lab4/knn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Цель лабораторной работы\n",
    "Изучить сложные способы подготовки выборки и подбора гиперпараметров на примере метода ближайших соседей<cite data-cite=\"ue:lab4\"></cite>."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Задание\n",
    "Требуется выполнить следующие действия<cite data-cite=\"ue:lab4\"></cite>:\n",
    "\n",
    "1. Выбрать набор данных (датасет) для&nbsp;решения задачи классификации или регресии.\n",
    "2. В&nbsp;случае необходимости проведите удаление или заполнение пропусков и кодирование категориальных признаков.\n",
    "3. С&nbsp;использованием метода `train_test_split` разделите выборку на&nbsp;обучающую и тестовую.\n",
    "4. Обучите модель ближайших соседей для&nbsp;произвольно заданного гиперпараметра $K$. Оцените качество модели с&nbsp;помощью трех подходящих для&nbsp;задачи метрик.\n",
    "5. Постройте модель и оцените качество модели с&nbsp;использованием кросс-валидации. Проведите эксперименты с&nbsp;тремя различными стратегиями кросс-валидации.\n",
    "6. Произведите подбор гиперпараметра $K$ с&nbsp;использованием `GridSearchCV` и кросс-валидации.\n",
    "7. Повторите пункт&nbsp;4 для&nbsp;найденного оптимального значения гиперпараметра $K$. Сравните качество полученной модели с&nbsp;качеством модели, полученной в&nbsp;пункте&nbsp;4.\n",
    "8. Постройте кривые обучения и валидации."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Ход выполнения работы"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Подключим все необходимые библиотеки и настроим отображение графиков<cite data-cite=\"doc:ipython,doc:seaborn\"></cite>:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from datetime import datetime\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.metrics import median_absolute_error, r2_score\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.model_selection import KFold, RepeatedKFold, ShuffleSplit\n",
    "from sklearn.model_selection import cross_val_score, train_test_split\n",
    "from sklearn.model_selection import learning_curve, validation_curve\n",
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "# Enable inline plots\n",
    "%matplotlib inline\n",
    "\n",
    "# Set plots formats to save high resolution PNG\n",
    "from IPython.display import set_matplotlib_formats\n",
    "set_matplotlib_formats(\"retina\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Зададим ширину текстового представления данных, чтобы в&nbsp;дальнейшем текст в&nbsp;отчёте влезал на&nbsp;А4<cite data-cite=\"doc:pandas\"></cite>:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option(\"display.width\", 70)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Предварительная подготовка данных"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В&nbsp;качестве набора данных используются метрологические данные с&nbsp;метеостанции HI-SEAS (Hawaii Space Exploration Analog and Simulation) за&nbsp;четыре месяца (с&nbsp;сентября по&nbsp;декабрь 2016&nbsp;года)<cite data-cite=\"kg:srp\"></cite>:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv(\"./SolarPrediction.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Преобразуем временные колонки в соответствующий временной формат:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data[\"Time\"] = (pd\n",
    "                    .to_datetime(data[\"UNIXTime\"], unit=\"s\", utc=True)\n",
    "                    .dt.tz_convert(\"Pacific/Honolulu\")).dt.time\n",
    "\n",
    "data[\"TimeSunRise\"] = (pd\n",
    "                       .to_datetime(data[\"TimeSunRise\"],\n",
    "                                    infer_datetime_format=True)\n",
    "                       .dt.time)\n",
    "\n",
    "data[\"TimeSunSet\"] = (pd\n",
    "                      .to_datetime(data[\"TimeSunSet\"],\n",
    "                                   infer_datetime_format=True)\n",
    "                      .dt.time)\n",
    "\n",
    "data = data.rename({\"WindDirection(Degrees)\": \"WindDirection\"},\n",
    "                   axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим полученные типы:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "UNIXTime           int64\n",
       "Data              object\n",
       "Time              object\n",
       "Radiation        float64\n",
       "Temperature        int64\n",
       "Pressure         float64\n",
       "Humidity           int64\n",
       "WindDirection    float64\n",
       "Speed            float64\n",
       "TimeSunRise       object\n",
       "TimeSunSet        object\n",
       "dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Посмотрим на данные в&nbsp;данном наборе данных:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "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>UNIXTime</th>\n",
       "      <th>Data</th>\n",
       "      <th>Time</th>\n",
       "      <th>Radiation</th>\n",
       "      <th>Temperature</th>\n",
       "      <th>Pressure</th>\n",
       "      <th>Humidity</th>\n",
       "      <th>WindDirection</th>\n",
       "      <th>Speed</th>\n",
       "      <th>TimeSunRise</th>\n",
       "      <th>TimeSunSet</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1475229326</td>\n",
       "      <td>9/29/2016 12:00:00 AM</td>\n",
       "      <td>23:55:26</td>\n",
       "      <td>1.21</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>59</td>\n",
       "      <td>177.39</td>\n",
       "      <td>5.62</td>\n",
       "      <td>06:13:00</td>\n",
       "      <td>18:13:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1475229023</td>\n",
       "      <td>9/29/2016 12:00:00 AM</td>\n",
       "      <td>23:50:23</td>\n",
       "      <td>1.21</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>58</td>\n",
       "      <td>176.78</td>\n",
       "      <td>3.37</td>\n",
       "      <td>06:13:00</td>\n",
       "      <td>18:13:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1475228726</td>\n",
       "      <td>9/29/2016 12:00:00 AM</td>\n",
       "      <td>23:45:26</td>\n",
       "      <td>1.23</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>57</td>\n",
       "      <td>158.75</td>\n",
       "      <td>3.37</td>\n",
       "      <td>06:13:00</td>\n",
       "      <td>18:13:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1475228421</td>\n",
       "      <td>9/29/2016 12:00:00 AM</td>\n",
       "      <td>23:40:21</td>\n",
       "      <td>1.21</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>60</td>\n",
       "      <td>137.71</td>\n",
       "      <td>3.37</td>\n",
       "      <td>06:13:00</td>\n",
       "      <td>18:13:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1475228124</td>\n",
       "      <td>9/29/2016 12:00:00 AM</td>\n",
       "      <td>23:35:24</td>\n",
       "      <td>1.17</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>62</td>\n",
       "      <td>104.95</td>\n",
       "      <td>5.62</td>\n",
       "      <td>06:13:00</td>\n",
       "      <td>18:13:00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     UNIXTime                   Data      Time  Radiation  \\\n",
       "0  1475229326  9/29/2016 12:00:00 AM  23:55:26       1.21   \n",
       "1  1475229023  9/29/2016 12:00:00 AM  23:50:23       1.21   \n",
       "2  1475228726  9/29/2016 12:00:00 AM  23:45:26       1.23   \n",
       "3  1475228421  9/29/2016 12:00:00 AM  23:40:21       1.21   \n",
       "4  1475228124  9/29/2016 12:00:00 AM  23:35:24       1.17   \n",
       "\n",
       "   Temperature  Pressure  Humidity  WindDirection  Speed  \\\n",
       "0           48     30.46        59         177.39   5.62   \n",
       "1           48     30.46        58         176.78   3.37   \n",
       "2           48     30.46        57         158.75   3.37   \n",
       "3           48     30.46        60         137.71   3.37   \n",
       "4           48     30.46        62         104.95   5.62   \n",
