Final_project_churn_prediction_week7.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### КУРС ПРОФЕССИОНАЛЬНОЙ ПЕРЕПОДГОТОВКИ\n",
    "### Специализация: «МАШИННОЕ ОБУЧЕНИЕ И АНАЛИЗ ДАННЫХ»\n",
    "\n",
    "https://datasciencecourse.ru\n",
    "\n",
    "### Финальный проект"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Тема проекта: Прогнозирование оттока клиентов\n",
    "#### Автор: Бородастов Михаил Дмитриевич"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. Цели и задачи проекта\n",
    "\n",
    "#### Задача проекта:\n",
    "\n",
    "Задача прогнозирования оттока клиентов - это:\n",
    "- задача обучения с учителем (обучение по прецедентам)\n",
    "- задача бинарной классификации\n",
    "\n",
    "Задча приведена на сайте kaggle -\n",
    "https://www.kaggle.com/c/telecom-clients-prediction2/overview \n",
    "\n",
    "В задаче необходимо спрогнозировать клиентов телеком оператора, склонных к оттоку.\n",
    "\n",
    "#### Формальная постановка задачи:\n",
    "\n",
    "Дано:\n",
    "\n",
    "- обучающая выборка (X_train) из 40 000 объектов (клиенты телеком оператора)\n",
    "- ответы на объектах из обучающей выборки y_train (метки классов - 'отток' и 'не отток')\n",
    "- тестовая выборка (X_test) из 10 000 объектов\n",
    "- каждый объект представлен в виде вектора из 230 признаков\n",
    "\n",
    "Необходимо: \n",
    "\n",
    "- построить модель машинного обучения (алгоритм), которая наилучшим образом предсказывает ответы на объектах из обучающей выборки (этап обучения модели)\n",
    "\n",
    "\n",
    "- использовать построенную модель для получения предсказаний на тествовой выборке так (этап тестирования (валидации) модели)\n",
    "\n",
    "\n",
    "- при этом, качество модели и качество предсказаний оценивается по метрике roc_auc (площадь под ROC-кривой)\n",
    "\n",
    "#### Цель проекта:\n",
    "\n",
    "Применить на практике знания, полученные в рамках пройденной специализации, для каждой стадии машинного обучения при построении модели:\n",
    "\n",
    "Стадии:\n",
    "\n",
    "1) Описательный анализ данных\n",
    "\n",
    "(первичный анализ данных - типы данных, доли классов, корреляции, определение наиболее важных признаков, визуализация распределения объектов в разрезе признаков)\n",
    "\n",
    "2) Подготовка данных для построения модели \n",
    "\n",
    "(обработка пропусков, работа с категориальными признаками, выбор метрик для оценки качества (основной и дополнительных), разделение тренировочных данных на обучение и тест, выбор оптимальной стратегии проведения кросс-валидации)\n",
    "\n",
    "3) Построение baseline-моделей и выбор оптимальной \n",
    "\n",
    "(определение наилучшего класса моделей машинного обучения для решения поставленной задачи)\n",
    "\n",
    "4) Оптимизация модели \n",
    "\n",
    "(балансировка выборки (отбор объектов), отбор признаков, подбор оптимальных параметров модели)\n",
    "\n",
    "5) Построение финального прогноза\n",
    "\n",
    "6) Оценка потенциального эффекта от внедрения полученного решения\n",
    "\n",
    "#### Итог (применение на практике):\n",
    "\n",
    "Применив полученную модель на практике, можно определять клиентов, которые, с наибольшей вероятностью могут уйти от мобильного оператора. Для таких клиентов можно проводить специальные акции, направленные на увеличение лояльности к текущему мобильному оператору и пытаться сохранить данных клиентов в числе пользователей.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2. Методика измерения качества и критерий успеха\n",
    "\n",
    "#### Выбор основной и дополнительных метрик\n",
    "\n",
    "1) Ввиду того, что мы работаем с сильно несбалансированной выборкой:\n",
    "\n",
    "- обучающая выборка содержит только 7% объектов, относящихся к классу \"отток\"\n",
    "\n",
    "нам необходимо учитывать разную цену ошибок (некорректно использовать метрику accuracy - долю правильный ответов), следовательно:\n",
    "\n",
    "- мы должны использовать метрику на основе Precision и Recall\n",
    "\n",
    "т.к. задача заключается в определении людей склонных к оттоку, то очевидно, что ложные пропуски (False Negative) для нас имеют гораздо большую цену, чем ложные срабатывания (False Positive). За ложные пропуски отвечает Recall (полнота)\n",
    "\n",
    "В качестве вспомогательных метрик будем использовать: <b> Recall,F-меру </b>\n",
    "\n",
    "2) Мы хотим настраивать порог срабатывания классификатора и иметь возможность более гибко настраивать нашу модель\n",
    "\n",
    "Для учета качества оценки принадлежности объектов к тому или иному классу мы должны использовать ROC или PR кривые и находить площадь под ними.\n",
    "\n",
    "В данном случае, исходя из того, что выборка у нас несимметричная относительно распределения классов, то предпочтительней ориентироваться и настраивать наш алгоритм на метрику - площадь под ROC-кривой, которая не чувствительна к изменению баланса классов.\n",
    "\n",
    "В качестве основной метрики используем - <b> roc_auc (площадь под ROC-кривой) </b>\n",
    "\n",
    "#### Основной критерий успеха:\n",
    "\n",
    "Результатом выполнения данной работы является - прогноз для объектов из тестовой выборки, представляющий собой вероятности отнесения объекта к классу \"отток\", который должен быть загружен на сайт kaggle, где производится расчет достигнутого значения метрики roc_auc.\n",
    "\n",
    "По ссылке ниже приводится рейтинг с загруженныи решениями.\n",
    "\n",
    "https://www.kaggle.com/c/telecom-clients-prediction2/leaderboard\n",
    "\n",
    "Чем выше позиция в рейтинге, тем более качественная модель построена и тем выше предсказательная способность модели."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  3. Построение модели\n",
    "#### 3.1 Техническое описание решения "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1) Для обучения использовали все данные из тренировочного датасета 40 000 объектов (orange_small_churn_data.train.csv)\n",
    "\n",
    "2) В качестве hold_out выборки данные из тестовой выборки - 10 000 объектов (orange_small_churn_test_data.csv)\n",
    "\n",
    "3) Удалили признаки, полностью состоящие из пропусков (230 - 18 признаков = 212)\n",
    "\n",
    "4) Пропуски для катег-ых признаков заменяли <b>отдельной категорией</b>, для вещественных - <b>средними значениями</b>\n",
    "\n",
    "5) Категориальные признаки кодировали <b>'частотным методом'</b> (частота вхождения каждой подкатегории для соответствующего признака)\n",
    "\n",
    "6) Отбор объектов с помощью метода <b>ALLKNN</b> из библиотеки imblearn.under_sampling с параметром kind_sel='mode' (удалили 3024 объекта из 40 000)\n",
    "\n",
    "7) Отбор признаков с помощью <b>ExtraTreesClassifier</b> из sklearn.ensemble с параметрами по умолчанию (Отобрали 82 признака, для которых получили максимальное значение качества по метрике roc_auc с XGBClassifier)\n",
    "\n",
    "8) В качестве основного классификатора выбрали <b>XGBClassifier</b> из xgboost (реализация стохастического градиентного бустинга)\n",
    "\n",
    "9) обучали по кросс-валидации на 5 фолдов по методу <b>StratifiedKFold</b> с shuffle=True\n",
    "\n",
    "- Необходимость стратификации (Stratified)\n",
    "\n",
    "Т.к. распределение классов сильно несбалансированно, то мы должны исключить возможность сценария, когда при разбиении на фолды, у нас окажется блок, в который объекты одного из классов не попадут вовсе.\n",
    "\n",
    "- Необходимость перемешивания (shuffle=True)\n",
    "\n",
    "Мы не знаем по какому принципу представлены исходные данные,  есть вероятность, что они могут быть отсортированы в соответствии с каким-то порядком , например разные по своей природе объекты идут группами друг за другом, по этому, чтобы снизить вероятность, что наша модель, на одном из разбиений может получить выборку состоящую из объектов, которые попадут только на обучение, но будут полностью отсутствовать в тесте или наоборот, мы должны перемешивать данные.\n",
    "\n",
    "10) С помощью <b>RandomizedSearchCV и GridSearchCV</b> подобрали оптимальные параметры классификатора"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.2 Построение baseline модели:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Импортирование библиотек"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import scipy.stats\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import StratifiedKFold, GridSearchCV, RandomizedSearchCV,\\\n",
    "StratifiedShuffleSplit,learning_curve\n",
    "from sklearn.feature_extraction import DictVectorizer , FeatureHasher\n",
    "from sklearn.preprocessing import OneHotEncoder, LabelEncoder\n",
    "from sklearn.preprocessing import StandardScaler, Normalizer\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier , GradientBoostingClassifier, ExtraTreesClassifier\n",
    "\n",
    "from sklearn.metrics import precision_recall_curve, precision_score, recall_score ,\\\n",
    "plot_precision_recall_curve, roc_auc_score,roc_curve, f1_score,auc\n",
    "\n",
    "import xgboost as xgb\n",
    "\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier, GradientBoostingClassifier\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis, QuadraticDiscriminantAnalysis\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from xgboost import XGBClassifier , XGBRFClassifier\n",
    "from imblearn.under_sampling import AllKNN,TomekLinks,NearMiss\n",
    "from sklearn.feature_selection import VarianceThreshold\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В работе будем использовать два набора данных:\n",
    "\n",
    "1. data_merge - тренировчный набор данных из 40 000 объектов.\n",
    "\n",
    "\n",
    "- Разобьем его на train и hold_out.\n",
    "- Будем обучать модель на train по кросс-валидации и оценивать финальное качество на hold_out\n",
    "\n",
    "\n",
    "2. X_DATA - тренировочный набор данных из 40 000 объектов + 10 000 объектов из тестового набора.\n",
    "\n",
    "В данной задаче финальнуй оценку качества мы должны проводить на тестовой выборке с сайта Kaggle, которая и будет hold-out dataset-ом для нас, следовательно мы можем использовать больше данных для обучения на кросс-валидации. (40000 - обучение, 10000 - hold-out для валидации)\n",
    "\n",
    "Финальное решение будем строить по выборке X_DATA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### работа с data_merge выборкой"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [],
   "source": [
    "#считывание данных\n",
    "X_train=pd.read_csv('orange_small_churn_data.train.csv')\n",
    "y_train=pd.read_csv('orange_small_churn_labels.train.csv',header=None)\n",
    "\n",
    "#установим имя для колонки с ответами\n",
    "y_train=y_train.rename(columns={0:'labels'})\n",
    "\n",
    "#объединим матрицу объектов-признаков и вектор целеых меток\n",
    "data_merge=X_train\n",
    "data_merge['labels']=y_train\n",
    "\n",
    "#замена -1 класса на 0, для удобства восприятия\n",
    "data_merge.labels=data_merge.labels.replace(-1,0)\n",
    "\n",
    "real_features=list(data_merge.iloc[:,:190].columns.values) # числовые признаки\n",
    "categ_features=list(data_merge.iloc[:,190:230].columns.values) #категориальные признаки\n",
    "\n",
    "#убираем нулевые стоблы в категориальных признаках\n",
    "categ_features_not_null=categ_features.copy()\n",
    "categ_features_not_null.remove('Var209')\n",
    "categ_features_not_null.remove('Var230')\n",
    "\n",
    "#заполняем пустые значения категорией 'w/o_cat'\n",
    "X_cat_filled=data_merge[categ_features_not_null].fillna('w/o_cat')\n",
    "\n",
    "#определяем нулевые столбцы среди вещ признаков\n",
    "null_real_features=data_merge[real_features].loc[:,data_merge[real_features].\\\n",
    "                                                 isnull().all().values].columns\n",
    "\n",
    "#определяем ненулевые вещественные признаки\n",
    "real_features_not_null=list(pd.core.indexes.base.Index(real_features).drop(null_real_features))\n",
    "\n",
    "#заполняем пустые значения вещественных признаков средними значениями\n",
    "X_real_mean=data_merge[real_features_not_null].apply(lambda x: x.fillna(x.mean()))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### работа с X_DATA выборкой"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_orig=pd.read_csv('orange_small_churn_data.train.csv')\n",
    "y_train_orig=pd.read_csv('orange_small_churn_labels.train.csv',header=None)\n",
    "X_test_orig=pd.read_csv('orange_small_churn_test_data.csv', index_col='ID')\n",
    "\n",
    "#переименуем колонку с целевыми метками\n",
    "y_train_orig=y_train_orig.rename(columns={0:'labels'})\n",
    "\n",
    "#замена -1 класса на 0, для удобства восприятия\n",
    "y_train_orig.labels=y_train_orig.labels.replace(-1,0)\n",
    "\n",
    "X_DATA=pd.concat([X_train_orig,X_test_orig], axis=0,join='outer',ignore_index=True)\n",
    "\n",
    "real_features=list(X_DATA.iloc[:,:190].columns.values) # числовые признаки\n",
    "categ_features=list(X_DATA.iloc[:,190:230].columns.values) #категориальные признаки\n",
    "\n",
    "# удаляем нулевые категориальные признаки + заполняем пропуски на 'w/o_cat'\n",
    "categ_features_not_null=categ_features.copy()\n",
    "categ_features_not_null.remove('Var209')\n",
    "categ_features_not_null.remove('Var230')\n",
    "\n",
    "#заполняем категориальные признаки категорией 'w/o_cat'\n",
    "X_cat_filled=X_DATA[categ_features_not_null].fillna('w/o_cat')\n",
    "\n",
    "#заполняем категориальные признаки наиболее частой категорией\n",
    "X_cat_filled_2=X_DATA[categ_features_not_null].apply(lambda x: x.fillna(x.value_counts().index[0]))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### определение функций , используемых при построении модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [],
   "source": [
    "#функция расчета коэф. Краммера, p_value, корректность использования статистики Хи-2\n",
    "def V_Krammer_func(table):\n",
    "\n",
    "    chi2, p_value, n_dof, expected_values= chi2_stat=scipy.stats.chi2_contingency(table)\n",
    "    n=table.sum().sum()\n",
    "    min_from_shape=min(table.shape[0],table.shape[1])\n",
    "    V_Krammer=np.sqrt(chi2/(n*(min_from_shape-1)))\n",
    "    number_of_cells_less_five=len(expected_values[expected_values<5])\n",
    "    cells_share=number_of_cells_less_five/(expected_values.shape[0]*expected_values.shape[1])*100\n",
    "\n",
    "    if cells_share<=20 and n>=40:\n",
    "        condition_of_using=\"Ok\"\n",
    "    else:\n",
    "        condition_of_using='Not allowed'\n",
    "        \n",
    "    return V_Krammer, p_value, condition_of_using\n",
    "\n",
    "#функция расчета корреляции между вещественными и категориальными признаками\n",
    "def corr_for_real_and_y(X,y):\n",
    "    X_and_y=pd.concat([X,y],axis=1)\n",
    "    CORR_X_y=X.groupby(y).agg(lambda x: x.mean())\n",
    "    CORR_X_y.loc[\"diff\"]=CORR_X_y.apply(lambda x: x[1]-x[0])\n",
    "    CORR_DIFF=CORR_X_y.loc['diff'].to_frame()\n",
    "    CORR_DIFF['Corr_sign']=['+' if diff>=0 else '-' for diff in CORR_DIFF['diff']]\n",
    "    CORR_DIFF['diff']=CORR_DIFF['diff'].apply(abs)\n",
    "    CORR_DIFF=CORR_DIFF.sort_values('diff',ascending=False)\n",
    "    return CORR_DIFF\n",
    "\n",
    "# удаляем нулевые числовые признаки в зависимости от порога, который определяет допустимую долю\n",
    "# пропусков(nan-значений) для каждого признака + заполняем пропуски средними значениями\n",
    "\n",
    "def get_X_real_mean(X,share):\n",
    "    DF_with_indexes=pd.DataFrame((X.iloc[:40000,:190].isnull().sum()/\\\n",
    "                                  X.iloc[:40000,:190].shape[0]), columns=['n_null'])\\\n",
    "                                .sort_values('n_null',ascending=False)\n",
    "\n",
    "    real_features_not_null=DF_with_indexes[DF_with_indexes['n_null']<share].index\n",
    "\n",
    "    X_real_mean=X[real_features_not_null].apply(lambda x: x.fillna(x.mean()))\n",
    "    return X_real_mean\n",
    "\n",
    "#Определим функцию, которая в зависимости от заданного порога, будет отбирать вещественные\n",
    "#признаки по степени корреляции (елси порог равен 0, то используем все вещественные признаки, \n",
    "# если увеличиваем порог, то признаки, которые почти никак не коррелируют с целевой меткой \n",
    "#- удаляются из рассмторения)\n",
    "\n",
    "#базовый вариант для случая, когда hold-out выборка отделяется от исходного датасета\n",
    "def real_features(X_real,y,threshold=0):\n",
    "    top_corr_=corr_for_real_and_y(X_real,y)\n",
    "    if threshold:\n",
    "        return X_real[top_corr_[top_corr_['diff'] >= threshold].index]\n",
    "    else:\n",
    "        return X_real\n",
    "    \n",
    "\n",
    "# вариант для случая, когда hold-out берется с kaggle для построения прогноза\n",
    "def real_features_new(X_real,y,threshold=0):\n",
    "    \n",
    "    splitter=X_real.shape[0]-10000\n",
    "    \n",
    "    top_corr_=corr_for_real_and_y(X_real.iloc[0:splitter,],y)\n",
    "    if threshold:\n",
    "        return X_real[top_corr_[top_corr_['diff'] >= threshold].index]\n",
    "    else:\n",
    "        return X_real\n",
    "    \n",
    "#Определим функцию, которая в зависимости от заданного порога, будет отбирать вещественные признаки\n",
    "#по степени корреляции, и в зависимости от переданного DF с категориальными признаками, будем делать\n",
    "#соовтетствующее разделение на hold-out и train выборки.  \n",
    "\n",
    "\n",
    "#базовый вариант для случая, когда hold-out выборка отделяется от исходного датасета\n",
    "def hold_out_train(X_real, X_cat, y, real_threshold=0):\n",
    "    \n",
    "    X_real_=real_features(X_real,y,real_threshold)\n",
    "    \n",
    "    X=pd.concat([X_real_,X_cat],axis=1)\n",
    "    \n",
    "    DATA=pd.concat([X,y],axis=1)\n",
    "\n",
    "    strat = StratifiedShuffleSplit(n_splits=1,test_size=0.20, random_state=0)\n",
    "    train_ind, test_ind = next(strat.split(X,y))\n",
    "    \n",
    "    train=DATA.iloc[train_ind,:]\n",
    "    hold_out_dataset=DATA.iloc[test_ind,:]\n",
    "    \n",
    "    \n",
    "    return train,hold_out_dataset\n",
