LYMXn6Oie/tocA9wItkq0st8fsecKGk7wGXtdj1bcAGEfEPSS+SXWq1xOc6YM/y3j9FdndVXQIMj4j/lsdLke/FBOAT4DsdqZuZmTUvRbTWu2I2+ZH0F+CEiLixK88fPnD6uGGJ+epcqvYNXX94Q+PVaImlGx4zxtzT8JjMNlvDQ+79kwsnvVEPOOXATRse85OX/jPpjeqs/0wdOnej7sY99FLDY85w0nENj/nhkYdPeqMeMN2m6zY03kpHncWYF19pdbiHZ262yZ5yksqnyTFIXUp6zMzMoL5dXVMESUvx+TONAD6KiJV6ozxdIWlDskuv6vmI2LIDz52Z7J5rab2IeLMe5au3iPgfOSWCmZlZtzRd4lNOgx7e2+Xojoi4nhz03JXnvskUXn8zM7OucleXmZmZNQ0nPmZmZtY0nPiYmZlZ03DiY2ZmZk3DiY+ZmZk1jaY7q8uaV9+B0zDTiGGNjfmdQxoar2bCE/dOeqN6Gzp00tvUmWbt6vWHu643JhIE2PuYaxoe85RDtmh4zH4/PqXhMQEGfO8bjQ/6zMMNDzntInM3PCZAvPZqY+N9+mmb69ziY2ZmZk3DiY+ZmZk1DSc+ZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNJz4mJmZWdNw4mNmZmZNw4mPmZmZNQ0nPmZmZtY0nPhMYSSNa/F4pKRTyv09Je1c7o+StHW5P1rSiHb22e76DpTps1idfN4yku6S9IikqyXN2NUymJmZdYQTn6lIRPw2Is7t7XJ0wpnAQRGxFHAF8MNeLo+ZmU3lnPhMRSQdIemAdtb3La0zj5ZWlv0rq7eRdK+kpyWtUbYfJuk2SQ+U26pluSSdIulxSdcAs1ViLC/pFkn3S7pe0pztFHkR4NZy/2/AVpOIu3bZ9yWlnEdL2rGU+xFJC3TldTMzs+bRr7cLYJ02vaSxlcdDgas6+NzhwFwRsSSApCGVdf0iYkVJmwCHA+sDrwFfiYgPJS0EXASMALYkk5algNmBx4GzJPUHTgY2j4jXJW0LHAXs1kZ5HgW+BlwJbAPMXZa3FRdgGWAx4C3gOeDMUu7vAfsA+1UDSNoD2ANgnkHTdfBlMjOzqZUTnynPBxExvPZA0kgmJgWT8hwwv6STgWuAGyrrLi9/7weGlfv9gVMkDQfGAwuX5WsCF0XEeOBlSTeV5YsASwJ/kwTQF3ilnfLsBpwk6TAyeft4EnEB7ouIVwAkPVupwyPAOi0DRMQZwBkAy882ONopi5mZNQEnPk0kIv4raRlgQ2Av4BtMbI35qPwdz8TPxf7Aq2QrSx/gw+ruWgkh4LGIWKWD5XkS2ABA0sLAph2I+1Hl/oTK4wn482xmZpPgMT5NRNIsQJ+IuAw4FFhuEk8ZDLwSEROAncgWHMhxOduVMUNzMrGl5SlgVkmrlHj9JS3RTnlmK3/7AIcAv51EXDMzs25x4tNc5gJGlzFCo4AfT2L704BdJN1Ndje9V5ZfATxDdi+dDtwCEBEfA1sDx0h6CBgLrNrO/reX9DTwJPAycPYk4pqZmXWLuwamMBExqMXjUWQSQ0QcUVk+snJ/7cpTvtDKU10fEW9QxvhExDPA0pVNf1yWB7B3G+UbS44BmqSI+A3wm1aWtxV3NDC6jXJ/bp2ZmVlr3OJjZmZmTcMtPtbjJJ0KrNZi8W8i4uzWtjczM+spTnysx0XEXr1dBjMzM3BXl5mZmTURJz5mZmbWNJz4mJmZWdNw4mNmZmZNw4mPmZmZNQ2f1WVNIz6dwCdvjmtozHHbfaOh8WoGnzOq4THj79c2Pubgxr6fAJ+89J+GxwQ45ZAtGh5z75//ueExdzrj5obHBFhpp5UbHjP+9WLDY9K/f+NjAgwaNOlt6qlP2+06bvExMzOzpuHEx8zMzJqGEx8zMzNrGk58zMzMrGk48TEzM7Om4cTHzMzMmoYTHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfMzMzaxpOfMzMzKxpOPGZzEka1+LxSEmnlPt7Stq53B8laetyf7SkEe3ss931HSjTZ7G68Lx/S5q2PJ5F0gtdLYeZmVlnOfGZgkXEbyPi3N4uRyeNB3br7UKYmVlzcuIzBZN0hKQD2lnft7SyPCrpEUn7V1ZvI+leSU9LWqNsP0zSbZIeKLdVy3JJOkXS45KuAWarxFhe0i2S7pd0vaQ5J1HsE4H9JfVrUVZJOrZS1m0nsXzt0nL1J0lPSrpAklp5DfaQNEbSmDc++mQSRTMzs6ldv0lvYr1sekljK4+HAld18LnDgbkiYkkASUMq6/pFxIqSNgEOB9YHXgO+EhEfSloIuAgYAWwJLAIsBcwOPA6cJak/cDKweUS8XpKSo2i/Recl4HZgJ+DqyvKvl/IuA8wC3CfpVmDVNpYDLAssAbwM3AGsVvb9mYg4AzgDYLmhM0S7r5aZmU31nPhM/j6IiOG1B5JGkslIRzwHzC/pZOAa4IbKusvL3/uBYeV+f+AUScPJLqmFy/I1gYsiYjzwsqSbyvJFgCWBv5XGlr7AKx0o1y/I5O2ayrLVKzFelXQLsEI7y98B7o2IfwGU5HAYLRIfMzOzKic+U7GI+K+kZYANgb2AbzCxNeaj8nc8Ez8H+wOvkq0rfYAPq7trJYSAxyJilU6W6x8lUflGi321pq3lMLEO8Pl6mJmZtcpjfKZikmYB+kTEZcChwHKTeMpg4JWImEB2RfUty28FtitjhuYE1inLnwJmlbRKiddf0hIdLN5RQHV80q3AtiXGrGQr073tLDczM+s0/0Keus0FnC2pluD+eBLbnwZcJmkb4GbgvbL8CmBd4BHgaeAWgIj4uJzWfpKkweTn6UTgsUkVLCIek/QAE5OxK4BVgIfI1qUfRcR/JLW1fNFJxTAzM2vJic9kLiIGtXg8ChhV7h9RWT6ycn/tylO+0MpTXR8Rb1DG+ETEM8DSlU1/XJYHsHcb5RtLtsJMUrWM5fHXK/cD+GG50YHlo4HRlcetls/MzKzKXV1mZmbWNNziY3Un6VTy1PKq30TE2b1RHjMzsxonPlZ3EbFXb5fBzMysNe7qMjMzs6bhxMfMzMyahhMfMzMzaxpOfMzMzKxpeHCzNQ1N049p5p5t0hvW0R03/aOh8WrWve2aSW9Ub9EL14B9+eWGh+w/08CGxwTo9+NTGh5zpzNubnjM8157u+ExAVaaeebGB33uucbH7I16AnzwQWPjxYQ2V7nFx8zMzJqGEx8zMzNrGk58zMzMrGk48TEzM7Om4cTHzMzMmoYTHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfMzMzaxpOfMzMzKxpOPExMzOzpuHEZwoiaVyLxyMlnVLu7ylp53J/lKSty/3Rkka0s89213egTJ/F6uTzfibpYUljJd0g6UtdLYOZmVlHOfGZSkTEbyPi3N4uRyccGxFLR8Rw4C/AYb1cHjMzawJOfKYSko6QdEA76/uW1plHJT0iaf/K6m0k3SvpaUlrlO2HSbpN0gPltmpZLkmnSHpc0jXAbJUYy0u6RdL9kq6XNGdb5YmIdyoPBwJR9jFS0pWSrpP0lKTDK+V5UtKZpQ4XSFpf0h2SnpG0Yhv13kPSGElj3nj/o0m/kGZmNlXr19sFsE6ZXtLYyuOhwFUdfO5wYK6IWBJA0pDKun4RsaKkTYDDgfWB14CvRMSHkhYCLgJGAFsCiwBLAbMDjwNnSeoPnAxsHhGvS9oWOArYra0CSToK2Bl4G1insmpFYEngfeC+kmC9ASwIbAPsAdwH7ACsDnwN+AmwRcsYEXEGcAbA8nPMFB14nczMbCrmFp8pywcRMbx2o3PdQ88B80s6WdJGQLXF5fLy935gWLnfH/i9pEeAS4HFy/I1gYsiYnxEvAzcVJYvQiYrfyvJ2SHAl9srUEQcHBFzAxcAe1dW/S0i3oyID0rZVi/Ln4+IRyJiAvAYcGNEBPBIpdxmZmZtcuLTJCLiv8AywGhgL+DMyupaH9B4JrYC7g+8Wp4zApimurtWQgh4rJKYLRURG3SweBcCW7Wz/9rjal/VhMrjCbj10szMOsCJT5OQNAvQJyIuAw4FlpvEUwYDr5TWlZ2AvmX5rcB2ZczQnEzsonoKmFXSKiVef0lLtFOehSoPvwY8WXn8FUlDJU1Pdl/d0ZE6mpmZTYp/JTePuYCzJdWS3R9PYvvTgMskbQPcDLxXll8BrEt2Lz0N3AIQER+X09pPkjSY/GydSHZJteZoSYuQrTUvAntW1t0OnEeO6bkwIsZIGtbBepqZmbXJic8UJCIGtXg8ChhV7h9RWT6ycn/tylO+0MpTXR8Rb1DGykTEM8DSlU1/XJYHnx+PU93XWHIM0CRFxFbtrH4tIj4XIyJeIMcQ1R6PbGudmZlZW9zVZWZmZk3DLT7WoySdCqzWYvFvIuLs1ravtmKZmZnVmxMf61ERsVdvl8HMzKzGXV1mZmbWNJz4mJmZWdNw4mNmZmZNw4mPmZmZNQ0nPmZmZtY0fFaXNY9PxzP+rbcbGnKdQ7draLzPSL0Tt9FmnrnhIcfd3tZk5D1rwPe+0fCYK+20cuNj9sJ7CrD3Ty5seMxT776o4THjvN82PCYACy/c2Hh92m7XcYuPmZmZNQ0nPmZmZtY0nPiYmZlZ03DiY2ZmZk3DiY+ZmZk1DSc+ZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNJz4mJmZWdOYZOIjaVyLxyMlndKZIJK+Jumgzhaunf0NkfTdDm47rp11wyR9IOlBSU9IulfSLpX17ZZb0nBJm3Su9PUjaUNJY8ttnKSnyv1zO7GP8ZV9jC2vyZ11KFvttR0r6XFJv5XU5udN0p6Sdp7EPnv19TYzsylfj1+kVFK/iLgKuKqOux0CfBc4rQ77ejYilgWQND9wuaQ+EXF2B8o9HBgBXFuHcnRaRFwPXA8gaTRwQESM6eRuPoiI4S2WrdpyI0l9I2J8J/f9bEQMl9QPuAnYAri8tQ0joiNXzhtOL77eZmY25etWV5ekeSXdKOnh8neesnyUpOMl3QwcU20latG68IGktSQNlfTnsp+7JS1dtj1C0lmSRkt6TtK+JfTRwAJlH8dKGlTiPyDpEUmbd6U+EfEc8H1g3xK/Wu5tJD0q6SFJt0qaBvgpsG0px7aSVpR0Z2lBulPSIpX9XC7pOknPSPpV5TXcqJT7IUk3lmUDS73vK/vqVH0kfb+U9VFJ+3X2dai1kklaW9LNki4EHpHUt7ze95X36tsd2V9EfArcCSzYzmfmCEkHlPujJR1TWuCelrRGG6/3WpXP0oOSZmilLntIGiNpzOsfftLZl8LMzKYyHWnxmV7S2MrjoUxsBTkFODcizpG0G3AS+aseYGFg/YgYL2lk7cm11gVJmwE/Ig+IxwMPRsQWktYFziV/3QMsCqwDzAA8Jel04CBgycq++gFbRsQ7kmYB7pZ0VUREB1+HqgdKzJYOAzaMiH9LGhIRH0s6DBgREXuXcswIrBkRn0paH/gFsFV5/nBgWeCjUo+TgQ+B35fnPC9paNn2YOCmiNhN0hDgXkl/j4j3JlV4ScsDuwIrAQLukXRLRDzYxlOq7+/zEbFli/Urkq/185L2AN6OiBUkTQvcIemGiHh+EmUaAKxHvobtfWaq+kXEisqurcMjYv1WXu+rgb0i4g5Jg8jX83Mi4gzgDIDlZ5mxK58HMzObinQk8flcV0hJYkaUh6sAXy/3zwN+VXnepW11jUhaCDgWWDciPpG0OiVBiIibJM0saXDZ/JqI+Aj4SNJrwOyt7RL4haQ1gQnAXGW7/3Sgfq3tqzV3AKMkXUIb3TXAYOCcUr8A+lfW3RgRbwNIehyYF5gJuLWWOETEW2XbDYCv1VpAgOmAeYAnOlD+1YErakmSpMuBNYC2Ep/Wurqq7q0kNhsAS0vaulLfhYC2Ep8FSlIVwJUR8VdJ59H2Z6aq9hrfDwxrY5s7gOMlXQBcHhH/aqceZmZmdR/jU/1F3WrrhKSBwCXA7hHxcm1xO/v6qLJsPK2XeUdgVmD5kki9QCYLXbEsrSQYEbGnpJWATYGxkoa38tyfATdHxJaShgGjK+taq4f4/GtWI2CriHiqC+VvK3Hrqur7KGCfMraoI56dRFIFrdcfJr5ebb3nRMTRkq4BNiFb+daPiCc7WDYzM2tC3T2d/U5gu3J/R+D2DjznbODsiLitsuzW8nwkrQ28ERHvtLOPd8mur5rBwGsl6VmHbE3ptJKs/Bo4uZV1C0TEPRFxGPAGMHcb5fh3uT+yAyHvAtaSNF+JUevquh7YR5LK8mU7UY1bgS0kDShJ5pbAbZN4TkddD3xHUv9SroVLjM7oymem5nOvd3lPHomIY4AxtN5FaWZm9pnutvjsC5wl6YfA6+TYkjZJmhfYGli4jO8A+BZwBHC2pIeB94FdWt9Diog3Jd0h6VHgr8AxwNWSxgBjgc786l9A0oNkC9G7wMkRcXYr2x1burAE3Ag8BLwEHFS6c35JdtucI+n75FlM7YqI18u4mcuVp3q/BnyFbDk6EXi4JD8vAF/tSGUi4gFJo4B7y6Iz2xnf01lnkt1OD5RyvU7r43Pa06nPTAs38/nXe/WS6I4HHic/C2ZmZm1S18b/mk15lp9lxrjzays1NGbf4Us1NF6NZhzS8Jjx2CMNj8lsszU85Lgrb214TIABi32p4TH7DP7CiZI9b+aZGx8T2PsnFzY85ql3X9TwmHFeR2YO6QELL9zQcCv9+gLuf+nVVod+eOZmMzMzaxo9PoHh5EDSUuQZRFUfRURjf/53g6QNyS69qtZOP2/tuTOT3XMtrRcRb3azXFP8a2tmZs2jKRKfiHiEifMCTZGqszR34blv0kP1nxpeWzMzax7u6jIzM7Om4cTHzMzMmoYTHzMzM2saTnzMzMysaTjxMTMzs6bRFGd1mQFo6FD6bbt9Q2OOv/SPDY1X8+GL3ZqloEsGLDNfw2N+8uBjDY85w0nHNTwmAM883PCQ8a8XGx6T555rfEx6ZzLBvVZu7PcRwKlX/LLhMQHe/+2ohsaLt8e1uc4tPmZmZtY0nPiYmZlZ03DiY2ZmZk3DiY+ZmZk1DSc+ZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNJz4mJmZWdNw4mNmZmZNw4nPFEbStZKGlPutXoxE0ihJWze0YGZmZlMAX6R0ChMRm9R7n5L6RcSn9d5vd02u5TIzsymXW3wmM5J+JGnfcv8ESTeV++tJOl/SC5JmafEcSTpF0uOSrgFmq6xbXtItku6XdL2kOcvy0ZJ+IekW4Hvl8TGS7pX0tKQ1ynZ9JR0r6T5JD0v6dlk+p6RbJY2V9KikNcq2o8rjRyTt3049dy/7fEjSZZIGlOWjJB0v6WbgGEkLSLqulP82SYuW7TaTdI+kByX9XdLsdXwbzMxsKuXEZ/JzK7BGuT8CGCSpP7A6cFsbz9kSWARYCtgdWBWgPO9kYOuIWB44Cziq8rwhEbFWRBxXHveLiBWB/YDDy7JvAm9HxArACsDukuYDdgCuj4jhwDLAWGA4MFdELBkRSwFnt1PPyyNihYhYBniixKlZGFg/In4AnAHsU8p/AHBa2eZ2YOWIWBb4I/Cj1oJI2kPSGEljXn+71Z5BMzNrIu7qmvzcDywvaQbgI+ABMgFaA9gX+HErz1kTuCgixgMv11qJyGRoSeBvkgD6Aq9Unndxi/1cXinDsHJ/A2DpypihwcBCwH3AWSW5+nNEjJX0HDC/pJOBa4Ab2qnnkpJ+DgwBBgHXV9ZdGhHjJQ0ik7hLS/kBpi1/vwxcXFqwpgGeby1IRJxBJk+MWHhYtFMeMzNrAk58JjMR8YmkF4BdgTuBh4F1gAXIlpE2n9rKMgGPRcQqbTznvRaPPyp/xzPxsyGyxeX6FtsiaU1gU+A8ScdGxLmSlgE2BPYCvgHs1kbsUcAWEfGQpJHA2q2Uqw/wv9Kq1NLJwPERcZWktYEj2ohjZmb2GXd1TZ5uJbt1biW7t/YExkZEWy0WtwLblTE2c5KJEsBTwKySVoHs+pK0RCfLcj3wndKyg6SFJQ2UNC/wWkT8HvgDsFwZe9QnIi4DDgWWa2e/MwCvlP3u2NoGEfEO8LykbUpslcQKsuXp3+X+Lp2sk5mZNSm3+EyebgMOBu6KiPckfUjb43sArgDWBR4BngZuAYiIj0sX1UmSBpPv94nAY50oy5lkt9cDyv6m14EtyBaaH0r6BBgH7AzMBZwtqZZQt9YtV3MocA/wYin3DG1styNwuqRDgP7keJ6HyBaeSyX9G7gbmK8TdTIzsyblxGcyFBE3kgf52uOFK/eHVe4PKn8D2LuNfY0lxwC1XL52W48j4g3KGJ+ImAD8pNyqzim3ltpr5anGOx04vZXlI1s8fh7YqJXtrgSu7EgsMzOzGnd1mZmZWdNwi4/1KEmnAqu1WPybiGjvVHczM7Me4cTHelRE7NXbZTAzM6txV5eZmZk1DSc+ZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNtX0VBLOpy0L9+scJg2dpaMyNvrt+Q+N9ZpbG1hPgk7seaHjM/nMObXjMj577T8NjAky7yNyND9q//6S3qbe+fRsfE+CttxoeUut/YW7WHrfXlu1NqN9zTv39/g2Nt9LPz2TMCy+rtXVu8TEzM7Om4cTHzMzMmoYTHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfMzMzaxpOfMzMzKxpOPExMzOzpuHEx8zMzJqGEx8zMzNrGk58zMzMrGk48bG6knSmpMXL/RcktXq1TElDJH23saUzM7Nm58TH6ioivhURj3dg0yGAEx8zM2soJz42SZIGSrpG0kOSHpW0raT1JD0o6RFJZ0matmw7WtKIDuz2aGABSWMlHSvpPEmbV2JeIOlrkkZKulLSdZKeknR4ZZv/k3Rv2cfvJPVtpex7SBojaczbMaEeL4eZmU3BnPhYR2wEvBwRy0TEksB1wChg24hYCugHfKeT+zwIeDYihkfED4EzgV0BJA0GVgWuLduuCOwIDAe2kTRC0mLAtsBqETEcGF+2+ZyIOCMiRkTEiMHyx93MrNn5SGAd8QiwvqRjJK0BDAOej4iny/pzgDW7EyAibgEWlDQbsD1wWUR8Wlb/LSLejIgPgMuB1YH1gOWB+ySNLY/n704ZzMxs6tevtwtgk7+IeFrS8sAmwC+BG3oo1Hlkq812wG7VIrQsEiDgnIj4cQ+VxczMpkJu8bFJkvQl4P2IOB/4NdkNNUzSgmWTnYBbOrnbd4EZWiwbBewHEBGPVZZ/RdJQSdMDWwB3ADcCW5cWIsr6eTtZBjMzazJu8bGOWAo4VtIE4BNyPM9g4FJJ/YD7gN92ZocR8aakOyQ9Cvw1In4YEa9KegL4c4vNbydbgxYELoyIMQCSDgFukNSnlGsv