       "\n",
       "  TimeSunRise TimeSunSet  \n",
       "0    06:13:00   18:13:00  \n",
       "1    06:13:00   18:13:00  \n",
       "2    06:13:00   18:13:00  \n",
       "3    06:13:00   18:13:00  \n",
       "4    06:13:00   18:13:00  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Очевидно, что все эти временные характеристики в&nbsp;таком виде нам не&nbsp;особо интересны. Преобразуем все нечисловые столбцы в&nbsp;числовые. В&nbsp;целом колонка `UNIXTime` нам не&nbsp;интересна, дата скорее интереснее в&nbsp;виде дня в&nbsp;году. Время измерения может быть интересно в&nbsp;двух видах: просто секунды с&nbsp;полуночи, и время, нормализованное относительно рассвета и заката. Для&nbsp;преобразования времени в&nbsp;секунды используем следующий метод<cite data-cite=\"so:dttseconds\"></cite>:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def time_to_second(t):\n",
    "    return ((datetime.combine(datetime.min, t) - datetime.min)\n",
    "            .total_seconds())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Radiation</th>\n",
       "      <th>Temperature</th>\n",
       "      <th>Pressure</th>\n",
       "      <th>Humidity</th>\n",
       "      <th>WindDirection</th>\n",
       "      <th>Speed</th>\n",
       "      <th>DayPart</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.21</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>59</td>\n",
       "      <td>177.39</td>\n",
       "      <td>5.62</td>\n",
       "      <td>1.475602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.21</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>58</td>\n",
       "      <td>176.78</td>\n",
       "      <td>3.37</td>\n",
       "      <td>1.468588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.23</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>57</td>\n",
       "      <td>158.75</td>\n",
       "      <td>3.37</td>\n",
       "      <td>1.461713</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.21</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>60</td>\n",
       "      <td>137.71</td>\n",
       "      <td>3.37</td>\n",
       "      <td>1.454653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.17</td>\n",
       "      <td>48</td>\n",
       "      <td>30.46</td>\n",
       "      <td>62</td>\n",
       "      <td>104.95</td>\n",
       "      <td>5.62</td>\n",
       "      <td>1.447778</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Radiation  Temperature  Pressure  Humidity  WindDirection  Speed  \\\n",
       "0       1.21           48     30.46        59         177.39   5.62   \n",
       "1       1.21           48     30.46        58         176.78   3.37   \n",
       "2       1.23           48     30.46        57         158.75   3.37   \n",
       "3       1.21           48     30.46        60         137.71   3.37   \n",
       "4       1.17           48     30.46        62         104.95   5.62   \n",
       "\n",
       "    DayPart  \n",
       "0  1.475602  \n",
       "1  1.468588  \n",
       "2  1.461713  \n",
       "3  1.454653  \n",
       "4  1.447778  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = data.copy()\n",
    "\n",
    "timeInSeconds = df[\"Time\"].map(time_to_second)\n",
    "\n",
    "sunrise = df[\"TimeSunRise\"].map(time_to_second)\n",
    "sunset = df[\"TimeSunSet\"].map(time_to_second)\n",
    "df[\"DayPart\"] = (timeInSeconds - sunrise) / (sunset - sunrise)\n",
    "\n",
    "df = df.drop([\"UNIXTime\", \"Data\", \"Time\",\n",
    "              \"TimeSunRise\", \"TimeSunSet\"], axis=1)\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Radiation        float64\n",
       "Temperature        int64\n",
       "Pressure         float64\n",
       "Humidity           int64\n",
       "WindDirection    float64\n",
       "Speed            float64\n",
       "DayPart          float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "С&nbsp;такими данными уже можно работать. Проверим размер набора данных:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32686, 7)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим основные статистические характеристики набора данных:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Radiation</th>\n",
       "      <th>Temperature</th>\n",
       "      <th>Pressure</th>\n",
       "      <th>Humidity</th>\n",
       "      <th>WindDirection</th>\n",
       "      <th>Speed</th>\n",
       "      <th>DayPart</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>32686.000000</td>\n",
       "      <td>32686.000000</td>\n",
       "      <td>32686.000000</td>\n",
       "      <td>32686.000000</td>\n",
       "      <td>32686.000000</td>\n",
       "      <td>32686.000000</td>\n",
       "      <td>32686.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>207.124697</td>\n",
       "      <td>51.103255</td>\n",
       "      <td>30.422879</td>\n",
       "      <td>75.016307</td>\n",
       "      <td>143.489821</td>\n",
       "      <td>6.243869</td>\n",
       "      <td>0.482959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>315.916387</td>\n",
       "      <td>6.201157</td>\n",
       "      <td>0.054673</td>\n",
       "      <td>25.990219</td>\n",
       "      <td>83.167500</td>\n",
       "      <td>3.490474</td>\n",
       "      <td>0.602432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.110000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>30.190000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>0.090000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.634602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.230000</td>\n",
       "      <td>46.000000</td>\n",
       "      <td>30.400000</td>\n",
       "      <td>56.000000</td>\n",
       "      <td>82.227500</td>\n",
       "      <td>3.370000</td>\n",
       "      <td>-0.040139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2.660000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>30.430000</td>\n",
       "      <td>85.000000</td>\n",
       "      <td>147.700000</td>\n",
       "      <td>5.620000</td>\n",
       "      <td>0.484332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>354.235000</td>\n",
       "      <td>55.000000</td>\n",
       "      <td>30.460000</td>\n",
       "      <td>97.000000</td>\n",
       "      <td>179.310000</td>\n",
       "      <td>7.870000</td>\n",
       "      <td>1.006038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1601.260000</td>\n",
       "      <td>71.000000</td>\n",
       "      <td>30.560000</td>\n",
       "      <td>103.000000</td>\n",
       "      <td>359.950000</td>\n",
       "      <td>40.500000</td>\n",
       "      <td>1.566061</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Radiation   Temperature      Pressure      Humidity  \\\n",
       "count  32686.000000  32686.000000  32686.000000  32686.000000   \n",
       "mean     207.124697     51.103255     30.422879     75.016307   \n",
       "std      315.916387      6.201157      0.054673     25.990219   \n",
       "min        1.110000     34.000000     30.190000      8.000000   \n",
       "25%        1.230000     46.000000     30.400000     56.000000   \n",
       "50%        2.660000     50.000000     30.430000     85.000000   \n",
       "75%      354.235000     55.000000     30.460000     97.000000   \n",
       "max     1601.260000     71.000000     30.560000    103.000000   \n",
       "\n",
       "       WindDirection         Speed       DayPart  \n",
       "count   32686.000000  32686.000000  32686.000000  \n",
       "mean      143.489821      6.243869      0.482959  \n",
       "std        83.167500      3.490474      0.602432  \n",
       "min         0.090000      0.000000     -0.634602  \n",
       "25%        82.227500      3.370000     -0.040139  \n",
       "50%       147.700000      5.620000      0.484332  \n",
       "75%       179.310000      7.870000      1.006038  \n",