    "\n",
    "# вариант для случая, когда hold-out берется с kaggle для построения прогноза\n",
    "def hold_out_train_new(X_real, X_cat, y, real_threshold=0, scaler=None):\n",
    "    \n",
    "    splitter=X_real.shape[0]-10000\n",
    "    \n",
    "    X_real_=real_features_new(X_real,y,real_threshold)\n",
    "    \n",
    "    if type(scaler)==type(None):\n",
    "    \n",
    "        X=pd.concat([X_real_,X_cat],axis=1)\n",
    "        \n",
    "        train=X.iloc[0:splitter]\n",
    "        hold_out_dataset=X.iloc[splitter:,]\n",
    "        \n",
    "        return train,hold_out_dataset\n",
    "    \n",
    "    elif type(scaler)!=type(None):\n",
    "        \n",
    "        sclr=scaler\n",
    "        X_real_train=sclr.fit_transform(X_real_.iloc[0:splitter,])\n",
    "        X_real_hold_out=sclr.fit_transform(X_real_.iloc[splitter:,])\n",
    "        X_real_merge=np.vstack((X_real_train,X_real_hold_out))\n",
    "        X_real_=pd.DataFrame(X_real_merge,columns=X_real_.columns, index=X_real_.index)\n",
    "        X=pd.concat([X_real_,X_cat],axis=1)\n",
    "        train=X.iloc[0:splitter,]\n",
    "        hold_out_dataset=X.iloc[splitter:,]\n",
    "        \n",
    "        return train,hold_out_dataset\n",
    "\n",
    "#Определим функцию, которая будет кодировать категориальные признаки соответствующим образом\n",
    "def cat_to_real_encode(X_cat, endocer_way):\n",
    "    if endocer_way=='One-hot':\n",
    "        X_cat_OH=pd.get_dummies(X_cat, dtype=float)\n",
    "        return X_cat_OH\n",
    "    elif endocer_way=='hash':\n",
    "        n=X_cat.apply(lambda x: x.nunique()).sum()\n",
    "        hasher=FeatureHasher(n_features=n)\n",
    "        X_cat_hash=hasher.fit_transform(X_cat.T.to_dict().values()).toarray()\n",
    "        X_cat_hash_DF=pd.DataFrame(X_cat_hash)\n",
    "        return X_cat_hash_DF\n",
    "    elif endocer_way=='freq':\n",
    "        X_cat_FR=pd.DataFrame()\n",
    "        for feature in X_cat.columns:\n",
    "            X_cat_FR[feature]=X_cat[feature].map(X_cat.groupby(feature).size()/len(X_cat))\n",
    "        return X_cat_FR\n",
    "            \n",
    "#оформим алгоритм получения оценок для соответствующих метрик в виде функции\n",
    "def estimator(clf, X, y, X_hold_out=None, y_hold_out=None, n_folds=5,random=0,\\\n",
    "                                              print_graph='No',print_metrics='No'):\n",
    "    \n",
    "    stat = StratifiedKFold(n_splits=n_folds,shuffle=True,random_state=random)\n",
    "    clf=clf\n",
    "    \n",
    "    Precision_list=[]\n",
    "    Recall_list=[]\n",
    "    auc_prc_list=[]\n",
    "    f1_score_list=[]\n",
    "    roc_auc_list=[]\n",
    "    \n",
    "    if print_graph=='Yes':\n",
    "        fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(12, 8))\n",
    "\n",
    "    for train_ind, test_ind in stat.split(X,y):\n",
    "    \n",
    "        X_train = X.iloc[train_ind,:]\n",
    "        y_train = y.iloc[train_ind]\n",
    "    \n",
    "        X_test  = X.iloc[test_ind,:]\n",
    "        y_test  = y.iloc[test_ind]\n",
    "    \n",
    "        clf.fit(X_train,y_train)\n",
    "    \n",
    "        y_pred=clf.predict(X_test)\n",
    "        y_prob_to_1=clf.predict_proba(X_test)[:,1]\n",
    "    \n",
    "        #площаь под PR - кривой - основаня метрика определенная на прошлой неделе\n",
    "        Precision_cor, Recall_cor, _ = precision_recall_curve(y_test,y_prob_to_1)\n",
    "        auc_prc_list.append(auc(Recall_cor,Precision_cor))\n",
    "    \n",
    "        #площадь под ROC-кривой - основная метрика , по которой сравниваются решения на Kaggle\n",
    "        roc_auc_list.append(roc_auc_score(y_test,y_prob_to_1))\n",
    "    \n",
    "        #вычисления дополнительных метрик\n",
    "        Precision=precision_score(y_test,y_pred)\n",
    "        Precision_list.append(Precision)\n",
    "    \n",
    "        Recall=recall_score(y_test,y_pred)\n",
    "        Recall_list.append(Recall)\n",
    "    \n",
    "        f1_score_list.append(f1_score(y_test,y_pred))\n",
    "        \n",
    "        #построение PR и ROC кривых и печать дополнительных метрик:\n",
    "        if print_graph=='Yes':\n",
    "            \n",
    "            #график PR - кривой\n",
    "            plot_precision_recall_curve(clf,X_test,y_test, ax=axes[0])\n",
    "            axes[0].set_title('Precision-Recall кривые для k-фолдов')\n",
    "            axes[0].grid()\n",
    "\n",
    "            #график ROC-кривой\n",
    "            fpr, tpr, _ = roc_curve(y_test,y_prob_to_1)\n",
    "            axes[1].plot(fpr,tpr)\n",
    "            axes[1].plot([0, 1], [0, 1], color='navy', linestyle='--')\n",
    "            axes[1].grid()\n",
    "        \n",
    "            plt.xlim([0.0, 1.0])\n",
    "            plt.ylim([0.0, 1.05])\n",
    "            plt.xlabel('False Positive Rate')\n",
    "            plt.ylabel('True Positive Rate')\n",
    "            plt.title('ROC-кривые') \n",
    "            \n",
    "    #печать дополнительных метрик\n",
    "    if print_metrics=='Yes':\n",
    "        print ('Метрика auc_prc по кросс-валидации для {} фолдов - {}'.\\\n",
    "               format(n_folds,np.array(auc_prc_list).mean().round(4)))\n",
    "        print ('Метрика Precision по кросс-валидации для {} фолдов - {}'.\\\n",
    "               format(n_folds,np.array(Precision_list).mean().round(4)))\n",
    "        print ('Метрика Recall по кросс-валидации для {} фолдов - {}'.\\\n",
    "               format(n_folds,np.array(Recall_list).mean().round(4)))\n",
    "        print ('Метрика f1-score по кросс-валидации для {} фолдов - {}'.\\\n",
    "               format(n_folds,np.array(f1_score_list).mean().round(4)))\n",
    "        print ('\\n')\n",
    "            \n",
    "    print ('{} - Метрика auc_roc по кросс-валидации для {} фолдов - {}'.\\\n",
    "           format(clf.__class__.__name__,n_folds,np.array(roc_auc_list).mean().round(4)))\n",
    "    \n",
    "    #оцениваем метрику на hold_out выборке, если ее передали в функцию в качестве аргумента\n",
    "    if type(X_hold_out)!=type(None):\n",
    "        \n",
    "        \n",
    "        clf=clf\n",
    "        clf.fit(X,y)\n",
    "        predictions=clf.predict_proba(X_hold_out)[:,1]\n",
    "             \n",
    "        roc_auc_score_hold_out=roc_auc_score(y_hold_out,predictions)\n",
    "        print ('{} - Метрика auc_roc на hold_out выборке (финальная оценка) - {}'.\\\n",
    "               format(clf.__class__.__name__,roc_auc_score_hold_out.round(4)))\n",
    "       \n",
    "        return np.array(roc_auc_list).mean().round(4),  roc_auc_score_hold_out.round(4)\n",
    "    \n",
    "    return np.array(roc_auc_list).mean().round(4)\n",
    "\n",
    "#Определим функцию, которая в зависимости от установленного порога будет строже или мягче\n",
    "#конвертировать предсказанную вероятность в соответствуюзий класс, т.е. чем ниже порог, \n",
    "#тем с более низкими вероятностями наша модель будет относить объекты к классу 'отток'\n",
    "\n",
    "def mistakes_balance_tuning(DF,threshold):\n",
    "    DF.predicted_class=np.where(DF.predicted_prob>=threshold,1,0)\n",
    "    \n",
    "    DF_mistakes=DF[DF.predicted_class!=DF.true_class].\\\n",
    "            sort_values('predicted_prob',ascending=False)\n",
    "    \n",
    "    #mistakes matrix\n",
    "    TP=DF[(DF.predicted_class==1)&(DF.true_class==1)]\n",
    "    FP=DF_mistakes[DF_mistakes.predicted_class==1]\n",
    "    TN=DF[(DF.predicted_class==0)&(DF.true_class==0)]\n",
    "    FN=DF_mistakes[DF_mistakes.predicted_class==0]\n",
    "    #Pr and Recall\n",
    "    Precision=TP.shape[0]/(TP.shape[0]+FP.shape[0])\n",
    "    Recall=TP.shape[0]/(TP.shape[0]+FN.shape[0])\n",
    "    #Roc_Auc\n",
    "    roc_auc=roc_auc_score(y_validation,DF.predicted_prob)\n",
    "    \n",
    "    print('TP (верно определенные клиенты из класса \"отток\") = ',TP.shape[0])\n",
    "    print('FN (ложные пропуски клиентов из класса \"отток\") = ',FN.shape[0])\n",
    "    print('FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") = ',FP.shape[0])\n",
    "    print('Precision = ',round(Precision,3))\n",
    "    print('Recall = ',round(Recall,3))\n",
    "    print('roc_auc = ',round(roc_auc,3))\n",
    "    print('общее число клиентов в валидационной выборке = ',DF.shape[0])\n",
    "    \n",
    "    return TP.shape[0],FN.shape[0],FP.shape[0], DF\n",
    "\n",
    "#Перепишем, определенную ранее функцию, включив в нее параметры , \n",
    "#используемые для оценки экономической эффективности\n",
    "def mistakes_balance_tuning_and_final_estimation(DF,threshold,\n",
    "                            income_per_cust_mnth=100,\n",
    "                            cost_to_hold_per_cust_mnth=10,\n",
    "                            loyalty_to_suggestion=0.5,\n",
    "                            top_customer_to_hold=0.25):\n",
    "    \n",
    "    DF.predicted_class=np.where(DF.predicted_prob>=threshold,1,0)\n",
    "    \n",
    "    DF_mistakes=DF[DF.predicted_class!=DF.true_class].sort_values('predicted_prob',ascending=False)\n",
    "    \n",
    "    #mistakes matrix\n",
    "    TP=DF[(DF.predicted_class==1)&(DF.true_class==1)]\n",
    "    FP=DF_mistakes[DF_mistakes.predicted_class==1]\n",
    "    TN=DF[(DF.predicted_class==0)&(DF.true_class==0)]\n",
    "    FN=DF_mistakes[DF_mistakes.predicted_class==0]\n",
    "    #Pr and Recall\n",
    "    Precision=TP.shape[0]/(TP.shape[0]+FP.shape[0])\n",
    "    Recall=TP.shape[0]/(TP.shape[0]+FN.shape[0])\n",
    "    #Roc_Auc\n",
    "    roc_auc=roc_auc_score(y_validation,DF.predicted_prob)\n",
    "    \n",
    "    print('TP (верно определенные клиенты из класса \"отток\") = ',TP.shape[0])\n",
    "    print('FN (ложные пропуски клиентов из класса \"отток\") = ',FN.shape[0])\n",
    "    print('FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") = ',FP.shape[0])\n",
    "    print('Precision = ',round(Precision,3))\n",
    "    print('Recall = ',round(Recall,3))\n",
    "    print('roc_auc = ',round(roc_auc,3))\n",
    "    \n",
    "    #общее число клиентов 8000\n",
    "    customers_number=DF.shape[0]\n",
    "\n",
    "    \n",
    "    #реальный объем класса отток\n",
    "    true_churn_numbers=DF[DF.true_class==1].shape[0]\n",
    "    print('реальный объем класса отток ',true_churn_numbers)\n",
    "    \n",
    "    # класс отток, определяемый моделью\n",
    "    predicted_churn_numbers=DF[DF.predicted_class==1].shape[0]\n",
    "    print('класс отток, определяемый моделью ',predicted_churn_numbers)\n",
    "    \n",
    "    #определим % лучших от класса отток\n",
    "    top_number=int(predicted_churn_numbers*top_customer_to_hold)\n",
    "    print('% дучших от класса отток ' ,top_number)\n",
    "    \n",
    "    #DF c % лучших от класса отток\n",
    "    \n",
    "    DF_TOP=DF[DF.predicted_class==1].sort_values('predicted_prob',ascending=False).\\\n",
    "                                                                    iloc[:top_number,]\n",
    "    \n",
    "    # число людей действительно относящихся к классу отток из исследуемых Топ25% нашей модели\n",
    "    true_churn_from_top_number=(DF_TOP.true_class==1).sum()\n",
    "    print('число людей действительно относящихся к классу отток из исследуемых ',\\\n",
    "                                                              true_churn_from_top_number)\n",
    "    \n",
    "    #число людей, лояльных к предложению\n",
    "    loyal_customer=round(true_churn_from_top_number*loyalty_to_suggestion)\n",
    "    print('число людей, лояльных к предложению ',loyal_customer)\n",
    "    \n",
    "    #прибыль без модели\n",
    "    income_wout_model=(customers_number-true_churn_numbers)*income_per_cust_mnth\n",
    "    print('прибыль без модели = ',income_wout_model)\n",
    "\n",
    "    #затраты на удержание\n",
    "    cost_to_hold=top_number*cost_to_hold_per_cust_mnth\n",
    "    print('затраты на удержание = ',cost_to_hold)\n",
    "    \n",
    "    #прибыль от сохраненных клиентов\n",
    "    print('прибыль от сохраненных клиентов ',loyal_customer*income_per_cust_mnth)\n",
    "    \n",
    "    #прибыль с моделью\n",
    "    income_with_model=(customers_number-(true_churn_numbers-loyal_customer))*\\\n",
    "                                                        income_per_cust_mnth-cost_to_hold            \n",
    "    print('прибыль с моделью = ',income_with_model)\n",
    "    \n",
    "    #разница двух сценариев\n",
    "    Difference_in_total_income_per_mnth=income_with_model-income_wout_model\n",
    "    Difference_in_total_income_per_mnth\n",
    "    \n",
    "    print('разница в итоговой прибыли после применения модели ', \\\n",
    "              round(Difference_in_total_income_per_mnth/income_wout_model*100,2))\n",
    "\n",
    "    return round(Difference_in_total_income_per_mnth/income_wout_model*100,3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### baseline-модель (data_merge датасет)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(40000, 38) 38 - категориальных признаков\n",
      "(40000, 174) 174 - вещественных признаков\n",
      "Метрика auc_prc по кросс-валидации для 5 фолдов - 0.2137\n",
      "Метрика Precision по кросс-валидации для 5 фолдов - 0.5438\n",
      "Метрика Recall по кросс-валидации для 5 фолдов - 0.008\n",
      "Метрика f1-score по кросс-валидации для 5 фолдов - 0.0157\n",
      "\n",
      "\n",
      "XGBClassifier - Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7429\n",
      "XGBClassifier - Метрика auc_roc на hold_out выборке (финальная оценка) - 0.7479\n",
      "Wall time: 53.1 s\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%time\n",
    "#заполнение пропоусков для категориальных признаков\n",
    "X_cat_filled=data_merge[categ_features_not_null].fillna('w/o_cat')\n",
    "\n",
    "#кодирование категориальных признаков\n",
    "X_cat=cat_to_real_encode(X_cat_filled,'freq')\n",
    "print(X_cat.shape,X_cat.shape[1],'- категориальных признаков')\n",
    "\n",
    "#отбор вещественных признаков\n",
    "X_real_mean=get_X_real_mean(data_merge.drop('labels',axis=1),1)\n",
    "print(X_real_mean.shape,X_real_mean.shape[1],'- вещественных признаков')\n",
    "\n",
    "#метки\n",
    "y=y_train_orig.labels\n",
    "\n",
    "TRAIN, HOLD = hold_out_train(X_real_mean,X_cat,y)\n",
    "\n",
    "#тренировочный датасет разбиваем на матрицу объектов-признаков и вектор меток, данные используем непосредствеено в DF-ах\n",
    "X=TRAIN.drop('labels',axis=1)\n",
    "y=TRAIN['labels']\n",
    "\n",
    "X_validation=HOLD.drop('labels',axis=1)\n",
    "y_validation=HOLD['labels']\n",
    "\n",
    "baseline_roc_auc_CV , baseline_roc_auc_HOLD = \\\n",
    "estimator(xgb.XGBClassifier(random_state=random),X,y,X_validation,y_validation,print_graph='Yes',print_metrics='Yes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "качество baseline модели : \n",
      "roc_auc на CV = 0.7429 \n",
      "roc_auc на hold_out = 0.7479\n"
     ]
    }
   ],
   "source": [
    "#для сравнения полученных результатов с качеством на baseline моделе\n",
    "def baseline():\n",
    "    print (\"качество baseline модели : \\nroc_auc на CV = {} \\nroc_auc на hold_out = {}\".\\\n",
    "           format(baseline_roc_auc_CV,baseline_roc_auc_HOLD))\n",
    "baseline()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### baseline-модель (X_DATA датасет)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(50000, 38) 38 - категориальных признаков\n",
      "(50000, 174) 174 - вещественных признаков\n",
      "Метрика auc_prc по кросс-валидации для 5 фолдов - 0.223\n",
      "Метрика Precision по кросс-валидации для 5 фолдов - 0.6258\n",
      "Метрика Recall по кросс-валидации для 5 фолдов - 0.0118\n",
      "Метрика f1-score по кросс-валидации для 5 фолдов - 0.0231\n",
      "\n",
      "\n",
      "XGBClassifier - Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7445\n",
      "Wall time: 54.1 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7445"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%time\n",
    "#заполнение пропоусков для категориальных признаков\n",
    "X_cat_filled=X_DATA[categ_features_not_null].fillna('w/o_cat')\n",
    "\n",
    "#кодирование категориальных признаков\n",
    "X_cat=cat_to_real_encode(X_cat_filled,'freq')\n",
    "print(X_cat.shape,X_cat.shape[1],'- категориальных признаков')\n",
    "\n",
    "#отбор вещественных признаков (удаление нулевых)\n",
    "X_real_mean=get_X_real_mean(X_DATA,1)\n",
    "print(X_real_mean.shape,X_real_mean.shape[1],'- вещественных признаков')\n",
    "\n",
    "#метки\n",
    "y=y_train_orig.labels\n",
    "\n",
    "X, HOLD = hold_out_train_new(X_real_mean,X_cat,y)\n",
    "\n",
    "estimator(xgb.XGBClassifier(random_state=random),X,y,print_graph='Yes',print_metrics='Yes')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.3 Эксперименты с baseline моделью (оптимизация модели)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Undersampling\n",
    "\n",
    "В ходе выполнения экспериментов с моделью были протестированы различные техники отбора объектов:\n",
    "\n",
    "1) Балансировка путем перебора гиперпараметра scale_pos_weight для нашего классификатора.\n",
    "- Балансировка исходного набора данных путем задания соответствующих поправочных коээфициентов для объектов меньшего класса (в нашем случае положительный класс - отток) ведет в общем к снижению уровня качества классификатора.\n",
    "\n",
    "2) Random Undersampling\n",
    "- Random Undersampling с уделением небольшой доли объектов большего класса приводит к увеличению качества как на кросс-валидации так и на отложенной выборке.\n",
    "\n",
    "- Рандомное удаление объектов - довольно грубый способ, т.к. мы не учитываем важность объектов для обучения.\n",
    "\n",
    "3) Методы из библиотеки imbalanced-learn:\n",
    "\n",
    "а) ALLKNN\n",
    "- Данный алгоритм позволяет поднять качество как на CV, так и на hold_out датасете, в сравнении с baseline моделью.\n",
    "\n",
    "b) TomekLinks\n",
    "\n",