4MWuVtLMzKZ+TnxskiLieuD6VlYt28q2a1fuD5vEfneoPpY0AFgIuKjFpq9FxN6tPP9i4OL2YpiZmVW5q8smC5LWB54ETo6It3u7PGZmNnVyi4/1KEkzk+NxWlovIt6sPYiIvwPztNwoIkaRY3/MzMy6zYmP9aiS3Azv7XKYmZmBu7rMzMysiTjxMTMzs6bhxMfMzMyahhMfMzMzaxpOfMzMzKxpKKLlZZDMpk4jFpwn7vn1jxoaM+67q6Hxasa/8lrDY/YdNnfDY/LWWw0PqcWXaHhMgHjt1cYHHTSo8TE/+KDxMQEGDmx4yA+uvbXhMQdsvUnDYwLstfsJDY13Ge/zeoxXa+vc4mNmZmZNw4mPmZmZNQ0nPmZmZtY0nPiYmZlZ03DiY2ZmZk3DiY+ZmZk1DSc+ZmZm1jSc+JiZmVnTcOJjZmZmTcOJj5mZmTUNJz5mZmbWNPr1dgGs8STNAZwIrAB8BLwA7BcRTzcg9szAjeXhHMB44PXyeMWI+Liny2BmZs3LiU+TkSTgCuCciNiuLBsOzA7UNfGR1DcixleXRcSbwPCy/ghgXET8up5xzczM2uKuruazDvBJRPy2tiAixgK3SzpW0qOSHpG0LYCkiyV9djlfSaMkbSWpb9n+PkkPS/p2Wb+2pJslXQg80pECSZpB0vOS+pfHM0p6QVJ/SaMlnSjpzlK2Fcs2AyWdVeI/KGnzNva9h6Qxksa8/s64rr1iZmY21XDi03yWBO5vZfnXyZaYZYD1gWMlzQn8EaglQdMA6wHXAt8E3o6IFcgus90lzVf2tSJwcEQs3pECRcS7wGhg07JoO+CyiPikPB4YEasC3wXOKssOBm4q8dcp5R3Yyr7PiIgRETFi1hkHdaQ4ZmY2FXPiYzWrAxdFxPiIeBW4hUxo/gqsK2laYGPg1oj4ANgA2FnSWOAeYGZgobKveyPi+U7GPxPYtdzfFTi7su4igIi4FZhR0pAS/6ASfzQwHTBPJ2OamVmT8Rif5vMYsHUry9XaxhHxoaTRwIZky89Fle33iYjrP7cTaW3gvc4WKiLukDRM0lpA34h4tLq65eYl/lYR8VRnY5mZWfNyi0/zuQmYVtLutQWSVgD+C2xbxu7MCqwJ3Fs2+SPZCrMGUEt0rge+UxmXs3BrXU2ddC6ZWJ3dYnmtq211snvt7RJ/nzJYG0nLdjO2mZk1Abf4NJmICElbAidKOgj4kHI6OzAIeIhsUflRRPynPO0GMim5qnK6+ZnAMOCBkny8DmzRzeJdAPycia1KNf+VdCcwI7BbWfYz8pT8h0v8F4CvdjO+mZlN5Zz4NKGIeBn4RiurflhuLbf/hBzDU102AfhJuVWNLreOlOOIFotWB/4UEf9rsfyyiPhxi+d+AHy7I3HMzMxqnPjYZEHSyeTg6U0mta2ZmVlXOfGxHtNiluaq9cpEhp+JiH1a20dErN0DRTMzsyblxMd6THWWZjMzs8mBz+oyMzOzpuHEx8zMzJqGEx8zMzNrGk58zMzMrGk48TEzM7OmoYiWl0EymzpJeh14sYtPnwV4o47FcczejdlbcR1z6ovrmJNn3HkjYtbWVjjxMesASWMiYoRjTh0xeyuuY059cR1zyovrri4zMzNrGk58zMzMrGk48THrmDMcc6qK2VtxHXPqi+uYU1hcj/ExMzOzpuEWHzMzM2saTnzMzMysaTjxMTMzs6bhxMfMzMyahhMfszZIUoPjDZS0aLm/tKS5GhRX5W+fyv1G133mRsVsdN2aVSNf5954Tyv/Kw35P+1NU9v/TL/eLoDZ5EiSopzyKOmbwKLAjcCYiOipadtnBX4h6QVgOPCdHorzmVo9JW0KfAN4T9J1EXFVT8eulGF+4HsR8b1GxCv1XQuYJiL+Vn2ve0rldV4Y+CAi/tmT8VrEHgF8DLwXEc82IF7t9RwC/Len48Fn7+kqwPTA/yLigQbF3AT4dfn/eaEnP0eVz9DiwFtA34j4d09/fitx1wRWBR4CnomIfzQg5urAXMB/I+KGeu3fLT5mragkPZsBuwEBbAV8S9KcPRTzBeBuYA/gjoh4qtoK00MxQ9L6wBHAicAA4AeS+vdUzFa8Cqwmaa8Gxvwy8DNJs/Z00gOfvc5fAy4AZu7peDWS1gCuBfYDflne6x5VSQiuknSqpFUkTdOTMSWtDPwF2AI4TdI2PRmvxFwVOAEYGRHPA4N7Ml7lB8pF5I+iEyQNL8t7+jtiU+AUMpE9BNhGUo/lDyXmBsDvyTzlOkn/V6/9O/Exa4OkrwJHAjtGxI+AP5MXzNu5ns3blSbzOYCXgAOBTST9X0RMKF8C09UrXivmBb5N/rJaGNg5Ij4p5ekxkoZIGhoR7wH7A/OW7r66f4m3ss+LgTvJuiOpb71jtoi/HPBz8rM0VtJskhbroVi1z9NQYAEyGdgd+CuwX08nP5JWAL5L1vdjYDtgg3onP5V6zka2Ln09IvYtcQ+VtHWd401TO9hLmpFMdE4B+kr6LnC7pBMlzVvPuJX4w8m6bQaMAxYjk58VejL5Kd896wKbAE+SrWqjImKCpIE9EE/l9f0msA15YemHgb/XK4YTH7OilS+O/wDzAPsCRMQ1ZHfXMGDbeh0sK60BlwH3RcSpwKHAAZI2K11Bh9Yr+WmlnoOAs4HvA1tFxIuSNgb2kjR9PWK2jF2Sqt+RrRCbAC+T3XsL9EQLTHmNV5N0rKQREfEp8G/gx2X9+HrHrClf4v8DHgBWlnQQmXidJGnzeser/EL/HdnSM0ep32XAn4CDJW1Y77jw2ft6NPBSRFwPHES+t+uTyXxdkp9KV8hGwOXAMcCqkgZGxF+AnwC/krRdneL1A1YDtiqv7ffI4+dGZDLyIfk9MSf5/VA3lf/XAHYGFgS2J5OCx4FTJa3UUy2XEfEh8D5wJvBLYIuIeKX8367YA/EiIt4B7gd2AY4nv5f+I2k3Sd2O6cTHjC+M6VlI0gIRMQZYjzxY/QQgIv5KftGeX6+DpaRlgcOBH9TGYETEtcCPyC/0v5BdXx/WI16t71zSdsqxLqcAL5BjI16R9BWyCf/2iPigHjFbxF4D+BTYB7iObFVbjxxHdaCkQfWMCVC6Jx8G+gO7Sroc+Bswp3IMTI+Q9CXyPRwMPEHW+VnywHlNKU+9Y64A7AmcDowFtpS0YDmYXE52t71Z77jFOLJeG0jaJCI+Ao4jx6RsQJ26g8rnaEWyNekgMqkbRnaZTleSn/3I5LYeJgBvk11Mvyf/N64pj7eIiLPIlokFgffqEbCS8MxW6vRQRDwCrAkcFRFPAs+TrTB1a+2p/DhZQNIy5fEt5Gf1rIh4STme6gSgXt+BtZhza+JQgk/J93fXiHhW0tLkj7Pu/wCMCN98863cyGRjNDmA7yhgLWBJ4Dbgpz0Uc1Xg1HJ/eqBPZd3swMJ1ilO7RM3y5EH4t2T33dnAjMCVwKWlrpvUuY612IsAtwOvA/OXZQsCKwBnkd0xs9U55kLlPd21su5Q4CTyQH14D36eFgGuBhYtjwdW3oNHgbXrHG82Mpk8r7LsNODcShn69sD7uhzwVWCZ8ln6P+AqYKOyvl/t/a5T3MHAfcB1lWUHlc/0V4HpWpaxDjFnLP8bfyETy2kr6zYF7gW2rPP7uQk57u9E4LKy7FfkOJ+vl9dg6R743G4EPFL2fxKZ4O0PXFG+Jx4FvlrnmBsD/yifm2vJ7ss/AJcA55Mtpl+rS6x6v2C++TYl3aoHgfKlfSfZErogZUAoORh1RbKba+bufpG2fH45CN4P9K8sWx3YoQfquz7wR2D18nhQ+VI5vPJ41h56rTcuX+LfKF/cr9EiqQNGAT+vY8zNyNaAK8nk58DKugFkYvs0sESd6zp75f4Pyhf5jOSv5hXq+SXeIu5AcjD+09XPD5nc/hEY0AMxNyJbHk4uB8Rvl/+lnYGbqVMS3cr/zTrAP8kzAmvLjiAT6FnqGRP4Uvle6FO+C34H/KisG0p2PY1orZzdiL0s8CA57m4f4LGyfBqyteUP1DnRKvtfimy1W4BsXdmLbJFemTzzdMXa/20d67oY2ZK2Ynl8FpkA9SeT6nWBZesVs64vmG++TUm38uV8euXLbeVyYJ62PJ4PuJVsyobKL7xuxKzF2phsUTqqHBB/Vb7k1ibPHnsG2LAH6rwj2Wy/T3ncB1gF+H0DXu9jgb0qjw8F/gUMqyzbo3wBdrtVohyQxgBLkwnBemSL1n4ttjsDWKWO9ZybbHm4jBzzsRB5Jsywsn4uYMHq56EOn6cR5fNba9XZnmyZ2K6ybb2TO5GJ8sXAemXZWuRYkK3LZ+tblISgTjHXLp+bbcpBeATZ0rJPZZsF6lzPzcjupD8BvyjLvkImPxeSXaiL1+s1LX/nJJOBr5MtWPcA85V1i5e/ferxGWoRf7byWf0PMLwsG0r+UPlJvT8/5e8A4A7grtrntyy/EjiynjFrN4/xsaYVEQ+R40vWKANQ7yG7uP5P0uDIU1RvJ8/kInK8Qndj1gZk/pT8VbUmcHLkWWNXAduSyck+kYNDu6V65oukaSLiAvKgcZyk1SJiAjADsKx6fhLBd8mkgHJ2zEnkWWx/kfTlss0A4KSoz/ipfuQ4gdcjzxy7jxwMup2kb5dyLAgsAXRrbqbK6yzgFeBgsm4/AQ4jW0H2AIiIf0eZAyXKN3xXlc/TJmSr3crAaEkrR8RFwHnAtyXtWLZ9rDuxWosdEePIsS/LSOoXEbeQY6e+B/QFzo4cK9dtklYDflMefosc79EP2BvYW9L3Srm6NVeRKqdpS1qIbKHckzzbck1Jx0fE38hBt0+SrYiPdydmTXk/1yb/N5YiuymPIVton1fOpXOkpDnK/263P0M1yrMM1yJbBq8G9pC0SES8RX5XzaU8s63b3xGS+pS6rksms98mB4ivpjwbEbIb/tPuxmpVT2RTvvk2Od9o0f8PnEO27MwIbE42I18JHEA24S9Y5/iHkQN5NycTq/nbKl8349R+UW1Gjvv4K3kgHkr+Ip9AdoGcDGxe5zrWYq9AfpmuWOL+k9LiQo5tOoJsIdirLOtyS08l5uxM/DV8CNnyMmt5vCU52PYPZPfFtMDQOtV1Y7Jr9IeVdXMBGwLXkwePZer8Oi9OdpstWD5PLwEfAeuU9TsAK/TA+/plJrYu7VLexzXL46XKa163bjVyrNS1wDbl8cLkGXmHlMdrUrpvuxlndjLRmYZsdbmbbNEaUNbPWr4rzmjtdalTPU8H1iiPv0u2iq5OtuI9TA90kZZY/0e2Ovcjv5+OIn8s7Fteh253WQJfavF4vfK/IbIVbTTZAvst8kfopj1S157YqW++Ta43skVhv3LA2B34VVl+HvkrZwgTT2E/EFikB8pwdDkI/p2JA3y3JH+59qMyuLmL+68Ojl6Y/FW6LDkA8wDg9LJuO/IMlG3L4351ql8t6diIHPNxYvkSO6gcOMaQY3leILuhfggcVKeYmwA3lYPVr8mxAd8HHiPHSfyjLLuYOnb9kEnPI2SC9xKZTC/M5xOF31G6TbsRZ15gJSpdK+UgtRrwQHl8IJnUdjsRaBG7VpfNybPFbicT563Ibsxzye6gh8g5dboTa1h5n2oJwFrloHg1MENZtnT5LM1ZxzquR3ZNDiUPxt8o9fwKOdM3ZHJ0D3nSQ70Snr7keJYTy2u7Z+UzvSd5Jt4oJg4Wr2f31jSV+6dRTgIor+/vyO7hLcuyLn83kUnd88CSlWWzlM/QMpXX/3ayq3jFen5+P1eWntqxb75NrjcyAXivHAyrg1AvKF+sM5bH9fxyqY3BWACYgzzoH17WrQE8RfmV3s04c5UD3+yVuH+prF+MbMqutQjsVA6SK9ch9izAzOX+ILKFadPyeCCZgO1Bnrk2F/mLel3yV+ZiXYxZHRC+CDmod2Wy+fx75OnbXyZbJb5TvsxXKgfMuer03k5TPjvLkqdsjyETgMtKmWoHsBOYOBNtpz9b5b0bUw5OZ5DdSLV1uwNHl/sblc/xV+pVv/K3D9lVeQslaSRbL39W6jkfOR5leHf+f0o972XiWYc7kknI6qXuPy+fofnL6/HletSzEn9wibNfebwj+SNl/cpr0b9OsWrJ5HSV/5OfkwnQCpXtuvSZaSNm9YfR8uSPoT3L4x2AYyvrVyRbTc+gjDHqYsxFyNaqnZg4ZmihUq9vkyeO9C/brlNe712B6ev53n5Wnp7YqW++TW636pdG+XK5lJx3Y7kW211CDgpVd79oKl9qq5ItDeeVf/D9yATosbJsDHVo0iUHJq5GDhQ8nGy9GkieWbN7ZbsT+fxg0B3oZssWefbH4WS3Wa1b6Sxgpco2qwBnVh7PVb5Ql+lizNnJ1pzaQXgJ4I+V9XOSv1g3qiyrvT5ditlKGRYnf/n3Iw/8t1cOjm+S40BmLJ+n4+jiqcdki8BFlQPUrGTyfEl5vDE50PanZCK5dMvPfRfjDiGTkBXK4znIFrUlK+W6kpxXph6v55fK+7NTefwDYGTlNV2FTKjHlrjr1yluyzPGtgBOBb5THm9PDr7doLuvacuYZFfoJWSivnL5Xzq23Oo26L7EmpYcR9iHTEaOI88CvJzs/t2VvITMVpXnrER2K87exZiLkD96bid/5A0or+3x5Onxq5fP9hKV52xKJu+D61n/z/bfEzv1zbfJ6cbnk565KV065cvtOSaekbJS+dulf/A2Yq9HDlSsHTgWJOcC2YocVDwH3fglVYmzKNlqtBZ5ML6aHG8yC9mN9lsy4VmzbFfXL9RShnVLzF+WL7cDyOSu9mt2E3KcxvTlcR+6MZ6J/MV4SflSXrC8no9SkoOyzQl8PslblPq19PQhE68Dy+P5ye6YhchurkupY3M9OfZj7crjn5fX95TyeAfyjKd6JNHV/5kfkYPCly+PjyMvJzBPebwFmfR2q4u27GtxKvPDkN1mt5JJ9Ell2cqlDL9orbxdrSvZSrh3qVu/8r/0O+DbZf1OVBL5Or2nG5BdpOuTScD1wNfIBOXk8j87Y51jjiR/9D1NZcwNOa7m2+RM28e0eE6Xzmglu2afKfvdhfwu3Lism4ZsqbycHCB/Wovn1n3qhc/23VM79s23ye3GxPlULqEM5iXPcPoHeTr5v4G56xzzV2RX0hqVZdsAv65jjGHkAf+blWVDyZarH5SD8DJk69Ip1HlwJJ9vOl+9fFn/tDw+mmyBOKaUsS6DFSsHq0XJX6pHkl1aazIx6Vu7HFRW64HP0qIl3oLkWUzLlOVHADeUg8oG1bLWIeZe5K/xr5OD1C8pB5bzKF2MLV+fbsT6XHcv2RrxDJmYLFNe87PJcVvP1g5mdX6NdyK70vqQrXc3kuOL+pBjbv5AJmX1SLg2IpOsg8mu4PvLgXntUs+9uhujlZiDyG6kRUp9HiQTr6vJFrxpqZzeXce4i5X38rMfJS3Wb0gmut2eOJVsmfxq7TNFToJ4EpUxYMBMZBf06PK39pmr21CDlrdaALOpmqRdyebzdSWNJr/UzoqIM8tpssOBv0XE092MU7uG0ErAOxHxhKTTyNaOxSLig3L9oJHkQNduX4ai1G14RHyvnIq7LDlAe37y7K0LgHMj4p1yyvGn1Ut0dDN2rb5zAe+WGEuRvx7/FxGHl+ntZyRfj7u6G7sSs29EjJc0D3mAfIHs/viUTEBeA/4eEVd2r5ZfiD9niTOeHCC+GdkV8lXgAzIZmiYiHqhn3BJ7D7JFaQBwfORU/lcAP468hEE9YsxFtkreTCaOf4iIdyVtS47n2ZA89XjdUpabI09jr6sy/cLHlcfHkpdu+XP5nK8FPB4Rr3Zh33OQXcP/ioi3yr6fjIg/lPWnk60hmyuv9P541Hk6gBJnFjKRu5DsZvoPOYbqNWCPiHit3jFL3NrFa7chL5z7rKRhwD/L/9Qo4JqIuLRO8Wr/szOQLWrzkpf9uKyyzTlkV/ht9YjZnn49HcBsMvExsJuk/ckL7p0DfF/SAHKQ6B31CFL+uTcm+7B3Kcu+K+l3wGuSjiF/cf2uHklP8RzwLeWFJ7clB34uQ/5yfJ88TfXLkg6JvDgn9Uh6avsp88j8EvhHmQ9pa/IMlJ0l/ZpsNn+9+pyuxqt8ga4HbC/pRXLswAHkWVxbkF+e27Z8Tldjtog/MPJ6ZueUmL8iBzL3IV/7P0TEo/WI1SJun4iYEBFnlMe112EEmeROqGO4WclTqOcnx9ysI+m/TJwB+zJyYsQL6hjzC1okPUuT3UJXlHUTyMSs0yQtSraSfQw8Jelq8gKg1WvEHUReRLZfvQ7+rYmIN8ocVjOQl0+Zn3ztD+mJpKfyOXpW0gnkGMDLJB1NJl57SXqfHH/3UL3ils+qSgL9B/KH3/qlPJeWRGwhskWz5/VUU5Jvvk0ON7Lv/Khyf2byi7PW0vk3sh99cB3jLUj+Sh5eHi/DxLFDR5FX6a5NvV6vM0Nqp+iPJQ/Ca5DNx8uSXUwL0WIQdx3rO1+Ju0p5/HvKGRrkGWXHU+cpAcizPmpniO1Ndkt8o7y/F5CtPQPrFOvLwJXl/mxkV+FIsitxW7KrZQvyjJUnqfN4jHbK1YeJZ6dt0wP7X4VMXrcnu4E2JCfY/B2ZZD1JHqzrdt2vduq5Vfmf2qwO+6sNtF2H/IFwcPmMbkIORq+dhbh6eW2/RA92uVTKdTLZ1fUk3ZzyoLLPuZg4nq7NOpTvjusp8/SQDSLd+v9pKx4Tv3tnJMfHLVZZN3N3Ynbm5q4um6pUfgnX/q5ONufuT/5D/408a+RB8gC2Z0T8sxvxFiETjNsj4l+S+gO/IM92mUBO4Pcu2Wx8uqSTyAPmApGz3taNpKGRs6zWHq9NjrHZNCLqfjVuSYuTTfPHk9dLerssvwK4PyJ+3rJMdYg5iDwQzxkRJ5dlK5EDfbcnk5O+kVexrlfM28huuk1L69ZqZHL5Z+C9iPidpH7kAPa76hW3A+WanzywPVbvrstyf0NywPS95CSTkF3E3yTnDLq1u/E6WKa1ASJidB26SUeSrbHLRMQ/SlfTFeTYsA3JwfC3ku/xgZFXYK+LWtdsi2V9oszAXFqiPoiIF+vxfkr6Pdm6/JXILvY29ylphsjWmM/K083YM0TEu22sa9lV/YXXpcc1KsPyzbdG3pg4ydlsZJ/5vuXxcuTkdZ+dktuNGLVTlD8h++hPIFsdRpItLZuQTcnf5vNzYxxPna8n1KJc/Uvsus98ysQz4lYlW1q+Qg7mrZ6JsytwQB1j1n6grU/Or3I4cFeL+p5L/VuWavOKfJnsfqieKn9Aqf8EYO8eqOvC5Bl5c1SXN/pWXvPzyUn05m2trFNaPUtdHiJbQrcs3wVDy7r5yVbMxetRnko91yC7vr/QyksdBme32N/sTLze4AXkCR0DquWplo/PX6i5+xcAzc/MveW7r9W6UYfpQrpVxt4K7JtvPXUjT3e9DlirPF6SHJdQne22Ll0S5LiD28l5ea4kW3t+CcxW1i9HNpnX/ayXNsrTn/y1+nfq0DVQ2e9slfuLkK0dXyuPtyK7eg4kZ0d+hDpcYJXPzyi7UDkA164KfRnZ9TKQnGTtQWCpHng9Ny9xfgg8AYxu8Tn7BbBunWLVDpKbkEnVoWTSPm8r2/YtfwdQpwtktlOudcku1L2ow+VUOlnPWqJd13qSl4J4mZzaYbpqrB54/TYlz/Jbs7X3sMX7OT3d+FFEdg3+is93IV3E56eRUCtxh1A587Qb8YeXz8pKLWO1ErNHXu+O3HyRUpviSV+4aN5LZAvAUZJ+QXaNPECeEgtARLxTj9gRcQN5gcsdImJz8ot0H/LCmz8nD9CHR8RfWyln3UXEJ+Svrf+LiKvrEbN045ws6fyy6Evkr/SdJQ2IPDNjP/Isp/mB70fE9d2JrbxQ4c8lrVAWHUkmXLX3cGvyzKILyHE3h0cdu7dKGfqSp1T/MSKOjYjFgI8lXQsQEXeTV4++qZt17V/2F5IWJich/DrZRdq//P1cuSK7CIaQA9jrObC5GkelXDeRY3tuj24MyO9iPT/tiXpGxGkl/gQmfqbq/jpKmpkcy7J1RNwqaTVJB0iaN0r3Tov388/Q9eNyZDfVj4H3JZ1cupy2B94hBzFPX15/VeIOJn8odkv5f1mRTGi/3NY2JeZMwLWaeEHSxuqtjMs33+p9I8ceHEuO55mTPNvlq2RXzAQyIaj71O/kP/vPyIHMj5Izoa5Htvxs0NuvS53qOpQcAPmL8nglclLEg6nTRVVbxFuMPDCdTH6JLkK2+BxAZcJHchD3LOV+3ZvOS/xvVB4vQ559c309YpITWH6fiV0r85Bz9GxCXg9qwbJ8XSqtlOQv9BsorZpdjD3J7qY6/q/0Wj0nUa49yXmI6jnRZO11rXUvHUeegfZ7skv8QnJyy/58vsWlbvUs/6/XklezH1iWnU/OlTOgst1g8mSELl3XrVLXweT4r2mYOFh69Rbbtqxrty/R0+XXp7cC++ZbPW/k2T03kV1PdwOHVdb1A3amzldZr+x/NnLQ9AeUWV7L8loTeq/1ZdehbrUvthXJJvQXgV+WZWuSk5H9lC7O7DqJ2JuTZxadUQ6ci5GnIX+vJ97LSl0XJE9t7kPOovs6pRuNHKz+ozoeoBYhx5z9kExABpDdhq9VDlhrkr/I5y6PZywH0i53TdDg7qbeqmfL+raxbu96xGjxun6NHI82O/kj7EjK9fDILtJzmHgZjsFkQlKP93Nxsnt9CHnpi1qLaO01vpSJs8jPQHYRd+tituRZjVeQE2puS7b6fptsvVq7xbYzkUlPXV7vLpe5N4P75ltXb+SYj9ov/T7kqeLTkb/grieTnWmpjE3p4fKsSCZcc9bK1NuvUR3rtjp5VeUNyATyBuCEsm49suWnroO1yXER95DJ1t/IyzXMXb7Y/0TOSN0TLU1fJVsGjyNbD6cnz2x6qhzIXmLi2LHutvbUfgFvTV5/6mSy5WV18lf4YeQEiQ9SZhov2+9BFy8qy+cv6rowOf5sXvJX+p2UQb6tlHFIKVOnZxLupXp+rtWlo8/p7nta9lObhXnVVtbVut2r9dyPOlxChky27iPHpP2RvBzENOSPhTNocYo6+WOiW9NclO+9e8gWpgvJxHXa8njf8n7PRGlpJ3/IrNfdunb7tertAvjmW2dvZFJzdTkw1a4GfhI5BfvVle2+RZ610eNJCNls/QdyPpkpPunh890dmzFxLqS+5CUyHqRcz6flwbIOsfuQY7TWLY+XI8/kqv2CXoI6XWS0RdwVSr2+THZTPkz+Mh9CnumzHDCizjE3Jk+fPoScw+VQMrlbhLxUws8oF1mtHMy7etXz3uxWa1g9KzG/Rl7j6wzyrKrWLs9Q94G25XtpH3IiyO2YeGmPBciup81b1LPbcyGRrTc3k4OLpyW7oi8mT9GfkWx9qZ7cUZfvqPLZ2ZNsnb0LmL8sn4380TBPi+175GrrnS53bxfAN986cyMnxVuAHGD7Z7KlZyA5l84NwH5lu13I680s1MCyrUgrv/Km1Fs5AK5DNs2/SOXinmSi+TD1P4W8dlA+l8r1zMjm9MfKQayuLT1MHKu1XvkcbUB2/6xD/nK+lC5eVb2dmCoHhj+SA18BliJ/nZ9AnZPJsv+Gdzf1Rj1LjNXIlpW5yC68P9KiO5bPt2bdTGlB7kody99al+D25JlUD5Bj4A6v1ZU6dn9X4i5WvhdvA4aVZQPJVtHDy+O6TJZaib0YOcnlKmQL+32UsXdky96l9OBFRrt781ldNsVQXgrid+RMyy+Tp6QuRZ5G/V/y19QWkq4i++23iYhnGlW+iLg3Iu5sVLwG+BJ5htbd5C/nv0paVtIaZOvBlhHxVL2CSZoO+KGkDcgDxUzK65BBXlTxPjIZqsulPipnYg0CiIgbI+JBsnVi74i4GfgnOZi5riJ9QCaUy5RLYTxCdhuOBPZUXteoLsrZNE+RB6R1yRaJAeT/0CPAD5TXkPsNcHpMnNRzO/I6YF26flIj69nizLp5yO7KRckWkAMj4qNyBhPlUhS1M6kuIc/Oe6MrMSMiymf2UEm7k8nkT4GtIuIosnV6PfJ760OozyVjStzNyC6mf5EJ3nGSZomI98jJReeXNC11PGOtvM67ACMjJ+x8iWzNm0/SumSid05EvF+vmHXX25mXb7515Eb2jd/BxCtez0o2784JXEM2oQ8hu5xmBWbq7TJP6Tfy1/JZwJfL40PKaz2aTHrqEaPapTYNOd7jwPJ4a/LL/CryzJtNeqCOG5C/WE8Gji7LTi8x1yDHbS1Tz7qSB+Xaa7ouOZNw7XM9jBwoWtcWprLvhnQ39UY9KzE3Iwfzbki2XN3LxFaQ7cr7XJuYciZyvqtuDbQlWwYfJ1s//lNex75kl/z65Piwuk4kWuIOJy8Xs2h5vET57N5LDi5+ljqfVVp5nQeU/431y/ftEeUzdAkTL/sx2Z7U0esF8M23Sd3IJuIJlGvYkF1dtzFxDMiXyW6v3+CEpyuv7+zkFZph4vW1vlQe/xS4tLLtdLQxC2wnY85Qe6/KF3ati2s+sgVitcp2a9LNWbZbxK59eY8gJ5bblPxFfmE5cKjcvwT4ep1f603JFoHfAheXZbuTXXs3kNMhdHvyx5b1pfHdar1Rz+XIU7aXL98ZV5GtwYuRY14erhyU+5AJUpcnnyyva1/ysjBrk13dY5iY7M1ATu65Vr1f37L/xcjBwnuSY7VuIsek3UgOyF+7B17frZh4rcHdgB+V+7UuvNqM+ZNt0hPhxMe3KeRWvkgfAJYmz/L5QVleG58xZ/nSm7W3yzol3cqX99fLgX5n8mybC8gk8izyV+UfmThosduDIssB4TRy3qV5yQG3z5Yv0pXJMRI79kBdZy6xZyyPVwd+VatXWXclE0/37XaC1yL+qmRSNx/ZzTMB+HtZN5BsOVimB9/rY8jWiNp4ntXILuKf1A5YU1I9yRalU8r96ciurdeYOGv66mRidznZUrlZ5bkCBtWpvruVGPdSpgMo/0s79mQCQHbR7k22hH+dnKH+u8B2PRRvh/Ia30VO7Lk5ORZuierr2lP1rWtdersAvvnW0RvZ3TUBOKg8rg1O/CqZEE0R/3ST260kO5uRF6LcuCybgeyrv4qcIfkndY75TXKg8g7kr+aVyPl5rgReKAfOul2tmRzrcXepz6lki8fy5GULlq5sdxqVU43rXOfFy+d0g3LwmKEcOO7ogVi91q3WiHqWes1DJubzVj7HfyJ/AA0uy2Yku5w+m+SyO98Tldd18ZJoDCJbDseQYwph4kSmG/XE56iVMtXmAxpR4tbldPFKXZcr7+VC5fVbrrzOPwLeJ1tJ6zp4usdfs94ugG++deZGzpHxJDCkPB5J/tKav7fLNiXdKPMNVR4PJVvUrga+WVk+R/nSu5NuztXDxAsn1r5Q/488fXxnJl5HaCD5K/aP5Azc3Z5bhexKu5P8hboW2X33nbJuX+Af5PiXtUt5uj2nSov4Ayp170OOM9m9Ev9f1HHm4ErchnY3Naqe5Jik5ykXxiW7ve8o92cvcc+kJD898LpuALxCdhXeSo4V2rbyut5Bi1PWe/JG/nBYnpySYPM67G+ayv/oRuW1Pq18XvYmB4vPUv6vLqRBCV5dX7PeLoBvvnX2Vg5SjwDfKV96S/R2maa0GzkO4IbK48vLl9vXyF9w32qx/Tl0Y5wNOfD8QirjDsgzjG4kB0XuyMQuij7k6evHdjVei9h/Bv5ZeVyb36U2HmEjsqXpkjodOOZg4oDazciu2VGUrhaya+lXZKvXlXRhUsAOlKHHu5t6o57lYHs7eW286vJrgGvK/dnJxOcc6jBHTos4i5LdwKuWx0eRrT1DySR9ASae1t2wFujynnY7LplUnlG+Y+cr/xO1sZTrlP+bb7TyvCmqtb0fZlOYyAt+9iUP1stGxGO9XaYpTUSsJ+keSZcCATwdEQeVU18HARuWU3V/L2kJsuvig26E7EO2puwl6WOyW+uFiNhG0lbkgbO/pGsi4vVyEcONJf0MeDfKt2sX67pFra4RsQ05GH5TYFFJH5EHyZ8BYyMviqnuxCMnq5tD0mnkaeOnka0hx5d455ItL9sAv42IJ7sRqy3/I5PJhcgzfAYDoyXdERGrkfPWdFdv1PMosmvrQgBJNwAnRcSmkm6RdGVEbC7pELKrdHwdYlJiLU62Jk1HJlVExMGSatcBXC8inq1t383PUKdEnr7+fHfilvpdTF5T7P6IeE3SW8CSkm6JiJslfQnYR9JVUZlWopF1rYvezrx8862rNybjCbIm5xuVWWrJ08VfabF+CHnQXLI87k8dLv1BNo/vTY6r+VOLdTuRg6prZ5OtSR1a8lrU9QHgCbKVcA5yssItye69per8Gp9Kdnv8srJsc/Issto4qtrYjLr9WqbB3Wq9UU9yPqdLyXEmv2yxbgxwXT3fy7LfNciW0DXJ1quDqZxBSiZka9Y7bqNu5FioG4Hdqu8VOfv9UUw8y3JpshVospiBuau3WuXMrImUCdw+LffvB56KiB0q6/tGTvDWN+r7q3kmcvKzNYGfRU4YWFs3Z0S8UocWl5Yxq3UdDXwYERtV1k8XdZgUsTKZ3fQR8YGk48lB1HsCz0fEBElbk2carQC83t3XVtIc5KnEL5TJ7PYF/g1cFhFXS/oJmcg+RXbxHRjdbHnpjXqWuNX38RZg4YiYszweFBHjyv2VIyfdrAtJy5KXorkjIv4iaWFy5vJbgDMi4s3KtnX97DaKpNold/aJiLcl9Y+IT8r/6/fJgdwflL9HRMTlvVjcbnPiY9akWhxI7iYPUJs1IO4s5GRyXyFPJ7+jATGrdb2XrOum5XG3D1aS+pQD/qbkmWr7RXbZ/Z5sMfsZ8FxJGGaLiNe6V6PP4p5Itl6dRk5MeDrZ6nMEsBc5sd7u5Flzv42IP3czXq/UsxK/zYS9XglsJVatrseRSePRwHkR8bGkBcmxabXWrk/rFbc3lBmsbyPPmL2mLKslP3OT/6tvAS9GxINTaoJX48THrIm1OJCMBXattsL0YNyZyZafr5DzjrzdgJh1r2v1ACBpLTLx2DkixlS2OQaYnzyoPNv6nrpVhlPJsTz3R8SPy7LNyYtlfi9yTNw05YDdpQPW5FDPSpweTdgrLVpzRMR/yrIDyTOnDgX+UVpDFybPLr23XrF7k6Q9yHm0ToqIsZVW383Ikw2+GxEf9Woh68TX6jKbipVB4C2vY/TZ48jBvP3K/eH1SHok9WnvcYn1Jjn4dc96JT2NrqukLwN7l+4AyAnzLgOekbSHpMslnR4RB5KDjQd2J16L2Cp/p4+IvchTjUdIWqC0VFxJnlV1RukOGw9dG4Tam/VsTYv3cWVg7tId1W2VpGdj4DJJR0s6JSKOIcemHUwOiu8bEU9PLUlPcQV5mv6eymtuTZC0Gjnx5WVTS9IDbvExmyqVg9UnEfGqpDXJg9XrwM0R8Y9Wtu8fEZ90M2Z1nMWC5Hwf/4yId9rYvk9EdPviib1R17Kf+cj5hv4EnEiOfzidnFflMvJ08p3IWcb/3d14lbgN7W7qjXpWWhs+10LVouXps5afOsT77LMoaQR5pt/25Pw8G5Knr0dpXZuNvEDne/WIPTmRNDs5num75MkAC5BdeVdO6d1bnxOTwQhr33zzrX43ci6OO8lrB61JXhl7L3Kuk2MoZ9tUtq/NgD0TeWDr9Cys5Pwmp5Cz5G5CntXzR7IlYk1aXOqiEnNGunEto16q65zAHuREbwuQ4zx+SM5QPJQyOSQTZ/BdqE7va/WCrmuR43dGtNjmGPKMp25NNtlb9SSnGpi93F+TbLXaHViwje27PWMwOUZq+0rcFciEcl3yNPXa/DiLVv9OzTdyLqRZmTjr9xQ1T8+kbu7qMpuKSFqEnIvjzMhm+PXJs6dOJU9N/Q85EVlt+9ov6yHkaapXRydbQ8pYh3PJ04z7k1PZ7xQR25GDbg8CFm4j5pVAl5rQe6muIhOCr5JJwT/JX8dfIU/V/xh4rXQV/An4cUQ805X6tYjb0O6m3qhneT8vAeYtLXfnAW+T40t2L91P1e37RjnzSNKJyjOTumI5MvHZUDnw/n0yKf4dmZQ/L2kd4FBJM0XPzLs0WYmIVyPi9Yj4V3k8dbT0FE58zKYSkhYlE4kZyCs1A7wBbCpp1oh4hTywrC1pXoAWicBPI+LGTsZchDyt978RcQ55QHydTICIiNpU90cq9anEvBQ4LLpw6nEv1XV24P8i4gHg18Aq5OzhL5FJwTpkkjAz8C45J8rVna1bG/qTl/j4VjnAX0smIX8nW1/+AMwkaa6I2D0iHu5qoN6oZ28ksTURcS056/Rm5GSLT5LJ+hvAKpK+SiZCF0fEf7tWQ5uceOZms6mAckbVS8izTmYHTlHOenwT8CVgO0nnk7POfgzUzorpT07gd2hE3NbJmAuRl6E4jpyR+WcRcaikfwErSnomIl4FrievqxRASBpUlh3Q2Zi9VddiNuD+0iowhmwZ2L+sO51MDs4hk5Tjog7jTyTNSR6QR5HdL6eXVb8luxSnjZz7aBly7M2A7sakwfUsSeyfy/5aJrFXlvpdAlwhad6IeLGVJLYrn6Na8rQR+Rq/Ts5Z82EpxzjgQDLp+klEXDNVjXNpZr3d1+abb751/0aOdVmn8vgw8iyNEcBWwPHk1cnvArZu8dw5uhCvH/mreMfyeB7yjJD9yQnzLiev+fNrcvDr1yrPXQEYPgXVdVbK2JbyuDZnzvSlLheQ3T/TkKeV12VmZCZeCfuqyv4XJMfa/Ji8tEhfcizKM5RrZHUjXsPrSSaqD5MtLXsDfyHnHFqSnDdnH3I81gLkNbrmKs/rT847s3YXYs5SuT8b2WpWu/bWNuTYtD2q70P1r29T/s1ndZlNRVqcnXIYOaPu0eTBZW7g04h4qYzhILrxBSBpYES8Vzu7pnQp3QP8nBwfsSE5tufeiLi9cqpwXX41N6quktYgryD/L+CXwGrkuJN/ksnB4uQg3NuAE+tUt9mBDSLivDLe5dvkQNvTySTzNOA64HxgXnL25q60YlVj9kY9FyEvU3JzeXwYeSmRo8h6rUZecDXI1qU/VZ772Tw7HYwlcnjHA8BtEbF3WX4x8NeIGFUe71/qeSBwYdRxUkSbPDjxMZvKtEgIDibHRvwkenDOkUq3wbzkL/OzI+KwFtvUvZugEXUtXWRLk0nBv8kWpRXJrqfnyCRvCeDjiBhbp5hLkXPvvEZ2NS1OtqbVkp+5ye6mq6lft1rD61mJ3eNJrKTBkZdj+BKZvF0REQdI+jZ5NtmVETFGebHOY8jLezxelwraZMWJj9lUqJUDybVRmWW3h2LWkp9h5AF6Dcostz0ct+51Vc5d81aUyRWVE+bdRZ4xdTvZIrI88E3y1P1fR33mJJoVGEleXuJd5ZXPXybHUS0J7FfKcQbZIjJTd5K83qpnG2XpsSRW0kCyC2uvkkDNSZ6F+HuyhfJoSoJFnp7/vYi4qa392ZTNiY/ZVEp1miCwkzFr3V51vW5SB+LWta6S1idP056pdM/9mWz1uIi8zthb5CUhViTPaHusTnEb2t3UW/Vspzw9lrCX1qJFgKUi4tKS/IwBTomIX0pakqznExFxVz1i2uTJiY/ZFKYyVmY6stW/zXlwWhxIupwctNZNpXau3K6Jswt3q3urN+pa2d9GZLLxDHB3RBxelq9HDoL9D3BkPbvveqlbreH1nER5eiRhL2chnkeOO9snIi4oyc9dwOUR8f16x7TJkxMfsymQ8iKU25IT1Z1BDs6c0GKbWtfTNBHxcTdi1ZKP9YC5gIERcXob27Z6qYHuaGRdW4m9Hnnqff/awOyyal3g5Yh4og4xer27qUH17NWEXXk9rw3JuYd+ABwVEX8oY37uJ+v6VKNbSa3xPIGh2RRGeeHAH5ODXT8sf6dpsU11grcrlDP/dkk5WG1Cnpr+BnCgpENbKVct5ozAPpKm7WrMyj4bWteWIic5/BrwtKRZYqIb65EMFAsAL1aSjT+RXVk/IS/p8UPywDwKuKYnDsyNqGf5HG0OnAVcImlTtXIB2/J+TpA0TXlel+tbYq4k6duRF6WdDXgT2Bz4kaQ9I+JlYN6IeMJJT3Nw4mM2hagcGOcnB7uuSg7I/FZEfChpaNmuX0kEBpOXNTg6ytTzXYjZp3S/bAt8nZy/59/kwau6XS3mEPI6Wfe194u+A3EbXte2RM7suzfwmCZeMqKe+/87OZ7mWUnXAw9FxPcj4j5yhua5ySTojp4cY9PT9eyNJLZ8dncBjpW0PdmCdiw5ueUuwE9KjB4dgG+TF3d1mU3mKl0EC0XEM5K2JC8ZMBDYJfJaQtuQk/ftFhHvlwPX5eQsxbd3I2Ztrp5TyZlsVwC+ExFPSdoaeL8cMCkHqz+R40G6NKdMb9S1E2XbFHgvIkb30P57vLupg+Woaz0r7+lOZMLzKdmKtUNEvCBpaES8pYkD4weT7+cRXf0clbhzl3gzkJfDeLr8PYTs7tqI7Lb9X3fqZ1Met/iYTebKQWMj4DJJ8wOjgQnkRGz9ygHzMOC8kgj0JccwHNnNpGc98rIQAGPJg9URJelZCfgpedYPkqYnJ9P7WXcOVo2uayfLdk1EjK4kJPXefyO61TpSjrrUs/L8BcvfccBuZEvPjiXp2QY4TdKAkvTMRF6+4vBuJj3TA3uS3bNfKnFfBp4ik6rFgGFOepqTW3zMJnPKyewuBXavHQwkzQH8gryswazA6RFxTVnXh/wl+243Yq5HXg/q21HmM5H0I7L15S/A2mQLy9Vl3bzkdaOe7mrMsp+G13VyI2lD8mr3i8YUflHMksT+ijwl/79kcvwMcCo5A/WJwEGR18HqCxwJ/L0erU2l5WgF8oy1K8hLX5wQEU9ImiciXupuDJsyOfExm8wp53bZPiK+Ww4O/SLiI008ZbzWVSDyf7rLAzTLPvqSE7rdGxGXSOof5crXpaXnU2BCRDxY+1UfdfoiaWRdJ2c93a3WCJNLEitpYfKq8tsDL0bEimpnKgab+vnq7GaTmUpXU+003v8Ca0tauxwIx0vamDwb6BTytOda8tGtBKTs41NJHwErS7oqykSEklYgx5r8s8X2XdabdZ2cVZKBKflq4EOAmyLitkoS+x9J32onia17y11EPC3pJ+RFVqcvy5z0NDGP8TGbzJREYG3giNLl9BhwArCTpF1Lq8gvyAntunW6L0wciyFpwfIrHfJsomnIy04gaRny7Kq6nu3T6LpOaaakpKfyOaodV6pJ7PjScrcx2foClSS2p9/XiHg/Iu6LiFt7aoyWTTnc1WU2mZG0OnkhymuBr5IH/keBYcD3yNPJL42Iq+rVIqCcp+d04AmAiNhIefHGNYA5yYTnZxFxRXdjtYjb8LpazylJ7LrALcBNwLeAlcnTyP8BnAQcHOVMQLPe4MTHbDIiaUFywOfpZcDnRuQv5MvJAa8AfcoZMPVKehYmT/E9MSIekHQD8ElEbKqcRXg48L+I+Ec9k4/eqKv1HCexNqXwGB+zyUCl+X0RYDCwnaS/R8R1ksaTE9hNC/whIj6Fuoyv6UOOwziSnCSvNlB5A0l/lXRvRKxIXsiResQscRteV+tZJYk9iIlnaN1MJrH9yST2apzE2mTCY3zMelElCRhaxjpcQx5AxgH7K6899Tfyek331xKBesSMiAkRUbv69kvAasq5c4iIjYF3JK3S3Xgt49LAulrPUsHnk9hpI+I64GRyduTdKUkPOIm13ueuLrNeVk5d/gnZsvJaRBwlaR1yduI3yYsp1uXCm5WzqNYD1iMnBryJbPH5IXA3cF1E/KMe8VqJ37C6Ws+pfI5mjog3y7LVgP8DXgSOj4iPJW1AXoB1THv7M2skt/iY9aIyGPQY4NvkgX8/Sb+JiJuBK4E5yKSkLsrBalPyDK1nySt+H0uegXM0sA7wVUnT1/vsl0bX1XpO5XN0laTfSDo4Iu4ALgG+DBxcWvBucNJjkxsnPmYNpHLF6XJ/CHnK+DfIg8UmwAbkKcDHl26fH0fEs92Mqcr9WYHNgC3JKfznIq+4fiSZjBwJjI6ID+owhqjhdbXGcBJrUzJ3dZk1iPJK0WuQA3c/BpYlB36+C5wNnBURN0g6iUxONoluXqNJec2i+SLicUnDKXOnkINOLyC7mOYAziG7vb4VZcLCbsZteF2t55TWm4/L/SHAisC/yMtOHAF8BxgF3BgR31eZnLB3SmvWPp/VZdZYHwCHkxdu3DoiXivJydvAbGWCt1mAr9RpnM1cwDeUF3/8OrBemcl2BPCviPinpFnIsT2/qkfSU9HouloPKEns6pJaS2IPAw6LvHzJLcCWkn7vJNYmZ+7qMmuAMhj0E/Lq0EPIq50PA4iID8gJ3tYlr3j+p3olAmU/r5JXqj4vJl5E9EFgaUlXk10Tl0bEk/WI2Vt1tR71AXkG3rnAHRHxWlneWhLrpMcma+7qMuthlTNgNimL7gBWIruZHomIUyQNIFtGXoiId7o710kl5mLAQGBJcgbde8irX/9T0iBgFeD1iBjb9Rq2GrdhdbWeU3k/hwI3k91b50fERWX9TuSA+KWAX0bE5b1XWrOOcYuPWQ8rB47Nyflp+kXE22RC8DdgKUlnA6OB8RHxTu05XY1XOVhtTs6kq4gYBVwPfIXsttiFHJx6R72SHmh8Xa3ntEhiVwbWBH5DDkjfu2x2GTn79noRcXm9zwQ06wlu8THrAZLmAL4cEWNKy8olwP7kxTZHAEsDVwHzAlsDN5cJ/boTcwB5qYlPJC0JXAhsFRHPlDE+nwKLkXOtrAIcExF/6k7MErfhdbXGKEnsT4FDIy81MRDYmEygpwGWAHaNiMd6sZhmneLBzWZ1VgaDrgM8KGlARIyT1BfYgzzF9y3yVO4vRcTh5MBiutPlI2lm8tTi08k5eQaTY3uGSjqU/LW+HLAC8ANgcES8UYcutYbX1XpOK0nst8kpCJ5TzuJdS2L/RSaxRzrpsSmNW3zMekBJCIaQ3UknkYNDtyHnyLld0qrAfuQEguPqkQRImo9s1VmIvDr2ieRpx6cAfyEvFPlUbXxGvfRGXa3+yvu4NTnw/aWIeF/S9eSFRqtJ7Hklia09z0msTVGc+JjVSWl1mZO82Ofr5Jw5PyDPaDojIu4r221KJgkHRcRfuhmzLzmW5qPy+Adkq85pEXGrpH6RF4ZciZyrZ9eIuKs7MUuchtfVep6TWGsGTnzM6qCcPXU28AowFBgE7BsRd5RkZFFyzM2t5e+5kVex7k731qJksjEv8AJwW0ScJ2kvcmzNBeRZXPMDfwL2r0fy0Rt1tZ7jJNaajRMfs24qCcgfyNmI/1AOJFuRv5i/QiYf3yFbYo4DHo+Ij7qZ9CxCDiI+k+yKWIYcsPx8RBwkaX+yy+sq4C5gjoh4qg5jehpeV+s5TmKtGXlws1k3lOn7bwN+WxIBkVejPqOc2fs7crK+K4C+wKe1bqluJD3zAjeQ17a6sCy7FbiXvGbSHhFxgqSfkWM27o2Ip7oTs7fqaj2nJLFnAr9vkcTeKOkrwMlkEvtNshVoZyexNjXwPD5m3RAR/wOOB9aRtHQ5IPQtScE5wIvAghHxEpkwPFKHsMsATwKfVMoxnkx8rgNWK8sOBY6OOl0zqZfqaj2gksTe1DKJBfYlk9ihZBL7AE5ibSrixMesmyLil8ClwDklIfgU6FsOFO+SrR9ExPt1Cnkz+Ut9PUl7VsrxKfAQOVHg0LKsrpeD6IW6Wg9wEmvNzImPWR1ExG/Iq1OfI2nZcibVGuSlGf5T51jvki07NwPDq8kPeb2kp4EeSzwaWVfrOU5irVl5jI9ZJ0nqW7qWPiciflPGupwi6SzgW8ARtfE13Yz5uXEVEfGupGvLw3UkbUOe2XUcedZNXa6y3ht1tcapvI/nSNot8irrTmJtqubEx6yDJM0bES9GxPja/DgttykHkmmA3wObR8TV3Tx7a1BEjIuIaCf5mQBsR86g/H8RcW0dzt5qeF2tZzmJNUs+nd2sg8pBYc2IWLA8bjUhKOvmjYgXu5v0AGcA10TEBWXZF/YnaQby+kkvR8TtXYnVSuyG1tV6Tu39Kffbex9/SM7T4yTWpmpOfMw6QdIfgQUiYoXy+AsHkuova0l9ImJCF2MNJE8v/jpwfpQLiraR/HxWjrZ+2XchfsPqaj3HSazZ53lws1kHKC8NQURsB/xT0pjy+FNJ/arble6hIZJW6U4iEBHvkacTXwTsKmnrsjzK2TfVmJ9Kmk5S/+4mPb1RV+s5EbEbMEbSfeXx597HmvJ+vlh72MgymjWSEx+zdlQSjNklzQ8QEV8HXmiZEFQSgcHANXRxDF01qYmIdyPiYuB8Wkl+qskHmSDN3cWq9kpdrWc5iTX7Iic+Zu0oCcZmwNXACaXbgIjYGviHpCfK408ricBlwIERcVtn49W6GCRtIul3kn4paanIK6pfAOwkabuyeZ9KzMuBEyLiuSmlrtZznMSatc2Jj1kragcO5am9RwGbAX8BdpZ0Lnz2K/oJSauVbQcBNwJHdnWQcS3pAX5GJjrzARdJWiny8hSXAyMlzVlp6fkLcFhE3Dol1dV6jpNYs7Z5cLNZhaTposyBo7x20Yzk1P2zAYcAu5Djbp4nz36Jsq2ARYDpI+LBbsSfBjiM7NpahLxK9l+BnYGdImJMSXpeKd0YPwJu72LrUq/W1eqv0mK4BnAqsBGwKXA6cGFE7Fy2uxw4LvJipIOA0cD+TnqsGTjxMSvKAX1zcvK2e4DdgAPICzSeCtwdEaOUVz7fC9g06jg5oaRFIq+gPhMwgJxV95sR8YSke4DZgaUj4p3Kc2eInMm50zHphbpaz3ASa9Zx7ss1K/6/vXuNlasqwzj+f4tQiFgoiIpYCBdrgVBNiGCxTRAQKhdDQBBENAokFEiEyiVasBUNapGLiRaoITEgyC2ALTSAVG2AQKAhCCoIlSpFpLQgkkovpH388K6BEeFDz+yZPTP7+X3pycw5Z62VttnP7P2u9ZbwsQBYDMwADpT0MkBEPAVMKgWh+wPHVhx6pgKXRMRxkp4od36eBFZFxH5ke4rr20NPmfNGh57ycz1fq3VHCS9TI+LtIfY5MsT+XNKS8rjrdGA88Bd4s+HoU7VM3KwmDj5m/+8PwGryNORHy2sPkDVxxwOXS3r0XX52o5QAsgdwOXCy3moG+SqwOVnrc3B57/Eqxnybnq3VusMh1mzj+FGXNV7bXZdxwGpJKyNiC/LRwN8knRoROwLbAY9LeqP1MyMcbydgb+B2SRsiYh9gmqSvlU/vm0laWy5WHwC2kvTkIK7VeqPcIbyavJuzSNK55fVPAlPIOp/LJc2vb5Zm/cG7uqzxShA4HLgbuD8iLpK0GjgF2D0i7gbmkXUQb7R+poMhx5INRbcuF6wXgb0j4lCltWVn1zckvVBV6IFa1mpd0rYbbxwwRtKJ5F2diRFxZfm25cAiYKpKG4paJmvWR3zHxxovIiYAs4FzgNfIux93SZpV7oacBjwy0u3ibeOMIv/PrS+B515ye/pc4Fjyk/nj5OOny4DzJd3ZyZjvMIeerNV6o4TY2eSH2FslfbsEoV8Ca8iC+DN85IDZWxx8rNEi4oPATPLR01GS/lEeRV0HPCjpnIrG2QL4FLlt+DAydGwoY18H3E+e2TMdeB6YL2lelY+ZerVW6w2HWLORcfCxxnl7mCiH8p0OPAzcXALBzsBNwAmSnq5gzE2A7wP7kG0lpklaWM5buZBsQnp1eRQRpfan49BTx1qt+xxizUbOwccapa249xDg0+TOqR8CE8ku6EuA2yQtaz8bpcMxR5UgMwG4Afg7+WhrQykengJcAlwDzFFFfZLqWKt1j0OsWTVc3GyNUoLA/sDFwDNkofE9wNNkKJkIHBMRo4F1nY5XLlYbImIyecryUcAr5Pb1j5Q53QecDCyuKvSU39vTtVr3tIfYiLgwImaT5zzNBXYBjo6IcZKWAlMceszenYOPDb132MkyGbhR0rWSTiFbQswjD3+bT9ZJrK0ihJSL1WTyLJ7blU1ETwPeB5wVEadExFJguaSHOh2vzrVa9zjEmlXHwccaIyImRsT2wEpgq9brki4gLybbSvq1pD9XNF4rhBwCnAVsWcZbTd7heZWs0ZguaXkVY7aN3dO1WvUcYs26wyc321Brq6/5LHl2zQxyq+/8cqrtAmAn8hPzlsBLFQ6/A/C8pAsiYiVwbkTcJ2mZpDURMRMYXb6uopC5zrVal0TERGAFGWJ3ab1e/l3tRgmxdc3PbNA4+NhQahXrliBwBFlfMxPYshTzTiODwRRgL+C88hiqkzE/DOwq6b6I+BzwrYh4gDwM8MxyB+bmiDhe0tISdNZAZ4cE1rFW6y6HWLPu8a4uGzoRsQ1wKnAFeVbOIvJR0zLgemB/Sa9HxA7AKmAbSUs7uetStqufSV6EbiDPVDkcOAP4ONndfE1EXEzW++xb0Y6xnq/Vuif+t8t6K8TeA2wuaUZETCJD0AoyxM6SdEdtEzYbQA4+NnTKybWjgDfIo/yfanv9SkmHlYLjA4GLVFozVDDuEeRZPd8DNgNeIA+YO07SsxGxl7Lz+viqdt3UtVarnkOsWW+4uNmGjqRl5cujgBnlwg95IvJLEfEFYA7wWJVBQNkA8i7gJDLwXA0cXELPQWUuY6vcalzXWq0r3kseQLgFsL2kT0j6HbAWWFFCz2SyMP71snXdvdTMNpJrfGzoRMQBwAnAL8g7LydGxCaSFpVD3w4Hjpd0b4ePt3YE9gSWSHqmvHwH2R/pfmB34NCIWEFuQ54p6V+drO0d5tCTtVr3lXqsncgQu29EXKXssdUeYr8DXOAQazZyDj42VCJiPPAl8jHPIxGxHAjguIh4GTgPWC/pXuj40/J25N2dbSPiJuDaUth8OrlN/VLgfOA58mI1v8rw0eO1Wpc5xJr1hmt8bChEdj4fDfyE7If1A0k3lvd2A75INgE9W9KrVV04ImJzYD+yrucJ4K/ALcCPyYMKXybDhyocs5a1WveUEHsucFUJseOBI4DdgJ8BHyX/Hc2rcZpmQ8HBxwZa66IeEaMlrY2IscAs4HXgOkl/LN/3MWBdqy6iC/P4ELAHcDbw/vL1dElzKxyjL9Zq1XGINes9Bx8bWG1B4BBgGrCcbNh4LdkLayVwq6THejyvI4EjycaRd1b0O/tyrTYyDrFm9XHwsYHT/qm3nGsylyz6XFW+ngP8FLiSrK/5kaRVPZjXKJV2AaU2Y32nn9D7da02cg6xZvVycbMNlIjYDjgpIq6Q9G9gW2CBpNvK+/sAvyX7GM0iTy/uSRBQW48kSevLn52Enr5dq228VuApoWcSWQfWHmK3Ab5JhtijI2KJ/z7NqudzfGzQTCD7FU2PiDHAauAzrTclrQAWAlsr20I8Uc80K9GktQ61EmLPi4hWw9g3Q6yk35D1PV8BdiVD7E0OPWbd4eBjg+Yh4CpgDHCqpIXA4ohYHBF7loMCDwLW1TnJijRprcPOIdasT7jGx/peROwMvFIe9xAR7wEeBF4DFkq6KCK+SzZtHAdcWlVRca81aa1NEhGbkg1Fvwz8U9LsiJhD3un5KrA9Wd/zdUkP1zZRswZw8LG+V+5s3AKMLfURtwPPAr8iD/B7EbhM0rqIGCPptUHd9tuktQ47h1iz/uTgYwMhIqaSO5ieAR6SNLO8fiBwDLkTZhawob3IeBA1aa3DzCHWrD85+NjAKBf+u4FNWychl7cOAF6Q9GR9s6tWk9Y6zBxizfqPg48NlIg4lDzldpKklXXPp5uatNZh5hBr1l98jo8NFEkLImI98KeImKCKu533kyatdZhJWhgRnweejoj2ELuwznmZNZXv+NhAiojDgP9I+n3dc+m2Jq11mJWTmq8BHGLNauTgYwOtScWgTVrrsHKINaufg4+ZWY85xJrVx8HHzMzMGsMtK8zMzKwxHHzMzMysMRx8zMzMrDEcfMzMzKwxHHzMzMysMRx8zMzMrDH+C6aPwc98Z/dSAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,1,figsize=(8,8))\n",
    "im = ax.imshow(df_clean.corr(),cmap=\"Reds\")\n",
    "ax.set_xticks(np.arange(len(df_clean.keys()))); ax.set_xticklabels(df_clean.keys())\n",
    "ax.set_yticks(np.arange(len(df_clean.keys()))); ax.set_yticklabels(df_clean.keys())\n",
    "plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\", rotation_mode=\"anchor\")\n",
    "ax.set_title(\"Correlation matrix\")\n",
    "fig.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "This correlation matrix shows us that the hillshade variables and the aspect show some correlation and that the elevation is somehow linked to the soil type. We can also see, that the wilderness area has the strongest correlation with our target variable."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "There are many more analyses that could be applied to the data before conducting the machine learning part. However, here we want to focus on the application of the ML methods. So, on this \"clean\" data set, we can train and test the model again:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "feature_names_clean = ['Elevation','Aspect','Slope','Horizontal_Distance_To_Hydrology','Vertical_Distance_To_Hydrology','Horizontal_Distance_To_Roadways','Hillshade_9am','Hillshade_Noon','Hillshade_3pm','Horizontal_Distance_To_Fire_Points','wilderness_area','soil_type']\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_clean[feature_names_clean], df_clean[target_names], test_size=0.2, random_state=42)\n",
    "y_train = y_train.values.ravel() # convert to 1d array\n",
    "y_test  = y_test.values.ravel()  # convert to 1d array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9631076650344655\n",
      "0.9032239907339351\n",
      "CPU times: user 4min 3s, sys: 532 ms, total: 4min 3s\n",
      "Wall time: 23.3 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "model = RandomForestClassifier(random_state=42, n_jobs= -1, oob_score=True, class_weight='balanced')\n",
    "model.fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "print(accuracy_score(y_test, y_pred))\n",
    "print(balanced_accuracy_score(y_test, y_pred, adjusted=True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "Which gives us a slight accuracy gain by ~0.01."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "In summary, we learned:\n",
    "\n",
    "- Look into your data\n",
    "- Clean up your data\n",
    "- Split your data into training and testing\n",
    "- Try different models\n",
    "\n",
    "In the next session, we will take a look at how we can apply these techniques in remote sensing tasks."
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}