       "max       359.950000     40.500000      1.566061  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим наличие пропусков в&nbsp;данных:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Radiation        0\n",
       "Temperature      0\n",
       "Pressure         0\n",
       "Humidity         0\n",
       "WindDirection    0\n",
       "Speed            0\n",
       "DayPart          0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Разделение данных\n",
    "Разделим данные на&nbsp;целевой столбец и признаки:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop(\"Radiation\", axis=1)\n",
    "y = df[\"Radiation\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   Temperature  Pressure  Humidity  WindDirection  Speed   DayPart\n",
      "0           48     30.46        59         177.39   5.62  1.475602\n",
      "1           48     30.46        58         176.78   3.37  1.468588\n",
      "2           48     30.46        57         158.75   3.37  1.461713\n",
      "3           48     30.46        60         137.71   3.37  1.454653\n",
      "4           48     30.46        62         104.95   5.62  1.447778 \n",
      "\n",
      "0    1.21\n",
      "1    1.21\n",
      "2    1.23\n",
      "3    1.21\n",
      "4    1.17\n",
      "Name: Radiation, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(X.head(), \"\\n\")\n",
    "print(y.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32686, 6)\n",
      "(32686,)\n"
     ]
    }
   ],
   "source": [
    "print(X.shape)\n",
    "print(y.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Предобработаем данные, чтобы методы работали лучше:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\matsh\\onedrive\\428a~1\\3188~1\\2425f~1\\9b69~1\\lab4\\venv~1\\lib\\site-packages\\sklearn\\preprocessing\\data.py:645: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "c:\\users\\matsh\\onedrive\\428a~1\\3188~1\\2425f~1\\9b69~1\\lab4\\venv~1\\lib\\site-packages\\sklearn\\base.py:464: DataConversionWarning: Data with input dtype int64, float64 were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Temperature</th>\n",
       "      <th>Pressure</th>\n",
       "      <th>Humidity</th>\n",
       "      <th>WindDirection</th>\n",
       "      <th>Speed</th>\n",
       "      <th>DayPart</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3.268600e+04</td>\n",
       "      <td>3.268600e+04</td>\n",
       "      <td>3.268600e+04</td>\n",
       "      <td>3.268600e+04</td>\n",
       "      <td>3.268600e+04</td>\n",
       "      <td>3.268600e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>5.565041e-16</td>\n",
       "      <td>2.904952e-14</td>\n",
       "      <td>1.391260e-17</td>\n",
       "      <td>6.956302e-17</td>\n",
       "      <td>-9.738822e-17</td>\n",
       "      <td>5.217226e-18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.000015e+00</td>\n",
       "      <td>1.000015e+00</td>\n",
       "      <td>1.000015e+00</td>\n",
       "      <td>1.000015e+00</td>\n",
       "      <td>1.000015e+00</td>\n",
       "      <td>1.000015e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-2.758117e+00</td>\n",
       "      <td>-4.259540e+00</td>\n",
       "      <td>-2.578560e+00</td>\n",
       "      <td>-1.724255e+00</td>\n",
       "      <td>-1.788859e+00</td>\n",
       "      <td>-1.855112e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-8.229646e-01</td>\n",
       "      <td>-4.184734e-01</td>\n",
       "      <td>-7.316829e-01</td>\n",
       "      <td>-7.366250e-01</td>\n",
       "      <td>-8.233591e-01</td>\n",
       "      <td>-8.683240e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-1.779139e-01</td>\n",
       "      <td>1.302504e-01</td>\n",
       "      <td>3.841386e-01</td>\n",
       "      <td>5.062367e-02</td>\n",
       "      <td>-1.787376e-01</td>\n",
       "      <td>2.279483e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>6.283995e-01</td>\n",
       "      <td>6.789742e-01</td>\n",
       "      <td>8.458578e-01</td>\n",
       "      <td>4.307058e-01</td>\n",
       "      <td>4.658840e-01</td>\n",
       "      <td>8.682924e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>3.208603e+00</td>\n",
       "      <td>2.508053e+00</td>\n",
       "      <td>1.076717e+00</td>\n",
       "      <td>2.602741e+00</td>\n",
       "      <td>9.814329e+00</td>\n",
       "      <td>1.797910e+00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Temperature      Pressure      Humidity  WindDirection  \\\n",
       "count  3.268600e+04  3.268600e+04  3.268600e+04   3.268600e+04   \n",
       "mean   5.565041e-16  2.904952e-14  1.391260e-17   6.956302e-17   \n",
       "std    1.000015e+00  1.000015e+00  1.000015e+00   1.000015e+00   \n",
       "min   -2.758117e+00 -4.259540e+00 -2.578560e+00  -1.724255e+00   \n",
       "25%   -8.229646e-01 -4.184734e-01 -7.316829e-01  -7.366250e-01   \n",
       "50%   -1.779139e-01  1.302504e-01  3.841386e-01   5.062367e-02   \n",
       "75%    6.283995e-01  6.789742e-01  8.458578e-01   4.307058e-01   \n",
       "max    3.208603e+00  2.508053e+00  1.076717e+00   2.602741e+00   \n",
       "\n",
       "              Speed       DayPart  \n",
       "count  3.268600e+04  3.268600e+04  \n",
       "mean  -9.738822e-17  5.217226e-18  \n",
       "std    1.000015e+00  1.000015e+00  \n",
       "min   -1.788859e+00 -1.855112e+00  \n",
       "25%   -8.233591e-01 -8.683240e-01  \n",
       "50%   -1.787376e-01  2.279483e-03  \n",
       "75%    4.658840e-01  8.682924e-01  \n",
       "max    9.814329e+00  1.797910e+00  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns = X.columns\n",
    "scaler = StandardScaler()\n",
    "X = scaler.fit_transform(X)\n",
    "pd.DataFrame(X, columns=columns).describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Разделим выборку на&nbsp;тренировочную и тестовую:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y,\n",
    "                            test_size=0.25, random_state=346705925)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(24514, 6)\n",
      "(8172, 6)\n",
      "(24514,)\n",
      "(8172,)\n"
     ]
    }
   ],
   "source": [
    "print(X_train.shape)\n",
    "print(X_test.shape)\n",
    "print(y_train.shape)\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Модель ближайших соседей для произвольно заданного гиперпараметра $K$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Напишем функцию, которая считает метрики построенной модели:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_model(model):\n",
    "    print(\"mean_absolute_error:\",\n",
    "          mean_absolute_error(y_test, model.predict(X_test)))\n",
    "    print(\"median_absolute_error:\",\n",
    "          median_absolute_error(y_test, model.predict(X_test)))\n",
    "    print(\"r2_score:\",\n",
    "          r2_score(y_test, model.predict(X_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Попробуем метод ближайших соседей с&nbsp;гиперпараметром $K=5$:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsRegressor(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "          metric_params=None, n_jobs=None, n_neighbors=5, p=2,\n",
       "          weights='uniform')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reg_5 = KNeighborsRegressor(n_neighbors=5)\n",
    "reg_5.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим метрики построенной модели:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean_absolute_error: 55.39857905041605\n",
      "median_absolute_error: 4.017000000000004\n",
      "r2_score: 0.8677873476991447\n"
     ]
    }
   ],
   "source": [
    "test_model(reg_5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Видно, что средние ошибки не&nbsp;очень показательны для&nbsp;одной модели, они больше подходят для&nbsp;сравнения разных моделей. В&nbsp;тоже время коэффициент детерминации неплох сам по&nbsp;себе, в&nbsp;данном случае модель более-менее состоятельна."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Использование кросс-валидации"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим различные стратегии кросс-валидации. Для&nbsp;начала посмотрим классический K-fold:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.83276085 0.5984654  0.83547149 0.75974839 0.76407458 0.81422383\n",