    "- Данный алгоритм показывает более высокое качество на CV , но более низкое на hold_out датасете, в сравнении с baseline моделью.\n",
    "\n",
    "- Наилучшие результаты показал метод 'auto'.\n",
    "\n",
    "- Алгоритм со стратегией 'auto' находит пары объектов противоположного класса, расстояние между котороми очень мало (т.е. объекты схожей природы) и удаляет в таких парах объекты большего класса, при этом сохраняя объекты миноритарного класса.\n",
    "\n",
    "с) NearMiss\n",
    " - Данный алгоритм позволяет поднять качество как на cv, так и на hold_out выборке, по сравнению с baseline моделью, при аккуртной подстройке количества отбираемых объектов.\n",
    "\n",
    "Наилучшие результаты показал метод ALLKNN, ниже приведена его реализация."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Undersampling - AllKNN\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(40000, 38) 38 - категориальных признаков\n",
      "(40000, 174) 174 - вещественных признаков\n",
      "(24894, 212)\n",
      "(24894,)\n",
      "XGBClassifier - Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7498\n",
      "XGBClassifier - Метрика auc_roc на hold_out выборке (финальная оценка) - 0.7492\n",
      "(29532, 212)\n",
      "(29532,)\n",
      "XGBClassifier - Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7455\n",
      "XGBClassifier - Метрика auc_roc на hold_out выборке (финальная оценка) - 0.7493\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 4min 53s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "#кодирование категориальных признаков\n",
    "X_cat_filled=data_merge[categ_features_not_null].fillna('w/o_cat')\n",
    "X_cat=cat_to_real_encode(X_cat_filled,'freq')\n",
    "print(X_cat.shape,X_cat.shape[1],'- категориальных признаков')\n",
    "#отбор вещественных признаков\n",
    "X_real_mean=get_X_real_mean(data_merge.drop('labels',axis=1),1)\n",
    "print(X_real_mean.shape,X_real_mean.shape[1],'- вещественных признаков')\n",
    "#метки\n",
    "y=y_train_orig.labels\n",
    "\n",
    "TRAIN, HOLD = hold_out_train(X_real_mean,X_cat,y)\n",
    "\n",
    "#тренировочный датасет разбиваем на матрицу объектов-признаков и вектор меток, данные используем непосредствеено в DF-ах\n",
    "X=TRAIN.drop('labels',axis=1)\n",
    "y=TRAIN['labels']\n",
    "\n",
    "X_validation=HOLD.drop('labels',axis=1)\n",
    "y_validation=HOLD['labels']\n",
    "\n",
    "#реализация различных видом ALLKNN\n",
    "cv_scores_=[]\n",
    "hold_out_scores_=[]\n",
    "kinds=['all','mode']\n",
    "\n",
    "for kind in kinds:\n",
    "\n",
    "    allknn = AllKNN(kind_sel=kind)\n",
    "\n",
    "    X_resampled, y_resampled = allknn.fit_resample(X, y)\n",
    "    \n",
    "    print(X_resampled.shape)\n",
    "    print(y_resampled.shape)\n",
    "    \n",
    "    cv_score, hold_out_score = estimator(xgb.XGBClassifier(random_state=random),X_resampled,y_resampled,X_validation,y_validation)\n",
    "    \n",
    "    cv_scores_.append(cv_score)\n",
    "    hold_out_scores_.append(hold_out_score)\n",
    "\n",
    "\n",
    "#визуализация полученных результатов\n",
    "plt.figure(figsize=(10,6))\n",
    "plt.scatter(32000,baseline_roc_auc_CV,color='blue', s=50, marker='o')\n",
    "plt.scatter(32000,baseline_roc_auc_HOLD,color='gold', s=50, marker='o')\n",
    "plt.scatter(24894,cv_scores_[0],color='blue', s=50, marker='x')\n",
    "plt.scatter(24894,hold_out_scores_[0],color='gold', s=50, marker='x')\n",
    "plt.scatter(29532,cv_scores_[1],color='blue', s=50, marker='s')\n",
    "plt.scatter(29532,hold_out_scores_[1],color='gold', s=50, marker='s')\n",
    "plt.legend(['test_score_baseline','hold_out_score_baseline',\n",
    "            'test_score_AllKNN_kind_all','hold_out_score_AllKNN_kind_all',\n",
    "            'test_score_AllKNN_kind_mode','hold_out_score_AllKNN_kind_mode'])\n",
    "plt.ylabel('auc_roc scores')\n",
    "plt.xlabel('n_samples')\n",
    "plt.title('score=f (n_samples)')\n",
    "plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "изначальная доля меньшего класса, по отношению к большему:  0.08\n",
      "после отбора объектов доля меньшего класса, по отношению к большему (в случае kind_sel='mode'):  0.088\n",
      "количество отобранных объектов : 2468\n"
     ]
    }
   ],
   "source": [
    "print(\"изначальная доля меньшего класса, по отношению к большему: \",(y.value_counts()[1]/y.value_counts()[0]).round(3))\n",
    "print(\"после отбора объектов доля меньшего класса, по отношению к большему (в случае kind_sel='mode'): \",(y_resampled.value_counts()[1]/y_resampled.value_counts()[0]).round(3))\n",
    "print('количество отобранных объектов :',y.value_counts()[0]-y_resampled.value_counts()[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Анализ результатов Undersampling - AllKNN:\n",
    "    \n",
    "Данный алгоритм позволяет поднять качество как на CV, так и на hold_out датасете, в сравнении с baseline моделью.\n",
    "\n",
    "В случае гиперапарметра kind_sel='mode' проводится более мягкий отбор, при котором, в датасете остаются все элементы большего класса, для которых большинство из n-ближайших соседей принадлежат к большему классу, а удаляются только те элементы, для которых большенство соседей принадлежит меньшему классу.\n",
    "\n",
    "В случае гиперапарметра kind_sel='all' проводится более жесткий отбор, при котором, в датасете остаются только те элементы большего класса, для которых все n-ближайших соседей принадлежат к большему классу, а удаляются любые элементы, для которых хотя бы 1 из соседей принадлежит к меньшему классу.\n",
    "\n",
    "Мягкий отбор выглядит более предпочтительным, т.к. мы удаляем небольшое количество данных, и как следствие теряем меньше информации. К тому же, при мягком отборе качество на отложенной выборке выше аналогичного качества для жесткого отбора."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Работа с признаками - Заполнение пропусков.\n",
    "\n",
    "Пробовались различные стратегии заполнения пропусков:\n",
    "    \n",
    "Для вещественных признаков - заполнение средними значениями и заполнение нулями.\n",
    "\n",
    "Для категориальных признаков - заполнение отдельной новой категорией и заполнение наиболее частой категорией.\n",
    "\n",
    "Наилучший результат показал метод - используемый в baseline моделях.\n",
    "\n",
    "- вещесвтенные - средними значениями\n",
    "- категориальные - новой категорией\n",
    "\n",
    "Ниже приведен скриншот с результатами без непосредственной реализации экспериментов для того, чтобы не перегружать отчет."
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Работа с признаками - Кодирование категориальных признаков\n",
    "\n",
    "Пробовались различные стратегии обработки категориальных признаков:\n",
    "\n",
    "1) one-hot encoding через DictVectorizer (sklearn.feature_extraction)\n",
    "\n",
    "2) one-hot encoding через OneHotEncoder (sklearn.preprocessing)\n",
    "\n",
    "3) one-hot encoding через get_dummies (pandas)\n",
    "\n",
    "4) Хэширование признаков через FeatureHasher (sklearn.feature_extraction)\n",
    "\n",
    "5) Частотное кодирование\n",
    "\n",
    "#### Вариант 1:\n",
    "\n",
    "Если не проводить предварительный отбор категориальных признаков (перед кодированием), то первые 4 метода генерируют из 38 категориальных (ненулевых) признаков - очень сильно разряженную матрицу из 0 и 1:\n",
    "\n",
    "Методы one-hot encoding - генерирует 63513 новых признаков, это суммарное число уникальных категорий для категориальных признаков,\n",
    "\n",
    "FeatureHasher , при задании n_features = количеству кникальных значений - также будет генерировать разряженую матрицу высокой размерности.\n",
    "\n",
    "С такой матрицей работать можно (используется формат Compressed Sparse Row format), но это значительно усложняет нашу модель.\n",
    "\n",
    "Поэтому, частотное кодирование в данном случае является предпочтительным, при котором количество категориальных признаков после кодирования сохраняется.\n",
    "\n",
    "#### Вариант 2:\n",
    "\n",
    "Если проводить отбор категориальных признаков по степени корреляции с целевой переменной (данный отбор можно проводился с помощью определенной ранее функции - V_Krammer_func), то можно выдлить только 15 категориальных признаков, для которых присутствует статистически значимая взаимосвязь с целевой переменной на уровне значимости 0.05.\n",
    "\n",
    "В данном случае первые 4 метода генерирует из 15 признаков 194 новых признака.\n",
    "\n",
    "Если для 15 признаков сравнивать различные способы кодирования, то наилучший результат из всех показывает one-hot encoding через DictVectorizer, но при этом качество модели падает, относительно качества без отбора признаков с частотным кодированием.\n",
    "\n",
    "В работе использовалось - частотное кодирование. Кодирование осуществляется с помощью спец. функции cat_to_real_encode(X_cat_filled,'freq')\n",
    "\n",
    "Ниже для демонстрации приведены способы оценки размерности матрицы, получаемой при использовании первых четырех методов без непосредственной реализации всех методов."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "63513"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#размерность разряженной матрицы для 4 методов без предварительного отбора\n",
    "X_cat_filled.apply(lambda x: x.nunique()).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "194"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#размерность разряженой матрицы для 4 методов после отбора по степени корреляции\n",
    "V_Krammer_list=[]\n",
    "p_value_list=[]\n",
    "condition_of_using_list=[]\n",
    "for i in range(len(X_cat_filled.columns)):\n",
    "    table=pd.crosstab(X_cat_filled.iloc[:,i], data_merge.labels)\n",
    "    V_Krammer_k, p_value_k, condition_of_using=V_Krammer_func(table)\n",
    "    V_Krammer_list.append(V_Krammer_k.round(4))\n",
    "    p_value_list.append(p_value_k)\n",
    "    condition_of_using_list.append(condition_of_using)\n",
    "\n",
    "CATEG_COR=pd.DataFrame({'V_Krammer':V_Krammer_list,\"p_value\":p_value_list,\n",
    "                       \"condition_of_using\":condition_of_using_list},\n",
    "                       index=X_cat_filled.columns)\n",
    "CATEG_COR2=CATEG_COR[(CATEG_COR.condition_of_using=='Ok') &\n",
    "                     (CATEG_COR.p_value<0.05)].sort_values('V_Krammer',ascending=False)\n",
    "\n",
    "X_cat_top=X_cat_filled[CATEG_COR2.index]\n",
    "\n",
    "X_cat_top.apply(lambda x: x.nunique()).sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Работа с признаками - отбор признаков"
   ]
  },
  {
   "attachments": {
    "1.png": {
     "image/png": "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"
    },
    "2.png": {
     "image/png": "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"
    },
    "3.png": {
     "image/png": "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"
    },
    "4.png": {
     "image/png": "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"
    },
    "5.png": {
     "image/png": "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"
    },
    "6.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Пробовались различные стратегии отбора признаков:\n",
    "\n",
    "Ниже приведены результаты в виде графиков, на которых отображена зависимость roc_auc метрики на кросс-валидации и отложенной выборке в зависимости от числа используемых признаков.\n",
    "    \n",
    "1) VarianceThreshold from sklearn.feature_selection\n",
    "\n",
    "![1.png](attachment:1.png)\n",
    "\n",
    "Данный подход показывает снижение качества при увеличении числа отобранных признаков. Это обусловлено тем, что признаки с небольшой дисперсией могут иметь весомое значение для построения модели.\n",
    "\n",
    "2) Отбор с помощью LogisticRegression с l1 через SelectFromModel\n",
    "\n",
    "![2.png](attachment:2.png)\n",
    "\n",
    "Отбор с помощью LogisticRegression с l1 с последующим построением модели XGBClassifier на отобранных признаках приводит к снижению качества относительно baseline.\n",
    "\n",
    "3) Отбор с помощью LinearSVC с l1 через SelectFromModel\n",
    "\n",
    "![3.png](attachment:3.png)\n",
    "\n",
    "Отбор с помощью LinearSVC с l1 с последующим построением модели XGBClassifier на отобранных признаках приводит к снижению качества относительно baseline.\n",
    "\n",
    "4) Отбираем только вещественные по степени корреляции с целевой переменной\n",
    "\n",
    "В определенной ранее функции hold_out_train при регулировке порога, мы делаем отбор для вещественных признаков.\n",
    "\n",
    "![4.png](attachment:4.png)\n",
    "   \n",
    "Можно наблдать, что снижение числа вещестенных признаков со 174 до 70 почти не меняет качетсво нашей модели.\n",
    "\n",
    "Для 63 признаков наблюдаем некоторое снижение качества на CV, и увеличение качества на hold-out датасете.\n",
    "\n",
    "Для 70 признаков наблюдаем повышение качества как на CV,так и на hold_out датасете.\n",
    "\n",
    "Опираться на данный отбор не следует, т.к. сам способ определения корреляции между вещественными и категориальными признаками является довольно грубым и не подвергается оценке уровня статистической значимости данного результата, но , как минимум, мы видим, что не все вещественные признаки являются важными для нашей модели.\n",
    "\n",
    "5) Отбираем только вещественные по степени пропущенных значений\n",
    "\n",
    "Оригинальные данные представляют собой очень разряженное признаковое пространство.\n",
    "\n",
    "16 полностью нулевых признаков\n",
    "\n",
    "147 признаков, для которых имеется не больше 2000 заполненных значений !!! т.е. заполнены не больше 5% объектов\n",
    "\n",
    "35 признаков, для которых имеется не меньше 87,5% \n",
    "\n",
    "3 признака, для которых имеется 100% заполненных значений\n",
    "\n",
    "Ранее мы определили функцию get_X_real_mean, которая отбирает вещественные признаки в зависимости от доли пропущенных значений\n",
    "\n",
    "![5.png](attachment:5.png)\n",
    " \n",
    "Отбор вещественных признаков позволяет увеличить качество или как минимум сохранить качество baseline модели:\n",
    "    \n",
    "Для 162 признаков (удалили 12 наиболее разряженных) наблюдаем увеличение качества как на CV так и на hold_out выборке.\n",
    "\n",
    "Для 43 признаков (удалили 131, оставили 25% вещественных!) наблюдаем почти полное сохранение качества, относительно качества baseline модели. (на СV - на 1/10 000 хуже, на hold_out - на 1/10 000 лучше)\n",
    "\n",
    "Удаление сильно разряженных признаков почти не влияет на качество модели.\n",
    "\n",
    "К сожалению, фиксация лучшего результата для отбора вещественных признаков и последующий отбор категориальных не приводит к увеличению качества. Пробовались различные стратегий для отбора категориальных признаков - по степени пропущенных значений, по уровню корреляции, по кол-ву подкатегорий для каждого признака(признаки с большим количеством уникальных значений).\n",
    "\n",
    "Вероятнее всего это обусловлено тем, что при данном отборе не учитывается взаимная важность некоторых комбинаций категориальных и вещественных признаков для модели.\n",
    "\n",
    "6) Делаем отбор признаков с помощью ExtraTreesClassifier\n",
    "\n",
    "![6.png](attachment:6.png)\n",
    "\n",
    "Для 107 признаков, качество на CV увеличивается (принимает max - значение), на hold_out - увеличивается, относительно baseline модели.\n",
    "\n",
    "Для 66 признаков, качество на CV увеличивается (принимает max - значение), на hold_out увеличивается.\n",
    "\n",
    "Отбор признаков с помощью ExtraTreesClassifier позволяет увеличить качество модели на CV, но при этом качество на отложенной выборке сохраняется.\n",
    "\n",
    "Наилучший вариант - отбор с помощью ExtraTreesClassifier."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.4 Подбор оптимальных параметров классификатора "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В ходе выолнения работы были построены две модели на основе data_merge датасета, для каждой из которых были подобраны оптимальные параметры с помощью RandomizedSearchCV и GridSearchCV.\n",
    "\n",
    "Ниже приведены полученные результаты."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "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>bsline_data</th>\n",
       "      <th>bsline_score_CV</th>\n",
       "      <th>bsline_score_HO</th>\n",
       "      <th>smpls_resampled</th>\n",
       "      <th>ftrs_extracted</th>\n",
       "      <th>score_CV</th>\n",
       "      <th>score_HO</th>\n",
       "      <th>GrSrch_CV</th>\n",
       "      <th>GrSrch_HO</th>\n",
       "      <th>GrSrch_HO_Kaggle</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>model_1</th>\n",
       "      <td>data_merge</td>\n",
       "      <td>0.7429</td>\n",
       "      <td>0.7479</td>\n",
       "      <td>ALLKNN (mode)</td>\n",
       "      <td>ExtraTC (60 features)</td>\n",
       "      <td>0.7465</td>\n",
       "      <td>0.7491</td>\n",
       "      <td>0.7469</td>\n",
       "      <td>0.7507</td>\n",
       "      <td>0.72068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model_2</th>\n",
       "      <td>data_merge</td>\n",
       "      <td>0.7429</td>\n",
       "      <td>0.7479</td>\n",
       "      <td>ALLKNN (mode)</td>\n",
       "      <td>ExtraTC (60 features)</td>\n",
       "      <td>0.7465</td>\n",
       "      <td>0.7491</td>\n",
       "      <td>0.7494</td>\n",
       "      <td>0.7473</td>\n",
       "      <td>0.72246</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        bsline_data  bsline_score_CV  bsline_score_HO smpls_resampled  \\\n",
       "model_1  data_merge           0.7429           0.7479   ALLKNN (mode)   \n",
       "model_2  data_merge           0.7429           0.7479   ALLKNN (mode)   \n",
       "\n",
       "                ftrs_extracted  score_CV  score_HO  GrSrch_CV  GrSrch_HO  \\\n",
       "model_1  ExtraTC (60 features)    0.7465    0.7491     0.7469     0.7507   \n",
       "model_2  ExtraTC (60 features)    0.7465    0.7491     0.7494     0.7473   \n",
       "\n",
       "         GrSrch_HO_Kaggle  \n",
       "model_1           0.72068  \n",