      " 0.85420738 0.79432111 0.74927049 0.28234327]\n",
      "0.7284886763598686 ± 0.16383980384698185\n"
     ]
    }
   ],
   "source": [
    "scores = cross_val_score(KNeighborsRegressor(n_neighbors=5), X, y,\n",
    "                         cv=KFold(n_splits=10), scoring=\"r2\")\n",
    "print(scores)\n",
    "print(scores.mean(), \"±\", scores.std())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.86645745 0.87359216 0.86201671 0.87253355 0.86825249 0.87149987\n",
      " 0.86931511 0.86595468 0.86938791 0.867287  ]\n",
      "0.8686296919477157 ± 0.0032664937392509145\n"
     ]
    }
   ],
   "source": [
    "scores = cross_val_score(KNeighborsRegressor(n_neighbors=5), X, y,\n",
    "                         cv=RepeatedKFold(n_splits=5, n_repeats=2),\n",
    "                         scoring=\"r2\")\n",
    "print(scores)\n",
    "print(scores.mean(), \"±\", scores.std())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.87288289 0.86965989 0.87846365 0.87005449 0.86511118 0.87199795\n",
      " 0.87614293 0.87572756 0.87447917 0.86520971]\n",
      "0.8719729406086248 ± 0.00427294746868573\n"
     ]
    }
   ],
   "source": [
    "scores = cross_val_score(KNeighborsRegressor(n_neighbors=5), X, y,\n",
    "                         cv=ShuffleSplit(n_splits=10), scoring=\"r2\")\n",
    "print(scores)\n",
    "print(scores.mean(), \"±\", scores.std())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Подбор гиперпараметра $K$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Введем список настраиваемых параметров:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1,  3,  5,  7,  9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31, 33,\n",
       "       35, 37, 39, 41, 43, 45, 47, 49])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_range = np.array(range(1, 50, 2))\n",
    "tuned_parameters = [{'n_neighbors': n_range}]\n",
    "n_range"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Запустим подбор параметра:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'n_neighbors': 5}"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gs = GridSearchCV(KNeighborsRegressor(), tuned_parameters,\n",
    "                  cv=ShuffleSplit(n_splits=10), scoring=\"r2\",\n",
    "                  return_train_score=True, n_jobs=-1)\n",
    "gs.fit(X, y)\n",
    "gs.best_params_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим результаты при&nbsp;разных значения гиперпараметра на&nbsp;тренировочном наборе данных:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvgAAAH0CAYAAABICFkFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAWJQAAFiUBSVIk8AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzs3XmYXFWd//H3t5esZF/IAnRDIAECBBKWALJrEAQGXPiNKAiCihsO4E9RHBHU3zALoiiOKIRFRxxHGYQgyr4GCIQlLCGBJE1CEiAhZN+7z++Pqm6qm+50J6mkuqrer+fp56buuffcc58E/dTp7z03UkpIkiRJKg0VhR6AJEmSpPwx4EuSJEklxIAvSZIklRADviRJklRCDPiSJElSCTHgS5IkSSXEgC9JkiSVEAO+JEmSVEIM+JIkSVIJMeBLkiRJJcSAL0mSJJUQA74kSZJUQgz4kiRJUgkx4EuSJEklxIAvSZIklRADviRJklRCqgo9gM4uIuYAvYG6Ag9FkiRJpa0WWJ5S2nVrOjHgt6939+7d+++11179Cz0QSZIkla7p06ezZs2are7HgN++ur322qv/1KlTCz0OSZIklbBx48bx7LPP1m1tP9bgS5IkSSXEgC9JkiSVEAO+JEmSVEIM+JIkSVIJMeBLkiRJJcSAL0mSJJUQA74kSZJUQgz4kiRJUgkx4EuSJEklxIAvSZIklRADviRJklRC8hLwI+KTEfHziHg0IpZHRIqI321hXztFxMSIWBAR6yKiLiJ+GhH9NnHO3hHxx4h4JyLWRsSMiLg8Irpv+V1JkiRJxacqT/18DxgDrATeBPbckk4iYgQwGRgM/AV4FTgY+Abw0Yg4PKX0botzDgEeAKqBPwHzgGOB7wPHRcRxKaV1WzIeSZIkqdjkq0TnQmAk0Bv48lb080sy4f6ClNKpKaVLUkrHAlcDo4Af5x4cEZXAjUAP4JMppTNSSt8GDgH+DByeHZskSZJUFvIS8FNKD6aUXksppS3tIyJ2AyYAdcC1LZovA1YBZ0ZEz5z9RwF7AY+klO7IGU8D8K3sx/MjIrZ0XJIkSVIx6UwP2R6b3d6TDehNUkorgMfJzNSPb+Wcv7XsLKU0G5gJ1AC75X2028G6jfXMfXd1oYchSZKkIpKvGvx8GJXdzmyj/TUyM/wjgfs345yR2Z9Zm7p4RExto2mLnifYGqvWbWTC1Y+wYNkaulZVMP2Kj+IvISRJktQRnWkGv092u6yN9sb9fbfynE6vZ9cqVq7bSEqwdkMD76zwGWFJkiR1TGeawW9P4xT25tT5d/iclNK4VjvIzOyP3Yxr5kXtgB688Gbm+0nd4lXs2Lvb9h6CJEmSilBnmsFvnG3v00Z77xbHbek5RaFmwPvPEr9hHb4kSZI6qDMF/BnZ7cg22vfIbnPr7bfknKJQO6BH05/r3l1VwJFIkiSpmHSmgP9gdjshIpqNKyJ6kVnTfg3wZE7TA9ntR1t2ll12cyTwBjA776PdxpzBlyRJ0pbY7gE/IqojYs/sW2ubpJRmAfcAtcBXW5x2OdATuCWllDud/TAwHTgyIk7JuUYF8K/Zj7/amvX5C6V2oDP4kiRJ2nx5ecg2Ik4FTs1+HJLdHhoRN2X/vDil9M3sn4eTCeVvkAnzub4CTAauiYjjsscdAhxDpszm0tyDU0r1EXEOmZn8P0XEn4C5wHHAgWTWzr86D7e43bWcwU8puVSmJEmS2pWvVXT2Bz7XYt9uvP+CqTeAb9KOlNKsiDgQuIJM2c2JwELgGuDylNKSVs55KiIOIjPLPwHolb3eFcCVKaWiXGNyQM8u7JBdLnPluo28u2o9A3foWuhhSZIkqZPLS8BPKf0A+EEHj63j/eUrW2ufB5yzmdd/BfjU5pzT2UUENQN68PKC5QC88e4qA74kSZLa1ZkeslULtTllOnWLfdBWkiRJ7TPgd2I1OUtlvuGDtpIkSeoAA34n1mwG36UyJUmS1AEG/E7MGXxJkiRtLgN+J1Y70Bl8SZIkbR4Dfic2uFdXulVn/oqWrdnA0tXrCzwiSZIkdXYG/E4sIprV4c9ZbJmOJEmSNs2A38k1r8O3TEeSJEmbZsDv5JrX4TuDL0mSpE0z4HdyuSU6zuBLkiSpPQb8Ti63RMcZfEmSJLXHgN/JOYMvSZKkzWHA7+SG9O5Gl6rMX9OSVetZtmZDgUckSZKkzsyA38lVVAQ1/d8v05nrLL4kSZI2wYBfBGoGuJKOJEmSOsaAXwRqm62Fb8CXJElS2wz4RaCm2Vr4luhIkiSpbQb8IuAMviRJkjrKgF8Eagc4gy9JkqSOMeAXgaF9ulFdGQAsWrGOVes2FnhEkiRJ6qwM+EWgqrKCnfvlluk4iy9JkqTWGfCLRI11+JIkSeoAA36RqLEOX5IkSR1gwC8SrqQjSZKkjjDgF4nma+Eb8CVJktQ6A36RyF0q04dsJUmS1BYDfpEY3rc7lRWZpTIXLlvL2g31BR6RJEmSOiMDfpHoUlXB8L7dmz7PXeIsviRJkj7IgF9EcpfKrFtsHb4kSZI+yIBfRKzDlyRJUnsM+EWk2Qy+K+lIkiSpFQb8IuIMviRJktpjwC8itQOdwZckSdKmGfCLyE79ehCZlTJZsHQN6za6VKYkSZKaM+AXkW7VlQzrk1kqsyHBm++tKfCIJEmS1NkY8ItM7oO2b1imI0mSpBYM+EWmJudB27rFPmgrSZKk5gz4RWbXgc7gS5IkqW15C/gRsVNETIyIBRGxLiLqIuKnEdFvM/s5LSIeiIilEbE2IqZHxPcjolsbx1dGxGci4tGIeCsiVkfEzIi4MSJG5+fuOo/cGfw5LpUpSZKkFvIS8CNiBDAVOAeYAlwNzAa+ATwREQM62M8PgduAg4DbgWuB5cDlwH0R0b2V034P/A6ozZ77c+B14HPAsxFx7BbfWCfUfC18Z/AlSZLUXFWe+vklMBi4IKX088adEfET4ELgx8D5m+ogIg4ALgWWAuNSSrOz+wO4Bvga8G3gBznnHAScDrwMHJxSWp3Tdg4wEfge8MBW32EnsUv/90t03nxvDRvqG6iutNJKkiRJGVudDCNiN2ACUEdmxj3XZcAq4MyI6MmmnQYEcH1juAdIKSXgu0ACvhwRlTnn7Jbd3p8b7rP+kt0O6uCtFIXuXSoZ0jtTrVTfkJjvUpmSJEnKkY+p38YSmHtSSg25DSmlFcDjQA9gfDv9DMluZ7dsyPazmMxvCfbNaXq5cQytlO+clN3e1851i07uUpm+0VaSJEm58hHwR2W3M9tofy27HdlOP4uz211bNkREL2Bg9uOejftTSi+RqfffB3g1Iq6NiCsj4k7gBuAPZEp0SkrzOnwftJUkSdL78lGD3ye7XdZGe+P+vu30Mwn4DnBeRPwypVSX0/YjMuU7AM1W5UkpXRQRM8gE/a/kNE0Fbk4pdWiKOyKmttG0Zxv7C6ZmoDP4kiRJat32eDqzMZinTR2UUpoMXEcmwE/LLnN5VUQ8SeYB28ZynPqmjjOuIVP7fwWwM9ALOCJ7vbsj4qv5vJnOwBl8SZIktSUfM/iNM/R92mjv3eK4NqWUzo+IKcAXyayOA5mZ+OOBc4HRwDs5p3wO+DpwdUrpypz9j0XEyWTq+a+MiJtTSivbufa41vZnZ/bHtjf27ckafEmSJLUlHzP4M7Lbtmrs98hu26rRbyalNDGlND6l1DP7c2RK6T7g0OwhT+cc3vgg7YOt9PMW8CqwA+8/J1AScl92NW/JauobNvnLEUmSJJWRfAT8xnA9ISKa9Zd9OPZwYA3w5JZeICImADXAwyml+TlNXbPbtpbCbNy/fkuv3Rnt0LWKgTtkbn1DfWLBUpfKlCRJUsZWB/yU0izgHjJvkm1Z73450BO4Jfdh14jYMyI+8PBqRPRuZd8I4Ndkau8vadH8aHZ7UUT0aXHe+cBOwFvAK5txS0WhNqdMxzp8SZIkNcrXm2y/AkwGromI44DpwCHAMWRKcy5tcfz07DZa7L8hImrI1N2/B+wOnAxUA+ellFr+FuCXwGeA/YCZEXEHmTfhjiWzPn898NWUUj0lpmZAT5554z0gU4f/oT0GtnOGJEmSykFeAn5KaVZEHEhmJZuPAicCC4FrgMtTSks62NUk3n/AtheZB2r/DPxbSmlaK9ddGRGHAxcBHwfOALoAi4D/Af4jpTRla+6ts2o+g++DtpIkScrI1ww+KaV5wDkdPLblzH3j/puBmzfzuivJfLG4YnPOK3Y1A99/0LbOEh1JkiRlbY918LUNOIMvSZKk1hjwi1RN/+Yvu2pwqUxJkiRhwC9afXpU069HNQDrNjbw9oq1BR6RJEmSOgMDfhHLfeFV3WLr8CVJkmTAL2rW4UuSJKklA34RazaD70o6kiRJwoBf1GoHOoMvSZKk5gz4RcwZfEmSJLVkwC9itQNyl8pcRUoulSlJklTuDPhFrF+Panp1y7yMePX6ehatXFfgEUmSJKnQDPhFLCJazOJbpiNJklTuDPhFriZnqcy6xT5oK0mSVO4M+EXOGXxJkiTlMuAXudqBuSvpOIMvSZJU7gz4Ra7522ydwZckSSp3Bvwi13wtfJfKlCRJKncG/CI3cIcu9OxSCcCKtRtZsmp9gUckSZKkQjLgF7mI8I22kiRJamLALwG1A3Pr8H3QVpIkqZwZ8EuAM/iSJElqZMAvAc1X0nEGX5IkqZwZ8EuAM/iSJElqZMAvAc3fZusMviRJUjkz4JeAwb260q0681e5dPUGlq52qUxJkqRyZcAvARUVQU3/3Fl8y3QkSZLKlQG/RNTkPGhbZ5mOJElS2TLgl4jagc7gS5IkyYBfMpzBlyRJEhjwS0bzlXScwZckSSpXBvwSUePLriRJkoQBv2QM7dOdLpWZv87FK9ezYu2GAo9IkiRJhWDALxGVFcHO/bs3fbZMR5IkqTwZ8EuIdfiSJEky4JeQmpyA70o6kiRJ5cmAX0JqB/qgrSRJUrkz4JeQ5jP4luhIkiSVIwN+Cal1qUxJkqSyZ8AvIcP7dqeqIgB4e/k6Vq/fWOARSZIkaXvLW8CPiJ0iYmJELIiIdRFRFxE/jYh+m9nPaRHxQEQsjYi1ETE9Ir4fEd3aOe+UiLg7IhZlrz8vIu6IiPFbd2fFo6qygp36vb9U5twllulIkiSVm7wE/IgYAUwFzgGmAFcDs4FvAE9ExIAO9vND4DbgIOB24FpgOXA5cF9EdG/lnIqI+DXwF2B09vyrgHuAEcC4rbq5ItOsDn+xAV+SJKncVOWpn18Cg4ELUko/b9wZET8BLgR+DJy/qQ4i4gDgUmApMC6lNDu7P4BrgK8B3wZ+0OLUi4EvAL8FzksprW/Rb/UW31URqh3Qg4ezf7YOX5Ikqfxs9Qx+ROwGTADqyMy457oMWAWcGRE92bTTgACubwz3ACmlBHwXSMCXI6Iy59q9ge8DbwJfaBnus+dv2Nx7KmaupCNJklTe8lGic2x2e09KqSG3IaW0Angc6AG0Vws/JLud3bIh289iMr8l2Den6RRgB+APQEVEfDIiLomIr0bEmM2+kxKw68Dct9k6gy9JklRu8lGiMyq7ndlG+2tkZvhHAvdvop/F2e2uLRsiohcwMPtxT+D57J8Pym43ANOBmhbn/Rk4K6VUNlPZNc2Wyiyb25YkSVJWPmbw+2S3y9pob9zft51+JmW350VEbYu2H5Ep3wHIXZVncHb7LWARcAjQK7t9BvgEmecD2hURU1v7IfOFomjs1K8H2ZUyWbBsDWs31Bd2QJIkSdqutsc6+I3BPG3qoJTSZOA6MgF+WkTcGBFXRcSTZB6wfTl7aG5ibazHXwOcnFKaklJamVKaQqZ8ZyWZ+v/hebqXTq9LVQXDs0tlpgRvvucsviRJUjnJR8BvnKHv00Z77xbHtSmldD5wLvAKcDqZlXfWA8cDL2YPeyfnlPey2ydTSm+16Gsh8BSZezywA9ce19oP8Gp753Y2tS6VKUmSVLbyEfBnZLcj22jfI7ttq0a/mZTSxJTS+JRSz+zPkSml+4BDs4c83cq1l7bRXeMXgA+sn1/Kcuvw63zQVpIkqazkI+A/mN1OiIhm/WUfjj2cTAnNk1t6gYiYQOYB2odTSvNzmhof2h3dxqmN++u29NrFqNkMvgFfkiSprGx1wE8pzSLz1tha4Kstmi8HegK3pJSakmZE7BkRH3h4Nbuufct9I4Bfk6m9v6TFtV8gswznXhFxXovzzgP2AmbRfNa/5OWuhe9KOpIkSeUlX2+y/QowGbgmIo4js2TlIcAxZEpzLm1x/PTsNlrsvyEiaoCpZMprdgdOBqrJvKW2td8CnAs8BvwmIj5O5mHcvYETgdXA2SmlslpKptYSHUmSpLKVl1V0srP4BwI3kQn2FwMjgGuAQ1NK73awq0lk1rQ/HfgmcBjwZ2BsSummNq49AxgL3ACMAb4BjANuBQ5MKT22RTdVxHbu34PIfnWa/94a1m9s2PQJkiRJKhn5msEnpTQPOKeDx7acuW/cfzNw8xZe+7x2DywT3aorGdq7GwuWraUhu1TmboN2KPSwJEmStB1sj3XwVQDW4UuSJJUnA36Jqh1oHb4kSVI5MuCXKGfwJUmSypMBv0S5ko4kSVJ5MuCXKGfwJUmSypMBv0TV5Mzgz1uymo31LpUpSZJUDgz4JapHlyoG9+oKwMaGxIKlaws8IkmSJG0PBvwSVptTpmMdviRJUnkw4Jew3DKdNwz4kiRJZcGAX8JqB+bO4PugrSRJUjkw4JcwZ/AlSZLKjwG/hDWvwXcGX5IkqRwY8EvYLjkz+HPfXU19QyrgaCRJkrQ9GPBLWO9u1Qzo2QWA9fUNvLXcpTIlSZJKnQG/xDWrw19sHb4kSVKpM+CXOOvwJUmSyosBv8TV5AR8V9KRJEkqfQb8Elc78P0SHd9mK0mSVPoM+CWuttkMviU6kiRJpc6AX+Ka1+CvIiWXypQkSSplBvwS16dHNX17VAOwdkMD76xYV+ARSZIkaVsy4JeB3Adt61wqU5IkqaQZ8MtAbe5a+NbhS5IklTQDfhmoaVGHL0mSpNJlwC8DzuBLkiSVDwN+GXAGX5IkqXwY8MtA7gx+3WKXypQkSSplBvwy0L9nF3p1rQJg1fp6Fq9cX+ARSZIkaVsx4JeBiKBmYG4dvmU6kiRJpcqAXyaa1+H7oK0kSVKpMuCXieYr6TiDL0mSVKoM+GXCGXxJkqTyYMAvE7U5Ad8ZfEmSpNJlwC8TuSU6c1wqU5IkqWQZ8MvEoF5d6V5dCcCKtRtZunpDgUckSZKkbcGAXyYigprcF15ZpiNJklSSDPhlpHkdvg/aSpIklSIDfhnJfdmVM/iSJEmlKW8BPyJ2ioiJEbEgItZFRF1E/DQi+m1mP6dFxAMRsTQi1kbE9Ij4fkR06+D5/xwRKfvz4S27m9LkDL4kSVLpy0vAj4gRwFTgHGAKcDUwG/gG8EREDOhgPz8EbgMOAm4HrgWWA5cD90VE93bOHwv8M7Byy+6ktFmDL0mSVPryNYP/S2AwcEFK6dSU0iUppWPJBP1RwI/b6yAiDgAuBZYCY1JKZ6eULgbGA78ADge+vYnzuwG/BZ4B/ncr76ckOYMvSZJU+rY64EfEbsAEoI7MjHuuy4BVwJkR0ZNNOw0I4PqU0uzGnSmzYPt3gQR8OSIq2zj/X4BdgbOBhs27i/IwpHc3ulRl/sqXrFrPsjUulSlJklRq8jGDf2x2e09KqVmwTimtAB4HepCZid+UIdnt7JYN2X4Wk/ktwb4t2yPiGDLlQN9JKc3crNGXkYqKoKb/+2U6c53FlyRJKjn5CPijstu2gvVr2e3IdvpZnN3u2rIhInoBA7Mf92zR1ge4CXgUuKada5S9mpwyHevwJUmSSk9VHvrok90ua6O9cX/fdvqZBHwHOC8ifplSqstp+xGZ8h2Alqvy/BwYAByTLefZIhExtY2mPdvYX5Rqcx60fcOAL0mSVHLyEfDb0xjMNxm+U0qTI+I64EvAtIj4M7CEzMO1BwEvA6OB+qaOIz4OnAl8NbduX22rGZg7g2+JjiRJUqnJR8BvnKHv00Z77xbHtSmldH5ETAG+CJye3T0VOB44l0zAfwcgIvoD1wEPAP+5RSNvfu1xre3PzuyP3dr+Owtn8CVJkkpbPgL+jOy2rRr7PbLbDj38mlKaCExsuT8irs/+8ensdhcydfnHAg0R0fIUgHuz+y9MKf20I9cvdbUDnMGXJEkqZfkI+A9mtxMioiJ3JZ3sw7GHA2uAJ7f0AhExAagBHk4pzc/ufhe4oY1TjiTzxeJuYAHw0pZeu9QM69ud6spgQ31i0Yp1rFq3kZ5dt0elliRJkraHrU52KaVZEXEPmbXwv0rmoddGlwM9getSSk31IBGxZ/bcV3P7iojeKaXlLfaNAH5Npvb+kpzrzgPOa21MEXETmYD/k5TSfVt8cyWosiLYuX8PZi/K/HW88e5q9h7Wu52zJEmSVCzyNXX7FWAycE1EHAdMBw4BjiFTmnNpi+OnZ7ct62puiIgaMnX37wG7AycD1cB5KaUt/i2A3lc7oGdOwF9lwJckSSoh+VgHn5TSLOBAMuvRHwJcDIwgsy79oSmldzvY1SRgA5kHbL8JHAb8GRibUropH2MV1OQ8aGsdviRJUmnJW/F1tmTmnA4e2+oTsSmlm4Gb8zCWs4Gzt7afUpX7oK0r6UiSJJWWvMzgq7g0n8E34EuSJJUSA34Zaj6Db4mOJElSKTHgl6Hh/bpTWZGpklq4bC1rN9S3c4YkSZKKhQG/DFVXVrBTv+5Nn+cucRZfkiSpVBjwy1RNTpnOnMXW4UuSJJUKA36Zqs150NaVdCRJkkqHAb9M5c7guxa+JElS6TDglyln8CVJkkqTAb9MNZvBX+wMviRJUqkw4Jepnft3J7LvE16wbA3rNrpUpiRJUikw4JeprlWVDOuTWSozJZi3ZE2BRyRJkqR8MOCXsdqB1uFLkiSVGgN+Gcutw39wxjsFHIkkSZLyxYBfxj681+CmP986ZR4z3lpRwNFIkiQpHwz4ZeyYUYM5bMQAAOobEpff+TIppQKPSpIkSVvDgF/GIoLLTh5NRXY1ncmz3uWeV94u7KAkSZK0VQz4ZW7UkF58dnxN0+cf3fUKaze4ZKYkSVKxMuCLiz4ykr49qoHMcpk3PDanwCOSJEnSljLgi749unDRR0Y2fb72wdd5a9naAo5IkiRJW8qALwDOOHgXRu3YC4DV6+v5t7+9WuARSZIkaUsY8AVAVWUF3z9576bPtz03n2fnvlfAEUmSJGlLGPDV5PDdB3L86B2bPl9+x8s0NLhspiRJUjEx4KuZS0/cmy5VmX8WL7y5jNuem1/gEUmSJGlzGPDVzC4DevCFI3Zt+vyvf3uVles2FnBEkiRJ2hwGfH3AV47enR17dwVg0Yp1/OKB1ws8IkmSJHWUAV8f0LNrFZecsGfT54mPzaFu8aoCjkiSJEkdZcBXq07dfzhjd+kLwPr6Bn501/QCj0iSJEkdYcBXqyKCy04e3fT5vulv88jMRQUckSRJkjrCgK82jdm5L58ct1PT5ysmvcKG+oYCjkiSJEntMeBrk7710VH07FIJwOvvrOR3T75R4BFJkiRpUwz42qTBvbrx9eP2aPp89b0zeXflugKOSJIkSZtiwFe7zjm8ltoBPQBYvnYjP7l3ZoFHJEmSpLYY8NWurlWVfO9jezd9vnXKXF5ZsLyAI5IkSVJbDPjqkOP2GswRewwEoCHB5Xe+TEqpwKOSJElSSwZ8dUhm2cy9qawIAJ6as4S7X3qrwKOSJElSSwZ8ddjug3tx1qE1TZ9/fNd01m6oL+CIJEmS1JIBX5vln44bSf+eXQCYv3QNv35kdoFHJEmSpFwGfG2WPj2quXjCyKbPv3zodRYsXVPAEUmSJClX3gJ+ROwUERMjYkFErIuIuoj4aUT028x+TouIByJiaUSsjYjpEfH9iOjWyrHDI+LrEXF39nrrIuLdiLg3Ij6er3tTc/940C7sNbQ3AGs3NHDl3a8WeESSJElqlJeAHxEjgKnAOcAU4GpgNvAN4ImIGNDBfn4I3AYcBNwOXAssBy4H7ouI7i1O+TpwDTAKeBD4CfB34AjgzxHxk627M7WmsiLzwG2jO15YwNN1Swo4IkmSJDXK1wz+L4HBwAUppVNTSpeklI4lE/RHAT9ur4OIOAC4FFgKjEkpnZ1SuhgYD/wCOBz4dovTpgBHp5R2Symdk1L6TkrpDOAAMl8MLoyIcXm6R+UYv9sAPrbv0KbPl9/5MvUNLpspSZJUaFsd8CNiN2ACUEdmxj3XZcAq4MyI6NlOV6cBAVyfUmp6cjNlFlv/LpCAL0dEZU7bbSmlh1t2lFKaDvx39uPRm3M/6rjvnLgnXasy/4Remr+cP02dV+ARSZIkKR8z+Mdmt/eklBpyG1JKK4DHgR5kZuI3ZUh2+4FlWbL9LCbzW4J9OziuDdntxg4er820U78efOmoEU2f//3vM1i+dsMmzpAkSdK2lo+APyq7ndlG+2vZ7cg22hstzm53bdkQEb2AgdmPe7Y3oIjoDXyCzKz/Pe0dry335aNGMLRP5vnnxSvX8/P7X2vnDEmSJG1L+Qj4fbLbZW20N+7v204/k7Lb8yKitkXbj8iU7wBsclWeiAjgemBH4D+z5Trtioiprf3QgS8U5ax7l0q+c+JeTZ9vfLyOWYtWFnBEkiRJ5W17rIPfGMw3+QRmSmkycB2ZAD8tIm6MiKsi4knga8DL2UPbe3XqVcCngEeBi7Z41Oqwk/cbykG1me9dGxsSP5r0SoFHJEmSVL7yEfAbZ+j7tNHeu8VxbUopnQ+cC7wCnA6cD6wHjgdezB72TlvnR8S/AxcCjwAnppTWtXfNnGuPa+0HcJH3dkQEl508msh+lXtwxiIefLXNvyZJkiRtQ/kI+DOy27Zq7PfIbtuq0W8mpTQxpTQ+pdQz+3NkSuk+4NDsIU+3dl5EXA18k8x6+CeklKwT2Y72Gd6H/3Pgzk2ffzjpFdZvbNjEGZIkSdoW8hHwH8xuJ0REs/6yD8ceDqwBntzSC0TEBKAGeDilNL9FW0TEtcA/AfcCH0sprd7Sa2nLffP4UfTqWgXA7MWruOWk8RDVAAAgAElEQVSJuoKOR5IkqRxtdcBPKc0is1JNLfDVFs2XAz2BW1JKqxp3RsSeEfGBh1ezq9+03DcC+DWZ2vtLWrRFtu0rwN3AKSmlNVtzP9pyA3foyjc+vEfT55/d9xqLVnS4SkqSJEl5UJWnfr4CTAauiYjjgOnAIcAxZEpzLm1xfOPKNtFi/w0RUQNMBd4DdgdOBqqB81JKLX8L8H3gPDK/IXgeuCSiZZc8n1K6fQvvS5vprENr+f2UucxetIoV6zZy1T0zuPIT+xV6WJIkSWUjLwE/pTQrIg4ErgA+CpwILASuAS5PKS3pYFeTgC+SecC2F5kHav8M/FtKaVorxzeumd8d+E4bfd4MGPC3ky5VFfzzSXtzzo2ZRyX++5l5fHZ8DfsMb+sZbEmSJOVTvmbwSSnNA87p4LEfmGbP7r+ZTCDv6DXPBs7u6PHaPo4ZNZhj9xzMA6++Q0rwgzte5n/OP5RWfrsiSZKkPNse6+CrDH3vY3tRVZEJ9M+88R53TltY4BFJkiSVBwO+tondBu3AOYfXNn3+l79OZ/X6jYUbkCRJUpkw4Gub+fpxezBwhy4ALFy2lqvv7dCrECRJkrQVDPjaZnp3q+b/Hj+q6fNvHp3DxMfmFHBEkiRJpc+Ar23qU+N25qiRg5o+XzHpFf749LwCjkiSJKm0GfC1TVVUBP/52bEcWNOvad8lt03jLh+6lSRJ2iYM+NrmenSpYuI5BzF6WOZFxQ0J/um/n+OhGe8UeGSSJEmlx4Cv7aJ3t2pu+fzBjBjUE4AN9YnzfzeVKXM6+g40SZIkdYQBX9vNgB268rvzDmF43+4ArN3QwLk3Pc1L85cVeGSSJEmlw4Cv7Wpon+7813mHMKhXVwBWrNvIWROn8Po7Kwo8MkmSpNJgwNd2VzuwJ78992D6dK8GYMmq9Xzm+qeYt2R1gUcmSZJU/Az4Kog9h/Tm5s8fTM8ulQC8vXwdn7n+Kd5ZvrbAI5MkSSpuBnwVzP479+X6zx1El6rMP8O5S1bz2Rue4r1V6ws8MkmSpOJlwFdBHTpiAP/5mbFUVQQAM99eydk3TmHluo0FHpkkSVJxMuCr4I7ba0euOn0Mkcn4vPDmMs696WnWbqgv7MAkSZKKkAFfncI/7D+cH5+6b9Pnp+Ys4Sv/9SzrNzYUcFSSJEnFx4CvTuOMQ3bhOyfs2fT5gVff4aI/Pk99QyrgqCRJkoqLAV+dypeOGsHXjtm96fOkaQv53u0vkpIhX5IkqSMM+Op0Lp4wks8dWtP0+dYp8/iXu1815EuSJHWAAV+dTkRw2cmj+fjY4U37fv3IbH7xwOsFHJUkSVJxMOCrU6qoCP7tE/tx/Ogdm/Zdde9Mbnp8TgFHJUmS1PkZ8NVpVVVWcM2nD+CIPQY27fvBna/wp6lvFnBUkiRJnZsBX51a16pKrjtzHONq+jXt+9afXuBvLy0s4KgkSZI6LwO+Or0eXaqYePZB7D20NwANCb5+63M8MnNRgUcmSZLU+RjwVRT6dK/mlnMPZreBPQHYUJ/44m+f4Zm6JQUemSRJUudiwFfRGLhDV3533iEM79sdgLUbGjjnxqd5af6yAo9MkiSp8zDgq6gM69ud3513CAN36ArAinUb+dzEKbz+zsoCj0ySJKlzMOCr6Ow6sCe/PfdgenerAuDdVes584anePO91QUemSRJUuEZ8FWU9hram5s+fzA9ulQCsHDZWj57/VO8s2JtgUcmSZJUWAZ8Fa2xu/TjN2cdSJfKzD/jundX8+lfP8msRZbrSJKk8mXAV1E7fPeB/OKMA6isCABmLVrFKT9/jDtfWFDgkUmSJBWGAV9Fb8LoIfz0/+xPl6rMP+dV6+v5+q3P8f2/vMS6jfUFHp0kSdL2ZcBXSTh5zDD+9yuHUTugR9O+W554g0/96gnmLfHhW0mSVD4M+CoZo4f14Y6vf4gT9hnStG/am8v42DWPcu8rbxdwZJIkSduPAV8lpXe3an75mbFcdvLeVFdm6vKXr93IF255hn/563Q21DcUeISSJEnblgFfJSciOOfwXfnjlw5teustwHWPzOaM3zzJW8tcSlOSJJUuA75K1gG79GPS1z/EMaMGNe17uu49TrzmUR59bVEBRyZJkrTtGPBV0vr17MINnzuIb310FNmVNFmyaj1nTZzC1ffOpL4hFXaAkiRJeWbAV8mrqAi+cvTu/P4L4xncqysAKcHP7n+Nz02cwuKV6wo8QkmSpPzJW8CPiJ0iYmJELIiIdRFRFxE/jYh+m9nPaRHxQEQsjYi1ETE9Ir4fEd02cc7eEfHHiHgne86MiLg8Irq3dY7Kz/jdBnDXBUdw2IgBTfsee30xH7vmUabMWVLAkUmSJOVPXgJ+RIwApgLnAFOAq4HZwDeAJyJiwCZOz+3nh8BtwEHA7cC1wHLgcuC+1gJ7RBwCPA2cCtwH/Cx7zveBeyOi61bdnErKoF5d+e25h3DBsbsT2ZKdt5ev49O/eZJfPTyLBkt2JElSkcvXDP4vgcHABSmlU1NKl6SUjiUT9EcBP26vg4g4ALgUWAqMSSmdnVK6GBgP/AI4HPh2i3MqgRuBHsAnU0pnpJS+DRwC/Dl7zoV5ukeViMqK4KIJo7jx7IPo16MagPqGxJV3v8oXf/sMS1evL/AIJUmSttxWB/yI2A2YANSRmXHPdRmwCjgzInq209VpQADXp5RmN+5MKSXgu0ACvpwN9Y2OAvYCHkkp3ZFzTgPwrezH8yMa52ql9x09ajB3XXAE42reryK7b/o7fOyax3hh3tICjkySJGnL5WMG/9js9p5ssG6SUloBPE5mhn18O/00vn50dsuGbD+LyfyWYN9Wrv23Vs6ZDcwEaoDd2rm2ytSwvt35wxfH84Ujdm3aN3/pGj75q8ncPLmOzPdLSZKk4pGPgD8qu53ZRvtr2e3IdvpZnN3u2rIhInoBA7Mf99wG1yYiprb20+J6KkHVlRVc+rG9ue7McfTqVgXAhvrEZXe8zNdufY4VazcUeISSJEkdl4+A3ye7XdZGe+P+vu30Mym7PS8ialu0/YhM+Q5A7qo8+bq2xPGjh3DX149gn+G9m/bdNW0hp/zicaYvXF7AkUmSJHXc9lgHvzGYb7LWIaU0GbiOTICfFhE3RsRVEfEk8DXg5eyh9fm+dvb641r7AV7djOupyO0yoAd/Ov8wPjt+l6Z9cxav4tRrH+ePz8wr4MgkSZI6Jh8Bv3GWvE8b7b1bHNemlNL5wLnAK8DpwPnAeuB44MXsYe9si2tLjbpVV/KjU/flZ/+4Pz26ZJ7pXrexgW/9aRrf/J8XWLN+c75jSpIkbV/5CPgzstu26tz3yG7bqpNvJqU0MaU0PqXUM/tzZErpPuDQ7CFPb6trS7n+Yf/h3PG1DzFyxx2a9v1p6puceu3jvPim3xklSVLnlI+A/2B2OyEimvWXfTj2cGAN8OSWXiAiJpBZDefhlNL8nKYHstuPtnLObmSC/xu0sjKP1BG7D96B2796OB8fO7xp34y3V/AP1z7GD+542QdwJUlSp7PVAT+lNAu4B6gFvtqi+XKgJ3BLSmlV486I2DMiPrA6TUT0bmXfCODXZGrvL2nR/DAwHTgyIk7JOacC+Nfsx18l1zrUVujRpYqrPjWGf/3EvnStyvwn05Dgpsl1fPgnD3P3iwtdTlOSJHUakY9gkg3hk8msU/8XMqH7EOAYMuUxh6WU3s05PgGklKJFP/9DZqZ+KvAesDtwMlANnJdSuqmVax9CZia/GvgTMBc4DjiQzBr8x6WU1m3FvU0dO3bs2KlTp25pFyohdYtX8c9/eYlHX1vcbP8xowZxxT/sw879exRoZJIkqdiNGzeOZ5999tnsQi9bLC+r6GRn8Q8EbiIT7C8GRgDXAIfmhvt2TAI2kHnA9pvAYcCfgbGthfvstZ8CDiLzxWICcCGZh26vAD6yNeFeaql2YE9u+fzBXPPpAxi4Q9em/Q/OWMRHrn6Y/3xoFhvqGzbRgyRJ0raVlxn8UuYMvtqybM0G/v3vr/JfT80l9z+jUTv24sen7cOBtf0LNzhJklR0OtUMvlSO+nSv5ken7sufv3wYew19//GRGW+v4JO/eoJL/jyNpavXF3CEkiSpHBnwpa00dpd+3Pm1w7n0xL2a1s0H+MPT8zjuqoe57dk3fQhXkiRtNwZ8KQ+qKiv4wpG7ce9FR/GRvXds2v/uqvVc9McX+Mz1TzFr0coCjlCSJJULA76UR8P7duc3Zx3Ir88cx7A+3Zr2T571Lif89FF+cu9M1m7wTbiSJGnbMeBL28CE0UO496KjOO9Du1JZkVkNdn19A9fc/xon/OxRHmuxzKYkSVK+GPClbaRn1yq+d9Le3PG1wxmzc9+m/XMWr+KzNzzFP/3hORatcBVXSZKUXwZ8aRsbPawPt335MH546j706lbVtP/25xdw3FUP8fun5tLQ4EO4kiQpPwz40nZQWRGcOb6G+y8+ipPHDGvav3ztRr77vy/yyV9NZvrC5QUcoSRJKhUGfGk7GtyrGz//9AHc8vmD2aV/j6b9z85dykk/f4x/+et0Vq/fWMARSpKkYmfAlwrgyJGDuOfCI/n6sbtTXZl5CLe+IXHdI7P5yE8e4f7pbxd4hJIkqVgZ8KUC6VZdycUTRnH3N47gkF37N+2fv3QN5978DJ/+9ZM8PHORL8mSJEmbxYAvFdjug3vxhy+O598/uR/9elQ37X9i9rt8buIUTrzmMf7y/Hw21jcUcJSSJKlYGPClTiAi+NSBO/PAxUfzjwft3LR2PsD0hcv5xh+e5+j/eIibHp9jjb4kSdokA77UifTr2YUrP7EfD33zaM4+rJZu1e//J/rme2v4wZ2vcPiVD3D1vTNZsmp9AUcqSZI6KwO+1Ant3L8HPzhlNJMvOY4LPzyS/j27NLW9t3oDP7v/NQ678n4u+8tLzFuyuoAjlSRJnY0BX+rE+vfswjc+vAePf/tYrviH0ezUr3tT29oNDdz8xBsc/R8PccGtz/HygmUFHKkkSeosDPhSEejepZKzDq3loW8ezTWfPoC9h/ZuaqtvSNzxwgI+ds1jnHnDUzz++mJX3pEkqYxVFXoAkjquqrKCU8YM4+T9hvLY64v51cOzePz1d5vaH31tMY++tph9h/fhS0ftxgn7DG32wK4kSSp9BnypCEUER+wxiCP2GMSLby7jukdm8dcXF9KQnbh/cf4yvvb759il/wy+cORufGrcTnSrrizsoCVJ0nZhiY5U5PbdqQ+/OGMsD37zaM4cX0PXqvf/s567ZDX/fPtLHH7lA/z8/tdYutqVdyRJKnUGfKlE1AzoyQ9P3YfHLzmWC47dnT7d339p1rur1nPVvTM57MoHuOLOV5i/dE0BRypJkrYlA75UYgbu0JWLJoxi8iXH8v2T9mZ43/dX3lm9vp6Jj8/hqH97kIv++3mmL1xewJFKkqRtwRp8qUT17FrF5z+0K2ceWsOkaQu47uHZvPrWCgA2NiRue24+tz03n4N37c9Zh9Zw/OghVFf6nV+SpGJnwJdKXHVlBacdsBOn7j+ch2Yu4rqHZ/Hk7CVN7VPmLGHKnCXs2LsrZxxcw6cP2ZnBvboVcMSSJGlrGPClMhERHDNqMMeMGsxzc9/j+sfm8PeX3mJjdumdt5ev4+r7ZvKLB1/jhH2GctahNYyr6UeEy2xKklRMDPhSGTpgl35ce0Y/3l6+lt8/NZffT5nLohXrANhQn3lx1h0vLGDvob353GE1nDJmON27uMymJEnFIHzj5aZFxNSxY8eOnTp1aqGHIm0z6zc28LeX3+K3T9TxdN17H2jv072a0w/cic+Or6FmQM/tP0BJksrAuHHjePbZZ59NKY3bmn6cwZdEl6rMG3JPGTOMlxcs47dPvMHtz89n7YYGAJat2cBvHp3D9Y/N4ZhRgznz0BqO2mMQFb4lV5KkTseAL6mZ0cP6cOUn9uOSE/bkf555k98++QZzl6wGICV44NV3eODVd6gd0IPPjq/hU+N2pk+P6nZ6lSRJ24slOu2wREflrqEh8fDMRdz8RB0PzVj0gfbu1ZWcesAwzhxfy97Dem//AUqSVCIs0ZG0XVRUBMfsOZhj9hxM3eJV/O7JN/jjM/NYvnYjAGs21HPrlHncOmUeB9f256zDXFNfkqRCMuBL6rDagT353kl7c9GEkfzl+QXcPLmu6eVZAFPqljClbgmDe3XljEN24YyDd2Fwb9fUlyRpe7JEpx2W6EhtSynxzBvvcfPkOv6Ws6Z+o6qK4Ph9hvDxA4ZzxB6D6FLlrL4kSW2xREdSwUUEB9X256Da/q2uqb+xIXHXtIXcNW0hfbpXc8I+QzhlzDAO2W0Ala7AI0nSNuEMfjucwZc2z/qNDfz95be45YnW19QHGNSrKx/bdyin7D+MA3bu69tyJUnCGXxJnVSXqgpOHjOMk8cMY/rC5dz+/HwmvbCQ+UvXNB2zaMU6bppcx02T69ipX/fM8fsNY6+hvQz7kiRtJWfw2+EMvrT1GhoSz817jzueX8BdLy5k8cr1rR63++AdOHm/YZyy/zB2HegbcyVJ5SVfM/gG/HYY8KX82ljfwJOzl3DHC/O5+6W3WJFdbrOlfYb35pQxwzhpv2EM69t9O49SkqTtL18BP29LWkTEThExMSIWRMS6iKiLiJ9GRL/N7OdDEfGX7PlrI2JuRPw1Ij7axvGVEfGZiHg0It6KiNURMTMiboyI0fm5O0n5UlVZwYf2GMi/fXIMz3zvw/zmrAM5ecwwuldXNjvupfnL+X9/fZXDrnyAT/1qMr99oo7FK9cVZtCSJBWRvMzgR8QIYDIwGPgL8CpwMHAMMAM4PKX0bgf6+TLwS2AV8L/Am8BOwMeBHsD3Uko/bnHOfwOnZ4+9E1gB7At8FNgAnJBSemAr7s0ZfGk7WL1+I/dNf4c7X1jAwzMWsb6+4QPHVFYEh40YwMljhnH86CH06V5dgJFKkrRtdKoSnYj4OzABuCCl9POc/T8BLgSuSymd304f1cAioCuwf0ppRk7bXsBzQAPQL6W0Lrv/IGAK8DJwcEppdc455wATgQdTSsduxb0Z8KXtbNmaDfz95be484UFPP76Yhpa+Z+pLpUVHD1qECePGcaH99qR7l0qP3iQJElFpNOsohMRu5EJ93XAtS2aLwO+CJwZERenlFZtoqv+QB9gWm64B0gpTY+ImWRm5ncAGn9Pv1t2e39uuM/6S3Y7aDNuR1In0Kd7NacfuDOnH7gzi1as4+6XFnLH8wt45o33l91cX9/APa+8zT2vvE2PLpUct9eOnLTfUI4aOYhu1YZ9SVL5yscymY2z4/eklJr9Tj2ltCIiHifzBWA8cP8m+nmHzAz+yIjYI6X0WmNDRIwE9gCeb1Hq83LjGCKie0ppTU7bSdntfZt9R5I6jUG9unLWobWcdWgt85euYdILC7hz2gJemr+86ZjV6+u584UF3PnCAnp1reIje+/ISWOG8qHdfXuuJKn85CPgj8puZ7bR/hqZgD+STQT8lFKKiK8CvwOmRsT/AguA4cBpZML8P7Y456WIuJpMGdCrETGJTA3+aDI1+H8AvteRm4iItmpw9uzI+ZK2veF9u/Olo0bwpaNGMGvRSia9sJA7XpjPrEXv/3JwxbqN3PbcfG57bj59uldz/OgdOWm/YRw2YgBVlYZ9SVLpy0fA75PdLmujvXF/3/Y6Sin9T0QsAG4Fzsppehu4EZjdyjkXRcQM4GrgKzlNU4Gb2ykLklSkRgzagW98eA8uOG53pi9cwaRpC5g0bSFzl7xfrbdszQb++Myb/PGZN+nfswsf3WcIJ+03lEN2HUBlhS/UkiSVpu3xJtvG/xdt92neiPgs8BvgNuCHwBtADfDPwC+Ao8ismNN4fAA/IxPsv0dm9n8psD+ZwH93RHwtpdTy2YAPaOthhuzM/tj2zpdUGBHB3sN6s/ew3vzf40fx4vxl3DVtIZOmNX977pJV6/n9U3P5/VNzGbhDV07cdwgn7TeMA2v6UWHYlySVkHwE/MYZ+j5ttPducVyrsnX2E4FpwJk59fyvRsSZZEqBPhURR6eUHsq2fQ74OnB1SunKnO4ei4iTycz4XxkRN6eUVm7OTUkqPhHBfjv1Zb+d+nLJCXvy3LylTHphIXe9uIC3l7+/hv7ileu45Yk3uOWJNxjSuxsn7juUk8YM5YCd+5KZN5AkqXjlI+A3rngzso32PbLbtmr0G00AqoGHW3lYtyEiHgHGZX8eyjY1Pkj7YMvOUkpvRcSrwAFkvhy4zqVURiKCsbv0Y+wu/fjex/bimTfeY9K0Bfz1xbeavTDrreVrmfj4HCY+Pofhfbtz0n5DOWm/YewzvLdhX5JUlPIR8BvD9YSIqMgN5xHRCzgcWAM82U4/XbPbtpa1bNy/fivPkVRmKiqCg3ftz8G79ueyk0fz1Jx3mTRtIX976S2WrHr/fx7mL13DdY/M5rpHZlM7oAcfy4b9PYf0MuxLkorGVi8pkVKaBdwD1AJfbdF8OdATuCX3YdeI2DMiWq5O82h2+8mI2C+3ISL2Bz5Jpo7/gVbOuSgi+rQ453wyb8F9C3hlM29LUonKvA13IP/vtH156rvHccvnD+b0A3eid7fm8x11767m2gdnccLPHuXDP3mYq++dyatvLScfLweUJGlbytebbEcAk4HBZF4wNR04BDiGTGnOYbnr10dEAkgpRYt+JgLnkJlx/18yD9nWAqcCXYCfppQuzDl+B+BxYD8y6+jfQeYh27Fk1uevB05PKd22Fffmm2ylMrB+YwOPvb6ISS8s5J5X3mbluo2tHjekdzeO2GMgR40axId2H0jfHl2280glSaUqX2+yzUvAB4iInYEryKw/PwBYCNwOXJ5SWtLi2LYCfpB5cPZsYAzQC1gOPAf8JqX0h1auuwNwEfBxMvX+Xci8MOsx4D9SSlO28r4M+FKZWbuhnodnLmLStIXcP/1tVq+vb/W4ioAxO/flyD0GceTIQey/c1+X35QkbbFOF/BLlQFfKm9r1tfz4Ix3uOvFhTw6cxHL17Y+sw/Qp3s1H9p9IEeNzAT+IX26bceRSpKKXb4C/vZYB1+Silb3LpWcuO9QTtx3KBvrG3jhzWU8MnMRj7y2iBfmLaUhZ45k2ZoN3PXiQu56cSEAI3fcoSnsH1Tbn27VlQW6C0lSOTHgS1IHVVVWMK6mH+Nq+nHhR0aydPV6Hnt9MQ/PyAT+3LX2AWa+vZKZb6/kN4/OoVt1BeN3G8CRewziqFGD2G1gT1fmkSRtEwZ8SdpCfXt04aT9hnHSfsNIKTHz7ZU8PPMdHpm5mClzlrC+/v1Xeqzd0MBDMxbx0IxFMAmG9+3OkSMHcdTIQRy2+wB6d6su4J1IkkqJAV+S8iAiGDWkF6OG9OKLR45g9fqNPDV7CQ/PXMQjMxcxe/GqZsfPX7qGW6fM5dYpc6msiP/f3r1Hx3nXdx5/f0f3uWh0syTrYssX2Y7txMQOJCEBksAG2C4spKGnfyQUlrAHSkvZA93dA+1S6KFlT2mXAmXpgYW2sKe0W6DsdoFmCyEJSSDBTuw4vsuWLMmSJY3uo7vmt388j8ajy8hOdH/0eZ0z57F+z0XPnJ9m/Jlnfs/3x5Ftpbx+TwWv37OFAzVx3awrIiKvmAK+iMgKCOfncu++Su7dVwlAa+9IOuw/3ZSYVYZzOuV4trmXZ5t7+dyj5yguzOWOneW8dlc5d+6qYE9VVMN5RETkhingi4isgvqyMA/dsZ2H7tjO5HSKYy19PHG+m8fPdXOyfXDWtoNjUzx66iqPnroKQEU03w/8Fdy5q5yG8rACv4iIZKWALyKyyvJyQty+s5zbd5bzu2/eR8/wOE+e7+aJcz08daGHrqHZN+v2DE/wTyc6+KcTXnWerfFC7tx1LfDXlhStxdMQEZF1SgFfRGSNVUQLeOetdbzz1jqcczR1J3mmqYenmxL8/GKCvpHJWdt3DIzx3WPtfPdYOwDby8Pp4Tx37ixnS6xgLZ6GiIisEwr4IiLriJmxuzLK7sooD9/ZQCrlONM5xNNNPTzTlODZS70Mjc+ebKslMUJLYoS/fbYV8Orv37nTC/x37CyjJJy/Fk9FRETWiAK+iMg6FgoZ+2uK2V9TzCOv28nUdIqTVwbTgf+55l7GJlOz9pmpv//Xz7RgBgdqirnTH8P/6h1lRAv01i8iEmR6lxcR2UByc0K8qr6EV9WX8Jv37GZ8aprjrQM87Q/peeFy/6z6+87ByfZBTrYP8tUnL5ETMg7VxXn1jjKObPMm7SqPakiPiEiQKOCLiGxgBbk5vGZHGa/ZUcZH3gSjE9McbelLB/4X2weYTrn09tMpx7HL/Ry73J9u21ER4bAf9o9sL6WxMkpIdfhFRDYsBXwRkQApys/h7sYK7m6sAGBobJLnmnt5+kKCp5sSnOoYnLfPpZ4kl3qSfOdYGwCxwtxZgf9QfYmG9YiIbCB6xxYRCbBYYR737avivn1VAPQlJ3i2uZdjLX0cbenjRNvArCE9AENjUzx+zqvRDxAy2FddzG0NXuA/vK2UutIi1eIXEVmnFPBFRDaR0kg+bz5QzZsPVAMwPjXNyfbBdOD/ZUsfPcOz6/CnHJzqGORUxyB/80wLAJWxgvQV/iPbSzlQEyc/N7Tqz0dEROZTwBcR2cQKcnPSIf39gHOO1t5Rjl7u5WhLH0db+jnTOYhzs/frGhrnhyc7+eHJTv84IW6pi3N4eym3bS/j8LYS3bwrIrJGFPBFRCTNzNhWHmZbeZh33loHeOP4X2jt9wN/H89f7md4Ti3+8akUzzX38VxzH3/JRcCbgOtQXQm31MU5VF/CgZpiwvn6b0dEZKXpnVZERBYVK8zjdY1beF3jFsCrxHO+a4hfNvd5Q3su99GSGJm338wEXP/7+BXAG8u/pyqWDvyH6krYWx0jL0dDe0RElpMCvoiIvCw5IWNfdTH7qn9q0hkAABolSURBVIt56I7tAHQPjXPscl/6Kv+LC9y8m3JwpnOIM51D/P0vvYo9+bkh9m8t5lBdnFvqSjhUH2dnhcp0iogshQK+iIgs2ZZYwbybd890DHGirZ/jbQMcb+3nQvfwvLH8E1MpXmjt54XWfsC7gTdWkMvB2ji31MfTQ3xqS1S1R0TkRingi4jIsivIzfGG4dSX8LDfNjw+xcn2gVmhv61vdN6+Q+NTPHMxwTMXE+m28kj+rKE9t9TFdROviEgWCvgiIrIqogW53LGznDt2lqfbEsPjnGgf4ETrTPDvp2d4Yt6+ieQEj53t5rGz3em22pIiDtXHubm2hJtr4xysLaYknL8qz0VEZD1TwBcRkTVTHi3g3r2V3Lu3EvDKdF4ZGONEq3eV/0RbPy+2DTA0p2oPQHv/KO39o/zgxc50W31ZETfXKvSLyOamgC8iIuuGmVFbUkRtSRFvvXkrAKmU42JPkhNt/ZxoG+B4Wz8vXRlkYio1b//W3lFaexX6RWRzU8AXEZF1LRQydldG2V0Z5YHDXm3+yekUZzuHONE2wIvtA5xsH+BM5yCT027e/gr9IrLZKOCLiMiGk5cT4mBtnIO18XTb+NQ05zqHebFdoV9ENjcFfBERCYSC3Bxurotzc93yh/6DtXEO1MQ5UFNMhar3iMg6p4AvIiKBtRKhv7q4kIO1xeyviXOwppgDtXFq4oWq0y8i64YCvoiIbCpLDf2dg2N0Do7xL6e70m2l4TwO1sbZX1PMQf9Kf0N5RDPyisiaUMAXEZFN73qh/6UrA5y8MsiZjkHGF6je0zcyyZPne3jyfE+6LZKfw/6a4vTQnoO1cXZXRsnLCa3KcxKRzUsBX0REZAELhf6p6RRN3Ukv8LcP8tKVAU5dGVywTn9yYprnmvt4rrkv3ZafG2JfdWxW6N9XHaMwL2dVnpOIbA4K+CIiIjcoNyfE3uoYe6tjPHDYa0ulHK19I+nAf/LKIC+1D5BIzp+Rd2IqxYm2AU60DaTbckLG7i1RDtQUc9PWmUeMct3MKyKvkAK+iIjIEoRCxvbyCNvLI/zKLd7kXM45rg6Oc7J9gJeuDHLSv9Lf3j86b//plOPs1SHOXh2C59vT7VXFBRmBv5j9W2PsqIiSo3H9InIdCvgiIiLLzMyojhdSHS/kTfur0u19yYl04H/Jv9J/KZHEzb+Xl6uD41wd7OanZ7vTbYV5IfZWxa6F/ppi9lXHiBXmrcbTEpENQgFfRERklZRG8rm7sYK7GyvSbcPjU5zuGOTUlUFOd3iPM51DC97MOzaZ4njbAMczhviAV6//pmov8HtX+4upKy1S6U6RTUoBX0REZA1FC3J5dUMZr24oS7dNTadoTiQ51TGUDv2nrgzSNTS+4DFm6vU/eupqui1WkMu+rTH2Zwzz2asbekU2BQV8ERGRdSY3J8Tuyhi7K2O8/VBNuj0xPM7pzNDfMciFrmGmUvPH+AyNT82r4hMyaCiPsKcqxp7qGHurYuytjtJQHiFX5TtFAkMBX0REZIMojxZwd2PBrCE+41PTXOganhf8+0cm5+2fcnCxJ8nFniQ/euna7Lz5OSF2VUbZWxVNB/89VTFqS4o0WZfIBrRsAd/M6oBPA28ByoEO4B+BTznn+hbbd85x7gZ+FzgEVANdwEngC865Hy2y39uBDwK3AcX+fs8Df+Sc+/kreU4iIiLrXUFujl9X/1q9fuccnYNjfuAf4lTHIKevDGa9oXdiOpX+cJApkp9DY5Uf+GeCf3WULdECje8XWceWJeCb2S7gaaAS+D5wBngN8DvAW8zsLudc4gaO80Hgy0AS+B7QBtQBDwBvNbPfc859Zs4+IeArwPuBVuC7QAKoAu4AjgAK+CIismmYGVvjRWyNF3HfvmtVfEYnpmnqHuZM5xDnrg5x1l92DIwteJzkxDQvtPbzQmv/rPbScB57qmLpOQH2VsVorIoRL1I1H5H1YLmu4H8ZL9x/2Dn3xZlGM/sz4D8AnwE+sNgBzCwP+GNgDDjinDubse6P8K7Gf8LMPuecy7zL6KN44f6bwCPOuVkzi/jHFRER2fSK8nM4WBvnYG18VvvA6CTn/Vr85zq95dnOIfoWGOYD0DcyyS8u9fKLS72z2rfGC73x/VVRGitj7K6KsrsySrHKeIqsKnMLfVf3cg5gthNoApqBXc65VMa6GN5QHQMqnXPJRY5TBXQCJ5xzhxZYfwK4GaiY+TbAzIqBdqAf2D0n+C8LMzt6+PDhw0ePHl3uQ4uIiKxbzjl6hidmXemf+QCQnJh+WcfaGi9kd6UX+hurojT6/46HFfxFMh05coRjx44dc84dWcpxluMK/n3+8tHMcA/gnBsys6eA+/GGy/x4keN0Ad3AHjNrdM6dn1lhZnuARuCFOUN93g5E8YbohMzsQWA3MAT8zDl3fGlPTUREZHMyM7bECtgSK+Cu3ddu6k2lHFcGRv3gP8y5q0Oc6RyiqWuYien5tfsBOgbG6BgY48nzPbPaK2MFfuCP+R8AojRWxSiL5K/ocxMJuuUI+Hv95bks68/jBfw9LBLwnXPOzD4EfAs4ambfA64AtcA7gZeAX5+z26v95SRwGtieudLMvgO82zk3cr0nYWbZLtHvu96+IiIim0UoZNSVhqkrDc8a3+/V7h/h3NUhzl8d5nzXEBe6hrnYncwa/LuGxukaGuepC7Nv0yuP5KeDf6M/zKexMkZFNF8394rcgOUI+DMD+QayrJ9pL7negZxz/8vMrgB/C7w7Y9VV4BvAxTm7VPrL/4g3Rv/XgFPAfuAvgF8FhoH3XO93i4iIyCvn1e73wjg3X2ufmk7R0jvC+avDXOga4tzVYc53DdPUPczEArP1AiSSEyQu9vLzi7PH+JeG866N7d8STf++rfFCBX+RDKtRB3/mFXfdwf5m9hDwVbxKOH8ItOBdlf994EvAG/BC/IyZ6fhGgbc552aK+j7rl808BzxsZp9wzrUv9ruzjXXyr+wfvt65i4iIyHy5OSF2bYmya0sUr/q1ZzrlaO0d4XyXN8znQte1q/5jkwsH/76RSZ5t7uXZ5tnBP5yf4/+OCLsrvd+1uzLK9vII+bmawEs2n+UI+DNX6ONZ1hfP2W5B/jj7rwMngIczxvOfMbOH8YYCvcvM7nHO/dRfN1Nf/+cZ4R4A51yHmf0CeCNebfxFA76IiIisnpyQ0VARoaEiwr/af22oTyrlaO/3xvif7xpOX/k/3zXMSJabe0cmpnmxfYAX22dHjZyQsb0szK6M0L9rS4RdquwjAbccAX+mnOWeLOsb/WW2Mfoz7gfygMcXuFk3ZWZP4NW0PwL8dM7vnl2g95qZDwBF1/ndIiIisg6EQkZ9WZj6sjBvvGl28L8yMOqH/iGaupJc6B7mQtcwA6MLl/OcTrn0zL3/j6uz1lXGCmZd7Z9ZVhVrEi/Z+JYj4D/mL+83s9ACZTLvwhtCc73Jpgr85ZYs62faM+vcz9y0eyDLPjPtzdf53SIiIrKOZd7ce+/eynS7c45EcoKmruF04G/qTtLUNUx7/2jW483c4Pt00+wbfKMFuemr/DPDfnZuibKtLExhXk6Wo4msL0sO+M65JjN7FO8K/IeAL2as/hQQAf4yswa+me3z9z2Tse2T/vJBfzKrExnbvwp4EG8c/08yfvdxvwznXWb2iHPuaxn7PALchFej/7mlPk8RERFZf8yMimgBFdECbt9ZPmvdyMQUF7uTNKWDv7ds7hnJWtlneHyK420DHG+bPdwnZFBbWsTOiig7KiLp4L+jIqKbfGXdWa6bbH8TeBr4gpm9Ea9k5e3AvXhDcz4xZ/vT/jL9anDOPWtm3wDeCzznl8lsARqAdwD5wOedcy/NOdb7gJ8BXzWzB/DKae4H/jUwArzHOffyZuQQERGRDS+cn7vgzL1T0yla+0bTV/0zr/4PjU0teKyUg9beUVp7R3n8XPesdUV5OeyoiLBjS4RdFdeC/84tEWIa6y9rYFkCvn8V/zbg08Bb8MJ1B/AF4FPOud7F9s/wPuAJvLKWbwZiwCB+gHfOfXuB333WzA4DnwTeCrwJ6MUrtfmHzrnTc/cRERGRzSs3J+QF8ooIb+LaOH/nHN3D4+nx/U1dw1zsSXKpZ5i2vlFclnqAo5PTnOoY5FTH4Lx1W2IF1674V1wL/vVlYfJyVOFHVoa5bH+tAnhlMg8fPnz46NFs82CJiIhI0I1NTnO5d4SL3d4Y/0s9SS52ex8A+kcWvsl3MbkhY1tZmJ1bvA8a28u9ZUNFhK3FhYRCGvKzGR05coRjx44dy1a+/UatRh18ERERkQ2tMC+HPVUx9lTF5q3rS05wscebtfeiH/wv9SQXHes/lVHhZ6783BDby8I0VMyE/zA7yr3wX63wLzdAAV9ERERkCUoj+RyJlHFke9ms9umU40r/KE3dXvi/1JNMfxDoGBjLeryJqZRXDrRreN66gtwQ28vDNJRfu/LfUBFmR0WEqpjCv3gU8EVERERWQE5GTf979s5eNzIx5Q/zSdLck6Q5MUJzwvt3Ijmx8AGB8akU564Oc+7q/PBfmBdie5kX+BsqIjSUR9IfBFTff3NRwBcRERFZZeH8XA7UxDlQE5+3bnBskpaeES75gb+5J8mlRJKWxAi9i4T/sckUZ68Ocfbq0Lx1RXk5bCsLe1f/KyJsK/O+BdheHqampIgcXfkPFAV8ERERkXWkuDCPm+vi3Fw3P/wPjEx6V/oT3hj/5oQ39Kc5sfjNvqOT01nDf16OUV/qhf/tfuhvKI+wrTxMfWmY/FxV+9loFPBFRERENoh4OI9D4RIO1ZfMW9c/MuEN9em5Fvpnfh4YzR7+J6czb/idXeM/ZLA1XkRDhR/+y8Lpcf/bysKE8xUl1yP1ioiIiEgAlITzeVU4n1ctEP77khO09I7Q4g/1aU4kuZwYoTkxQs/weNZjphy094/S3j/KUxcS89ZXxgrSV/sbMr4B2F4eIV6kSb7WigK+iIiISMCVRvIpjSwc/ofHp7ic8MN/xoeAlsQIVwayT/AF0DU0TtfQOM82z5/TtDScx7byiBf8Z135j1ARzddNvytIAV9ERERkE4sW5LK/ppj9NcXz1o1NTtPWN5oR+r1hP5d7R2jtHWEqlT39941M0jfSz/HW/nnrIvk5s672N5SH/W8BVOt/OSjgi4iIiMiCCvNy2F0ZZXdldN66qekUHQNjNM8M++nxvgG4nBihpTfJ2OTCk3wBJCemOdUxyKmOwXnr8nNDfpWf8LwPATUlReTl6Kbf61HAFxEREZGXLTcnlK7z/7rG2etSKUfX0PisMf/p4T89IwyNT2U97sRUigtdw1xYYKKvnJBRW1JEXWkR9aVh6suKqC8LU+f/e0tU9f5BAV9EREREllkoZFTHC6mOF3L7zvJZ65xz9PnlPjPH+8/c+LvYRF/TKcflXm+IEMy/6bcgN+SF/7LwtQ8Apdc+AMSL8jbFBwAFfBERERFZNWZGWSSfskg+h7eVzls/ODbpV/iZM+4/MULn4Niixx6fStHUnaSpO7ng+lhBLnVlYepLi9Kh3/sgEKautIhIQTCicTCehYiIiIgEQnFhHgdr4xysnT/R1+jENG19I7T1jdLa593o29p77d+DY9mH/gAMjU9xumOQ0wuM/Qcoj+RTV1pEXVmYnRURPnr/3mV5TqtNAV9ERERENoSi/Bwaq2I0VsUWXD8wOklbnxf622Y+APSN0trrfSgYnZxe9PiJ5ASJ5ATH2waoLytSwBcRERERWUvxojziRXEO1My/+u+cI5GcmBP6/W8Dekdo7x9lcvpa2c/60vBqnvqyUsAXERERkcAzMyqiBVREC7h1gbH/0ynH1cGx9AeAjTwTrwK+iIiIiGx6OSGjpqSImpIibl/rk1kizRQgIiIiIhIgCvgiIiIiIgGigC8iIiIiEiAK+CIiIiIiAaKALyIiIiISIAr4IiIiIiIBooAvIiIiIhIgCvgiIiIiIgGigC8iIiIiEiAK+CIiIiIiAaKALyIiIiISIAr4IiIiIiIBooAvIiIiIhIgCvgiIiIiIgGigC8iIiIiEiAK+CIiIiIiAWLOubU+h3XNzBJFRUVlN91001qfioiIiIgE2OnTpxkdHe11zpUv5TgK+NdhZpeAYqB5iYfa5y/PLPE4svGo7zcv9f3mpH7fvNT3m9dy9X0DMOic27GUgyjgrxIzOwrgnDuy1uciq0t9v3mp7zcn9fvmpb7fvNZb32sMvoiIiIhIgCjgi4iIiIgEiAK+iIiIiEiAKOCLiIiIiASIAr6IiIiISICoio6IiIiISIDoCr6IiIiISIAo4IuIiIiIBIgCvoiIiIhIgCjgi4iIiIgEiAK+iIiIiEiAKOCLiIiIiASIAr6IiIiISIAo4K8wM6szs6+b2RUzGzezZjP7vJmVrvW5ydKY2YNm9kUze9LMBs3Mmdm3rrPPa83sB2bWa2YjZnbCzD5iZjmrdd6yNGZWbmaPmNn3zOyCmY2a2YCZ/czM3mdmC76vqu+Dwcz+q5n92Mxa/b7vNbPnzeyTZlaeZR/1fQCZ2cP++74zs0eybPNvzOyn/nvEsJn9wsx+Y7XPVZbGz24uy6Mzyz5r+rrXRFcryMx2AU8DlcD3gTPAa4B7gbPAXc65xNqdoSyFmb0AHAKGgTZgH/A/nXMPZdn+3wLfAcaAvwN6gbcBe4F/cM69azXOW5bGzD4A/HegA3gMuAxUAQ8Acbw+fpfLeHNV3weHmU0Ax4BTQBcQAe4AbgOuAHc451oztlffB5CZ1QMvAjlAFHi/c+5rc7b5LeCLQAKv7yeAB4E64E+dcx9b1ZOWV8zMmoES4PMLrB52zn1uzvZr/7p3zumxQg/gnwEH/Pac9j/z27+y1ueox5L6916gETDgHr9Pv5Vl22K8MDAO3JbRXoj3IdABv77Wz0mPG+r3+/w36tCc9mq8sO+AX1XfB/MBFGZp/4zfl19W3wf74b/n/wvQBPyJ34+PzNmmAS/cJYCGjPZS4IK/z51r/Vz0uOE+bwaab3DbdfG61xCdFWJmO4H78f4o/mLO6k8CSeBhM4us8qnJMnHOPeacO+/8V+51PAhsAb7tnPtlxjHGgN/zf/zgCpymLDPn3E+cc//HOZea094JfMX/8Z6MVer7APH7bSF/7y8bM9rU98H0YbwP+u/F+798If8OKAC+5Jxrnml0zvUBf+T/+IEVPEdZO+vida+Av3Lu85ePLhAEhoCngDDeV7sSfDN/Dz9aYN0TwAjwWjMrWL1TkhUw6S+nMtrU95vD2/zliYw29X3AmNlNwGeBP3fOPbHIpov1/Q/nbCMbQ4GZPWRmHzez3zGze7OMp18Xr/vclTz4JrfXX57Lsv483hX+PcCPV+WMZC1l/Xtwzk2Z2SXgALATOL2aJybLw8xygXf7P2a+savvA8jMPoY39jqON/7+brxw/9mMzdT3AeK/xr+JNxTv49fZfLG+7zCzJFBnZmHn3MjynqmskGq8/s90ycze65x7PKNtXbzuFfBXTtxfDmRZP9NesgrnImtPfw/B91ngIPAD59w/Z7Sr74PpY3g3V8/4EfAe51x3Rpv6Plj+C3ArcLdzbvQ6295I30f87RTw179vAE8CLwFDeOH8t4B/D/zQzO50zh33t10Xr3sN0Vk75i9VxkhAfw8bmpl9GPgoXqWsh1/u7v5Sfb+BOOeqnXOGd1XvAbz/8J83s8Mv4zDq+w3CzF6Dd9X+T51zzyzHIf2l+n4DcM59yr//6qpzbsQ5d9I59wG8oilFwB+8jMOtSt8r4K+cmU9o8Szri+dsJ8Gmv4eAMrMPAX+OVzbxXudc75xN1PcB5v+H/z28IZflwN9krFbfB0DG0JxzwO/f4G432veDSzg1WXszhRVen9G2Ll73Cvgr56y/3JNl/UylhWxj9CVYsv49+P957MC7MfPiap6ULI2ZfQT4EnASL9wvNOGJ+n4TcM614H3IO2BmFX6z+j4Yonh9eBMwljnJEV5VPICv+m0zddIX6/uteMNz2jT+fsPr8peZFRHXxeteAX/lPOYv7587s6WZxYC7gFHg56t9YrImfuIv37LAutfjVVR62jk3vnqnJEthZv8J+G/AC3jhvivLpur7zaPGX077S/V9MIwD/yPL43l/m5/5P88M31ms7986ZxvZuO70l5lhfX287td68oAgP9BEV5vmwY1NdNWNJrwJxAPva3oH/BIou8626vuAPPBmq65eoD3EtYmunlLfb54H3tjrhSa62oEmugrEA6/izbz3eWA7XkVEB3w8o31dvO7N/6WyAsxsF15nVgLfxyuHdDveDKjngNc65xJrd4ayFGb2DuAd/o/VwJvxPsU/6bf1uIypyP3t/wHvTf/beFNXvx1/6mrg15xekOuemf0G8Fd4V2m/yMLjKJudc3+VsY/6PgD8IVl/glfLugkvvFUBb8C7ybYTeKNz7lTGPur7ADOzP8AbpvN+59zX5qz7beALeH8nfwdM4E2CVId3s+7HkHXP7+P/jDcy4xJeFZ1dwK/ghfYfAO90zk1k7LPmr3sF/BVmZvXAp/G+qikHOoB/BD7l5t+MJxtIxht7Ni3OuYY5+9wFfALva71CvCs5Xwe+4JybnncEWXduoN8BHnfO3TNnP/X9BmdmB/FmoLwLL6SV4M1keg74v3h9Oe99XX0fXIsFfH/92/BKqh7G+6bnFN7stn+9mucpr5yZvQFv1uFb8S7mRYB+vOGZ3wS+uVBYX+vXvQK+iIiIiEiA6CZbEREREZEAUcAXEREREQkQBXwRERERkQBRwBcRERERCRAFfBERERGRAFHAFxEREREJEAV8EREREZEAUcAXEREREQkQBXwRERERkQBRwBcRERERCRAFfBERERGRAFHAFxEREREJEAV8EREREZEAUcAXEREREQkQBXwRERERkQBRwBcRERERCRAFfBERERGRAPn/+HE63/auTMAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 250,
       "width": 380
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(n_range, gs.cv_results_[\"mean_train_score\"]);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Очевидно, что для&nbsp;$K=1$ на&nbsp;тренировочном наборе данных мы находим ровно ту же точку, что и нужно предсказать, и чем больше её соседей мы берём — тем меньше точность."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "На&nbsp;тестовом наборе данных картина сильно интереснее:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 250,
       "width": 380
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(n_range, gs.cv_results_[\"mean_test_score\"]);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Выходит, что сначала соседей слишком мало (высоко влияние выбросов), а затем количество соседей постепенно становится слишком велико, и среднее значение по&nbsp;этим соседям всё больше и больше оттягивает значение от&nbsp;истинного."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверим получившуюся модель:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean_absolute_error: 55.39857905041605\n",
      "median_absolute_error: 4.017000000000004\n",
      "r2_score: 0.8677873476991447\n"
     ]
    }
   ],
   "source": [
    "reg = KNeighborsRegressor(**gs.best_params_)\n",
    "reg.fit(X_train, y_train)\n",
    "test_model(reg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В&nbsp;целом получили примерно тот же результат. Очевидно, что проблема в&nbsp;том, что данный метод и так показал достаточно хороший результат для&nbsp;данной выборки."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Построим кривую обучения<cite data-cite=\"doc:sklearn\"></cite>:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_learning_curve(estimator, title, X, y, ylim=None, cv=None):\n",
    "    train_sizes=np.linspace(.1, 1.0, 5)\n",
    "\n",
    "    plt.figure()\n",
    "    plt.title(title)\n",
    "    if ylim is not None:\n",
    "        plt.ylim(*ylim)\n",
    "    plt.xlabel(\"Training examples\")\n",
    "    plt.ylabel(\"Score\")\n",
    "    train_sizes, train_scores, test_scores = learning_curve(\n",
    "        estimator, X, y, cv=cv, n_jobs=-1, train_sizes=train_sizes)\n",
    "    train_scores_mean = np.mean(train_scores, axis=1)\n",
    "    train_scores_std = np.std(train_scores, axis=1)\n",
    "    test_scores_mean = np.mean(test_scores, axis=1)\n",
    "    test_scores_std = np.std(test_scores, axis=1)\n",
    "    plt.grid()\n",
    "\n",
    "    plt.fill_between(train_sizes, train_scores_mean - train_scores_std,\n",
    "                     train_scores_mean + train_scores_std, alpha=0.1,\n",
    "                     color=\"r\")\n",
    "    plt.fill_between(train_sizes, test_scores_mean - test_scores_std,\n",
    "                     test_scores_mean + test_scores_std, alpha=0.1,\n",
    "                     color=\"g\")\n",
    "    plt.plot(train_sizes, train_scores_mean, 'o-', color=\"r\",\n",
    "             label=\"Training score\")\n",
    "    plt.plot(train_sizes, test_scores_mean, 'o-', color=\"g\",\n",
    "             label=\"Cross-validation score\")\n",
    "\n",
    "    plt.legend(loc=\"best\")\n",
    "    return plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 277,
       "width": 399
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_learning_curve(reg, str(gs.best_params_), X, y,\n",
    "                    cv=ShuffleSplit(n_splits=10));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Построим кривую валидации:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_validation_curve(estimator, title, X, y, \n",
    "                          param_name, param_range, cv, \n",
    "                          scoring=\"accuracy\"):\n",
    "                                                   \n",
    "    train_scores, test_scores = validation_curve(\n",
    "        estimator, X, y, param_name=param_name,\n",
    "        param_range=param_range,\n",
    "        cv=cv, scoring=scoring, n_jobs=-1)\n",
    "    train_scores_mean = np.mean(train_scores, axis=1)\n",
    "    train_scores_std = np.std(train_scores, axis=1)\n",
    "    test_scores_mean = np.mean(test_scores, axis=1)\n",
    "    test_scores_std = np.std(test_scores, axis=1)\n",
    "\n",
    "    plt.title(title)\n",
    "    plt.xlabel(param_name)\n",
    "    plt.ylabel(\"Score\")\n",
    "    plt.ylim(0.0, 1.1)\n",
    "    lw = 2\n",
    "    plt.plot(param_range, train_scores_mean, label=\"Training score\",\n",
    "                 color=\"darkorange\", lw=lw)\n",
    "    plt.fill_between(param_range, train_scores_mean - train_scores_std,\n",
    "                     train_scores_mean + train_scores_std, alpha=0.2,\n",
    "                     color=\"darkorange\", lw=lw)\n",
    "    plt.plot(param_range, test_scores_mean,\n",
    "                 label=\"Cross-validation score\",\n",
    "                 color=\"navy\", lw=lw)\n",
    "    plt.fill_between(param_range, test_scores_mean - test_scores_std,\n",
    "                     test_scores_mean + test_scores_std, alpha=0.2,\n",
    "                     color=\"navy\", lw=lw)\n",
    "    plt.legend(loc=\"best\")\n",
    "    return plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 277,
       "width": 387
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_validation_curve(KNeighborsRegressor(), \"knn\", X, y, \n",
    "                      param_name=\"n_neighbors\", param_range=n_range, \n",
    "                      cv=ShuffleSplit(n_splits=10), scoring=\"r2\");"
   ]
  }
 ],
 "metadata": {
  "authors": [
   {
    "name": "Лещев Артем Олегович"
   }
  ],
  "group": "ИУ5-24М",
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "lab_number": 4,
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  },
  "title": "Подготовка обучающей и тестовой выборки, кросс-валидация и подбор гиперпараметров на примере метода ближайших соседей"
 },
 "nbformat": 4,
 "nbformat_minor": 2
}