       "model_2           0.72246  "
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF_result=pd.DataFrame({'bsline_data':['data_merge','data_merge'],\n",
    "                        'bsline_score_CV':[0.7429,0.7429],\n",
    "                        'bsline_score_HO':[0.7479,0.7479],\n",
    "                        'smpls_resampled':['ALLKNN (mode)','ALLKNN (mode)'],\n",
    "                        'ftrs_extracted':['ExtraTC (60 features)','ExtraTC (60 features)'],\n",
    "                        'score_CV':[0.7465,0.7465],\n",
    "                        'score_HO':[0.7491,0.7491],\n",
    "                        'GrSrch_CV':[0.7469,0.7494],\n",
    "                        'GrSrch_HO':[0.7507,0.7473],\n",
    "                        'GrSrch_HO_Kaggle': [0.72068,0.72246]},\n",
    "                      index=['model_1','model_2'])\n",
    "DF_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Лушая модель была построена на основе X_DATA датасета, ниже приведен поэтапный процесс построения модели и подбора оптимальных параметров."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- baseline-модель (X_DATA датасет)\n",
    "- Делаем отбор объектов c помощью ALLKNN kind_sel='mode'\n",
    "- Делаем отбор признаков с помощью ExtraTreesClassifier, подбираем оптимальный порог так, чтобы получилось максимальное качество на CV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(50000, 38) 38 - категориальных признаков\n",
      "(50000, 174) 174 - вещественных признаков\n",
      "(36976, 212)\n",
      "(36976,)\n",
      "используем 82 признаков после отбора\n",
      "Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7486\n",
      "Wall time: 2min 41s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# baseline-модель (X_DATA датасет)\n",
    "\n",
    "#кодирование категориальных признаков\n",
    "X_cat_filled=X_DATA[categ_features_not_null].fillna('w/o_cat')\n",
    "X_cat=cat_to_real_encode(X_cat_filled,'freq')\n",
    "print(X_cat.shape,X_cat.shape[1],'- категориальных признаков')\n",
    "#отбор вещественных признаков\n",
    "X_real_mean=get_X_real_mean(X_DATA,1)\n",
    "print(X_real_mean.shape,X_real_mean.shape[1],'- вещественных признаков')\n",
    "#метки\n",
    "y=y_train_orig.labels\n",
    "\n",
    "n_folds=5\n",
    "random=0\n",
    "\n",
    "X, HOLD = hold_out_train_new(X_real_mean,X_cat,y)\n",
    "\n",
    "# Делаем отбор объектов c помощью ALLKNN kind_sel='mode'\n",
    "allknn = AllKNN(kind_sel='mode')\n",
    "X_resampled, y_resampled = allknn.fit_resample(X, y)\n",
    "print(X_resampled.shape)\n",
    "print(y_resampled.shape)\n",
    "\n",
    "#Делаем отбор признаков с помощью ExtraTreesClassifier\n",
    "\n",
    "clf=ExtraTreesClassifier(random_state=random)\n",
    "clf.fit(X_resampled,y_resampled)\n",
    "\n",
    "feature_importances=clf.feature_importances_\n",
    "FI=pd.DataFrame({'features':list(X),'feature_importance':feature_importances})\n",
    "\n",
    "#определили оптимальный порог 0.0005101412048828447, при котором отбираем 82 признака\n",
    "best_features=list(FI[FI.feature_importance>= 0.0005101412048828447 ].features)\n",
    "\n",
    "#строим лучшую модель до отбора признаков (передаем признака классификатора по умолчанию)\n",
    "param_grid={'base_score': [0.5],\n",
    "         'booster': ['gbtree'],\n",
    "         'colsample_bylevel': [1],\n",
    "         'colsample_bynode': [1],\n",
    "         'colsample_bytree': [1],\n",
    "         'gamma': [0],\n",
    "         'learning_rate': [0.1],\n",
    "         'max_delta_step': [0],\n",
    "         'max_depth': [3],\n",
    "         'min_child_weight': [1],\n",
    "         'missing': [None],\n",
    "         'n_estimators': [100],\n",
    "         'nthread': [1],\n",
    "         'objective': ['binary:logistic'],\n",
    "         'random_state': [0],\n",
    "         'reg_alpha': [0],\n",
    "         'reg_lambda': [1],\n",
    "         'scale_pos_weight': [1],\n",
    "         'silent': [None],\n",
    "         'seed': [0],\n",
    "         'subsample': [1],\n",
    "         'verbosity': [1]}\n",
    "\n",
    "\n",
    "stat = StratifiedKFold(n_splits=n_folds,shuffle=True,random_state=random)\n",
    "clf=xgb.XGBClassifier(random_state=random)\n",
    "\n",
    "grid_searcher=GridSearchCV(clf,param_grid,n_jobs=-1,cv=stat,scoring='roc_auc')\n",
    "\n",
    "print('используем {} признаков после отбора'.format(len(best_features)))\n",
    "\n",
    "grid_searcher.fit(X_resampled[best_features],y_resampled)\n",
    "\n",
    "print ('Метрика auc_roc по кросс-валидации для {} фолдов - {}'.format(n_folds,grid_searcher.best_score_.round(4)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Подбор параметров по сетке:\n",
    "    \n",
    "- Делаем грубую прикидку по RandomizedSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "используем 82 признаков после отбора\n",
      "Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7509\n",
      "Wall time: 2min 35s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "seed=np.random.seed(0)\n",
    "param_grid={'base_score': [0.4,0.5],\n",
    "         'booster': ['gbtree'],\n",
    "         'colsample_bylevel': [1],\n",
    "         'colsample_bynode': [1],\n",
    "         'colsample_bytree': [0.5,0.7,0.9,1],\n",
    "         'gamma': [0,1,2,3,5,7],\n",
    "         'learning_rate': [0.03,0.05,0.07,0.08],\n",
    "         'max_delta_step': [0],\n",
    "         'max_depth': [3,4,5,6],\n",
    "         'min_child_weight': [1,2,3,5,7,9,12],\n",
    "         'missing': [None],\n",
    "         'n_estimators': range(140,220,10),\n",
    "         'nthread': [1],\n",
    "         'objective': ['binary:logistic'],\n",
    "         'random_state': [0],\n",
    "         'reg_alpha': [0],\n",
    "         'reg_lambda': [1],\n",
    "         'scale_pos_weight': [1],\n",
    "         'silent': [None],\n",
    "         'seed': [0],\n",
    "         'subsample': [0.5,0.6,0.7,0.8,0.9,1],\n",
    "         'verbosity': [1]}\n",
    "\n",
    "\n",
    "stat = StratifiedKFold(n_splits=n_folds,shuffle=True,random_state=random)\n",
    "clf=xgb.XGBClassifier(random_state=random)\n",
    "\n",
    "grid_searcher=RandomizedSearchCV(clf,param_grid,n_jobs=-1,cv=stat,scoring='roc_auc')\n",
    "\n",
    "best_features=list(FI[FI.feature_importance>= 0.0005101412048828447 ].features)\n",
    "\n",
    "print('используем {} признаков после отбора'.format(len(best_features)))\n",
    "\n",
    "grid_searcher.fit(X_resampled[best_features],y_resampled)\n",
    "\n",
    "print ('Метрика auc_roc по кросс-валидации для {} фолдов - {}'.format(n_folds,grid_searcher.best_score_.round(4)))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Качество модели удалось поднять, пробуем дополнительно увеличить качество путем полного перебора некоторых парамтеров с помощью \n",
    "GridSearchCV\n",
    "\n",
    "- фиксируем все параметры и меняем 'subsample'\n",
    "\n",
    "Качество удалось увеличить('subsample': 0.8)\n",
    "\n",
    "- делаем следующую попытку - меняем 'gamma'.\n",
    "\n",
    "Качество удалось увеличить('subsample': 0.8,'gamma': 1)\n",
    "\n",
    "- далее делаем следующую попытку - меняем 'n_estimators'\n",
    "\n",
    "Качество удалось увеличить('subsample': 0.8,'gamma': 1,'n_estimators':180)\n",
    "\n",
    "- делаем следующую попытку - меняем 'gamma' , 'max_depth', и 'learning_rate'\n",
    "\n",
    "Качество вновь удалось увеличить\n",
    "\n",
    "- фиксирем параметры лучшей модели\n",
    "\n",
    "#### 3.5 Получаем предсказания для  тестового датасета с kaggle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "используем 82 признаков после отбора\n",
      "Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7528\n",
      "Your submission was successfully saved!\n",
      "Wall time: 26.1 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "seed=np.random.seed(0)\n",
    "\n",
    "param_grid={'base_score': [0.5],\n",
    "         'booster': ['gbtree'],\n",
    "         'colsample_bylevel': [1],\n",
    "         'colsample_bynode': [1],\n",
    "         'colsample_bytree': [0.5],\n",
    "         'gamma': [0.5],\n",
    "         'learning_rate': [0.04],\n",
    "         'max_delta_step': [0],\n",
    "         'max_depth': [5],\n",
    "         'min_child_weight': [3],\n",
    "         'missing': [None],\n",
    "         'n_estimators': [180],\n",
    "         'nthread': [1],\n",
    "         'objective': ['binary:logistic'],\n",
    "         'random_state': [0],\n",
    "         'reg_alpha': [0],\n",
    "         'reg_lambda': [1],\n",
    "         'scale_pos_weight': [1],\n",
    "         'silent': [None],\n",
    "         'seed': [0],\n",
    "         'subsample': [0.8],\n",
    "         'verbosity': [1]}\n",
    "\n",
    "stat = StratifiedKFold(n_splits=n_folds,shuffle=True,random_state=random)\n",
    "clf=xgb.XGBClassifier(random_state=random)\n",
    "\n",
    "grid_searcher=GridSearchCV(clf,param_grid,n_jobs=-1,cv=stat,scoring='roc_auc')\n",
    "\n",
    "best_features=list(FI[FI.feature_importance>= 0.0005101412048828447 ].features)\n",
    "\n",
    "print('используем {} признаков после отбора'.format(len(best_features)))\n",
    "\n",
    "grid_searcher.fit(X_resampled[best_features],y_resampled)\n",
    "\n",
    "print ('Метрика auc_roc по кросс-валидации для {} фолдов - {}'.format(n_folds,grid_searcher.best_score_.round(4)))\n",
    "\n",
    "predictions=grid_searcher.best_estimator_.predict_proba(HOLD[best_features])[:,1]\n",
    "\n",
    "\n",
    "output = pd.DataFrame({'ID': X_test_orig.index, 'result': predictions})\n",
    "output.to_csv('BEST_last.csv', index=False)\n",
    "print(\"Your submission was successfully saved!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.6 Финальные результаты для построенных моделей"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "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>bsline_data</th>\n",
       "      <th>bsline_score_CV</th>\n",
       "      <th>bsline_score_HO</th>\n",
       "      <th>smpls_resampled</th>\n",
       "      <th>ftrs_extracted</th>\n",
       "      <th>score_CV</th>\n",
       "      <th>score_HO</th>\n",
       "      <th>GrSrch_CV</th>\n",
       "      <th>GrSrch_HO</th>\n",
       "      <th>GrSrch_HO_Kaggle</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>model_1</th>\n",
       "      <td>data_merge</td>\n",
       "      <td>0.7429</td>\n",
       "      <td>0.7479</td>\n",
       "      <td>ALLKNN (mode)</td>\n",
       "      <td>ExtraTC (60 features)</td>\n",
       "      <td>0.7465</td>\n",
       "      <td>0.7491</td>\n",
       "      <td>0.7469</td>\n",
       "      <td>0.7507</td>\n",
       "      <td>0.72068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model_2</th>\n",
       "      <td>data_merge</td>\n",
       "      <td>0.7429</td>\n",
       "      <td>0.7479</td>\n",
       "      <td>ALLKNN (mode)</td>\n",
       "      <td>ExtraTC (60 features)</td>\n",
       "      <td>0.7465</td>\n",
       "      <td>0.7491</td>\n",
       "      <td>0.7494</td>\n",
       "      <td>0.7473</td>\n",
       "      <td>0.72246</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model_3</th>\n",
       "      <td>X_DATA</td>\n",
       "      <td>0.7445</td>\n",
       "      <td>-</td>\n",
       "      <td>ALLKNN (mode)</td>\n",
       "      <td>ExtraTC (82 features)</td>\n",
       "      <td>0.7486</td>\n",
       "      <td>-</td>\n",
       "      <td>0.7528</td>\n",
       "      <td>-</td>\n",
       "      <td>0.72370</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        bsline_data  bsline_score_CV bsline_score_HO smpls_resampled  \\\n",
       "model_1  data_merge           0.7429          0.7479   ALLKNN (mode)   \n",
       "model_2  data_merge           0.7429          0.7479   ALLKNN (mode)   \n",
       "model_3      X_DATA           0.7445               -   ALLKNN (mode)   \n",
       "\n",
       "                ftrs_extracted  score_CV score_HO  GrSrch_CV GrSrch_HO  \\\n",
       "model_1  ExtraTC (60 features)    0.7465   0.7491     0.7469    0.7507   \n",
       "model_2  ExtraTC (60 features)    0.7465   0.7491     0.7494    0.7473   \n",
       "model_3  ExtraTC (82 features)    0.7486        -     0.7528         -   \n",
       "\n",
       "         GrSrch_HO_Kaggle  \n",
       "model_1           0.72068  \n",
       "model_2           0.72246  \n",
       "model_3           0.72370  "
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DF_result.loc['model_3']=['X_DATA',0.7445,'-','ALLKNN (mode)','ExtraTC (82 features)',\n",
    "                  0.7486,'-',0.7528,'-',0.72370]\n",
    "DF_result"
   ]
  },
  {
   "attachments": {
    "111.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAAA80AAANICAIAAAByqZrJAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAP+lSURBVHhe7P15fBzVmfcN6493e977M8+Tz2fe+9bM3MnjZJ6ZeDKZxMMk4JnkCU7uSXAmJCjJBEyG4GYzIixRSACFLQphEdhxGoONDDJuEMZtxNJGFjQIyW3ZstqyZLX2tizbLa8tL7jltb3I1nudpapOVVe1WlJJbrV+31wxVafOuc5ap351+nQr78y58062O9Y/DAAAAAAwvfn8ehhMGslji2BOY9DZAAAAAADpsCgt2HQ26GwAAAAAANewKC3YdDbobAAAAAAA17AoLdh0NuhsAAAAAADXsCgt2HQ26GwAAAAAANewKC3YdDbobAAAAAAA17AoLdh0NuhsAAAAAADXsCgt2HQ26GwAAAAAANewKC3YdDbobAAAAAAA17AoLdh0timgs0+ePBmJRHbv3n3u3DkZNGFcuHChsbFxcHBQnk86ly5dotyPapw5c0ZeuNyIglnKIwJd6RcXXRE0Zgh5MpFQg8iuOnqUyk+1kBc427dv37FjhzyZXA4cONDa2mopzyRw5sLF+KlzQ27nS24Pnz4/KqdjSDI2XMlIjH85ko4ezWT0Xrx48dixY5MwK6pQdrKIZia5GGC0iAFmO66o72gg0bNPnmcAeaNOD4fDNM/QOJShztA8OeYJeWwP5cv+KLcorfHbzyPDK/cNL+1nB5ZLqXZd8/Cvuoe/u8UanqH9unt4S2L4973WcNW+XD98V+fw/d3Sbm23RhiDUdXu6xr+183DN7YObzo2vKzfGiGNZVLmy2VZrbNPnTr12GOP/Y//8T++9a1vXXHFFTNnzgwGgxMqHfbs2fOd73ynurpank86NB8VFhbmKfzkJz+h1wx52ZkjR45QtKGhIXnuNqJg8+fPP378uAzSAhsaGuS5A/39/T09PfLEgQxdZcgfOfJkIlm1apXsJ85XvvIVv98vnlj09HqIc1kkyIsvvjhv3jy1syaHd3oP/Z9lm/afPCvPXYLcfvPN5uPnRjG8x5BkbLiSUeqNf8MNN9CNIy/bQUJn7ty5bt0yGULZyfKZmeRigNEiBhg9TUhSyyCNFStWzJo1a/v27fJ8JA4fPnzbbbfRE/maa66hf8ln+oFK0Dw55gl5bA/ly/4otyit8RjJ5dbjw7ruuXBpuPYok7mWaKqtOTh87tLwkpg1PEN788AwvTx9dJgdl+5kpz9pMUUgI00cN0/zgxeGvbut0UZlDceGTw4x1f7n3cPnLg63nbBGUI2KRAWj4olTtczZZtmrs0msPPvsswsWLKC7mk5JXm/duvXb3/72hg0bRIScRMyGNCuJU2qEqqoqmgSbm5tFiBP0nKOElFyeu42uA55//nn9VUcEjviIzWSSzdBVhnCZPUk6W292apYdO3Zce+21ahNNN6Czx4YY//qNf/bs2ZdfflkfWrZcFp2tQ7KMZFbm4gxcXsQA+4u/+Iva2loZxCHZTUI5c509NDT0+OOPP/PMMzRE6ZSeUK+88sqvfvUrcerEeHT2FMWitMZjnxxh2poUpFjobR5kpyv3WaOpNk6dTSL+9naWHR3r2leNQCZ09u4zcj17Wf/wiQvDx84P39lhjZm5qXl52kZYj6doFJmSiFO1zNlm2auzI5EISZb9+/fLc87HH39Mb9KnTp2i2YFu7wMHDohwkjXvv/9+KBSi44sXL5IWv5tDB+JTrXg8vnLlyp6enscee+yFF15wSn7y5MmysjL9s37KyOfz/eIXv6A5IhaLUQjNVm+88Yb++h4Oh1999VXxLKR/KfKuXbvEpbEhZkP9cUtQ2ShHfSKjf6moonaffPKJWDqlYtx7773f/OY3n3zySVF422jjQRSMZlh6uOqPdhGon6Y2F1FdXU0hP/jBD7xeL/UChdh2kHBFp4FAwOKBsE1CUOO0tbU9/PDDt9xyC9VXn+spOSGO9+7du3Tp0p07d9IxldDv9994442UpKOjQ0QYD6rOFtAYo3G7e/duOqa66wsqpIpIOf3sZz+jglGRRKBt+amVaBCKtiKoQ1evXi0WxW3LT63R2NhII4Sc0FuZPiT0VE6tRGWjhBs3bqRwui/UBh8zaXR2//HkPZ9s/6eVYfqXjmXo8HD74ZOeD7opvPDjaO+x0zJ0eLjryKlbPuz+59e2eJv3rO4Z0LUsvcHU9n/6o3fbvv560wvb9p4+LwfDstZ9b0UHKPK/rWpuHTgh5O+Wg8d/Emi3xHTyQNgWxuKZQpzKNmZSb/y+vj4aSPSvZVISUxkFWnS27d1HXS8mPRozr732mjpQx0mqzna6SW3nIlELmi1LS0tp+NEgPH369FtvvUXHK1asoLqItMAtxAC7/fbbLZqYZPecOXO+853vUFemearq0MDzeDzq292JEyeam5vJp9NApWOhs2nKotzV4ZHJMFDd2s51ROrsainM4cOHX3zxRboL6F+xckeI3Pfs2fOnP/2Jbpz33ntPbxnb6XpUWJTWeGzHKaZo9e0iJGRbj7PlWzom8b3pGNslQsf0Lx2LBV2hsyv2D3ecGD56nm2omN8mk4skr+8f3p8c3nV6+Mk+tuOi6ySLRoJeSNtHe1lC724WmcLPXxreeVrmqJvQ2VQ2PeSdOMv0xf7hW9tZCel0/dHhI+eHF+5iInjF3uHYmeEj54ZDR9nbgp7q2Z2sGBT+9sHhzZrOFh70HKnwVIVPz7OYJOjJG5UweooVjIpHMSm+XmaR5KEoWw5PXBjuOTn8VJ8MJIcUmTxQsdX6klEWVDAqBjULRUv/ccFoLXt19uuvv15SUkK3ujzn0B1y/fXXR6NREhD333//u+++K8IPHjxIL+VdXV30tr148eKbbrqppqaG5ojrrruObldyQpMITSUU/vbbb9ODxym5+uiivOi2fOqpp2iiWb58+be//e329nZyRVpWPA4pL9Iu+koA/fvTn/5Un6TGRurjliBZTzc8/UuzwG9/+9uf//znNB9RqX74wx+S6hIRXnrppR//+Mf0HnLkyBGnaONBLxhlQW01MDCgB6Y2F2X3jW98Q4RTa1M/LliwYNOmTTQjO3WQcHXzzTc/88wzeoNTWvKgJiGoXgsXLqRAulRZWUkKnvzQ04I6lF42yA+F0+RI0AG9p82bN6+uro6yEM1C0agfaY6mhML/eKAGoWKLTAViaFGR6FgvBj3AqBjUODRI6MWMjkUDUvmpodasWUPlv/POO3//+9/Tk4PiqAqGmlFk4VT+tWvXfv/73ycP9GpKrfTmm29SZdWCpWklKgk1JjU4lY0et+OX2k46u2XgxGdf2nT/+h1v9sRvru76/PKGHVzFfrDr6P/xfIjCa2Kf3vlx9B9WNO7hEnzz/sG/XrZRxP9FddcVrzXpWvbZLf2UnPTxqx0H/uX1pv9c23FuiM0S86o6v/Byw8/WdpDwPXrmPJXkb5ZtFEqaYn5lZXj+B10ipu7h9a6D33ur9dv+bcKzU2EsntOUbcxQj1B/qTc+DQC6O+jWtuhpCqcbjQJHnKwonAQQDTCa6GioUNc/++yz43/lFlhGqdN97TQXieRFRUXBYJBeAGbPnk1TBF0SNzhNAuIGB24hBhhJWPGkE4HUO6RZSY/S4KEeSX0s0jNFDCQd6pfHH3/8d7/7Hc1pMkjDaaDSMQ1smqPuueceMTzoWExTmQwD1a3tXGc7u6qp6GX1u9/9Lt0X+t0h9LfInUpFRSLoWHwU6TRdjwqL0hqPNQ0O07z1sSIKdVMluCp8SWfTe8zpIbbeTHbh0vDBs2yFWCQheUqalQ7OXmRxjp0f7js9fOgcS0LSk+IsiTHFTE5GpbMpPmX0yl650kzOKRdS8M/tZMqenFNJtg6yS3phKDJlJNQwKXIqidDZ6lr1o9uHj19gp9tPsUJSFqTgU3W2XmZKsrSf5S7cHuZJ6L2CwskhteSJIVZsckUPA5EF1eXAWZYFSe3OEyy+aAe3LHt1NokA9akjoJfdO+64Q9w8dLPdddddQjHQsVjn3r17N93b+io4ndLtSvMF3TAkI1pbW0U4YZtcvTlpunnkkUf0xxKd0vxCt72ekBQtHdDkRbcoRaB/SXaP8/GQ+rglqFRiHqS3eXq31r/bQZPR3XffnUwyKUBl1nVVmmhjRi8YNQi9aYgHtggUzUWX1JUS/ZMHcUnITcKpg4QrvcFpsnvllVdEg1PF9RVigtLS7E+PCmr/goKCpqYmEX78+PFbbrmFeoeOmbz94x+FyP7www/JGwVSM95www30ksajMz9iuXc8UNX0ZtehrEUPimLQAVWBshbrKFQjehmjaorZXC8/FY+kyaFDhyiyqmD0nnUqP2VBL6UikDpdPP/0gqVvpYceekg0uGh/MZLHg63Opta/46OeB0I7WDdop8Ub+ug4cfZC68AJEf5p8jzp5rV9h89fvPTTQLsen8RxwXttQssePHWO4tTtkc94cbpxX4KOSQ3/4J2IUNIEleS/edfrMbd/evrvyzdTzL0nzv7dK5urdh4R4UfOnCct/lrnQTq2LQwdq57TlI0HjBHR/mLYUI/s27fv0UcfJYVKN5Q6KRE0MIR8UcOdJityqK9W0M1II1Acjx/LKHW6r53mIkpIE7LQcFQkKuSf//xnESccDs+fP5/eycUpcAUxwGi00OsQIYYBzaIkuzs7O8XzhUKog8S8Qcf6Y5GlV/j000+Li4v/+q//+vrrryelS+NQhKsDktAHKh3TOPzOd76jDw/qd8pXfy6nHwaqW9u5jpykzq56KnJLDyy9yvSv+mRRVQFFphqRW1uHIk7mWJTWeOzG1uHeU2zapBnoyLnhqkPGbmnSi046m6g7KqOt2s/SvsU1KEUYvMC+JSnCSf6uO8SOfxFh0lMkVzWrupdDNYvOpiLR8amh4cd6pUomVS1eDCiEwvXCPLdrOHmRFeY7Yb7CrW01IeU9cE7mpepses2gl4E/7GDH5JDkNYn+/9xmikOml5ncsgVyzS0VbH+SGR1QZJLm5Xulq/4zsvUW7mK6/L0BFv7l+uGyPTI7tyyrdbZ+R+nQXXfzzTeLW47uUroZ6JagO4FuG/EWLt6DaYKgN1eiqqqKXmTpNqZo6lOBsE2u35wUSKL5ueeeE36Il156ibKmAtBdR9NHf39/S0sLJfzoo4/oCUePQ4o/fpmiPm51qFT6PEjTBJ1SCUmT3XrrrfqcSCH6cZpoY0YtGL3c03xUV1cnAvXmWrNmjYhM6GvwdEyphNwknDpIuFIbkJqXYtKsZ3lP0BeMKRX1oJhqBTRHi2masiM1SfqSiiSmV0IsrVH/vvfee9SJ4oPLcUJVS21byl00FB0QdECFpBaj6tDrB/WLKBKVnx4kugTRsYxVvWedyr927dovf/nLVPEe/kGNCNQLlr6VRPEEdCyKPR5sdTZp1isrtpKwfrMnLuxnazt+9G4baVa6eubCxc37B5e17vuvdZ3/x/Mh8kDxr6rY+uEu+fwmXu04ILTs+j3HSCWXRfYLPyvaD5B6pqsUh9QwmYhPkJ9/WNF46LRsEMrle2+1LmnZSx7+8dUwCXQRTtz5cZRMHKcWhgJVz2nKJs/HhBj/4juFggceeIAEDV3SJyURkwaGkC96eJrJqrm5eebMmU888QTdTZZROk4so9TpvqZLtnORJbk69vQKilPgCmKAUS8IbU2PPwqkeYAEKL2K688Xy2MxzYRA+ps6+p577vnKV75CM5LoZduBSsfkhyZtfXaiSY9GC41JijPiMFDd2s51trOrnoruAppmSbWLyIT+ZLHkrudo63C0WJTW+I2U8UeHmS6k0iQusFVeCiRp66SzSXTq+7Mp7YkLw41ckqpJVD2tJs9cZ1+4xK6SDqZSiZVmumRRwG8eYOqWTt+NM6s+xApDp6JULYMyGhmVUOSle7h+G1v8JmGdupHDSWenuq05wkr4+14WWfgX4XpTkB04y6R2xwm28i324bho2auzSSGlLg+TvCD11tfHVsJo9NN7Ks3dNDXQLSHWOynVD37wg2fN7Nixw3JHEbbJ9ZuTVB1pu3vvvVe64JSVlZ08eVLoPHqu0N1Oao+KRNGoSHQz62uuY0bMhpbZjYotdqTQ1FZUVPStb32L5iOa2vx+v3hoURwqs36cJtqYsRRsw4YN1157LTUsBerNVV9fL64Shw8fFvM1HVMqKokId+og4V/1EI1GyQP1CCVRdxDRAZ2ST5o677zzztOnjR299GImMqJ/Z82a9Zvf/Ebf4iIgqUrPBhpXV1xxBc2kY/g00AIVw9K2oiI0POiYiiHKQ1CnVFdXUyuR7qHeodPU8gssY1XtWdvyU4NQZK/X+/3vf58eQqIN9YKlbyW9eAQd6/07Zmx1NonaL5Y3/uCdyPwPunR7snE3iVp/z8D//nyIdOqvanvf6I5f8VoTeaD4JIVJEMv0w8MUTWjZtX2H85du/Pm6TtWVvups0dn/+sbWwbNyIYpGD4n7JzbvpsjfWNWshxOPb9olEtoWhsJVz2nKJs/HhBg2ov3FC5X+bVp9UuIRbfRHmsmK4u/fv3/58uU0Wj73uc/RVKDfR+PEMkqd7munuciSXB17egXFKXAFMcBotNDo+tWvfvXuu++SmqRnltgtSQ0u+oKGR+pjMT11dXX0LKD4+oAU4Wo/Uueqc7j+uXQmw0B1azvXEamzq56KqnnTTTfpHwMS+pPFkrueIx2nOhRxMseitFy0Zf1MUDZxHZlGZycvGr/FcXs7W9/t5D/f4a7OHjgn1XPFfrkVhMyigMnPxWG2eNxz0jAS35ZoZHpe+iW1VBazJNfLnOqWbRy/yH7DRPcvwtWmmN82XP8p27gydMl4YXDLsldni3tA3UFL99gbyjcCiaampgULFtCUob8r6x82iQg6ljtKkJpcvaWfeeYZej7xiFZIXpOG++Uvf0kOKeEjjzzy8ssv69tex4P6uBVc4L+7Ij7nohfxefPm6WuTJObEQ4uOqcz6cZpoY8ZSMCrMwoULb7/9dpqYRHOxNVLtIz9CXUalVLqec+og4V9tcCo2TXOkIfTlBxGur0+oS+aEWNWj5z0dU3aEaDpyos+SFy9eFHM9DaEHHnhALfDYoKpZ2pbeQH74wx/SU4eORTFEuP7GSG1C1aEhRCWn1qPXJxFOZaM6UpkphF6r9LFKzxI9C9vyU4hY26YDejZQm5AfvWAjtpIIJ+hYHXhjw1Znnzw/dPXqFrHqrCLWmP+0dY84/TR5XkhbEX9Ji/H1o9+s3yG0bMvAiX9Y0Sj2dltI1dl/vWyjHlPfB9Jx+CR5oH9F+PmLl370btsz4ZhTYehY9ZymbPJ8TFjuL7p9xFfT6JgGvP4hHqGrhEwmK31sEM3NzRRfHwnjxDKjOt3XTnORJbk69ihQlzvALcQAE6NFbAih155f//rXNJNQU1OD632R+lhUicfjr7zyito7NKIKCgr6+vqcBiodU+eKD1jEJX29LJNhQKg6O3Wuo9PU2VVPJdbFKEREIPQniyV3PUc6TnUoTjPHorTGbAs62J6HjhPGgu6vutnGjyiXnqQUD51jWz7o+K5OtllZSFIhbVfxTclkz/F9ETV8z7EqLsevs0V8i6UqYBL9emH+dbPcTyI8xM6wnR50KvZ7iLx0D1Tr3lNst4wu4uk1Y+U+Fjk1F1Fm4ZZcCbfkoesk+wTg7s50Ovv7TbJUYvuKHu6KZa/OphvpzTffpNdWejzQrUVSiUS2/t04Ad1jN9100xVXXEEThwgR208pplAk9KpNt/qRI0csd5QgNbl6S9OjjpJ0drLnK01G9CR77rnnxE1OfmbPnq0/V2hq+Ou//usVK1bQ8ThRH7d0vG/fPlKKpNvEV54pX5pZ9uxhUuDTTz+l2VA8tOiUykzTgXiepYk2ZtSCCSiv66+/Pk/76Vy1uahZ6CVEX5CjVCQKxduRUwcJ/9/61reEh/379//kJz9Zu3atJQlBB+LJTZKU3nCoX8izcEUFEN/kY/qRK8jDhw9TU+jfudG/O0glvOOOO9TqjA3yINqWCkO1qKqq+sY3vhEMBsVVvRgbN24k6SyWn6lPr7vuOmo0S/k/+eQTUS8a6lRfenOj8B07dlA7iCxsy0/hd911l6ggNc5rr70mIqsFG7GVBHQ8/gax1dnEivYDXyxv3P4pU71Hz5wnqfpUY4wGx8/Xdd6/fscQlWx4eHnb/v/34johbZ9v2avHr9+byF+6UWhZsR/6F9Vd4kdC6NL/9crmpoPsZ8JTdfb/Y1Hd/A9YTPJf0rDry6+G46fOqR4o04qu+OfKNrUdOpmmMBbPTmUTV8cG9RT1l97+otcIOqBHPr1m0zENDHp/o+6mWZGpj5EmK0r71FNP0QRCB2KAkR6iUcpzGC80GilH+lecppl4beciS3J17FEgdLbriAEmRgtNMtQp9MwSDz5qalVnpz4WVcTsVFJSIhYvqN/pWPSp00ClaNS5lJ0YHmJ8iu8eZDIMCDHOKQvbuc52dtVTUaD47r7YHa4+WSy56znaOhRxMseitMZsJBPbT7BF1qZBtlfkj31sE4X+xb66o+yTushxpqT7TjNtretsCid5unAXW9Xec4Yp3ef48vZodTapczp+i+tXOtVNjW8xiwL+Scvw3iT7TqR3N9sHQg7PX5Kym+JQ1SiEkohwi86mOFRTqm/z4PCD0WH/AfbCQA1CzUK6mdTz7jPsp0VIu6tlFr+ESA5vb5dJKDklcdLZZXtY2sZjTGr/YQcrav8ZN3ePZK/OJujx4Pf7v/KVr4gNi3R7pP7FFlK33/3ud8XyoYAkKSngGTNm/O3f/i3ptjDfmGW5o3QsydWbUzwtrrzyyi996Us0R9ATQnwxghBzlr4aSg85kt30rzgdD8KzqC9Bk11paameLzXISy+9RIX52te+RlVbvny5mGjoEk1tRUVFlITevNNEGzOiYBYdRlWmxrE0FzU7ZU0zqZ4jTVikyCmwpaWFTm07SPivrKykmNTjFJmqQBXhDmQSCiTo+a2vyenfyPnc5z737//+79u2bRPhTD9qCpKEKb2eUdloan7vvfe+/OUv/+u//isleeyxx8SjYjxQg4ieIqhGt99+e1tbGzWFuKoXQ/QIFVJkrVeNyn/fffellp/ahFqSfFJrUJuI7iO3tuUnLU5vYjRUZs6cSQd0SoFUML3TM2klgo4t/TsGSJjmLaxVTShUUrdPNu7+//55/f/+fOj/+ae6n63tSPCdG83x43/78mYK/2/e9f+5tuPrr8slZIpfVNdLMSn+l1Y0PtfUr2vZAyfPXvtOhC5RErrqbd4jmjtVZ//bqubiDX3k/P+zeD3lQnmJS7qH/9ef6v7nS5s+2i0lnVNhLJ7TlG3MpN5fNFl985vfbOY/nC+6mMYDhdAYsOgPiuA0WdG/t912Gw0MGh50VdxrrpA6o9re105zkSW5OvYoEDrbdcQA0/UiPfjoeSrWZaipVZ1NUDdZnqoqoqP5nJf3F3/xFzTY4sqPkKYOVAqnzqX5auHChTQUaYTQmBTjM5NhQOjj3Haus51d1VR0unr1akry93//95Q7CXQx/Vpy13O0dSjiZI5FaY3HbmyVUltAqjEwwFQjXfK0sYVbuiKE+ID2RUaSmyQo1x9lclMk0X8tZLQ6m1TpIf6phth2olvmOpvs193D+/nOcoKKFDoqf+iaqhblX/Ek9ifZL4ek6myqafUhVh6C/qH66mvbpKTJm9gho5aZ5HI4IVuM/qH3EPHNUSedTVnUam1FHD7HNL2I44pltc4W0COE3rCFYsickydPUipd8YwZEmc0GY029wnl3Dn2Z3KpYPLcgQyjuYtortRPG1Nx6iDR3bYeKAkhTxSodzLva5oxMyyh64geSZ2yx19+0Wi2jaMzqlwmCFKoB0+dO3PBNCapQIdPn7fVqRTT6a+4U/wM/8A7ObH9u+jkITU8TWEspCnb5cJpsqKBMWldb5uXGPmTPBeBMUMqXP+BDidomJEkHe1ESqnST1Mj4jTXOc2uOmLmTBPBwogO02NRWuM30o6/6mabQ4TCztAo1d2d2fLXW0jRij+rnho+Yr0oFaWlmJbwNEZ1pyT0ryXcycaQRYY2BXQ2AAAAACaaRCKxefPma665Rv+NbTA2LEoLNp0NOhsAAAAAwz6f78c//nEd/8NeMgiMCYvSgk1ng84GAAAAAHANi9KCTWeDzgYAAAAAcA2L0oJNZ4POBgAAAABwDYvSgk1ng84GAAAAAHANi9KCTWeDzgYAAAAAcA2L0oJNZ4POBgAAAABwDYvSgk1ng84GAAAAAHANi9KCTWeDzgYAAAAAcA2L0oJNZ4POBgAAAABwDYvSgk1ng84GAAAAAHANi9KCTWeDzgYAAAAAcA2L0oJNZ4POBgAAAABwDYvSgk1ng84GAAAAAHANi9KCTWeDzgYAjJ28hbWwVJOtAwCYlliUFmw6G3Q2AGDsWPQlTJhsHQDAtMSitGDT2dzU2TAYDAaDwWAwGEw3i2BOY+l09ofBYN369TAYDAaDwWAwGIyM5LFFMKcx7BsBAAAAAAAgI1xbz4bOBgAAAAAAQAc6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMByHmSiXgikZQnE0siFo0l5DEAEwGNsfhoRzPGZdYzUV2Ers8tkgl6nE3O08wtslJnJyI+nz8UM7UkC7MEAQBGIBEN+X0a/mB4RH2SjIV8vsiYH0uJiN/nD8fFCcn7uOrJep4RuPNzjGQs6PNpQ4QRD9PYVLuYBQQdu5yP0EBUDiQ2pOxjmq6YxiVQ4e1pZrLuN9OE4GIXoetzk2Q8HFQeZyFtEsh6slZnE6abHU9bAEZLPOz3BSKatk7ESHPrAsWBcepsFaursbnGnZ9rJKIBn98YB1xmq8qbCaORhqmG8+jAuMkQcVteDh3q4lxjAl2fk/B5I6R9NJGMR+h0irw/Za3ODgQCfnVNw3LrJOOxSDgUCkeUT4SS8SidJROxKLsiwuWJ5XMjHhpMCQYgp0h53iT4LcKOxM3CAxl0R8iP4rVnXyJmvUUyu780TxQ7Egr4fMGwyMh6LnC6ExW/eGjmHExI63M7OwmFQobyZuvdxonjRE8HbJQE6VkRimijVyP1ijYuteQZPCf0vGNm5zmGo9plTWY0iOnMsWl4Q4obWrugd5eA+WGJUicE7YrExpWMw2Ynnr3qV4MlQ9fnIKnj1OhYhugzGjCWPrMN10dCJKR/yMtHgs3DyAWyVmfTawtfjNM/HlSetuyCn5otRg0YZAt2Ig6LEgiwdqLm8rM7LURtGKW7mSIZn2WzRT2eOhalC9b9KQDkDPwTemM9W8WsXpUpjB/K20jeR6O5v3RP1sdo6mPV8U6Msw96g+w5Sa6ZBIPOzjHYErbsU7ZIxSd7ba7nI0iMpjQTPU/NHowZ6mxlhGf2nEiw1bJgKEqjk8XS8s5BlKaxwPpGVlw5dGwa7YaWE4d2QztMNSkTgloOxRWL5A9KHcDjyM7j/ZU6NaDrcxX+XNDXs83IPhPPDKXPjHAaFaY+ZiMhEOBDjIaJNuAmSBZmsc6merJG0u5z5WZVt8Gzm0gufvCm03U5zQv6R5HKncb6KhTjoQwjNQA5SJwJFJ/Pzx4u6vu8ej+ZbhB+qHxqz05FxIzuL+XQdMwwnzvdiXw5U/PMcqWMXJ71wOWGCW3ex2xMsKVt7b/KJep754lefxaYxrGK4wjPZByzQajcBCxWzo5BXmuaIAyMb3Gwu4+6hUXRpItj07Ab2vgEmpIEhTp27AjTsenM7ErOQeyIxzGuKO9rKuj6HIXvfOSPM9LHyscclmcGPfRoDNOB9VlCfSm7lo0EU4qJlYXZrbPZoVTa5luHWjARj8dj8p3WuIn0KKZb2Djhn1hGYpRWECP/uItAbpNMxGIR/g0Sf1DeE+b7SblbTDcOod47I99fpnBnV4TTncgjGROgtaQgJ2C9zJ9k+jo2eySybufbSJTuHmGidx4djiM8g3HMD0JRMTIZUSVWrsEry1Z1ddQlQ3oC+xnaLenYNNaO03HsCNOxesbnBtWV6bVM6QiH3kfX5zQJmg/CoQB7nInOtIwKHR5uepaQ0BYDyzxEJlwWZrvOZidcaceNMGo9P3v/DkdoRohFgqO4iXgc/omSgv0HEQDkGvw+sJtnlLuFHSrLOcozLpP7yxRuikOYzp3uRGsiS0lBbiBXDWlwaQtH7BkYjicoXJNYNBRGnOidR4f5ijKsTBdMw00/4Qcm6UkonwblFKYmSIFftTaYXdM49oRjR1hyNs5SXOkB5hROeTrmaLpg8qWf8IPp0vVTGyYN/ewLvJZRoZMSzgLEfGMeIuxsQmVh9uts2Zx+ensRYeO5ifiRSUUAkKsk9I9uNZiWsZ1norqI4feK8hGa8VlbJveXKdwUh0hJY3snWhazmCBTSgpyBDauQuEw09ZaSMTvZ7vxtVGR0URvjqSSUfLUMSlOxOKsCM11TE1gJR6mTolG2b96i9s3DXPDVY9E/80+c0coU40lZ+PM6sqYgyxldeh9dH0uwneDmDqb9RnvTsszQ/9dR2s4SyDmm9QhMqGycCrobBYQMN6p+U0nb0L+2y6juYnExBGUXw3jySe0fQG4fLBbxRcIi41s7LvVQb8+3NktwX6ANCnC9Q/l2b1Cd4j4Mj7dIEaSjO4vUzjfPklZiDPrudOdyN+rQ7zQVLaQ8YYNcgn2DCSUj3XZ8KBBqY2ejCZ6Hsn2V+HNV5RxmdE4FqnlmhbfFpqz2ovXWt0pYfz4NPsiKtcpFEffOuLYNCy2/NY17y8tAfNvN9VYJgS1K1RXPAuplkzdZelKA3R9TsIGBXti8PGS5H8ZQh2T6jNDf5bIcHZsGpPWkeP0MHKJqaGzRXvpYfRKzKdo1ugxY/tURjcRQf3A9vYIB+YXJAByC5qNtMEuHnYynM0m4jf//QGah4wbhN0rPEC7xfTnVUb3l+VWE38lR5+zLOdOd6IeTje4umMM5BJcVisyOzUko4leDmOTI47pijIuM31OyK8QM6bQX8QYPbzWZkTz8C2bWsVNXylzahr1hlbCHaYaQp0QLFOHrStzHHNXKqDrc5NETPQrR1sMkhh9pjyzCFO4PlZSR47Tw8gVslJnAwAAAAAAMMWBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPfJUp2dVJBBGjLUQIZLkolYNBY3/wUnGdGMvGbF+QoAUxV+U2Q+sHEPgMmAJmvrXJ2NTJPbAb0BLhOJeIz/zXQ7hFxTkRcEzqOWS0GLV+lBQV4QpCvHuMhKnR0LFKh4irxhrSEtl4hS41qwtNBTVMIo8nhK/FHRYKlJOIEYv2omES5VHI6GZJz6VB4DkDUYd0WRp9C4kdLAbgLbuwMAt0hEyooLCotHMSwvF9PhdkBvgMtELFji4c+n4kJPSWrPsv62osVKhL3Go600qKRNRHwlHj6emVf97+nbSEFd7Y1QjnGStTpbqSk1p6ckJJrDckkh6vcUB4y/Sh8PlhZ4w6ZXkwxEdAZRHKCCjS0hABNHLFBEd4U8YaNbu5HSgEcZmGD4PKuNRHV+z0Jy/3ZAb4DLBD2firQFUTop9vii4tgB6n5PWYTHT4RKPMoCrL+oSFPa5itxEtB+W6/s4SjTjLYco2Uq6Gx+b5VF+KGjzqa2tVxIxKLmj8r5dDIanZ2IhoI+b4nXFwwbAp59HhEJUXBpmT8kgpOxULCspKDI6w8GtVcnE8l4WHOkFCnFzzBF0w458UgwGE1bYACcSUbKzFMM/yRNHpuQY9Hro/FnepSxoesvK2WDVBvaaUZp6pC2YuNPQDcbD6dbZDgRUXPQrvhDuBdyA/YgVMdlLGg8JE0kYuFgWamYOGUQw34+JRyGil18NmnTCKQrctSpjpxvhxEH+NQj495wqLxjbzhcsA0Wtzx1Cr9i6T6n3nAaH2CKEPWZFn6SYS8JvTT3laKMLQ8Jpgw15WZ5yjksgdJI0iT7qMsxaqaCzk5GfBmuZ5elVdFmEW2LEiUWLC4oLgtFYrFYJOhla+U8Bgv3lATCsUQiFiorFsHUsUFvQYk/HLVIewb7dKOwNEjXouSosFBOaXZ+SBepi/BUpSxeWQBZTyzA3uvZYypIkLB1mDrYCzyNRTZCQ2WlXq/2ysp0Oh+kRNhfLBcJHEep3ZA2Ye+PYIto7GajEoQDVADjRo0FSwpF0ehKiXEbgqmLGEDqxBbz28zMbLTQoIjGE/FIoETfzsDmU22sBEv1+VQOFZ+44C8p1FZolfg0uvVBx6d6r4/HZ+EF+gPc8XYYeYBPRTLsDYfKO/aGw2PPoTfYg73E6/XyVmc3ui6BnHvDYXyAKQO90XlMN5H5LcqKqoytOL4csvFj3NoGqmQfZTnGQNbqbL65hlNU5A3q7y3sUkkZUw0S5ZUmHvaVeCghW52I2kgKRUQ7oUZJxg0f7JVftDuL4dd7IKlHoYLZ+7YMAHoBD7MzBz8kWnQ3kNlgXPAnqN9fUuKlGyIc9DrIVDa4lY/J2CdD+iSTUL5iQi/98llGRzaj1OnWUHHwZx7pUZ9H+mcu1d1fdAW3xJSH9ap5IDKVldKvpik1mZBHNJ+qUfVTtgalDJVEVH62SO+a6vyrL96yQpQE9VVQI5rz7ZDJAJ96ZNgb9pV36g2nx55Tb7A8tc+sCZaZUFTOvcGS2IwPMHUw9SCD9bvjjg16StgJZgZ7F7MqcLZHm6nHkoC2H0SBDSsjwejKMRayVmd7g/QCyyGBUKy/rJovESkfgyfjlKKsRN3+LmGNZ27OFKxREvFYNBIOBpmA1+7uYHFBYWkgHDF/MTWlqyTUoerEYuDgxxDadGTZYA7AaGCjWVmmozGnbGIzIL1rmqPYs0x58LKvbUfDoaDfW6wPcdMo1ce9w5C2kOov5R7RdbwhxQXszcFpRQNMFdi4TFF2+ijSYeuVHvaOqP6eABsrvpCc/Imwv0Q8Ea2jWMIeqKa8NAFomeqNMqW5HTIb4FOMDHvDtvKOvZFySwsce4M/2JUrehHS9Ibt+ABTiJSBxsai8TJngg0d+6k/wT4MNY9gDnvORCMhX0lhiXUoWiT7aMoxNrJWZ5vajbWx8dpr06Q2kKYw7ToTjWczf6goUWKBEk9hSZk/GKSZJRpU56JELMy2ARYXFhgfVqV0lYRe4B3ztPPDEvAlAchsMF5oUJoXpuw+EaZpyvzubjx42Ye8niKvLxgMhaOxsJLYaZTaDmkdB3+aMx2tmKxo5unO/iNtMLXQ36M0rO9TGslElO/lZb8fJYYkWxBl34NRYZ9ppg4VQcr8q41uy9NAH/RpbgdG+gE+Jcm4N1Ir79Qbqc0ucOwNumKrs0fojdTxAaYQykuTIOU9TMeijHWSbFuI/g1GW9gnXaZRlyLZR1GOMTI1dLbxeE29JGDfauEfTOnwD81T7mm7m19BjyJS612R0hGCeLBEe9vWJgYrzJFporD7/U/FD3lin73Fre/xAIwaNpzVwWc9F7CHpTqF0dATg93y2a/peTzSKDUNaYmTP8vNxuLJK5YU7s9/4HLAxqHa4fQQNXVzKmyI8CgpwkuDD0h1AhbzLEtoGoc06vhjwfI0YKdiZDneDmbsBvjUZPS9oVfesTccHnuOvWF9sOuP0wx7Qx8fYCph6V0ms2zvKTbv211g+0VSXgkTEcvPUdBgUnOxkeyZlmPMTAGdTe+sgRJtXc5yOxqwpin2682biPiNT7kllpnVDiMKtby+ey8ZJWcy21igUP+MgsL1/fNUMHW3X8iv/1BI1Gf84CBLITrUyQ9BT4zi4mLLYjwAo4cePx79rqCHosd2VMWCxmyVZJ/jyE1SdDd4fNp0Ew95tV3TnNRRmmZIC5z98fkyGEskk0m6c0uL9XuXTbD6RMoq4LRDD0wp2KQoxyVbkUp9WIrRYnR9SJ+71fmUDQltXdk8VCi+1I6m+dfYyGl5GrBTOV6dboeRB/hUJYPecKq8U2+Ym11/7Dn1Brm319nOveE0PsAUgt+0IfHjEXx618aAqqAcFrNZfP3nKFUouu50OMkeNEpqlmOqhnYsh0tkrc428BSVlOnfdjRf4mgtkmR/kUOGsb9tY3x5UmKZWe1QovCfOi8oLOY/Wx5W9o3wbAqLS4roqvLj6GyOoYx5F1KvFSidST3NIhcWsmKFNDHu4IfgySGzgRuIbwfze6LE5/TrV3KwF/G/UhFTJQeNUU9RMQ9X940QNqPUeUhLnP0lY6GyUva1Z3bfGk9Z7bb2FBaST9uvtIApSSzIJkWG/VAh+Kik6b+Y+l/9to38wjuNVhoSSngyGiixGyoyvnn+ddbZzrfDyAN8qpJBbzhV3qk3HB579r3hrLPT9IbT+ABTCNaJQrapfagqKKaBbZQxDQQrxg1MY0zTgmysGonZe57tALcvh1tkpc4eFyl/S9OeNIJdxdFZZrmo2KcYvR8ARk8mo8xpLNoH0/SnftZmMOKQHnHI658mG+A2yUky6FbnUemQdrTxnUiTwSg9TRUyqZdD5UfbWKNuw9FmAKYUE9KHYxgaEzSWck9nAwAmlmQyEQu5+GfjkuzzZu0zq2QsqG8TAwAAAKY00NkAgFERD3nF/g557gIkrr3F/GM70zYxAAAAYEoDnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO6TjTr7l+9cCYPBYDAYDAaDXRaTknTcQGfDYDAYDAaDwWCGSUk6bqCzYTAYDAaDwWAww6QkHTfQ2TAYDAaDwWAwmGFSko4b6GwYDAaDwWAwGMwwKUnHDXQ2DAaDwWAwGAxmmJSk4wY6GwaDwWAwGAwGM0xK0nEDnQ2DwWAwGAwGgxkmJem4gc6GwWAwGAwGg8EMk5J03EBnw2AwGAwGg8FghklJOm6mus7+2m0v/dNN5VekhE+QfW3BaksIDAaDwWAwGCynTErScTN1dfYV19/zNzOv/MznhX37b77zzD+nxHHZbnroM5+/6Uu3pYSb7PWv3rjknxdYAmEwGAwGg8FgU8SkJB03U1Vn33jPX5K2/v6SK7iiveKmRz/3j1f+5TefmdiF7Ux09m3PfPbzV/7dTSnhMBgMBoPBYLApYVKSjpupqbNf+uJXrvzL7ywxBd7467/8/LVf1AXubUtmfvfOz379li9ct0hocR646IvXPvHV28q/dO2dn/3mPTNvYptA/vmnj86YfdOMa7Xl8AUvzbz20S97VhvhenKLzhZZzL7zbykLGbLoi9+9Jf/zV/7N1Y9+8bqXviYCF1B293yOSnLtM1/FOjcMBoPBYDBYlpuUpONmSurs2574m89/O3XN2Ng8fdNDf/X5b//NN3/9xWsf/cLXv/2Zf7zzy0LgMqF87V99/aYvUPjs733m8z/+3Ddv+pvvPEri+G9mXvmXVy9iypgvSH929i08+a8/R8mveOirmltDZ8ssHv3ST59grkScVJ192zOfm3nlf599z8yfPvPFb97wlzNvGWnbCQwGg8FgMBjsspqUpONmSursEfZvvD7z6+pqN1v8zr/2JXbMdba25s3CP6NFu+KnN33m80IrM52d/30en9mSL8y88m+ue50dG/myLKRPZszVZ3/K45j3jfzTd7/9mW8+Ixe2f7n6S9+88i8NzzAYDAaDwWCw7DMpScdNDurs25753Oev/MKNRshXr732M19/gm3tMCV8/Uuzr/yMpnotOltdLP+n71z5GbHUrSfnWXzu2iX/dKO0L35Tc2VK/tLfkUb/7jN6tC9998efmT3Bm8hhMBgMBoPBYOMxKUnHzZTU2VzLqkpamNw3kiKUmYaeyTX0mHT2V797pRTHis6mOGxzyLWKXZ+6nv3S333+yny+fcUwfd82DAaDwWAwGCwLTUrScTMldTbftmFZGP7a9Xf+5edvmMk0NF9FFjs9mLHdGtYFaRaeTmfLTSA8GtuFIqIZyVkWn/up3W9pm3Q2y0LZXgKDwWAwGAwGy3qTknTcTE2dzSTvX33+yv8++9Ev3fT612576UvX3kKnf6WJ5q9e++PPzLyF/5zI6iuu//Vfff7bcvE7Y539l/9IyV//2oLXv3rdLX/5+R/LLd1KcpbFFXd+if9iyddueuJzuuxesGgGyfRrXxI/csLU/8wb/u5GsdS96G+vgOyGwWAwGAwGy26TknTcTFWdTeZ55gtf/578OzX/eMMXflqu7MdY/U/fv+G/i0sz6ZK2OJ2xzv67ny5iP1QikosNIdbkShaf/x4Jaz33r157Ewv/yqPM2y/f+efrbsmXf0+H/TEdbM6GwWAwGAwGy2qTknTcTGGdLW3115x/lDrNJUdTN35k8FfWM/xL7PiD7TAYDAaDwWBTw6QkHTdTX2e7baYN1jAYDAaDwWCwaWZSko4b6Gyr3bboC1+/wfi7kjAYDAaDwWCwaWVSko4b6GwYDAaDwWAwGMwwKUnHDXQ2DAaDwWAwGAxmmJSk4wY6GwaDwWAwGAwGM0xK0nEDnQ2DwWAwGAwGgxkmJem4gc6GwWAwGAwGg8EMk5J03EBnw2AwGAwGg8FghklJOm6gs2EwGAwGg8FgMMOkJB030NkwGAwGg8FgMJhhUpKOm2zU2QAAAAAAAEx1oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPfJep194Uzi1Fl5DMbJ0T3BnkF5bOJMNNIXOyVPsoXkqXF0/dnE4KmkPB4fYxiBrOTn5fGEgrsDAAAAyGKyXmd3bs1b3hmXJxNO8sCe0O4T8iTniFatz1vSGuXHppoO9hYurC3cctkrfiIa2RPX1HH8441j7/pDnQULN/oPybNxMfoRyEruj8mTCWVy7w4AAAAAjArobBOTp5AuN1lZ01ipIo6hs0cGOhsAAADIYqaQzj4crGwJ7uGBxOAeX2VUaJnYlhZf52A80lq4dGNhZXf01HDyQK/vjYYCX0tgh9wMkehs8245nNzd7fVtLPBt8UWOiXAV8lO8fH3e0kZvZVtYbq8YSuzoLnujYc6KLWUbDiREmMqFAwGKfFSeUfzohpYy6fx8PNJWbM0ufS2OUvLCpY1BwyGDXzqRiLQVrdjgeaNVrxTHNhfifLyzkyorSi7WiFkjfLCHapFSU0upDgSqtniWNBRVdUcHh0SYuQGNhk3BqdbUSoPhD7jbD/riF+QFA9YUjXMXrp/nb6GMKEDo7CjPkWrhi6g7XpxqrWHR2XY1kp3rbyxY2lj8Qa9p28yhPf5KFu6lpjNpWdvxwGt3gEZCY4EvSjGFznYYbE4jyr4wNqPCsWwAAAAAyC6mkM6OlS6sLe3kgQQTUlvD/DDsr52znCRUbyjSW7a8Np+OfW3BSF/wg8bZmtjiom2jZ/nWQKQvtKWtcEntvI/NYpak5O6+gH9D3orWkLZ7Ib6hYcaSRt+WPvLs863PX96domlOBFfUejZoOy6SfcULN/iYYD0bqdwwYynPLhItXVo7u3IPd5m2Fks3Fvg7g8reCQGvICk2VqnQhq0kRks7z/ArTrkMJzY05FPJI/HY7l4KF5XVV4hTaqqU6lC3Z+H6wg+oMWUDBpQGVBp2vXeHLlh10tR6w7wVDaUbZHj+iqj1pSV5NBJp9SzcUExx+IYWluPi9Uqta4ubxV5kJZctbUVLags+SOkWVWcrNQpUbpyxcEuIq1jWRAs3lG7YE9mxx/+G0rkHOucp8T3LjWV1ZTxEvctrZ7zRp9du7tKNxR/3hnqOUkiaweY0opwKYx0VzmWTHKXK1ual2opum7dEAAAAAEwkOaKz897o5YqHCZGChQ1BufR5LLC8toiLMyZ9FjYKgcXY0zaX9JZMY8CiGZ/4U47rvT26moz7lupSzyDZ3KhrnWTzlvylbSw9K15DQFfyp3qLpP5OWwufVgsz7JJwy6EcR8qFJSnQXyQGj8X41/K4/tMko7WmolRDlHDWe/tEKJ3GPtiQz9Uki6/t7U6JppG21oVbxLuBjKb1kQpFM+8bMXIcppLIAjMpqXQlO03pSpaFcDUUeW/9XEOIi1O2Xj586nD0gFakJBV1Iy/52dAbtUr8M3SqNdo+72Kqjj4exCkdmGsnSm4/2Cimw4iyL4xlVKQpm85QIrJ1tkVkL9kaNlbxAQAAADBJ5IrO1iWjEk5wuck+tecSU12NjvuWrPfukCc6JvW5p23Owq0RecIgiWYjLtkatpB0TAaJ7Lj4VrNjy96FbNk7s1qYoUv5lVw+C5gOYzrVORe5PuqpioZ3H0tqmzR4I6TX2dQstaU9IpCzo2XW4hZqBDUtYU4uybTW7NTQ0wqmcKccWS5LW9gitzS2Ci7eLgwMnc06uuhjPTJfyNffypKnYjv6QhvavP4G7aMP68DgleLFoPGwuNFn5NtdvES8xVlqx4tqWrAnRc59ph9RNoWxjArnspkwS22IbAAAAOAyMZ10tq6uGMf8y03aSGCSjz1b85d2qloypl41YIu7bImUCW65mhuvs8RkcXhJMquFGbo0x7Q1QupR51zYcWJPn59tSl7PNiTwfSaqcjXV1CjVHub5gAjksM8HWAnVtIQ5uSTTWmvlT8EU7pQjy4XtLG9RzdhcLjB0NtWI7/lW47NN6kOxuoYZdOkNOo2GdnRrL0sU36zaI1tkMWg8LN5YrPqpZPvmU2qX2jhH/Ut5BMcR5VQYy6hwLpsVTWpDZAMAAACXjyzU2adinX1RbVOBsmLHl1p1NaPpP8KkRdLo7CXG1gv1o3kVk0IajBaatjew5ep5dULFmulpmbW0LUpF1VcxtWVgDX2dOLNamGGX1JcEfVnUMRdCV1dDiS2NeSlr0mYtqCtFY0VcoC9Rq2kJc3INx/K4qbOZWlW70hZDZxt7h8zwNWZ9CwcfD3p8tQXYfhVRjKPd8xY2pu41SqkdL6rtYHMcUU6FsYwK57LZMJTYsUf50icAAAAAJpss1NlM6s32x5hgvXDYv1zfB8xEyaw3ouynKgb3+VbU5o9WZy+snfcx35g7fCrsX2/zVTwRzaeHs/2vsiRygbDB8ksgGkwnzV6i7t5mW2+17M7HPt6Yv0QI0MxqYYZdWri+NMI3/PI2mS3bxCmXoxRH++7d+dgHUrCqytVcU0MpkrDO1xdBB2OlS6SwU9MS7NSmtE7lyVxnbyjbIf/Ci3OOai68XxZvDVt+wMTQ2eYa8dabV0ctwz7QKNoi9lCfiX6wMV+Lz79CqsXf08m+ViiLoY4HdmmeTGKnsxfWemwGm9OIciyMZVQ4lw0AAAAAWUcW6mxSiewnMsTu0hkr2qL698kORIuWiPCGQMRhJTiNzl7e6q/ckL9w/Uxyu7xVqFYrp5iyJFErd8GeOuD3UZLaGQtr8xc3+HYb33WzwFYWF5q/jXc05l3Oq0C2dGvoqLaymEktzPBLnQES5YvXk7c5PvlTgIy0uVD82VTyJY3BAyzcpFxNNVWV4pnox41zuEO66vlA/iags+o1Y1+eDHW22D5RKz4WSJejzIV1Zf7ijWU9Kf2i6GyjRovXU1fOfUO2XrKntYAlZ048dZ36Vg2KH6naSMWYuZiabmt4i/Lbedp4oEt5CzcUbRBvMnY6298bth1sDiPKqTApo8K5bAAAAADIMrJSZwtOnUgkNdU4bgzRNoa/iZ3BHwAnnW3zFUmCajHuP8FtiC2nNrHPZSg59tzH93fL3aj1yLBcRugXBarRCf0roRrnkzaBnAtnEoMOr1U0HpwuWXAabPYjyrkwFtKUDQAAAABZQ9bp7D9s3s3Xet00cqvrbMulcRp5Tg4Oxne0zdO+oGaJMH4jn7rOtlyCwcg++9Im/fifX9vCRiEAAAAAsoAsXs92lfiGxjn8b/W5TvSDjXOWNHhNf63QZSJVG+ZU2S2WAwAAAACAbGW66GwAAAAAAAAmE+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPdxTWfH+vdcvHhJegUAAAAAAGAaQ8KY5LFFMKexdDp7/4H4qVOnpGMAAAAAAACmMSSMSR5bBHMaS6ezj588uWfvvpMnTw0NXZTuAQAAAAAAmGaQGCZJvDvWT/LYIpjTWDqdTXb8JJPt/Xv2kl8YDAaDwWAwGGwaGolhksQkjC1SOb2NoLNhMBgMBoPBYDDYGAw6GwaDwWAwGAwGc9+gs2EwGAwGg8FgMPcNOhsGg8FgMBgMBnPfoLNhMBgMBoPBYDD3DTobBoPBYDAYDAZz36CzYTAYDAaDwWAw9w06GwaDwWAwGAwGc9+gs2EwGAwGg8FgMPcNOhsGg8FgMBgMBnPfoLNhMBgMBoPBYDD3DTobBoPBYDAYDAZz36CzYTAYDAaDwWAw9w06GwaDwWAwGAwGc9+gs2EwGAwGg8FgMPcNOhsGg8FgMBgMBnPfoLNhMBgMBoPBYDD3DTobBoPBYDAYDAZz36CzYTAYDAaDwWAw9w06GwaDwWAwGAwGc9+gs2EwGAwGg8FgMPcNOhsGg8FgMBgMBnPfoLNhMBgMBoPBYDD3DTobBoPBYDAYDAZz30bQ2YMnTh44cLB/z57dsX4YDAaDwWAwGGwaGolhksQkjC1SOb2l09mDJ07s2bvv+PETZ8+euwAAAAAAAMC0hMQwSWIS3CSPLYI5jaXT2fsPHDxx4oR0DwAAAAAAwDSGhDHJY4tgTmPpdHasf8+5c+elYwAAAAAAAKYxJIxJHlsEcxpLp7N3x/qlVwAAAAAAAKY9JI8tgjmNQWcDAAAAAACQEdDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPlmps5MnjxxNaHb8pAxNT/IkRT55Tp6pnDx+5Kjw8WlD9bbq3af5cQYYCV1lDG5Zg2TYDuPk3MnBxJHBpDybaE4qyCBbTvbVVjcfkScTzwilGStjcjtBZUll0jICAAAApg1ZqbPbmvIW1ipWd8PaPSOJgF1PL6y9LmijxfYF6/MWNjXQ0aHOGxbWXrV2nwgfESOhq4zFLWuQ+lUH5NlEcmTV8tq81bvk2cTS9851KvPvW1zdJy+ZOdKw+LrrHqzOtONMHOnr7Nw3KgV5pOHp6657x74k42BMblkTPd0wCW8YE1RpAAAAYFqTtTq7fkVUrGcfaa6uz19Yu6AxIa/ak4HOJgZHsSoMnT3BcBFZu+8k50hf7eL5181f1uzQQWNcbB2DfoTOBgAAAIArZK/OVmTloVVLpfg70hZZ/Fa3Lgf6GpsXV8e4DBE6O04hC5bWL1gdqT0gNz8oujZe/VbzijZNryeP1FaHFywNza+IVO8+LgMVnAXxuSPRzmUVm64uDy8LitwF547sjq5Yvfm6pZsfXNvZeVTZxHKgdwWPv2Lbp30mt/aueL129UU7ny4P3bf+iGyQ3VR+7ry6d5+sHOfovnfWhucvqV/wVnuDVmtG8khDcNt95VTB5hXbNN9HYyveaq7uO1K7dvN1S7Y1s6Dkvm0RHi3ScOjQZOtsRURyrbeKOpdtE6lu3neks3rZ04+zNW5616qu7SOpva+hurq6U0lCF7Tzk0eaa1cte/rpZauqG8QKNl1dtfi+6657fJlMzjlJTlYsfnzxiuoGPUwljeR0SpqateRIZy2PT7lb3Nq6EvU8sq/hncWPr2AbZWQTUTj3XqtUnSDvlCtz33zElGuf5ttIkNKmhHPpAAAAAOACU0Jn80XWiihJCYv2bVhdm7e8ne8nYDp71pLQNeXhxW+R6KxlHvYypaskUda8B3c9vaQ2f8mmB99qfrA8lL+w7uk2VakwHHT2ORLK+eSnIrIq2EwZ5S9vl6pl/SYR/k5j+4P0YrCkSa7N7o5ct7B2xnIqWPN9y0M3LA9pbh1dsXotrrtqcf19bzUvI4nMGqTuqiWh+aubF6/edPVCI+aFve03aM4fLK/LWxhaFhX6fs9iXsHFweg7azddtbD2hk94xQ+0U2FmLQ5d4wsvXt3ZSWXgHxdcV9FM7Xbd0vobqOkup85e0SkP7rvv8cdJt65qoO41Ynaumn/d47Vamr7q+667j4vGk83L5l+34OkVtQ3NDasen3/dg0wz2ujsIw2LZbyG6sWPz5+/OHWx2ElyOiW1zZpxpPZpfoFeBVaQGl78uO7WyRWr5333PXgfKWTx+sADHnz88cWrqqkq5NwocN87D143n1+oXkEXFmifBPRV8/BqKk31YjqUCVLa1Ll0AAAAAHCHqaCzuU69pjpOh+l1dl65vny3Z/Hi2nyuF211dl81id3NtYMslJRS7erQ1RXdWlqJvc4+2fvgwtr567VF8UOd8xfWPt1G0jZ5pC9W3XVIhu9lcvbpLjo6XVtRm7c0okkYvmAs3Dq64vUyiicaRInZ1TxLxrQ4T7xTrjXCyUPN23o7ZXHO8YbqZA3FdfY11Vrjnozep57yYl8mnU3CjyQx09BcExpq2hSz7x1DaBsy+8LJfX3NfVoCln6+EI0W0cxk+uIG7Y3qZMNiLZ6Cg852TOqUdecKtRLcq3Tr6IrVcz5709BgAUpUVmER9Ujt42pMpSVMG9L7VrHkrAw8f6VNnUsHAAAAAJfIWp1dO3NJ6GqyxaQ4a2eUdwoJkF5nq/uzm98iAcoWfe109pF3MtgdYa+zu5ryF4YeXN9bu01Y54NLlHwH46Ruq4Pb+Bo5SWEKilGmV+tClqSP7tbZlVIvjnWBn/m8Lkgi2ur8yPpNTKBLmXWapH/ttu4Vb4WZehYOuc7mBePwMqzYLc/ULToTD1eV9z3OuW8+6bwF6tqrKvpURc4kpRCIxpGESczmBrHwK9ObXXGRvqqhU6Nh1YOmFXVOSu6cEZKmZq2+EDAMt86u1HpyrAEsKTvn3ww1CXJTW5D0J7e11SueXqClt9TKuXQAAAAAcIus1dmGAG3eq63OjaSzF+grvkx4hPIWNzc76OxMdiHb62xWtrobVjcvfsswvuf7XN8nm2bQ68HSzQ++tW1VMHy1lLPWFwDDraMrW529qfqoPCOxvGJJ7ax396Q6v7AtTIr8HVLgfGMM+Z9f0by4untZuYPOtir4yf4epKE39SVhG9FnkpuapDRJSxKeC5hoX7aqurqhYZW+CcLsivm5j2+1MDA2bkvsJadzUoesrRLZcOvsypqGB6hfDWXL2GzdOqWMLECsdPe9w3aRPMi2ntQ2VJMcd9DZTqUDAAAAgFtkrc5WxZ8BF6lhTfuerPaZdDaXngKxg4JtBbHT2XwfhbHd4sKRvt7arkMWvWWvs5lOrVvcJfZAE0n6H0c4N+8bYXLW2FzO4Ts9hFtHV7Y6W4nJd5jwlwrmPL+iV3N+rvndOuXtQm8osW/ETmfz0/uatN8U51tZLtO+EYMU0WeOycTm0w3NJLP1MEsEI73ZFTuT2yucsZecjkmdsuZHypIz2yxiXLEvRUqLsAAlKttiInS3ZT1av8BzZd8lFbB9I8KhUTCOc+kAAAAA4BZTTGeLrclXVXR3Hog3rN00U98OwWVu/sL6xW1H2E8BBjdfpS1v2+nsCxeizeyrgdV79h1N7ItG2O9qGxpdwhNuWiY3dXDrI4eJ6vLa/KVNtXtPXkie7GsMXyUVMNsRPmt17xHSyicPVVfU5Wlylvupe7AxfuRoguJfTWWW5XFyZauza2csb244kDhyYM8qHznfVM0lPf/yJXN+cvCkqIjYbM03kNSvYH+UJ3mEh9vrbPEasCRcvTdx5Gi8uoJtd8l2nS02J89X9zizte37VnXy0yPNqx7UNxufbF5GSav75O9xsO8szn/8HbGNn0dMFbw898XVcpmdw/Y7OyV1zJr7eXBV8xH2q4XNqx6nAhuFsi+Frc6ev4CX/+SRvoZlD+q/fMi/7riYL4Mz73TCVbOorxDgJ/lPJdrr7DSlAwAAAIBLTDWdTUJibf0MJlVrr3lrV7WhR5mGfrpRbJZgdp32p23sdfaFc0eamq7hMcmu9hm/FajDE8oI0oQAHdy3oryO6VFmofvWsy9oEifbmjWHoQermxU5+2nt6pAo84zyzgajPI6u7HR2U3VbE9foZPXLukTliJOd1fx9g9t1a/UfBzz0juZ5Rnn7igonnU0Ro08vlcnnf7JrkveNjElniwiKzCb6qh9fQGEkSx9fVVutpO+rZpuUDT29r3Yx3wzOo66wyV44NyF9OSR1zPpk5zvyyoLFDX2mStm7stXZ7zTzrSAMprjlFdW7qTRHGjTPlGntCiednbZ0AAAAAHCDrNTZ44L95fCTSXkyEs5/rX1E7P98+ukjRx1yp/iDDn/yfRR/iZ3828akcLs/wcP+iHomf2ee/7n1sbVDNnEy478dnnnMFOyTOjtMk9VoSuEU1T48Y8+jKQIAAAAARkXu6WwAAAAAAAAuP9DZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO6TjTo7b2Et7LKb7AwAAAAAADAmoLNh9iY7AwAAAAAAjAnXdPa+AweOfvopDAaDwWAwGAwGIyN5bBHMaWyE9eyzAAAAAAAAAI6b+0akSwAAAAAAAKY90NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPvkjs4e6B+QRyANA/39g/IQOJIjrTSV74nBgYEBmz6gYBd6ZqC3nejFjAEAAGBiyUKdPdhbU25DTS89X9k1fmBloL2yvLI9s+fmQFN5eZMrj1hWnFF6osxtK+A69mXrr6+oqJ+QNyLHrhmRjFKSZBy977E2ttpKg6pYHXslLwOjuSfSk0njj/auGuytkne2QGnWgfaqiorKqqqqyoqKKsXnQFNVVU1TU02lUqv++srR3svULuS2t9+uUqaajneiMA2dUZPpiM9kkI+vJAAAAMZMVq9npz7nXNE54318Gthr2bRk8lR0hTGUbVyMvWsySTm2ZnOhsc3NOPZKTmUyasbR3lWOTgeaKiubtMmETvQ3noGmCpEBvQaJA+qOUavsNDlbL422ShbMQ2e0pCmlmQwijq8kAAAAxs6U1NkDA73t9TVVNfXKB78UpJ8NatftPxm2uB3s722qr6mpb+o17RWgYHsndEHL3PL8svNERekfoH8pCV+Ek09FHmDr264s9oWxuiacyyYZ7G9v111rsZsspVBwKJBNeUTXGJFs+kDtI14UcWZJSbEsWTJXVeWVNU1K2e38c0Tni3A7CWJqAeZFbQ95oh3Qf5tqKsur6rU9Bs7jzwLFSGk3WWGbrqfYFE0WPCWJqFG9FqplrjigIOOEFV6cGcF0xIdKk1Fq7oZOjMwICrQWm8XLpPH1u4o3nh5MUGQjmob98CT668ur1D6jeFJoW3U2uVBXu21IHb2scFp1rIVKqamskv296nBnaFBG5qHDcL6beA7GfWjb6ArOg5wupFY5tSSp0QAAAEwEU1BnV1YydUD0NlVVVMiVL0Wr0VO4kiKQPOmnCCk6y+yWfUTNP0Mmd/VGbP6RO3MyMNDfXlNhLJqx9TU9fk1NTdUInli5qqpqatrZx9Q8WlM5nSo1UHwrHtQLDoWxce1UNgOj7vLT84FBembXV9ouC/ICGSVN3zhKD7DjigpesH4WQ3aSGoPQ9YEaznZssFIRvVQdGYHlU86rJkrJ/LMSsIIZ/i0XmqgJTBlKSK7pgSTqjB0LrBxC0umtNMAqXl7Tru0xYFFsx58J20pwr/ZdzxuA0PrOnITS8JqzEKV+lnj6XgrKXOpU7pYfigyYqGIZk+bSMjONM/tiZ9r4ppGlbOxg+0NsdipRpCpqVrFPWtF6bFhqBeKwsssAbd9IFXuvpPD0Ktv+dhpgI1JWx8iVk1JTVqVKylH4qKd2k81t8q1MGyrWoeOUxvY+TCmKitoB5kFu34MpJXEYnwAAANxnCupsffGWnWpakl0QzwtKpDw4Bm0eIYpbemKpGoBONd2kPIQpF/1z6io1Pt9imt6TpcAic3W9TpdFogYikGHoJcfCWNvCqWwKet1ZauWykYGBuUB8mTJNeYweYNJB3Rc8INczlRgMvaNM4YOKb1LBWripV5lGM/wop/1NRpsR1IRqhhoUS4ayFdKmJq3ZKDt5pIwQczuxM9vxZ8a+Emm73tJ3chyyJMoFKq/asEY83RkJbj0Lo115sfWcyanhhVwaCezbniUwmtGp8U1tpo1RcxlV+EshSdheJtbpfUXLwJQXg5VdbRvxPUgKTa+yReXlCaHcTilZGJgvsSopWRtN4jxtqLAiGGVMM0EoFTGK5VjKdIPcoQctmTh2NAAAALeZgjpbeSoYzw/lAj146AlO0k55mJgw3LJnTDtbF5KwNSTF+8BAP1twq6+qMOKrz0rKVdNZTp4sBeaZq1LJ8JnyvDNnlloYi2vnsqkYdSeBVF7B1uqcmsnq0ExKeYziUM7mKkosBdaFREobDTLXvfzTbi1cj0ww/3JNl8NW63hJUzJOaVOBVjNqAvov6RZ+ptTXaCVeOKMZLUU1XzSTWgnyal6q1XJMbQCtIkpBCAo2O6AKGBVmSrLKvJJs9IiagdqW/JLq06btM2t8tahGQVOKbA9VxHizMArHsGtjChMLv4OsCLYDOKXrje5NycLAfEmtEqFfZL6dpw0dU8kd0zjdh06l1MeGhrWmNj1o24Y2HQ0AAMBtclBnszNSgPU11p8r0NDdskRs26IJvmY7wLYEVNWzTZz9/CNcPb4mxTh6/o6ezOUiUh6fMkB3paPHtC+MxTU7sy+bCVOTDg70sv2ilRX0oDeViWGfnuHcOFp57T+KNhfYqJ8STocVFZU1bCsp+e412kptNraqy3auqvAFc3VtlqOmUhGaS8hsdsaECy+1uMzSaXU3N4OlCuaLGk6VIK/2xWNu7PtOKQiR0rCm/NlJub4JhGABIr5xxDC1iuGCjuzb3pQgTeOrI0su16aIQicotYhIKZQaE6lKnZzzuDQOmXTtrRfdaMKoloZRCVP9zZgvmapkXGS+7acNM2oZ0qaxvQ+dSplmkFMm9j1obg3HaAAAAFwnN3W2Dgmq1Oe84TblmSXhS0TymPnWHv1MZKjuKN4InlLKxTJXZYEuI0xrk4R+wakwFtfOZVNJbVLGQLvNHloq0GgaRykOxUhVPgRPqRRYnhkpLW2gxNcjEyy+rX/rBVY2JUOF/vqKpt72Knmxv76qt7e90oiqtBLzaTSYUkmG+aLEsRKOXc/cmJpaT2PpLmvDKg1EPkik9VPeenyjsOZiq23JL/EUzm2fWeNbisqF9oCjzFZ3ITHUVrLujTHnRwGirehAJKGclYwljreTuToWzJcsVdIvOt3sFvSmZWSURr0PnUpp7QDWQLJYjj1oLolzRwMAAHCbnNPZdGAsClmiSxS37OlrRGDfJuMX2ENRfxT1N1Xq8ekZZcTn67ojeEopAWWurLKz0moPPZMH5YJTYayuncumoNedyRn9Mnm1rBkyrAWi3NiRQ3nU4pCMNTmXNeGSQB6y76jK+EZKVmtdQwzQie6Ryq3sizb5Z4nsMmbhtvuzCX5NFz7mM8I0QlRZw84Uj+xUL4cG82ZbCebVWHBm0fR2qSlXljLZ1mXpQCkIx1RxJR67IMqvjzy1sMYRg5waJ+wSj8/KY9/2LMHIjW9X1Er7NzUGK7yeARt/+okYdnoVlMwIGuRaRLomDiknNYrE6XbiJdUvWDDX1FIlI53JN6uKqYwalKlp6NimcbwPzUVRUTuAVUy5jxx60FQSx2jsh0mMXyQBAADgBjm4nj3YX08as7KKfQxbpf8amoLZ7UA7PWlIErB/pAiUPngo+2K+vlXDfIXEgJ4/w86TpcA8c56OSphaQOGBoAvGN6scCpPiOl3ZNJS6y+LydtIrboGrYWuB7MtjLg7lzvyyhCSMdOc8T4ISDvRrskVNqbUhr4L6kTZFYik1gTDAfrKBl8HShnonmLJIgeWpCy86qzItKJtGCHtlIX3MI1vanJ2mtrFTJbjXXpuu517b2Y9iiCtGb5iHKkNW3NywTHkpaVLe8czF5kNQO1Gq4Nj2LNKIjZ9S1P5689c+rQxyL3w0GDceRwwwfsV6fygfOvBolWxLuhJmQlTIMnp5SZ1SmGtqqZIpnT7OUkqvoA4ddmqbRoam3IeWRlfRHVkGuRae2oOmkjhE62+qsPn0DwAAwHjIap0NQA6RokQ1zDI4V+A/6yKPAQAAgGkJdDYAk8M00tmDg3wLBKYEAEDusa/n3qWhq58PXV3e0nRchgHgBHQ2AJPDQDv/AyupsA0Qpv0RU5z+pir9z04CAECO0dqUt7CWW33FPhkGrBzeV1/XumhN86JAW1XrvoFTMngaAp0NAAAAAJAZ0NnpObynYmUoXzaRbqF7P9wzcHaw6Z3QjDW7ZMzpAXQ2AAAAAEBmjE5nnxn89OjAYc0+dX1d1+z/+BkZfLk42PPA81Jb5y+uY7trnq/7omyu2que5/r7zT4ZWXJqUC//KKpwqr+ZrZe/2HxIBowZakD3+8UAOhsAAAAAID2aom0IC9VIOru8m4ekU2mHKsqkyhQ2Y2lTzUEX1XCq/3BFJt/3ObyvpiPu9n7FuFaY0JOq/D11tGndphl6Ia06u+9J/RK3L5a3tn8qrzkx0LA5//nN5Q099TtPyKCxUv9mapHcBDobAAAAACAtn/bca5aDijU3yUipMB38ow910XmsPVCfV9bmnmCy+D/R3xC+amHdou6RpHxrk6vFYAyGw3y7SKh8pyV3tl3E2Elip7OfbJUnZ08dqqqo++qbfenfAfo/rHdLHENnAwAAAABcZgZbm67WxaJh9RWxNKLWooPPnj3Ycf3C+sqD/PjUoZp1zfeWh26uaKnoOMyDzvY3t7wY1uIf3lW+pqVG253SH25Z1HBAnkhS/J8900satLxD+yr6sfa6lgdW1v9oZbi84YAQrwOtrYtWhvIW1z+wprlqJw+yizZKjlaVswbJr9huST7YLPS3Zul1NtHRlL+wSb66sCYK37G0/o43W2v2ic8Njtava36grC5v6eZFovyslbraD/aVV2y6OhDjcdiuElkd08r6ofoPm+9Yyhq8Ulv1Fzp7oLmVOuJHK5v1cLeAzgYAAAAAGJkUqZ1eZBMpOnhf248Wbqpiojq26Pnaq1a2VnXHmhparl9Y92QH3wLR2pSvrTQPhjdTLlpy5uresGWbRKrOJt3ZdYfMgmnf/KVNla2xpta2B56vvb6GxRyM7aoJbMp7PlzZvL3dOdqo4ftqBm020Zh2YKdESNHZ1AKL+UcEx3dRE31vTVdN8/aqdZuvXljPV8pP9Hdsr3wzlFfeQuGs/KxJQ98r3/Ri3faa3qMUoWlNaAZVp3l7TUPrvc/X/mideFNh1aQGr+k90N7Q9D2t75jOfj50s5bLVWw9nkd3CehsAAAAAICMUKT2iCKbUHXwiYF9+yor6vLLu/hi84mB3gP6D6D2rgvlix/iOL79AfkNyzMkAR94MyzjM/W8ucb6i912OpspV+Hh1GBsX6+215mp9rIOqdVM+0aco00Gis4+NTgQ2/ViWe1V77Bl6YG6TfkVPfryMj+Vi+WmfSNcZ7/Yq/XFwY6bqaH0Hd7sNMzbjTeLWO8mb/sOCMXPdPZKPZcTNRWp7TkuoLMBAAAAAGw50VvXwn4HWlpLTewMk9qLN1XSQW/Hi8al5vJmufdDgelgZf279qqVbcqX/E4N7ttX39xTEWi+d6m+oYJJvTvqjrIF78Wbaw6T7GbymslfKdBV0uts4szgpweamrdXrmt+oLzO0NbW/dkO0UbBYNOHRlMsCvSk/EHgE0017FJFt2VJnulstYm+t6ZPvIcwybump6Z5uzS2Bt8q3KbobOO3X1hDLW2u0lM1t9wsl6j5sv3zm16s29W+z9gcYtmfTafQ2QAAAAAAk4BVBeY9HxY/GDLY23q95ZKNPFV1MN+3wFdqGZ/2Pfl87Yylmx9Y01LRsKsmYAjHgbpNeRU9g93NX2XCmuQm+14j6b+bmfi2YKezme4UK9/xSlKWi+vvpXeAuu1NNWEHne0cbRSYGiq/rM2qs1mp6BLVRQZosITavhG2l4a/YxCHKstqv7dS0e7MukbU2f019WLrtmr6NvTe5rYXKzZdvZgEd1M9f+GBzgYAAAAAuCyk6GxuVz1fJ77bl7+47ip2UEfSbSSdffbszla2LZgrQiamjW8rsn0jhtpjW0SaygMhkZBiXr2ubdFiQ0oqpOrswfo36+T+ClLqYqMzh28IsdPZaaKNAqWhnm9usmzC/nRPuVjXt/Gs6myWe/7ipnq+PaZpTe3Vcmu1lTQ6m32NUlv2TkHf53OsamXtV/k7D3Q2AAAAAMBlwV5n63ZH3YGqitovVvT01tnKU4sOHqypqJV7jlub8pe2tAs9Kv68i6H2WKp8/Qt5fCU4f6mtdlT8Hz82sG9f1ZpQPltx5xf5wnbVYS4uhdJVdfbzWu5poo0CU0PNKGuqaD3Av/h4oKmuma9kk9nuaDfpbCpN+dLa7wl5LV5LdopSHqP3hy9qf0sync7mHq7/UPwwy5mB5qarSLiTDxFNFODUAaqmaDfobAAAAACAy8KeF7U/cDiizVjZM/L+aab2xO9bx6v43ye/iu9hqFD2jRBMRxprzPvKn3da2WX+9QLkL667+c2OdqGYGYNNgfoZC2u/uLguf+Gmyg/VDSFsrwgl4Zs00kTLHCaX88taK7krvUiKhRZ1GLuiFSw6m33T9KsLN1WxVwVSyUyjU71Y8cpamrSt7Wl1Nr239C1izcL+FGX+4voXZb5n6F3oavLzPPN29Uq5BQU6GwAAAAAgFzl+bODT8f5FwxE4NThw2FbgmskwmiODTXU9vWzp+czAzp4XV4YUtW1R/6OF/yzgGP6kPPuD6qlte8LhxwcnCuhsAAAAAADgLlwfp/uj9NOCaaezB/t724l+9tI20M//M41BCwCVwYGBAZsBQaHyaGpivumndF1wx9ph6tQB3tm9WojlFAAAJpcs1NmDvTXlNtSwP4XJrvGDMdJfX1FZ39vfz8UEnVTUj6nUA03l5U2TNHGreQ26LBKMFjBysW9iFjqOGpP78XTbaBlnaZ2g1jfqMFFjwNTHbmfSXy/vJonhe7C/vrK8orKqqqqyokLpK2rKypr69voq5VYZaK+c4N50tRHUm56KXl7ZPiZnoyuGq7eqMZ7HPmdlAe7elsaconYqHVdUNWlTvPV0wjBNDQAAoJDV69mpTzY2U4/jEe/WRO+2/EmHkpe7zykTRi72TTzOrI1n4qQwMQ1lrsQEjQFz0V3OxLFdTJ1OJxVV8oSORfhgb5WMMNA00SrbUs5xNoJjpUfJqIrhVqYCd71dNtythvl21LCE2kdynUnKBgAwFZmSOntgoLe9vqaqpt7yYeBgf29TfU1NfVOv3WerlKipprK8ql77BHmwXx6wI+ZqoFd4NaVmnznzYOWjx9SCMSiD/gHhRFsxcyoRhXOfTbpPozAc5kucaXlRBK34SiKrGwWzR/bpqX6mXTKiGDUSTaynk8hHJG94mwZO3/LcPfMpG9jyIa5dCxODPDPb6GnblGKneaDbphatLStnyU7CLlaVV9Y0WVrMsUpp24NhEyelj9NmYpsLq4plEBrIbkiBRHRFk3L/DrRXyvVBakmRQtPZA02a3h4Zhx4coXFG2Qj2eWjQVfNNTwEyImspp05nyVJaVr9FLKRUM6UGPMymnOmKwFLwBBRsjGfjjh2p/FpaGU/gdE9psLispZu0xNKVZVJMrQwvl+qSkqVkkX4SGWFYcET5Rb7KYJaZsUJoNymdW04FdrlQxN4B4bpehjpGY5mJptUvsISmqQEAABSmoM6urGSTHNHbVFVhyIOB9irxCSFdqK9SP/qWDA6wFOU17RSDOzXc01FVTU0NT829ah/N8k8kKbeBgYH+9pqKSlka20cuK1sVuWlnn1KyAKcSyc8yBwZpaq+vlD7NLo0Hq3HBKL7pI1GzG5WBJiNPvmFAO2PeeQ1NLSCP2EWbxmOhWhNRdSqVBh6x5bl71sBGv+nFdWhh7cN+evz1U3Tdp2ObNsnPh1k45VSlNKaBmlophKhcvSxdve3OAipHO/UvSyyusRarrDIapFxbAs6kPZxKktLHjpk45MKqYhqEJmgUkHri+1WZspChPNxcSsrXcCj2jVSyETMalW3fgw4VV8m8EahwYksIhdPgMb0rCKw3Pe9qnty509neDCoijUgSnMZAZ8WwKa1dNW1uVbtB7lwEysx+PBuFcE7seC843FMG0idPS1U3T4r6GLWtDIUqdw29lYkZxoRwbzuJZHLPUHK9t5uYH2WMisOBfuMmTT11zIUald0zerw00dQG0Z4SKVMDAAAoTEGdrazTselcRqA5WJ3Y6TT1ocuTGw4N9+zISM1i6WdigubwVT89vrlghKVsjiUyF+Ks9G92qcRRLphS2rsx0a8LbZLcTU1NWnmY3OJH5hYQR8yv9mQxYKFqdZQV0Exanrk3JCJ/4huy1Ag2tbBSiEHtyCEvls5cOJv+4XWoUX7mXy8Eu6D0HDWWjdC2lolVScnTkKoZtIdjScQlo+iOmTi3hHkQmqCLFRVsMZALlYpyVe+Zm4vyVbuLJBX/zyhUtqW1ZA+mqbhKho1A2s5USvOphsmZKACLxI4cOn1QGZJKbvotomJbTUumhO0gdyyC83g2CuGUOG1au6IqcJ96FIpv3AfkX7/iUBntSB4ro1PC3FvKJqNlcM+w2UztXho6WnWYX61m5kqaT51yYY2qXMgsGstUP7PkCgAABlNQZysTGjsVEdjjkK1ZabAFhtSJz4jPMNxbMrLMmoMDAyRM2GeQFfbxBZayOZeILQhVsGUR5XllcakUVLlgKb6dGzNUBv4ooKj0X3pQ8TMtVPVtHFmqIWGh+lOFQS7FQzijlif3xiObYZSBkdrCFKGioqqmvdf0+xdOeZm9sdLarWez1KaSacksVTY3swFVQoml9AtDv+hURhXHklgzd8rEMRf73rPH9J5hrrC5qgJtX/Yg9VWaMadj14NpKq6SWSNQL8vlTQFbQx7JmdE+lpaytgGrZa/YdCJjWYqhYTtQU9wRqYPcsQjWVjHGs1EIp8TOaR2KqmD2aRoERuE4NncsV82yaIbkVmE+TGXTJhGn0azCets6hcjiqcU2Fdp86piLuWczjGZybckVAAAMckRnswO+B1MldbOcHp9juHecQQfYx4NV9WyDXz//dNA2vsBctvQlGhzoZfv/KitIKWs5qS6VgioXzMUnUt1YEA9dIbPZGXtSDbBVbnHZ8G0cmauhkZq1DGD/Hbnlyb3Zpx7g0MIMuQu0kq4bRbPLi4VbtIWltIzUQK0Q7IpSPNvUhLkSRotxVF8jtYdzSazXHDJxzoVdUAqZFhZXNJshWSQpAZQ1H0OUhn9A3l5jFj32WHvQXDmGUXEVczynRmBLuGxTrIq+D9fA5IydiNTGEceIRUcV5Jhat5eGZK9eQEsxFFIGquqO4TDInYrADhzGs1EIh8Rp0jJsiqpg9mnqG8ON4x1LUfjKb4oi1jAXhZAB7D8jziHqgjrHKB5Lr5XUVGjTqXMuRqMSGUYz52TJFQAADHJlPTtlFrbFiM8w3DvNoFxuiDDCWBlKLRhhKVtmJRpol2twhnOGUlAlL3PxVXQ3VvrrK5p627WP+/vrq3p7lR9ls2sBSzUkLNRUHV2JZVRP5l4tn77i5dTCJiiSeHA75UXhpic7JbDxoy/BS/RCWKrs1MxUCSWW0i8M/WIm7eFYEmvmTpk45mLfexL5dTENo7VFKh7I6be0JzWo/CidWpYnoRSO2dig9WCaiqtk1ggslu2gN2NyJmqqpVaqoMeyFJEPUB7LUgw79IFqzpT7kMfskrnZU4vAetfUTORAv6IdZVj+sxQuU5gwiqpg9mmMNkL371QZBhfaA04ym/uwnUQyuWdYYssUohVPLbap0OZTx1zMPZthNItrU64AAGCQKzqbTbvKmu5Ak90XA9X4hOHeaQZlU67+yOhvqnSIL7CUzbFE7CGqJyanUmgwbSO/7T7AVtS0DJS8WA76o8bBjQVKUlGhPzfMZ6pv48haDQELVX5XmX3VSitJJi1P7pUVNFYM+Sh2aGEWQ/eplMixTdvV71SxBTetzVRMqZVCKBkw2KlNalYJ04ZY0xig0xHKqOJUEpG5ISccM3HKxVIVM+xipbbxleeq+yYhTVdEOta5Jh/99XpEKoHMx9iqPWj348HMvRFDL5VzxVVY/EwagZdau2LKUYE5UyPpJbHvdObGuMnoRItlKYaAxda9KC7ZoXqr2k4jSnwGO5UZOI5noxCOiemIWoV/zXWQyl9VWSkusHC7oiqYQ5WGFpe4G6fKcFiHVNrukmcwH/aTSCb3jKm3WWVGvT/bKRdLz2YYTXVNx+rU0NuUshwPAJi25IzOJgba6aFI8zz7p8bu8W2Nr7t3nEEH+9kGROaTVGLvaPaNcBxKJIOr+H4PPVjkRdAkT89HLQNTXuyRW64tRNm7scBKZWgZeoSoK0p2LWBTDYK3XD9/4vPqmPMbseV5i/LXBx6pSvv5LK3WKS0swmXllOiOeSl+TM1nQaQmyKv+VStLlXld7VLTBdZBQhtYxgCvoeZixPYgbEvCUPs4XSa2udj3ngFPwzG3Kr9CnvgVc5EVlU0MtJNwq6qqNLYdODWXUw86Vlwl00YYYD8IIRvBUiENU/mM9rG0lBJLa1c+lEbcN+JYTaUGDoPcuQimJOp4NgqRPnFTTRXBfptPuReciqpj9mlqaMO/Q2UEbOXdSWYLH06TiN1otqLHoZwH+vXiqcU2FTrl1D6X1J7NJJq1edjNw6cGNskar2oAgOlOVutsAEC2Q3oDoiJ7YauzZhE5kfCfOJLHAAAAoLMBAONhcHQ7tcFEwz5x0LYAsfX+SpvfyJsQBgf59jU8BwAAQAE6GwAwDhx/JQ5cJgZ769neELYFqLJGl9wTTH+T2KciTwEAAHCgswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA94HOBgAAAAAAwH2gswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA94HOBgAAAAAAwH2gswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA94HOBgAAAAAAwH2gswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA94HOBgAAAAAAwH2gswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA94HOBgAAAAAAwH2gswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA94HOBgAAAAAAwH2gswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA93FTZ8NgMBgMBoPBYDDdLII5jY2gs4cBAAAAAAAAHOhsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeZDjo7Kf9LR8bhZSNph7w2odhmkw0tAqYuEz2o3Byfk3SfATNsfktFXpxQHLNxLf9pO6KSiXgsnrh8tU8mYpFQwO9j+AOhcCx+2crCh7OKDB8jSWrYbB1V467ctCXndXYyGvBHxehIRHzB2GUeKPFwiR3huLw+YSSj/pISfyQhTwUscOKzBrnKOAYVPShHfpy4OT6ZL20mAJMGHyKpTEJP8Kk2YJNPLOR1Z8alHKbfiErEgj6vlwtcn7fEG7Dc/ZNAMh4OiPxNBMKXQ5+mPtBHHhLp5j4SKSW+yW/TTHBrBs1o6s8xcl5nx8NebUqNBcuyR1RO9mNfPu98phZggdnTJGCqMY5BlZFGcXN8Ml/TTxVlEUyRTOZkIxSQ9S2QxkGAhbtRkmmosxMRvzcQ1Zs0Sc/USX6oJiJCZAfCMa0YyXgkyJe2L4PoH8uonqLjxq0ZdBreNjmts+m9KRYN+32haIwRDfnYYTw73hXtBm0yFgkF/f5gKGJadk8monq4MrXQPENX6NU+EIyw98NkPBoO+unEds2e5xcOB72qKDLrGMo+HOTuovr7ZiIWJud0heUf5hNsgk4C/MRUSPuyg1wmg0FlNzDY4A36S8oCobDysKRRLkefPrCkKxrltkPLccw5+TJuOIoQNkY5mATsFIlDD1Jwykw06hmPPc4DAasMZELR7/eZRqjdaMkkNy4YEkzlWVJz0rjlKSKiBGwHBJ2xNpgC4zEWtDy1zIuTTv1p0xR2beF8R2skokJlh6zXY/R0J5R3gMnBblRz5JNTGxxaeVmtTXMfj5fgUxz7tJ1dVwaCU4PysSgaVAZNPOYZ1ATrYK2caoFSyplS/WlCLuts6mPWpyQJGSGaclnvarf0ZSZl0MbDPq8/zF8JIkE6lK/myVjQy8LjBL0peOUHoTx5IMhfISiYPVECPDG9WHjtVhi0/MidIYpYoDyh7Et8NB1QNrFIQI9D04iPP1tYsQJebyAQlCdBylQrv0PZQY4z0qBSBwYbmFqkJBtjJfwFWDx3k1EacmJgsXjeUEzE46M8oAwt4znq4NrkSyYx7hhxSA9rdVEOTAopisRh1nCYicRYGMWMx1VwjGS1HEyceJiEd1QZoWlH3gi5scmRRqeR2nYQprglXxFyxUe+bAMSWiTIfEb6bIW/pjhs0GB10epMTyr9CWZ/Pzq2hRbP7ilC2XM5LZazqcnoVDRtUgrtyX72OOpsMTiMQeOVe1atcx+L5/f72Tl/m2OtIlvIoUGYJGANmkiQRjd9vDCxqEVTcZAo9uVMmfqnB7msswm6x/1yGGbD7mwFy6Blk1HQeB6wU+3xQMNUHBCxoNxxyJIb27jioRLjSRMPe202eBn5xUNl6rNLS5dMGOOe2krutmHTiF6uBEXX25CKKK+kKTvIaUYYVOxIm3QZNO/qj0EaWMbwJ/WjPh/1U+ZAHVrsKS8fq46uafirqktPopWVi+zU+wNMNBZF4jxr2M9ErP9GNeOJEcamKWOgxEJsmDBXMnG6kTdibqxGamp1EGbiVpyH9DaZEl8zS8ZJM5V4y/gipbqUbWpoEllstZoOnO5Ha1s4jwcDTUzrGJtHSJMKUla6JxY+BsRKnkBbjGYXjPqwyuqPTjEyxbElHo8oLjo2CE+utVsyaYymCcYomhVbieJYTlP1pwk5rrPpJtemsazanZ0yaNnncWH24iph73zKHJSI00txJBwKaIvI5uSmkcsupdZUTUBzUhlvFmtULZ8gzaQinM0CegxzdD3T9GUHucsIg8qYcgVJCtAeNuqQJTFVJhe8OGwBkT9hmCvTyzEJFnEXO7o2hJlES8LLGgobC6RgcjFNJZnOeMZMpI418/BRRpyKFkVoaxbCBgkbNUb8tCNvxNwo2PRMUQdhJm4JumtK+Jrl5fzxjrGQiLOtIGVealvRAlrdLTjej9a2SD8eBBadra3kJqIhbZn7sujsgNLV+i4adkGpuDqETMPJEs9oFccGYevEXl+A3nImd8ykDF6VVIniWE5T9acJuauz4xF219GbN7/7fD6aEsouyzcl7DEPWrYYYHorZojtW/Qa6y0L0LUITcbRkExkTm4aueySaWbjWBP42LuyEpWFeP3BEM8mFlEeJYYzs2ctU+eygxwn/aBiA8NYrGPEteFrjB5xTCM8JEeORGyJJVcmD+wTFebb2TX5MhZRGEYSf0mJLxjmn1xjdE4+pqmE96D9rOEwE1nHmnKijzgTehTjTUuuCBrx0468EXPjwepQUgZhJm4FSVKsoaCf/3jHlBuY7O7ibzGpd6TA8X60tIXzeFAx9o0wAqFINEqNp/76iPoxwmRgHtUKlgvqEDINJ0s8vVXSNkgyQZo2GGCiZvKWDewGL8NBojBsy2mq/jQhh9ezk0l6/S2h11tOlC1n03/lxcuOZdDSdGS7Bqy//QvYOy5PZE5uGrnsUuq9Z71JuCgKh7U5jxwr64OJiPZFIdMsYPasZ+pUdpDrpB9UdGoavKa1LXXIskcKX+6zwvybhpY+TB1d62uKGloSvazssYA17cnHNJWwU9tZw3EmMo+1DGY8I4rYq0D/yiFjxE878kbMjdVITW0MwszcWmB1tVsSziLYl/bMJaQbUdSZDsz9KZ61jvejtS0ye4rQzSsFNSGWrk2L3JP+pmIZ1QaWC+oQMg0nSzyjVTJqENa8polwAnEYvE4SxYRaTlP1pwk5rLPZwMjW3dmpg5Y9C/RJIkmCRYxKiubVC56IBLwykTm5aeSyS6l3XupNQnkYP3BFN7W+EsAmLiNcmQXMno1MHcoOcp0RBpV5YJj1LY0e5SkS41+klI5YRONR7NU9sJGl3wzOrlVfShKlrDQvQGlPOqaphHCYNZxmIvNYy2DGU6Iw3etlWxzkQFLipxl5I+ZGwXydTlxxHISObnll9SRUWSlZs5hYiO7HiLYPIBmnJ5I22ZvvyIj+/QmH+zGlLRzGgxXtd/2IFJ2dTb/rZ7mgDiE6NuY+SzylVRwahGKY21ltxInE2mESVh4HiWJfTlP1pwk5rbNpFtM+yoqHskz7pQ5aupP4F0zK2D8B/WdR4nS3sa+dlDHB4da+EQEb/FpU/ucH2CYb9uSI6p/WmmYBs2c1U4eyg9xmpEGlDwyCbUPVZmMGEx+kyHXdHOXjj48gf1B+PsoHXIw9m+UFdWSlc819sQs0GLUHlamszKf+EACTgUVQEPazhsNMZO0/5cQ8L2mYorC/TSN1nzW+48gbMTceHKdxLBL7tMScDNwyxPTOtjVahne2IuoloZopN1GS/ayEzR1pez/atEWGTxGKpu4U0ciWv1NjDA51UJqGkDr3WeKZWsWhQcSfCioTu2JNg25CYUUzIwtqL1Gcy2mZ+qcDOa2zpyT2e1smacPLODfWZNG+HJBNZDounAfQWC5gME4JbDtqsntvnPk5pc7Q7dQbq1RipyI7XMi4jplEZL86niV/d33isW+QjNtzknAqTraV87IAnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPm7qbBgMBoPBYDAYDKabRTCnMaxnAwAAAAAAkBHQ2QAAAAAAALgPdDYAAAAAAADuA50NAAAAAACA+0BnAwAAAAAA4D7Q2QAAAMDYOXv27OHDhw+Oj0OHDp0/f156BADkCtDZAAAAwNgZv8gWkB/pEQCQK0BnAwAAAGNHymQ3kB4BALkCdDYAAAAwdqRGduaAhjx3RnoEAOQK0NkAAADA2JEa2Q7S1ps3b/7www+bmpo++OADOujr65PX7JAeAQC5AnQ2AAAAMHakRk5h//79GzduPHr06PHjx+m4ubmZ1Parr77a09MjY6QgPQIAcgXobAAAAGDsSI2cwsqVK2Ox2MDAwObNm+m4oaGhpaVl/fr1f/7zn2WMFKRHAECuAJ0NAAAAjB2pkRX6+/vr6+tJZL/yyivV1dWksJctW1ZXV9fT07Nu3brf/OY33d3dMqoZ6REAkCtAZwMAAABjR2pkzv79+0lDV1RU7N69+/z584sWLVqzZg3J68OHD7e0tAwODra3t/t8Pr/fLxOYkR4BALkCdDYAAAAwdqRGPnhw7969DQ0NJLU7OjpIT1NIY2Pj8uXLd+7ceeLEiSQnEom8//77Xq/X9udHpEcAQK4wXXQ2zW7yaGqTTCQSuVETAADIDaRGPniwrq5u165dXV1db7755sqVK19//fU/cerr67ds2dLT09PX1xcOh2tqahYvXkyKXCZTkB6zkmQiFosn5Ikd4kVCRV7QIA/RmL2HNJeGk/GRM5ZHOQNvDodaJVh7QAhMGaaBzk5EAyWegoLScLrbdIoQLs3LK/DH5RkAAIDLjtTIBw/u2LGjk9Pd3f32229ff/31t95667333vv444+vXr2a5HVLS8vWrVvXrl3r9XpJhctkCtJjtpGM+ks8hcUlJSVFnkKv/cM0FihIIRCTF4djwdJCTxE5KCku9JQY4USaS8OJsFdeo4xLg6ZrkkSoxJTTlCdGkoW3NWsOX0Rt7ETEJ6+l6YgJIRkL+lRCMQj9TMlxnU2TQzHdm9GwHzobAADABCA1cgodHR3l5eVPPfXUnXfeef/99z/77LO/+93v6NTv97/44otlZWUynoL0mF0kI2WekqB88JDy9RTbCl4zsWBxkS5+YwFPkT+qKbOYv6hI95DmEpPQHl1M0uO8qFgrhU4iXFpUWlqSOzo76vcU65VJhEo9ZRGtcaiuBaUhTcmwM+PaREMKH9J6jOS4zk7EovzTFXrRzhKdHQt6vT69KImwz+sVk4q4EIsGvIUFniIeJclPCgq9AW0SEjo7RnFZeLEvpM85evJSz5wiORXFQ77iwoI55C3o9HkcAACA8SE1sjObNm26/fbbn3jiieeff/43v/lNcXHxI488snDhQnlZQXrMKpig80XlCRELFnlGkrVMmisPXe1RLEiGvcYTOc2l4Xg4qIq7mL/AGzZOCSqaxxeNUwFzRWcnosGgsoRNFSzU3zuScdPWGuqHSat2MhYyr6xrsO5L0L+RUDAUFqWjoHCITsxbW3goxXHcHZS75LjO1sgenW1ekY77C/LySsPskF2YNXtOQZHXW1QwIy9vtsdTMJdOij1z8vPyC4O89DLSXE+xdmGeX9xm/Mqs2TPnUZIykt/JcOnsvPw5LKK3aG5+3gzpAQAAgKtIjZyWF154Yd26dR9++OHKlSvff//9F1988bHHHpPXFKTHrCLqszw/Se+O8EA1LWZbSYbLPCX6sqyJNJe4rlTWugm2qYSt6LIXgVzR2WbYa4fTmnUsWOLxK68/E0oiGghG4/FYlFAFNNPfpKkjsThdjAR8wVAoJE+C/oCmzCmWP0DCmxELB/zTbGUcOnuSSa+zvREePBzxzsrLKwqJoUhzmhopzxPQasKi5RtX8rUUlMY3Vz1lEXXnAAAA3ENq5LRUVla+/PLLr7zyykMPPUQHixYtuv322+U1Bekxm0hVsfRATbugbV6UNiBlyDYdF5UGU3RWmkuSWKBY2UJBJCNlXGXblTA3sN+iw/Zos+3ZxWWTKGoSkYA/SAo6RlI7FPTrSpnpbKPLqGiBqFaoRMQvr9CREczSBP0R6/6fXAY6e5JJq7P1C0q4cyQiUppP5+w2tPgNzDN0OoMJ7Tm+HJyIAADgMiMkcnpqamo2btxIavvJJ598+umn//CHP9x///3ymoL0mE0wFSs/N5WMoLPZurPt5WScLYeGg97ilK/wpbnE4IrT7DPq02Pmps5mLxa233RMsFXlSNhfUmj50uikkYgGpFLm69mqzjZO9K0mXFer9aDkqu7OeaCzJ5nx62x1A0jMN0euUzv6FbA18fxSLGgDAIDbSI2clrq6unXr1vn9fhLZPp/vhRdeWL58ubymID1mFSnbRFI2kqg4LWarsJ8IUXd8K9hcSkTKiklkG+tGRNTnKQnGxI8HJuMhehOI5NRv+8WCJSP/nAhrBKdNNhOLvnCdic5ORCwbRfQr0wTo7ElG6GHtHTRaNmfUOlvZ/5EIeDTdbdbZdJqfN9dYv1YiAgAAcBOpkdMSCoVWr1791FNPvfLKK++9996f//xnEtzymoL0mFUw5ax+/zBln7SK7WJ2ImL6PqOq1NNcEtgu69Ij3YZcWdO2Wb1nkDwNmvZbsHeSyakz+7hB/VpjPCwFdSY6m8UOq+XWU08ToLMnGa548+eVhWPxaKB47oz8Uevs/Bk05cUTiXjEXzjb9A1JRWcPR7x0bR6N8mQyEQsW04m6jQQAAIBLSI2cls2bN3d2dpLIXrZs2SLOCy+8IK8pSI9ZRtTnKZYf/SdJ9yqP00Q06A8ZUs9hMZv9+ojmgM7o2afHSnOJiIe8GfxKdE7tG2G/Rmwjshn0KPd4dYXK1Ljzl01dhu3P1nOOGycZ6WwS1sZ3Irkrk+zOeaCzJx163Z9D6jovf6YnEDZL6Ix0dmkwXMod5OXNML6JbNXZNJhDpXNn8Gh5+XMm7W4EAIBphtTIafnwww9JZ7/55ps+n+9lztT5/WyCBG+Rhy8as78Wo0kpUuB+T4Hxo9aOO7OZ1PKxPxjH8BTJX7OVOF5i+tmKnftc0tk2K/WKeEmyv+kjgwstf8NmgknGwkE//xM1/kBI312dmc6mKzy130//D4Sm2y/7TROdnXskM/z765nGAwAAMCakRnZm3759H3/88c6dO9esWfPee++Vl5cvXbr0iSeekJcVpMcsZfxboNN4yL2/nT5hJNFWUwnobAAAAGDsSI3sTDAYPHbs2MaNG6uqqtavX79kyZKXX355CupsAMCogc4GAAAAxs6hQ4ekTLZj7969W7ZsuXjx4v79+0ltv/322yS7n3nmmXfeeUfG0Dh8+LD0CADIFaCzAQAAgLFz9uxZkshSLKfQ1dVFCvtTzpEjR+h07dq1f/zjH19//XUZg0Ni/fz589IjACBXgM4GAAAAAADAfaCzAQAAAAAAcB/obAAAAAAAANwHOhsAAAAYO8HDnTdHVjyxo+rIuZMyCAAAONDZAAAAwNi5YdvyguYXf9KytP7TXhkEAAAc6GwAAABg7FQdaru5bcXjvWsHzh6XQQAAwIHOBgAAAAAAwH2gswEAAAAAAHAf6GwAAAAAAADcx02dfSyRgMFgMBgMBoPBYGRYzwYAAAAAAMB9oLMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAACAsXP8RKKja+u2zqbOHe3bWutbtnwUad/U3rU10rmlZ9f2Hfv7u3d1UQjZtsiGrU21rZ1N7T1tPbt7evf09g3s6+nvjXQ2RVo3nDmblB4BALkCdDYAAAAwdvp2R7c0fti7p697R0ukbUOkbX1X77aOnuZIZ2Pv3p274vt6+jq7trdG2hsiHY3N4WBLZGPn9pbtsWjf/l3dfe09u7v3fhrf1hrq3d0lPQIAcgXobAAAAGDsdO1s30F6en+sa3tzR+emlvDa9taabU0ftrTUbN/TR5K6PVLfEY1si2zc1riuZUPltq0fR7bVt/ds3b53x/Y9vV072nbsj0Z3dkZjndIjACBXgM4GAAAAxk6ko3HHvr49x+KRjk3d27d1djW21L+1bfN77e31XdGWrp4tbc3rW7d8tK3hva0bA+FPKrbU+Vu31nVHt3bv7Ojt39G3d0d0Z9v2/uj2XdDZAOQa0NkAAADA2Onoadm27ZOtDe+0Ntd2d4V7d3e0d27q6d0W7W3t2tHW2rZpW1Pt1sYPWhrWRbbWNK9fs/nj15o2Vkda1kd3d/fu7Ir2te/Ys33nvr7O7i3SIwAgV4DOBgAAAMZO46a121rXb928bku9vz0S6t4V7ehobI1s3LE/1ru3r727qXHDOySyG2pXbdm8bmtoTfij1xo/WbNlfWXX9rbO7qaenR079uzYsbunNxaVHgEAuQJ0NgAAADB2Qh9V1K59aeMnFZGure1djZ27+xrqqzv6unp2dnX2duw7keje2XPw5HG6uqm2MtzwQWPN6obgaw3VL4c3vNvcFCSpHd3ZvmNXR3vHZukRAJArQGcDAAAAY6dxy8cbP/FvCb29bVt9a7R985b1G2oD0Vhve2d4R/+O2MD+Hfv6+w8d2Bnf296xZef+3e0d4d49uzt6WmoqF4bXr9666b2evm090a0dnZukRwBArgCdDQAAAIyd5kjD5o1VWzZVN35Y0da2uTEUaNr8cVfnZrLujk19Ozt37du5I9azc8/Onf29PTva+uP79x0/1rlre3u0Nbz1k/r1axo3vdvVu62tbaP0CADIFaCzAQAAgLGzdfOHW9dX1n/4xsbqlU0fvdHSsr554/uR5pro9q3bW2ujnZt27mrr3xs9OLA3tq+vb3dP/5HDu+L7ov27dh06UB984+MqX131yoYN720JfyQ9AgByBehsAAAAYOxs21oXrn0r/NEbTcHXmj9a2d7ySduWD3q6Nne31XY0vL0jumnvPhLZe2J7ew8cjfcf6N81sK9v3+6d8QPbOhrbult6dm/f2lTb1r6lY/d26REAkCtMC52diEVjiWz5e7bJRAo59ad2ef3sauR4YfLIgiIAAHKOloaqcNXy5g9f7W7buO0jX3dXY09HfVdToH2Tv7WmvHvr2p29DQfisQMDe/YfGdi5qzu2d2ds3669Rw83bvl4a8sn2yIN2yIbt25d37OrR3rMSpKJWCyekCd2JFOQFzSScRsPMqqCvGDAnuHpcrZLk4uI5jEhr0wSeIaOhVzX2bFAiaewuKSkpLjQU+KLpJsiJoe4vyDPQoE/Li/mAOFSViVPwNrSMd9cdqE0LM8vA8lQESvCXF9MBgAAgAs0feJvfP+l8PsvRTa+2xNt6m77ZHvH+khoTUf9mmhrbXvoje7IJzv6mnf3d+/s79m5Zzvp7P74vp3xva2dm7ds+6itfXN7R2Nbe0N71u7PTkb98lFa5Cn0hm2fpLFAQQoBfbZNhL2FniJywDyUBvXw9KnokRlS0tnmnAiVWNLkKIlwqWgfhVL7vpgAEtGQ3x8IEgG/PxS7/GJq6pDbOjvq9xTrd18iVOopi1zuNzGus4uD7KVQJ5feDoXOThGzEe8sHn4ZdXYiWJiXN2vWLAhtAICrhD+qaFhb1rT+ra5t69s2r20Lrenc/G5r3esd9f7e7nBn+P3t7Z90d23cGevZva93Z393LN4f+/Tw/sTRgyeO7z06sH1XV3d0a2f3ls7uJukxu0hGyjwlQbkeRILZU2zoZEdiweIi5fFb4tFFMon2omLNnQVTKqbCi5VnOOnMkpBF4FFgUWlpyXTQ2RZYr0yazE5EA4GI3mXxsD8Yw7J2puS0zk5Eg0FlCZvux8IMZoeJhetsZ7VJg7msyDPHU1TmVxffZXBBYbEvxId6MhrwesvEMScW9Hq9rHb8QMcn7kJzoIgoiId8xYUFlKEvaLygmqKrF/j9JQrolSWxwHT27Nmz8+aoYjYZKs7nEteoeSIWZBlTjcoCUc0/y9cXjlOZeBZKDomwTyk1P9NPmStRVI5DFycCnry8olCYBL9aNrNjXgKlcayVFRmb0aLbtmRqEqVH5KEV2zHg1CVmPywz42ykvgIAuMLmD15r+fiN3t5Iz/Ztfbu7Dn460N5U3dUc7NhUGVm/qru9fnvP5p0727t7tsZJXh+O7zt0cO+h+MHBTw8eP7Zj767tOztJZ/dsb6F/pcesgulbn/IXdGLBIs9IstYiAuPhYEgRZjF/gTdso9Os0jHqM2Us9oDKYw4VzeOLxqmA005nW95IJph42Kd0YDIW0rYHUJfEE/RvJBQMhUXnUFA4RCdxUw/zUIpj6cHpQG6vZ5tIhr0FWbKe7aCzY/55+XkzCop8fr/XMyc/f56f30TJSOnsvPy5xf5g0FfEggMkmrh0NVZm2XrxLG+EjpjQnTWvWFNkYspigXMLeUjxPF3uJsPM8RwPi1w0l7IuDIobQI1OBcnLLw7xdot4eXyvPxgo89AhL4kZlrQ0QJpWlIbDJG5+sc+n1zwRLBQVDQT9xZTx7FLRLyzxnIKCOQVFrEQFMygH0QSmZkuGi/niOD8VruYW8gordbMiZXZSaSmOpT948cWZXWU10Vw412hkprMdW1JkoEVlyeQuIZaR3YYhhzHg2CVmPywzeZZBXwEAXCFSu6ql+tWOzqaWzesOHIlHu5t6usPtm95u+bAssuGtHTu29UQ+2Xdw18EjBw4cO7InfnDPwf37Dx/Z/+mR2MCBzp4mLrLJmulf6TGrILVrXjclmTzCQmp6EUhCvchuQcSaKhZgqjuZIA1HhCLWz37ZXhT2UGcvAtNMZ0/qYjZBytof1p4hyglT3KSpI7F4PB6LBHzBUCgkT4L+gLZSxBIESHgzYuGAX33nmgZMG52d6WddEw3XdQoz5hb5o3zIceVsbGzm4lRIyRi9G2prvqFiTQiGS/N1haWIR1sBF9ZTKcKSbZnOZ9JTYPEho1M0SpBfysKT8YhREO4zVbyJpKwumhLk+ZBj5kd6TUTDwYjWF4Z/nlgpPC+g8GKUmr115M+eLQU13wyi18GIZYVdka7MQtuSxKh52spaGtm5JUX1tKhKMosHieMYUApGWJrM8MNqI84y6SsAgCu0rivbFnhxW/WKtm11rfXvdLdvaK2vjDS819VW3xl+P9q1ed+hfTt2NMcOxvYciu/eF9u5u2f3gV39h+J7jw4cOHa4O8pWsvl69hbpMZtIVbGxQEHaBW22qOUsAmOBYtstnKmpKB+vz1dS4g2G6ZnhLfEU+8TTkkNSk6tsuxLmOuxNZbJrHI8GA2x/dsAfMJ4tTGcbW0gSEZ/x+XQiom0uoSMjmKUJ+o09KNOAaaKz2Y3t8N2NSYbrOn1/dizsK9TWGiOkm+cUB+htUOAvnq0oKLGSWuyZPUNflGdyTkRgR9pmCDsBp4pJ/TgemKdIVMLwwnx4yngpAmWeOXOUpYckaWQK9fHl5pSMDD3InAnByEQoW3i3KNpkjIRg0F/GV2dluLXsMd8cKVj1xMxbgT+s+2L5zKYXKN6cUWPJ3Iyi2IVXqV6lY1lXosxjKqNTZS3q1rklzVGVY3aoM2O2fOw4jwEW365LWLiRIlA8RynXSH0FAHCFzRsC4Y9WNa57Nfz+y03ryrq6mjqbP4p8UtFat2pbbUV324a+ne3R6Nbtu7v6Dx2IHTqwc//uxi0fR3e0b98R6Yv1dLOVbLKW7p5m6TGbYCpWfq4mGUFnpxOBfMXL9mJqKhLSBQVq5JhfWQeP+vSvRU43nZ3+PWZC0FakE8lkQl2S5uvZqs42TtglvqDNdbVaWrbZW9HdOc900NmxYInTF6QnH4vapCHIfgWDCW2mmIwNHwLzzlu2Y2DGzHn6SgCTc0w9MeGobyFRtJwOy1RbYNULkFISY5WU+ZCbFCjPubNllslwKWliEoVFVDJSe3bajSXlTrVC6YJbyS8W8MzIy5/JdnuU+f2kDWW4tezGErhMzAIsmj0Z9dOrioIINsHKYkFT3dzTjNlzJLNJkEq/6SprLmialhRR1f04Mhk/9EWVly1WIBZqPwbYFZsu4eH5M2Xp58yZqX3GkUlfAQBcoXFLXX3du5uDq7asK9/y8aqtDdVbPvGHg76W9W9t/bB8Q92ajvZNXV3hzq4wE9PdW3bF926NhCLtm7t6msR2Efq3O7o1S/dn04RmVnUpG0lU0ojARKSMfa1RE2IqtqnYb5GY9DMFyEhRn6ckGJO/bRcP0ZtAZJr8th+1wuQvZscj/qCqjBNRKZ0z0dmJiGWjiH5lmpDzOtv57fnykKLJ+MZhJsTYFWM/gxOmHQBCaIdIRRoiStFyOiyelsgoANt4YuzwNgoifBhl1FaVecpiPZiVJFW7GUnZGvKs4tIibQOJmjFLqWVsiGl+QV9ppptRewWRiT2FhfmzzcvbIlZ+vii3GqzAqp9f6JNrvkSg1GZDCkcvPg93rKyogXLq1JJ8Y4smudVkZg/qmr/9GNALxtG6xOqHpedn6YsPAHCT8IbApvdf3vj+y/Xvr9j00erQe6989LZ36/o1kS3BjqZPVq0pbWr4ILIt1NmxuWdn1449ff1HDg2cP72jP0o6O9rbQjq7O9pE1tWdjftGuAZWv7botL2a4ywC032sbJ+KrVOr34NkBRErDEyBp5JFT/qJI817zITBhLG+O5uhB2Sis9mXKE2pzV+qzH1yW2cno37lR4GyAiGAxL6RWIQpvnxtay/7Sl/+3FI+/pKxYPFsXUqxbwryXzNKRMooubY8SpMN2w+Rr27ptQgvIhENFCk/ZscLIBQb/6rcPLoTkskEz0/bY8x8aHtb4hHfPKEi+V7oAh+/Peh2olAb7aboQS4383TdaGTMdO+sYl6hZJzlqyVhifPz54htIKJEsq48MfkqFpO9pRJa5ZRgBS6zzaGs3cROD0sSvfjpK2ttZIeWTMZpRjQ2rKjJ+GHKerbjGODxU7vEWhJWG36WUV8BAFwhXLtm8/p360OBTaHq+o/XbHr/lY/X+bbWv9++ra6jvWFZ+QMtW9d3RRo6Wzd2dm7t7Nyya+/OWHxP/PTx7mhLV/fWzu5wZ1cjqfD2tk3SY5YR9XmK5Uf/SZLLis5jC5sh4xnrKALZz2A7fqzsLB2VjEl1lxV5TD8/ojGd9o1chsVsgp5GPmNBOxmPBOQXITPS2ey3r/TvRFIkLfF0Iad1ts0r72S/BqYgBaPGjNker/Gba0maVGaSJGLkz9FvJrZLbYYIzZtBIkypAheMqsxOkYDs3NjTbRGWiVDpXOlZyY+nMZgxV1u7iAW04s3w+H1FdtqNJdW8i5/z05ZnlYxpVpXZkqj2kRSV4bzsvqBWW6XYPLHxVUPdF1sMn1WqrbSY6qZhI7NlTNZsliRK8dNVlhfUVHe7lmS+6cRY+FGSsUMd48uwjmPAFF/pEnNJWI7yLIO+AgC4wqaG6sbwx5s3BxsaPmgIVjTWVdbUvNUcDnZ1hjs6wi++fG99wwdbWuuj0ea21vq331kUadvY3b2lLxbt6tna2dXU1bm5vWNje6S+raVOesw6SCcXefhDlP2RGe1xQkLY7ykwfgvbSQQyIWzFiJdWOsbDvhKRsafEvCpqMH109uVYzJYkoqGA38dRvgiZmc6mK7Fw0O/z++n/gWn3R25yez17asL+sqkxjenYBvM1Y11+jhHyPAoHDsUbNTZ+DNE4uiJNHKOsrGvFdqmRXXIDAEhHw6Z14aZPmraup3+31lSEGz9qqF+37sNlbW2bIi0bXnr1/jfXLFq78b3WyKb6je98UlfR1rGZ5HV3z9buHrYtm9R2dxcp8oaOtnrpMUu5fFugp83ma5B7QGdPWeKRYJBtODCWc6c85sVZAACYCmxprA6HP9q6tW5bpLGzu7kz2t7WuWVd8M9bW+oi7Q2vv/m4/73HmiMbIpFN1VV/3ta6oTWyoWXrR50dm0lbd0dJajO13dEV3tEnP/0DAOQM0NlTlUSwaM6cOZ5Sx6+jTEEiZewH66CzAQBTicSxw60toXDTJ127t2/fv7e7b3vX9kgo9FpjU1Vbx2b/mqeaWoLtHY0NmwKhjavbOxra2zZFWkNbN78f3vhua2R9Z0+4o6uR1PaZMyelRwBArgCdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAMHbWHRn+nw3DN3YNf3qene5JDv9by/DM8HA7tlsDMO2BzgYAAADGDknqvLrh/23D8AN9w68dHJ7XyU7JvrVNRgAATFugswEAAICxA50NAHACOhsAAAAYO+uOmveNnB3+BvaNAAA40NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAAAAgPu4qbOPJRIwGAwGg8FgMBiMDOvZAAAAAAAAuA90NgAAAAAAAO4DnQ0AAAAAAID7QGcDAAAAAADgPtDZAAAAAAAAuA90NgAAADB2Ljlw8dLQ8MVLF4aGLg6dv3hh6MLFM+eHLjQ3Ny578Sa/f+mBgf1DFy8MM6O4Fy5dHKL/SI8AgFwBOhsAAAAYO1xU23CR/2/o4pmOtpZN4c0fB9e8s+b2t974ycmj97U3/3T5srkdbZsvXjrPZDbF40iPAIBcATobAAAAGDuqPhZyWcC189DQ0MU33/zdqldvbm26p2H9Ldsa7y594jsdbT/q7vzJm2/8pLWl6cLQ2UvD9D+G9AIAyBWgswEAAIBxoUvkixflyjSDDi9eGLp44UXvXYf2Fr+96qFDexuiW6sGdgVff/kXlatuW/zsNaXP/PR88sKl4SFKKjwAAHIJ6GwAAABg7AhNPUxK+xLbY61zkU5JZw8NlT59s3fhr4/0t1w41j70aXTDh75Dva8e37N60TPXvb7yP7q72oQkpyTSIwAgV4DOBgAAAMaOkMgktNl2bJUhCrh4fujcn58r7WgKnhxoO9rffOJga7T94093rznQ+cazT/3HsrIf9O5ol/GhswHIOaCzAQAAgLFzceji0IVLu3ZFt2wJS73MIOF9/tKF4QsXz5UvX/TBO6/072j8qKri0/3bli76fSJWuafj5SceufknP7nq7Llz/CdH8HsjAOQg0NkAAADA2Bm6eGHnzh31oXeXL114ie3P5svb7N+hSxeHBo8fq3jVe/5Y5/PPPXrsQORiouv8sa5z8XcPdr+yYvkDDz/8uwsX2Y/6XcTv+gGQi0BnAwAAAGPn0tD5urqac8d7qgJvsF/L5oqZbc7mivvUqZPLljy14ePViX0dA/3NNeve9L3y7Pn4mqF9b5+IVz715O9FfNLY7P8AgNwCOhsAAAAYOxcunA83NhzY9eG6d7zJ0yeGLl7g34AkY/+7ODT0x5LisiXPvvj8wrJXlr3+2suvv/z0hYP+fdt9Lz5794bQevUnSqRHAECuMC10diIei8WT8uRyk0ykMFlFS4R9ZcFYZrmxYqbG5IUfd3GT0UCZL5yQZwAAMKV5efmSipXeFS+WBAOl5UsfP3bs6AW2ZVusUrMF7mVLX+jc9lHz5ndX+144/+n2CwMfnOx/+6H7bxs8cZxE9pCMydS29JiVJBP0JE03cSdTkBc0kuxZnOpBRtaQoQwZoiAv6CRi0fSFmrrY1FbAJM2k6YbR4o5KyClyXGcnIr4ST2FxCVHkKfRmgbaL+wvyLBT44/LixBIPzMvLm10WladpSYaK8vLm+GLyVJAIFublzbUEjoFIaX5efmEQQhsAkAusfn1pb/sn/TsaP67ydTaVf++7/+vphc9XVX/Y0dnT0dWxqWHLmrffv+YHP+xurTmf2H7+yKZT+94q/u3177/1x0dLntja0sF2mjCRzdS29JhtJKN++Sh1fpLGAgUpBPTHRSLsLfQUyWdxaVB5jCTCpTK2oFR3n97jcDykuMyGx7t7JKKBEo/aFBp6nYsLPSVKW2QFiWjI7w8EiYDfH4o6dEgyFvJFcqmzRiSndTa7fUtDWn+ys7LI5X7N4jq7OMje+HQmr0zJjPNKhorz82Z5I/KUkQh4XJHZROblAACALOfZJx/burnqo7Vv/P53d8d3+m+/878W+TseXdnwG2/VAy9+ULJi85OrW3/5xCu1VYvvvuvnH7/9dHXlY2vf/EN8x5tX//yhf7+r7KE/vHDo6DG224T9tZosJBkp85QE5XoQCWZPsaqTHYgFi4t0IZgIlXh0JUyivahYc8fUtKdEf06nw+SRkhUXG/7p8Z6Zk+yHmqeY3kSiYX+KzqYW8Pg0FUMN4Ln8ksYgHvEHDP0cD/sdPj2Hzk5rU01nJ+PRmNKbsWCR+i58eeA6uzQsz3QSYZ/X6zduGX4eiIrzeMhX5JnjKfKFEzxcmeLoUnFhQUFhsU8f0iJKlP7Dr+heWAN4vWy/RjIa8HrLQsYqOrugemWkCG0usz0BrUFFzqxUQb2RZdbxUBldKRNJ6W4ro9JTQfzaraVUQhQpFg2wOLyGPAYAAEwZ/vTs01XvvLzq1UWrX3vxrVcfX7To/lufCT62qu3J1e1/8Hc+ubrzj/72B5Z9+Ml7f9rdUb521WN9214+2rf6pl/e9YPiwHcfqf6Ph9/74S2PdUd3Dl/MyvVsJmJ9ygeh9Cj1jPQoZdJcUYnxcDCkqK6Yv8Ab1k6jPjoZeea3eKRUapn4/pHceHxQTfg+11jAqrMtbyTUIjYr3peNZELdv2NR08l4LBwKhsLUR9DZaW1qfw8yFizx+DPaNDGROOjs4Zhvbl5+cUjTymyHhtC0kbI5+XkzCopICRcVzJ03b46ePBkunZ2XP8fj9ft9RQUz8maX8omL5zB79lyWpNhDifWNKeFSecw1tLEyHfHOsixdc8KlqtBWZbaWczEr1FwqntwEwrOeNWvO3MJib3GA2joemJdP8coCoUCZh5KIeEobsCLNmj1nLnPF6pA3S1QCAACmCivKyw7ubUue6F30dPFbrz/89AuL/uPRwH88FvzF4k13L9t8+x/XPOXveGp15/PeP5yJrz7c9dpTTz/y7/OL/+Oh97/7aNX3Hv3wmoc/+N4jH/zg9ie379wlPWYVJGnNco5k8gj6zrz0bIWEepG+sMNUvD8cC7PtBkGmw+yxeCTJSUVIJiIhliwUycGPSFN0diJUYl4qZE132dcO7SEx7Q9ra3mJSMAfjMTiBKntUCgIne1sU1Nnsz3abP9WcVk2vPhxjTlj9hyDIvH5GbugCW0us7kmtWyJpuktT5OoTJkr6piJZS6ieQ56kiRzIIW2rrOFhtbkt4PMNl/gMlvKaf5OUKS9EyjRzFnLRMUh7SwaiuhVVXS24SpaZrxFAADA1ODYiWPFD/zmvl/+/DbPT4ofWjDn5sf/79uWfeOmP/747uf+6F09/5e/e8bf9qy/beUrj58/uKpi5e9/9PDa7/3uw2seef/7v1s3l+zhKrL/+F11wd2LpcdsIlXNkQJMu6CdDHvTCHHLfgcS3YVFXl8oHA2HfKWFnhLlo1adFI9UBK+PHu3eIKULeks8xT79k9scIUVnpwpv6pls25eeiNJ7T8AfMN6YkrGgIbn5qQ8629mm6Hp2IhaNRiNhf0lhFnxtQEjRQq+BtleCiVIhtJm4lpKbfWVQ/ToiF9pcifJvNXrK6N1QUkaall1RVCzDJK6NtW1dlvMj6zceJRYFLfUwz9mQ2YoHS9Y8Kl+ML2Mvs3oCi87WimSTHgAAsh72WyEXLw0NDSWTyf7+/q7ojr5d/YMnTrFf0h66+NRzzz/5VtfvX234OFB64cCbi5c8/h+PVH33ESav/+W/Fv570Rvff3jd9x95/3sPf3jNw+9Lj9kEX3A2PSFG0NlsudrpMt/dnSatkI7G00WQ4pFtmShQ/cT8yhJ5bjA1dfZwgi9bR0MktcXPvMXDPkVmC6ENne1oU3vfCPv0K8OvW0wgaaSkJq8VmW3RoWpyfmQS7ASbZyw5OOhsXWjHfHOcv9zILjKhzXxqRUqtAhP/+aUkx1OuEPGw31tYMGdmfl5e/hxlZ4uIBp0NAJjqcJ3N0A80SHxffGbRkifXdPzW+37rpqXnDvj/UPp7EtZMZxev/drPS79w5Y9/+9Qr37qr/D9+9/73frdWeswqUraJpGwkUUmzmJ2IlLHvLlpVtInU1XNbj6Q5zdFSROiUx1ZWm154UjaSZBOJaIB/J4ttx7bobOzPTmNTTGdTbwbFRgVJNgzKdFKS/ZhefmkwQDJbi8GjG18+5LtAdImar1wxsOTgpLOl0A6RkjbCUhCbUwLkUpfZImfr7hC+pcSucvovgPKleJ6VEg06GwAw1SF5nQqJbvZzfUMXF/75xWfe6rjvOf+O1vLkAf8Df/jD3IfXXfPIuv/16zVfm/fk9392e1uk7Zq7X77m4XXfuPNl6TGrYDpXXWM2ba+24ryYzfaLpK6+Wp/Tqc5tPTLNqX4PkpUxm358wwVS3xwsbZNddbZ+EVXX14mIP6D+xl88jH0jaWyqrWez38Dx6l9yZp9XOX6YNWkI4axs9wiFIvrf0GHfTyT0TdEEae/8vPy5pQGKGCgtmDN7lq5EI97ZebML6Y0xOZxMRHzzpOzOVGeLbdaUm51Y19EiKUUSOefPo1slmUxQI9OJ2EZi1ck88Tw/7wFqfjqxynHobADAVEcq6xT4+vbQ8ldWPrVq672lb/VFXjl7YPX9vy8ROvv//uVrV934xP0PlSx4+PnvF3/4jTvLvnb9H6THLCPq8xTLH4xKklxW9F8iGvSHzNLPdlmZ/e6z/RaHOH9Oi6dgkqJZPnV29KiUiR4vZUUe08+P5AA2K/T85xFlW2WHojFg20ECuppJxiMBbVc2W9k2hFgk6PdDZzvb1Ns3Qj1fWih+3b6gsCQb+pZLSTOK9OVLxWZNK24n/o1JT1kkalKiiVDp3BnSywx6seXVy1hncxmsrpbbIgpsKZKac/4c/V5P1cn0/s1+8YTDSshvNSUadDYAYKojVLV+IOF/eebCpYuf1Hz024cf+K97HuluXH4m/uYDT/z+ew+/9z3S2be98C/X/ern82/9X7d6r/S8+PV5T1z1gwXCYfZBArjII56kpervIkf9ngLjt7CdFrPZ6rMVI14yGiiRz2mzc8LJIyMe9rE/5kJ4SsxbE3IC250wRluxOmeZWk1EQwG/j+M3/XRMMh4O+n3+APvzNbEE9o2ksym7P9vxz5dOCdTC61uhFez/RHoG8A0f6hcaRwvlnFHiMZcQAACmIEJmX7x0/tLQxd27t9etfbav/eXTe1afjb/pf+2pm3/12H2P/OG50gfeX/PMP/7jzFlz77zi53/4lx8/vHzlKpk+S5nQJ+lYnU/px/sYmY51nrpMD509hYn55ubnzy2ld8BEIhYum5fvzi9MxyPBoK9wNn6uGgAAxolYv9bRQ4ZIaQ9dOnf2zIolvz1/oPLCgVXn99O/b5wbWJM8uPrcgTeO7lr9XMmtqysevea6n/zwP28+f/6ccAgAyBmgs7OeRKTMM1vs0Zgx2+ZrJGMhESxi+1BKHb/LAgAAIDOEqtZRQ0hmD1248OMbPc/++al31jy5o638yC7/vp7XGj750/Lnf3Pbglv++PhtZ/e9/erSXz71x/svZuffgwQAjAPobAAAAGDsXLhw4fTp0yc4Jzny+MSJ4ycHTx4/UfibJ4pXbLmm6NV/m/f7b/z0wW9c/7tv/PyJK//z4X+5/vF//8F16wPP+cvvf+mFZ4eGhqRHAECuAJ0NAAAAAACA+0BnAwAAAAAA4D7Q2QAAAAAAALgPdDYAAAAwdpbvH/7vm4a/Hxney3++qfXE8FeahmdsHv7wKL8MAJjGQGcDAAAAY+e/bRjOq7Oxz22WEQAA0xbobAAAAGDsPL9v+P+3afjfW4d3n2GnW44Pfyk8/NmG4XcP88sAgGkMdDYAAAAAAADuA50NAAAAAACA+0BnAwAAAAAA4D5u6uxjiQQMBoPBYDAYDAYjw3o2AAAAAAAA7gOdDQAAAAAAgPtAZwMAAAAAAOA+0NkAAAAAAAC4D3Q2AAAAAAAA7gOdDQAAAIydSylcvDg0NHQ+eWGosbll8QtL7/vNA7fe+cv7fvPg4heWbd7afO782aGhC5cuXhq+SFFlEoH0CADIFaCzAQAAgLEjNbLC0NDFQ4eP/PGZxZ477rl1wb233HHPLQt+SXbrnXfTv0+Vlh46fJgpbLPIJqRHAECuAJ0NAAAAjB2pkTkXGZfih4/e/9Cjty6479YFv5J2572akea++7cPPXz4yNELQxdkMg3pEQCQK0BnAwAAAGNHamTO8KWhC0MXn17ovXkBSeq7SVVL45r7lgX3cWNq++nnnjt7/qxMpiE9AgByBehsAAAAYOxIjcy5MDTU3Np26x13k7C++bbbf3zDjbcsuIsJbtLWC4rmL7jnP2/+xQ2/+MUtC+723HH31tY2y84R6REAkCtAZwMAAABjR2pkDunsJS8uEzp75le++otbb/9BwU9vWfDL+YV3exbc91/z7/jbL/7Dz/7r5/Nvv4t0tnfpS9DZAOQ20NkAAADA2JEamXPh4sV7i+4XOvuK2bM/+7f/109u+Pkvbl3wo5/958233XmzZ8Hf/f0//N0/fumWBXffsuCee+9/gO/nNsS29AgAyBWgswEAAICxIzUy50zy7G0L7rp1wT0333r3T2/8+Y03e558dlHhvUWf/dz/vPueX/3x6YXzbvb89MYb599+F+ns2wrvPnPmDHQ2ADkMdDYAAAAwdqRG5pw8ddKz4JeeO3/185tvu/bHP7218K5nFi2uDLz/u8d/T//+4Zlnb7/r7mt//J833nwr6exb7rj31KnTMiVHegQA5ArQ2QAAAMDYkRqZw3X23aSzPbffdfNtCxbcfd8jJX/84U9+9tf/5xfo30f/8CSFUDj/OW3obAByn2mks5PJpDy63CQi/jKv11sWisuAcZBMJBLZUi8AAJiGSI3MOXX61Pw77iKdfeuCe/jv+kmbf3uhcir+YI3Q2adkSo70mJUkE7FYPCFP7KCHrAV5QSMZt/UgI2vIUIYMUZAXzDiFT20ca0XNGI3hwT9lmC46O+rzFBSUhtNNEZNFxDs7L3+Op7jYFxn/fRL3F+TllYblGQAAgMlGamTO6TOn+b6R+7jOJpN/nsZzBxPW4u/UTD2dnYz6SzyFxSUlJUWeQq/9kzQWKEghEJMXhxNhb6GniBwwD6VBPZxdKZWxBcaDOr1HjUSoxDZ8KpOIBkpsNYvREcWFnpLLUOlEJODz+SLZoKWmDNNDZ8f8RaWlpdmhs7kyLnZLGUNnAwDA5UX+/fSLwxcvDp86l7zj7rtfX13JlTT/5WzS1qa/Bymk9r1MZy+459TpM0P8i5DMxcXs1NnJSJmnJCg/fyXB7ClWdbIDsWBxka4DSQt7dHlOWrGoWHPH1LSnJJTJs9nkUYNUOj3eS3JIZ1PzFNObSDTsT9Es7GXFeMshYePxR+XJ5ED6PxAKBaGzR8V00Nk0GOkupjfjy6+zY0Fv8bxZeXlzC71eL5upKMDrC8eigVLPnCI58dBILivyzPEUlfnNgzke8hUXFhQUFvuCMbkUbtbZiVhQxigLRPV7UctCOPWZG0H4ZOHBmHLBrgyJsI8KHY2HyihBWUSc+41FeX4eiCadwmVJ4pSlKIkb+2YAAOAyc/HS+YuXLly8xOTyqbPnbr3zrgV33/fMIu9vf/foLQt+yYS1VWczEzr75OnTQ5RwmFR6tupstuDsU/RcLFjkGUnWMmmuPHHj4WBIe2oRMX+BN6ydRn10Ynos2WLxKKCieXzROBUwd3R2IhaNs7ZJ1Sz0tmJ6I0nGJ3f7SDIWDEQSyVhI0dmstAn6NxIKhsJCRFBQOEQnvBY6PJTiUJFlyLQh93V2LFDM372zU2eHS/PyZs2aPXNekdyvHfPPy8+bUVDk8/u9njn5+fP8YvZIhkv5dhOv3+8rKpiRN7uUz1KKzk4EC0XKQNBfPDefYgipy7OYPWeup9jrZSnzZomkhk+6UEx5aT6dysDzmjVrztzCYm9xgGbdmG9uXn5xSN5LlH9enidATewUzkoyp6BgTgFVlhdFrx0AAExZzl04t33HjkDVupWvv/HGmndvWXD3/Q89Uhmoeql85R9LF91WeN8td9xnEdma3VOxes3KilXvVVVv79t1YWhIeswqSAibH54kk0d4mtouPeuQUC/SV8SZiveHY+EgQyo1G+w8suXdMnrOMRe5o7M1UjQLVZdVMzFiW00MpK9JZfP/GjqbnZCmjsTi8XgsEvAFQ6GQPAn6eXQZyx8g4c2IhQN+9Z1rGpDrOpvuTfkJV1bobMK8As2kZ36RJkmHk6HifKFJOYmAJy+fx2XSdZY3IoLZHu9ZeQV+kuWqzo6GgxFtqqF5kFLy+OYsomVztARcDusXkqHSOXOKggnnMvC85vqU2YyFaIKay+lCSu8czkoiis0xy3EAAJialPsqPLcX3nLH3bctuM9zB9k9JLUf/+Mz9/z6gVsW/NJzx92e2y3y2rBb7rj3lgX33coS3rPS95r0mE2kqlh6mqZd0E6SIHR+3MYCxR6mjiUkuguLvL5QOBoO+UoLPSV2H3TaeUxGyrjKtithLmCjs0uLy3zeotJAOBIJB5zaamJgUjnMc0vR2frn68OJiM/4KD0R8csrdGQEszRBf2TSCp4F5LbOToRKirW7L3t1tqE8hyOl+XlzigP0RijwF8/ml+OBeXl5njIZSpR5hBOzNxq/kVAo6C/jy9My3JyFkYD7tNko7lQGa14MLsK5VGZyWlnDtg23VJaE9hz15QEAAKYid9//wM13sO8+ehbwpesFv7p1AWlo0/cgnczDbf6Ce8juKfqN9JhN8AXn0ehstlztdJnv7k6TlmVmbCnRsPMY9elblaeJzqZWKDB2uTu11YRgqGwbnW2sTpPONk70eFxXqxVhu7wV3Z3z5LLOJpVdRN3MfwsomYz4C0pD8cv/6z9pdTY7nTWv2KvCNlTzRHyziQJbp1e8xQKeGXn5M9mujjK/v1hftnbU2amymeNUBtsEmow2y2yH8BSd7VACAACYQvTs2Pn0c3+67c57PLff5bnjl1JeM6k9ot1zy4K75t9+52133k0etu/YKT1mFSnbRFI2kqikWcxORMqKSWSnfQrbaGY7j1GfpyQYk0/3eIjeBCK59tt+duvZ5pBJe7+Ih/0hfZdKguln7Tmeic5ORCwbRUxKfRqQwzqbDcFUnGeHScKsVi3Sk120XeENl6p7OQwMb8KTdsspEtZJZ3OfyjaQRDQUisSTzmUwl1ySDBXl5ZcGAySnTZfswllJlFgsSt68gKG7AQBgCiK+AXn6dHLg0JGDR47yn+27x/a7j1ZbcM+hw0cOHT58+jT76+vkRHrMKpjOVddNTdurrTgvZrP9IqlfeCTNFVQ3EaQ6t/NIGtSG3FrTTv0M3vrLLGk7wkVIPtsg16pH1tkk033aYjiHzs3CO8fJ5fVsM6lj9vJgVqvWJd5EsDA/f24pH4PJWLB4tiZ52a9uzy70RxLJ4SSN5XlSdhve2JbtWcX8HkzGWcIRdTb3mT+P3kwTiTh3OZvn5VQGW53Nt3MTKdLcJpzr7Pw5xcFYIpHgjvPlhm4AAJiyXLp0/tKloYtDZBdPnj1zy4JCrrOdvvuo2IJ7zpwhhT0kLGt/Pzvq8xTLj/6TJJeVZ2kiGvSHDKnnuJgdD6m/SKdCjyuPNyR+myJJ0Sw/8pdmeVxn0tZ1JxUbzcJ/HlFrq1iwpCCzH0R0F9NqdEY6m62H69+JpEgBfQ/K9AA6e7JJr7NpbNKMNjOfVDKRP0d5jU+ESufOEOF5MwrK5Hui4Y2mI3mdpKzPq4c762yTT8pLfzG2L4O9zmYZUFSbFfCUcF4SX7CM/egJwSqh3ZEAADBVkb+fzTlz9ozyd2pShLUwvtTNf/Lvl8lkki9kS6THrIMEcJGHLxqzPzJjTNxRv6fA+C1sp8Vsu0+XjXjJaKCkUASanRPOy+MK00ZnM/EaLHVqq0liDDqbrsTCQb/P76f/B4wtKNOE6aOzpxaOf059hL+zPqY/w+6QyPU/6W4ofnLtqmcAALhsSI3M4Tr7bu3vrqcobGEmnX12KuhswYRugc7Jv50+QaCtphLQ2WDSSF28BwCAKY/UyJzTydO3jEJn3zWldDYAYNRAZ4NJI1LGfqEbOhsAkFNIjcw5debUfPYbf9DZAAAGdDYAAAAwdqRG5pw4fdKz4O5bFvzqNtLZTj/tB50NwLQBOhsAAAAAAAD3gc4GAAAAAADAfaCzAQAAAAAAcB/obAAAAGDsrDsy/D8bhm/sGv70PDvdkxz+t5bhmeHh9pP8MgBgGgOdDQAAAIwdktR5dcP/24bhB/qGXzs4PK+TnZJ9a5uMAACYtkBnAwAAAGMHOhsA4AR0NgAAADB21h017xs5O/wN7BsBAHCgswEAAAAAAHAf6GwAAAAAAADcBzobAAAAAAAA93FTZx9LJGAwGAwGg8FgMBgZ1rMBAAAAAABwH+hsAAAAAAAA3Ac6GwAAAAAAAPeBzgYAAAAAAMB9oLMBAAAAAABwH+hsAAAAYOycPXv28OH/f3vvHxzFleB56t/7ZzfiLrqJ2/uhjYu77d3ZneF2bru5nes79dx1hzqit1uzMcvyh5eptmGtnjY7NTCMy7jtE9jYxVgNZSNDCWjIAwPJCI8KhFG1NBZlYckljEwZJKpaEqgshChJyFRZyJRG4NC99/Jl5suXmSWVVEKq0vcTH5uq/PHyd+Y3X71Mjd9bHGNjYzMz7O9JAgBKCORsAAAAYOEsPmRrkHJ4iQCAUgE5GwAAAFg4PCYXAl4iAKBUQM4GAAAAFg7PyO6M6PDv7vASAQClAnI2AAAAsHB4RnaCZOvOzs7m5uYrV65cvHiRfBgYGOD9nOAlAgBKBeRsAAAAYOHwjGzj7t27ly9fnpiY+Oqrr8jnq1evkrR99OjReDzOh7DBSwQAlArI2QAAAMDC4RnZxrFjx5LJ5OjoaGdnJ/nc0dHR3d196dKlffv28SFs8BIBAKUCcjYAAACwcHhGFvjiiy/a29tJyD58+PAHH3xAEvaBAwfa2tri8fiFCxe2bdt28+ZNPqgVXiIAoFRAzgYAAAAWDs/IjLt375IMfeLEicHBwZmZmdra2jNnzpB4PT4+3t3dnclkrl+/riiKqqp8BCu8RABAqYCcDQAAACwcnpHv3btz505HRweJ2jdu3CB5mnT55JNP6uvrb926NTk5mWXEYrHz588HAgHH14/wEgEApULJ52ztzKbDOxYn2XQ6zZdA+DgP8hsaAABAHvCMfO9eW1vb7du3e3t7T506dezYsePHj/+a0d7e3tXVFY/HBwYGotFoa2vr3r17SSLnownwElck2XQymUrzL07w66wA76GTTc1dAv9kIZ1MuIxH5imRzFViEeOyMuhaJMuMa3rRUOI5OxmqshBK8h5FSEqtKivzR+nHqL+srEpNsc5zI4wIAACgwPCMfO9ef39/D+PmzZtnz55dv379s88+u2XLlldfffX06dMkXnd3d3/66afnzp0LBAIkhfPRBHiJK41sQq3xVPtqamq8nupA1DHYypdbinnJTUcD1R4vKYCW4A87XYrTkRrLKIxURBjPMuVk2M/7+Ko9NcV8cbeTToRqPFVVfnlN07VYxbbDci1zOhZSFCVWorc2S0Np52yyT1YpCf6l2EHOBgCAFQjPyDZu3Lhx5MiR3bt3P//881u3bt2zZ89LL71EvqqqWldXFwwG+XACvMSVRTYW9NSE+RWHXFY9PsecbCUZ9nmNIEgitMcIySS0e316cQbpqN/r99dYc3Yy5POZhUT9VTURXkgy5PGqCb1SN6l6vfOYp6KArB4fuRNJRFU5Z1vW4mwq7PcEY0+3Wpvk/1AkEkbOzosSz9lkr1xRd7mpSDAQMM4GiXDA/JaOKoFAiJ82yM4c9HoqPN6gauzP+eXsVERhBSjRdNKas3nZVdUBNWoUwaYeTpB/fNVVpJc+IxqkMNbZp4ST+LEKAABEeEZ25+OPP960adOuXbvefvvtbdu2+Xy+l19++a233uK9BXiJKwqab8UKq2TYO+eFlUZzISWmouGIcPEgF6VA1HItIdPwKIkUmZJYckKx1pQJjUTIx5RZQjYasFf+Fin6kiVD0iLR7aCK652sHpffFpaGbDIciqWzyYiQs+ncpsn/Y5FwJKptHNIpGiFfhO1DYF3JMCXbzMed0s7ZdD8NJ8j2J0TI7sE7LyOxwNqyCkU7UuhnMy9HfWVlnhDdAZPqhjVl5VVeRVUDnoo1azZoR1YeOTupVLESSIz3VlVu2FBh5Gy9bNLH51lXtqYiGGPdWeHr1lXSPj4yUbP8bNRPh/MEVFXxVpWXrfNbT48AALC64Rk5J/v3779w4UJzc/OxY8fOnz9fV1f3yiuv8H4CvMQVBYlz1sBHYvIcodZSmW2DBHVr7TNtDkGrZmmSFEZLhmhWz6bnvIZno0GPUdNdKjjmbMtKJdtBbmazhJB8TVI2+9fM2fQLydSxZCqVSsZCSjgSifAvYZUNzodSQyR4U5LRkCrec60CSjpnp2OB6uqaYDgai0XDgRqPT7HU0i4LNFxrETapVJRVVlaWVYfprmh0z0Z8a3jgpqRDnrI1LCTPO2dnI96yskoe5tmhSOI8G5GVbfahZfNJscKNPtlwtT6BpFJZtjaghXGCOfsAAAAoPCPnpKGh4dChQ4cPH37xxRfJh9ra2k2bNvF+ArzElYQt3tEEmLNCO3ftcjLks7Z3yMaCLGXbJkWmE1CUmppAOJpwvoYnw6ypstdfgj+12nI2+yFBMdYcbcDz9HI2jcraL+C2nG2u+3RMCSWM8BJTeR/yyexMxwmrsdWUIkq7PtvKymjBFfWvYdGaJtvqcJSEbW8ky0K3Vs8dI/0rfCFyU6ih+tZpyXbeOZuVYIRputz6iHKfNM3TdPJi4RRjAqnQBhLFg3xeCEESzc3hAABg1aNF5Ny0trZevnyZpO3XX3/9jTfe2Llz59atW3k/AV7iSsLWXGGunE2rq916s9bd1p4JxXi+0ZqzSf6uqhIHtl/Ds6kEgWRwn9vTmcWLPWfTbjWe6pqgGlYCXm8gpMp3QEuFmbK1aG2tzxZztvnFGI7lanFBaCtvIXeXPKspZ9PA6X6P/bRgtc3eSIpEXJpkY4G1a/xRGoV5rTGNuGs3+AIiCp3reedsua85om082lhlQ4jmabecTbtXVvP54Cz/zQoAAKwYeEbOSVtb24ULF1RVJSFbUZT9+/fX19fzfgK8xBWF7cppa0gikqMyOx0L0sca9SDGSCiemnCSvQQwm01FSKSPGa+zI0nTmiOdoqcGfVdJybz0QMNtYdNJem9BWzmThXbfDoUkFVUjRrPqNM3PeoyYT85Ox6SGIpakvgoo6ZydjlkevaCH/9N+OtcBGrTX+Hw+XrMc9a9Z6/VWG40zaLIVGmqYzDtnswFZLTWDNRbRRqR91viMPqwViDYtoXCKOQFa/W62YgEAACDBM3JOIpHI6dOnd+/effjw4cbGxn379pHAzfsJ8BJXFPTSKT62aGteLeJemU3bi9gqnWmUtsMLoNXbYngWruHS1X2O7F+U2HN2MhqOCquWxOyn8xM9ic8O8LrquXM2iemK+dIFAvlu3XglTknnbPrEs0//vYLcUvk8K+J+l7XWKNPDNKvfLuNtsCmk95o1lX62G2aTYd86exTOnbO1Idf5wsl0Ok0KqFij52xWNutDaw5iygajtbZrziZZfF3ZumqyFrOzWXIQkVEQuwEAwIRn5Jx0dnb29PSQkH3gwIFaxv79+3k/AV7iCiOhGJfSLInLQv5LJ8JqxMx6rpXZ9DXYc7fssDUFFyZMega9xjXccnWfzSYsM1Ua2HM2XT367wHZ+a3RJcBSGz2vnE3rw41nIslAIeFdZ6uBks7ZBLIrej3sBtlTY72jWkZYqDWCtRa7zfpnes5QPN8h2Zjl7wq9bmD+OZuUGfFXlrPxyz2hqJihxbLLq4L6ru+es8XCLKMAAAAg8Iyck+bmZpKzT506pSjKIUbxvD+bYF5K6R+ZMa9WCdVTZb4L260ymwZEGbfhpO6pqEL/YgtBvoabfao83hJszujUbiRL/ziPtswuf+xn6VlAziZ9ktGwqqgq+S9kNkFZJZR6ztYoyj+4Pp+/lk7zsISZv3MUsIC/xI4/3g4AAE7wjOzO8PBwS0vLrVu3zpw509jYeOTIkXfffXfXrl28twAvcYWyfJdS9ykX5eV9kazGZS5iVkfOLl1o+pXAAQgAAE8RnpHdCYfDDx48uHz5clNT06VLl955551Dhw4VYc4GAOQNcjYAAACwcMbGxnhMduLOnTtdXV3ffPPN3bt3Sdo+e/Ysid1vvvnm+++/z4fQGR8f5yUCAEoF5GwAAABg4UxPT5OIzMOyjd7eXpKwv2Tcv3+ffD137txrr712/PhxPgSDhPWZmRleIgCgVEDOBgAAAAAAoPAgZwMAAAAAAFB4kLMBAAAAAAAoPMjZAAAAwMIJj/dsjP1mV3/T/X94yDsBAAADORsAAABYOP/xs/qqq3X/vvvd9i/7eCcAAGAgZwMAAAALp2ns842f/+bVvnOj01/xTgAAwEDOBgAAAAAAoPAgZwMAAAAAAFB4kLMBAAAAAAAoPIXM2Q/SaQghhBBCCCER9dkAAAAAAAAUHuRsAAAAAAAACg9yNgAAAAAAAIUHORsAAAAAAIDCg5wNAAAAAABA4UHOBgAAABbO9PT0+Pj4vcUxNjY2MzPDSwQAlArI2QAAAMDCWXzI1iDl8BIBAKUCcjYAAACwcHhMLgS8RABAqYCcDQAAACwcnpHdGdHh393hJQIASgXkbAAAAGDh8IzsBMnWnZ2dzc3NV65cuXjxIvkwMDDA+znBSwQAlArI2QAAAMDC4RnZxt27dy9fvjwxMfHVV1+Rz1evXiVp++jRo/F4nA9hg5cIACgVkLMBAACAhcMzso1jx44lk8nR0dHOzk7yuaOjo7u7+9KlS/v27eND2OAlAgBKBeRsAAAAYOHwjCzwxRdftLe3k5B9+PDhDz74gCTsAwcOtLW1xePxCxcubNu27ebNm3xQK7xEAECpgJwNAAAALByekRl3794lGfrEiRODg4MzMzO1tbVnzpwh8Xp8fLy7uzuTyVy/fl1RFFVV+QhWeIkAgFIBORsAAABYODwj37t3586djo4OErVv3LhB8jTp8sknn9TX19+6dWtycjLLiMVi58+fDwQCjq8f4SUCAEqFVZGzs6lkIpnO8m/LTDYtsKB5SkeVYDhZsMVZ/Aw9FQq81AAAUCB4Rr53r62t7fbt2729vadOnTp27Njx48d/zWhvb+/q6orH4wMDA9FotLW1de/evSSR89EEeIkrkmw6mUyl+RcntBsJEd5Dh1yL3UoghZPLNP8iQXvlmLBtMiWCff1x2KrCxbBoKPWcnU2oNZ5qX01Nja/aUxNK8s7LR0qtKhMpr/SqiVxnLhup0IaysnXBBPuSTkQisdRiDrjFz9BTwbLUAACwcuAZ+d69/v7+HsbNmzfPnj27fv36Z599dsuWLa+++urp06dJvO7u7v7000/PnTsXCARICuejCfASVxrmldTrqQ5EHa8QyVCVDfOam44Gqj1eUgAtwR8Wr8XJsJ/3sl+mk6EajzGabcLZVDTos0ymREgnyHJXVflta5quDyPRKLGnf6lOx0KKshwTLmJKO2fTI9s8MpOq16Mud1BjsdYX5tXHyahSvY7kR380r6icNW5lo/6ysio1xb8thILM0FPBXGoAAFg58Ixs48aNG0eOHNm9e/fzzz+/devWPXv2vPTSS+Srqqp1dXXBYJAPJ8BLXFlkY0FPTZhfZ8hl1eOz5GRnkmGf18i/6UiNx7gWk9Du9enFkeTo8aoJ/eROLtNes/CE6vGF9J8xszGFjGZOmJRZXaNEY2F/aeVssnp85E4kEVVtOVtbH/xLOuL3BGNP96pI8n8oEgkjZ+dFSedschjWRITdIZta/h9bWKz1R/k3SjpcvaasUmHHTjIcCCjRZCIUqK7yeBV6jGXZl6rqQEg/E2kDpWlLikB1ZVnZ2g2+QICdfYzR/Z4KbzhFRw0EI2YKp/21IU1yzZBjAfp3csAFvZ4KMmdq1BiCzlMgnEhFgtVVFcGY9l01TwXi91RUpQV4vAHFmAQfn/zjsy60sdS5hiE4zJU4TQr7bhkJAAAWDs/I7nz88cebNm3atWvX22+/vW3bNp/P9/LLL7/11lu8twAvcUWRjvqrFKGKKhn2euZKtjSaCykxFQ1HhGZ/SbUqoNflpJMJ4SfZbDRgpksymCVp0rRuTjiVYGmPzl1J5Wx9hSRDcs5OJ8JhIeGSJa+exw1P4cgmw6FYOpuMCDmbzi1JI8lYJByJag1/SKdohHyx/tLOupJhXFsHlS6lnLPJbTc9/NLJaJjCd4Flxh5rZ2ejPr1SmlZPr11XUeUNBLxV5WVl6zyeqkryxeepWFO2pjrMlkCvw6Z50Zqz2ehr131nAxmFxuFsxGdEeEIssLZsbSDGv3FyzpBrAUl1w5qycjqfZN7Wla2hmZrCSlu7tqKy2hfwhciZOalUlq3xRfQzari6rMwTIosRC9CxPAE1HArSAjaEWCpm469bV0lLZgtt1NabNffuw7jMlds8AABAAeAZOSf79++/cOFCc3PzsWPHzp8/X1dX98orr/B+ArzEFUVCkQKfnH/tWCqzbZCgLlRai2SjQY9ZO2bL2SRS2yZccjlbx56zrdBbkqdZn03yNUnZ7F8zZ9MvJFPHkqlUKhkLKeFIJMK/hFU2OB9KDZHgTUlGQ6p4z7UKKO2c7fcFlYDXH4rGYtGQv5ocv1oeW0acYq3ZjQVlPQnTVFvm5fGQnHD08czEaf2sfVujj8GI+tcY/R1j9hwzxArQg7ZRgBS/0yGPnlzZmGYfAu2yhpdvRtxsKhaJGM3Aaa7XgrZ1/Cwdns++lLMdhsk9V3rQZvPA71gAAGDx8Iyck4aGhkOHDh0+fPjFF18kH2prazdt2sT7CfASVxL2IEsSYM4KbUultI1kyGdv75AMszbHXr/4tDstx2ivwkZ0aLK8SnP2fJvvFAoalbXfiG0529xk6ZgSMq7r6ZjK+5BPZmc6TliNGVt1FVDKOZvcM1eZTcK0w9H4rWq5YCFRirVJpYJEURqALbHZMqjwJXfONr8xaDbWOtFPFWIC1sg9Q9poWngVPpHwLRbFsivL9w6l0byrBW1zMEY2TX9cCpE7IVp1r82kNL6wQOZHt2Hc50qbBxa0aUezbhsAABaNFpFz09raevnyZZK2X3/99TfeeGPnzp1bt27l/QR4iSsJeuVULVeOOXI2ra52683ioUPPbCpBiIYDPsvTjvQZK9q0MKwGa7w1oVAQOZtB71VsT4UuIWbK1qK1tT5bzNnmF2M4lqvFeaWtvFfg2xaWjBKvz7bupivheHQKojQPatW5lqBsGVT4Ig5kGUH6pqEHbZqdzcpek9wzRNDjtRmz7dMxKqQdSjOCdjbiNSJuNuqvWFNWvo6cQQNKKOjRy5PGFyZkfnQbxn2utEWi00bMBgAUGp6Rc9LW1nbhwgVVVUnIVhRl//799fX1vJ8AL3FFYWsmYmtIIpKjMjsdC/rM5xqdSUdqLI3BSYGpJIng7NkqMmHbJXz15exkuMbtnS9LRCqqRoyGt2man/UL7XxydjomNRSxJPVVQCnnbLKbCi29CO6Nwp4e9iDKWipLTa8ZlkGFL+JA1mxpS5oULWhHSMy29aLkniEKTehVapjndQodR0yrdBJaixR7aQQWbv2RMInZvBcbzGcMZjZQl8YXFsj86DaM+1wRWMj3h0N0TqTZAwCAxcAzck4ikcjp06d37959+PDhxsbGffv2kcDN+wnwElcUNDmLPwXnvJK6V2Y718GmY5YnJC0hnvQLi/WeCdXhBRurLGe7/SCwlJD47ACvq547Z5OYrpgvSyCQ79ZtXuKUdM5mTycHItpDr1lyD1hlid3LAguJniA551JCQfocX7nxE5wlKFvypPBFHIhlSXLK016jYhndgD0GuGaNy7N/uWeIYS+AJOc1Zet84WQ2m03FlA1Gu2jLPBvQltO0ACP1shrzKoUdaeTIJOPPkaHFj67DuM4VxTYPAABQEHhGzklnZ2dPTw8J2QcOHKhl7N+/n/cT4CWuMBKKx8d/+s/SVtJm/ksnwmrEvFy4VmanIuJLdgXoi0n0ssm3hKX0ZMhrRPN0TKWV4eyzyGrK2fSVf087ZNuw1EbPK2fT+nDjmUgyUEh4R9lqoLRzNt3SYX+19sJ8+mp8fQdYPlhINChf5/GL7ZQsQdmSJ4Uv1oHCXvrGDe3JPksfE5qT9UcCbeSeIQ1tGEsBWXLi/Q6dMB2pKqgfQpZ5NmGPKFoibjKkj17uURXvnBla+Og+jNtcadBHOhGzAQCFhmfknDQ3N5OcferUKUVRDjGK5/3ZBJKTvR6HK2lC9VQJ78J2qcymWVjGHC4VVegfZaF4vNKbZ9NGP7c/y7KKcjbpIuN4U7O0LCBnkz7JaFhVVJX8FzKboKwSSj1nc0r177LOD/bmDctrSAoF/Zvt8yqW1rvbH8Kc9+j5sUTFAgCAEzwjuzM8PNzS0nLr1q0zZ840NjYeOXLk3Xff3bVrF+8twEtcoSzplTRX4Tihg+JlleTsVUsqFg7Tv/C4dtn+wGOW3MaG/JXm6wUBAKCk4BnZnXA4/ODBg8uXLzc1NV26dOmdd945dOhQEeZsAEDeIGeXNOmwt6KiwuNfxqc/Y8HKioqqaudf/AAAoOgZGxvjMdmJO3fudHV1ffPNN3fv3iVp++zZsyR2v/nmm++//z4fQmd8fJyXCAAoFZCzAQAAgIUzPT1NIjIPyzZ6e3tJwv6Scf/+ffL13Llzr7322vHjx/kQDBLWZ2ZmeIkAgFIBORsAAAAAAIDCg5wNAAAAAABA4UHOBgAAAAAAoPAgZwMAAAALJzzeszH2m139Tff/4SHvBAAADORsAAAAYOH8x8/qq67W/fvud9u/7OOdAACAgZwNAAAALJymsc83fv6bV/vOjU5/xTsBAAADORsAAAAAAIDCg5wNAAAAAABA4UHOBgAAAAAAoPAULGeXvfUhhBBCCCGE0FAKzDlEzoYQQgghhHC+SoE5h3Pn7JGH0xBCCCGEEK5ykbMhhBBCCCEsvMjZEEIIIYQQFl7kbAghhBBCCAsvcjaEEEIIIYSFFzkbQgghXJzphyOjY/cWx0hqdCQzJZcMISxmkbMhhBDCxbnokK1BypFLhhAWs8jZEEII4aLkMbkQSCVDCIta5GwIIYRwUfKM7M6IDv/ujlQyhLCoRc6GEEIIFyXPyE6QbN3Z2dnc3HzlypWLFy+SDwMDA7yfE1LJEMKiFjkbQgghXJQ8I9u4e/fu5cuXJyYmvvrqK/L56tWrJG0fPXo0Ho/zIWxIJUMIi1rkbAghhHBR8oxs49ixY8lkcnR0tLOzk3zu6Ojo7u6+dOnSvn37+BA2pJIhhEUtcjaEEEK4KHlGFvjiiy/a29tJyD58+PAHH3xAEvaBAwfa2tri8fiFCxe2bdt28+ZNPqgVqWQIYVGLnA0hhBAuSp6RGXfv3iUZ+sSJE4ODgzMzM7W1tWfOnCHxenx8vLu7O5PJXL9+XVEUVVX5CFakkiGERS1yNoQQQrgoeUa+d+/OnTsdHR0kat+4cYPkadLlk08+qa+vv3Xr1uTkZJYRi8XOnz8fCAQcXz8ilQwhLGpXQc6+n7ktaxvmqXt7eDQ+nJE6Uu+T7qO32ed42xF/Yx/7nDE6WqQDOxXi6mhTsO5kLxsl73EhhBA6yzPyvXttbW23b9/u7e09derUsWPHjh8//mtGe3t7V1dXPB4fGBiIRqOtra179+4liZyPJiCVvJIcjfX2xXNcQOVLLVEaZrTrM7cSaOExuZdUGtU6wHT8Vl9XKV7L4r3X3Zbr9nBfe++o1HHFmInfGo3LHVe1JZ+zR5t2/bTyJ1Z3tS/3TtC+o4yw8eiw1L0v8EPaY0cb+fzF0T8tK/tuXTvpfutEJe9oMab8tKzs9SZrx1zeavhZWdkfvn2dfM57XAghhC7yjHzvXn9/fw/j5s2bZ8+eXb9+/bPPPrtly5ZXX3319OnTJF53d3d/+umn586dCwQCJIXz0QSkkleIsebajc9s2bbjlRee27hN7ZP6Mue42napr6x/bjsvIXhFuAqPtgZ5r22/3LhxjzlKvO11ucCfvN5kXDdvte58buNmMtbWzet3NHTpYxW9vQ3bntn8Al+uI61iTuht2qEt8o7tG5955Wiv0GsFGP+s9aDScLKx6aR64mCzVktoN9PVbF2oUrfkc7ZkpvXtjTvalv1GUMvZZT8IWs9WnbW/x7rrkTrD7/sLlbOJ93m9OHI2hBAWSp6Rbdy4cePIkSO7d+9+/vnnt27dumfPnpdeeol8VVW1rq4uGAzy4QSkkleEvU0vPHOklVc29x3dutHfOWcVsvVq+9mJ9VuNKDx6dpdZAg3Tu1r1bE3Dulvhtzvr1pvBve/k1o0Bfcgudfv6t6+4BLvi8vrBZ7YbATre/LqwXH1Hn9ly8DN95fSe2PzLphV0dzF85ahq3D7RMM1/PJdFzp7DIs/Z5GTxy5Vw10tz9h9+93tl3z8izEzm7NY1Zb//ByRqa5G6q7F2Z5CdUyw5O9Oq1O7cc4LspjwrD18/umdz5U8271SvG9UA5GzV1Xjkhed+Wvncdr/Zve/kntoAO/EhZ0MIYaHkGdmdjz/+eNOmTbt27Xr77be3bdvm8/lefvnlt956i/cWkEpeCXapG7c1m/VTJO/O/bOw9Wob/6zpZKdZAsnWGxt5NdPtW9e7hNTV1bil0rm+nATrLWYNbm/D+h1GOp8euX/FL1Z1F6/D1082CpX9w+07njPC9GjXZ1+Y9xL323eupEWOtZ1ovSV0uT8aE9q93L7V19TcdLatL46cPZdFnbNXSGU2kebsHWrD+rI/2NmpdxwmX9e8EDxiROqmXWVlPzkRI73MnJ1p3fOjb317/UF2rmFZ+Uc/+MlPN++p3fnLH32LZPc9V1hpoyefW1P2P/50c7DhpLL9B98u+8Nd2g0xnW6l8oU+LnI2hBAWQJ6Rc7J///4LFy40NzcfO3bs/PnzdXV1r7zyCu8nIJW8Ahw9u+OnliYKJPz9JHeNVe6rbaZpj1uldd/JHRsPfiZ1pFors6fjza9Y4zitCF9p7SgW7+222kqXevrbbXWWO41ldrRVae26T/N0+2fX23uF+wGysTobDjZe6br1RYym7dazjcjZuSzmnN3btHlFVGYTWc5uoxXY39raqu2OXcEflf1+bWuvWXVtz9ldyvpvfftHO/XTk5azA/zMwqrDeQX5aHtbU6t+xukig3379Vb6GTkbQggLL8/IOWloaDh06NDhw4dffPFF8qG2tnbTpk28n4BU8gqw76hUb0pzdm2umtScV9t4W+36rXKDh3jnkW07Xtn8zHa/czrvO/lLoTKbVrH/1Jrjac7euSLq0QrncPvOZ7aflG8eyK0Ibbq9eVcTybXWXsvoF03B1tbOppPN17tI1G5uOmi0Ibnfd1Jpp0nG+BpEzs5l8eZsegO9MiqziVrO1hpka09D0icgaXNtoYmIlLMr/3T9t8r+2QvNZh2AlJWlr7d7r5wl+/rb29d/f43eHTkbQggLL8/IOWltbb18+TJJ26+//vobb7yxc+fOrVu38n4CUskrwHxzds6rbW/DC884jTtM60Fb205se87h8T5arWttqbIKcjZtB++0RJnYZ9fbP2s/uWe7+MzoMnu/72zwyNHPzLmNdzbwr7faA200dehmulCfndOizdkrqDKbqOdsEq+/z+K1Ebjdc3bZd39a+d2yb/3pCWMp3HN239Fnysu+/c9+8Nz2nW+fOLj1/0LOhhDCpZNn5Jy0tbVduHBBVVUSshVF2b9/f319Pe8nIJW8AqQRVmusyB1u3Zaj3UiOq21v07bnclaEEz87YmsOIVdmE+nTk4ql3YjcuKW47Tu5Y/Nctw10kQNObWyWw9HWoLV+ffhKgL1y5HZvq5yz0T47p0Was1dUZTbRyNlac5HtO36pNyBxz9kvkI69J372bfMtJa45u+31sjLztEjbjSBnQwjhkskzck4ikcjp06d37959+PDhxsbGffv2kcDN+wlIJa8Euxq3bNYfWyTmaDSc62rLWkHYojBJXU2W5+dsId5emU2laV5ofEIfCqzT34hS7I427dn4gv1h0OErJ60vy2sPrpxgQ9Nzk7AdzXg9fOWg5SUNXzSh3UhOizNnr6zKbKKZs9njjwT9gUj3nM1zubL+W2Xf29FGW4+45mxaO/4HL7DHw2/fanrhu6R85GwIIVwqeUbOSWdnZ09PDwnZBw4cqGXs37+f9xOQSl4Rkuz7TO3ZW6zVIo3LZu1yV/OJk0JrAder7f3rB7faQza1q3H7+j30ETr2lUbMzZaI6VCZzRw9u2PjzmbteTv7WMVrpl3Z7hCyifev+J/ZflB/bcvtzxpeWFGvWLnVflC9wv/S0P2+s4rx+pHRdvXEWf0df/HOpoMKcnYuizFnr7TKbKKQs/nr/GrZc4pz52y6OLu+V/btzUdvuedsusg/+u9pfC/71ve3B/YY3ZGzIYSw8PKMnJPm5maSs0+dOqUoyiFG0bw/m6W6bc+xvxTzzCvay2GZtOmC8IJn16ttlyr+rRmmWUWd6Wp8fSPvvtn6J2xcKrM1718/umMzG2vjNu0duCVgb4O5irhCmL7VHtixUeu+/pe1tkckl9l4b/tJ5UggeCSgNDWJf7Hy/hdNjScCSsNRhQTuUbQbyW3xPge5+sRfVocQwqciz8juDA8Pt7S03Lp168yZM42NjUeOHHn33Xd37drFewtIJa8sbX/2vIDa/6b6/My4tGApYVfhIhexyNkQQgjhouQZ2Z1wOPzgwYPLly83NTVdunTpnXfeOXToUPHlbAhhniJnQwghhIszNcpjshN37tzp6ur65ptv7t69S9L22bNnSex+880333//fT6EzsjomFwyhLCYRc6GEEIIF2f6IYnIPCzb6O3tJQn7S8b9+/fJ13Pnzr322mvHjx/nQzBIWB/JTMklQwiLWeRsCCGEEEIICy9yNoQQQgghhIUXORtCCCGEEMLCi5wNIYQQLsrfpvo9nyuv94X70nibKoTQFDkbQgghXJTPX3+v6mrdf+g+SAK31AtCuJpFzoYQQggXJXI2hNBR5GwIIYRwUaLdCITQUeRsCCGEEEIICy9yNoQQQgghhIUXORtCCCGEEMLCW7CcrRUEIYQQQgghFJVis5vI2RBCCCGEEOahFJvdRM6GEEIIIYQwD6XY7CZyNoQQQgghhHkoxWY3kbMhhBBCCCHMQyk2u4mcDSGEEEIIYR5KsdlN5GwIIYQQQgjzUIrNbiJnQwghhBBCmIdSbHYTORtCCCGEEMI8lGKzm8jZEEIIIYQQ5qEUm91EzoYQQgghhDAPpdjsJnI2hBBCCCGEeSjFZjeRsyGEEEIIIcxDKTa7iZwNIYQQQghhHkqx2U3kbAghhBBCCPNQis1uImdDCCGEEEKYh1JsdhM5G0IIIYQQwjyUYrObyNkQQgghhBDmoRSb3UTOhhBCCCGEMA+l2OwmcjaEEEIIIYR5KMVmN5cpZ9d3VzX1bGjosnbvqiQdG7vLLR0XKSvTUJ5igVSumZPgXlsrDbPElrcMh/qHquvl7hBCCCGEsLBKsdnN5arPvhmemp19PKWoZsfK/kezs7OJnm6jSyEcSZBCRbKTwQZpmHl60x9PBducZq+HLIzElCINI6n2BeMp/0Vb9wVZ3jaRekynmp2a8CJqQwghhBAupVJsdnP52o20PUiTXDiS5LXX9UMxkhQz45XGAIWR5uz0yJBWzVwdy5CJzmZSFfJg85EVNdRn667l7Jno1Xzqs9niJ3ps3fO3vGU8yUK2BqI2hBBCCOGSKsVmN5exfXaHf+TJ7Ox0mNbpap9nIm1ar67qeCYxNZPNPor2JzfoqdHXP5UYGdnAR+9RRqYS/QPk84aeTGJiPNCSipFRxoa1gXXlcKxMkCj6wMu/8gmlp6ai/QNmxFcSytBkMvuEdI/EEvROoGkkOjGdpTH2EZmWzxhSk+VsfeZFB0ITU9GevsAgKy2TUVs6aPePxhOZGZaJ+RJJi+ChXx8EjMp+OnUyJ5/rxZpKIVsDURtCCCGEcOmUYrOby/ocpDqSoHXYqcqLEymSDnnd9jVlgmTuJ6mJqcTEI1r9PDXuYcNbI3JfhMTeiRHy2TtEM+vsYzbK4BAvnGvN2YpWa67VZxsTehAZoRPKToywqM1KfjwdG0yFx2i2Tg3eXGjOZq1WHj+heVpbllnWVMaWs+VF+ChDppWM8zYqGwan9RsSi44hWwNRG0IIIYRwiZRis5vLmrPf+rAy/ogk3TTNtZMBLRdqEbP/mjZAeWySpMZYjFYD58zZT6JdrKpYliXd2SdZEnaJNEyP84cF2YS0kolsQiwrs9CfjGtPTHZUX00FW7TPuduNsDxtyKvV2dT1xjDlVyfJFHlbEWu7Edsi8KVjNx49tC371IR2s2FoCdmPyQ2DBl1GDURtCCGEEMKlUIrNbi5zzuYPRNJcy4N1NU2cj0Lm85FDMRIZR5Lkc86c7fbcIU262YkJNZ4ihobMems2oeko686c1IMvK3l2Jjn2IBQbqFQsReXI2enMVILWwTNZgxZ5FDFbO+RsyyJ4aB32lEKCchPL/f3y85d0bWhkM4GrtDT2+UHgKmuDzkj1OzQ1gRBCCCGEi1GKzW4ue86WI6YtcZp5esE5WwzHlf0kv9J6azaW1jrFNPQRG4y2z55KZbVK4ieJHu0eIHfOdm03srCcrVWrJ3o6KuhrWB6pwotZNCu1PE1CdgMvjcLWj95rwW9WgRBCCCGErkqx2c0Vl7PLaWZ9EtXyLpHFzdRgD/lsydna+0nyz9llV2lDFBJw2YSERs/1XWuNVhZKd4X2uZ6leT7Rp5uzteYiE+OhDF1Mx3eKl6vX+Hxac7alF4QQQgghLKhSbHZzxeVsHqCzmeDFa5WNA6HME9qMhNXLsqYUs+mxEV9LMky7zzdnG+1G1P4Ma9PMHkZkE8pOPQhe7F6r3gyOkEJY7GbtttNDSZJT114coTMzNc5eckJbsMxmxn32d/axe4PEoNEEhZhkzandczZrDZIeG65mf5fHcREqaON1itGI3FVbzoYQQgghhEukFJvdXHk5mz3hR99DovH4Ufgj/Y841vfxeD07mxpKzb/diEg2O8VfridNaHYmFuth1cYd3sFHpGyO0PrC2691nw438S5cmrMltPlxz9lvdQdosidMBtwWQU0l2QDBOWumkbMhhBBCCJ+WUmx2c/lztptrG5bsz6RbpRNy+GPv9A+2V6lzVSQvqfXDxjOgc8gT+YL/BA+EEEIIIZyvUmx2c+Xm7NVtt79nPDwx49LsG0IIIYQQLptSbHYTOXtlOhx9/CSbnY7ypiwQQgghhHClKMVmN5GzIYQQQgghzEMpNruJnA0hhBBCCGEeSrHZTeRsCCGEEEII81CKzW4iZ0MIIYQQQpiHUmx2EzkbQgghhBDCPJRis5vI2RBCCCGEEOahFJvdRM6GEEIIIYQwD6XY7CZyNoQQQgghhHkoxWY3kbMhhBBCCCHMQyk2u4mcDSGEEEIIYR5KsdlN5GwIIYQQQgjzUIrNbiJnQwghhLBo/OHfXjs3MJ748uuRh9MFcXYVk55+3HE3/e/+7nNpJcM5lWKzm8jZEEIIISwOf9wQk1Ly4uWRc3XzJ43XpVUNcyvFZjeRsyGEEEJYHF64dV9KyYuXJ83VTXQkI61qmFspNruJnA0hhBDC4vB3hWsuYsiT5uomM/1YWtUwt1JsdhM5G0IIIYTFoRSRCyJPmqseaVXD3Eqx2U3kbAghhBAWh1JELog8Zq56pFUNcyvFZjeRsyGEEEJYHEoRuSDymLnqkVY1zK0Um91EzoYQQghhcShF5ILIY+aqR1rVMLdSbHYTORtCCCGExaEUkQsij5mrHmlVw9xKsdlN5GwIIYQQFodSRJ7LR4PDX/9u+OvBL6XuFnnMXPVIqxrmVorNbiJnQwghhLA4lCLyyMOvO3sf2Toyhx6+fHryfzjOfbnj6xvxh3VND7sz8pA8Zq56pFUNcyvFZjeRsyGEEEJYHEoReWTg4TPHJ/fGbC/Vzkztfc8M2dz3Jv+/7qnuQTmX85hpJ+sI77l8uM1DNhkJx9L8izM5519a1TC3Umx2EzkbQgghhMWhFJFHrj/8v1mGfua3U78TG4fEJteKCVu3qmlq0BhGl8dMmWSoygl/NHeUXSjpZCKRcgzB1j7pqL+qKpTk3wTS0UBVlS+c4l8dyMaCueZfWtUwt1JsdnOZcnZ9d1VTT6Uide+qbOqpUjusHZ+mZAY+r6iXOs7b+kSwf7g639EXNtYKc21Djg3XVdmSVGIDjgOUq59vaOqxem2tMQDZTz4aVroStl2FjuiLpYIfCdtLuWYrStyaHRWNA8GeZHVDl95l4dLZbuwut3Wfx9akO7l9cSQrr6bUj65JHQtvSex7T8fypqHw2FRiYiLwntyrWF2NW7+rOp5JTExF4wlbL+KSnKnWNiT8PcM+6XQhn6yEonhfaxc3yXTJPJPTmuPpaAmcx6mJnGndTnG01+Iv8VJEHrk/9bIQo/1tU51DtLr6d5fNjqZnHnZLozN5zLRDa38JqQhJtmqMf1ui+mzX+Cz3cR+QMMfM0VsH5OwCKcVmN5erPvtaKDM7OzFSIXSs7H80O/tIbTC7PHVHErMzkTap47y9OpmdfRK7auue24WNtcJUJmbTQ31SR2rDcIwe9U+yj5+Q/yd65BO0d2hGO7wFphTWq7ztAb0rf0zGJf/MxGLGuB3ewUekVFIm7fX4UbiNnbh7ptjoIvrWrO8LT9EZYLMxmxrqW+Q1ic529oHX1n0eW5PsY7OJHqmjZE+ILMrU+Aa5e6EtiX0vby8m1VifeOaZh0Oxx7PpiQm1f9hhoxepq2/rb+ifnn08He0fV7q6pV5Lc6a6pkzw0w49X2UmvPpdTQW92InwosjJrTo+xTKQ0cXVytgkG5Kf1manMoGCXT1v+uPDXlXqSJzPqakvknU7xdFezleKfJQi8siXj7ojk9+T8vTxye/bG40Qzz78nTQ6k67AXEjJNhUNhyNJnmhpWw3zWzoWDocTPMemExE16PcH1Yi1MUeWFKAEagJKOCqMpwa8VVU1QaEwhr0PnxtaCi09HDWquunUzRlLx+jkyWSEySNnF1ApNru5fO1GLk6k6Flev7Wtp1eyZH/uG+WldnE5exXrkrO7SffZTGoDu7pU9kxlZx+Fcl4J6MVsasJDPw9ESbLRA3FlnFyWJgPaYB9lsnQzaXtOV2DsCZmEPTZVkNj9eDLAJs1iccbHqljIRTE9+yTatag6FdecPbfzydlwKSU7Rt7bDmeGUpAethMjUkfmkpypKsjnx1OKVk49jZipwR5zdIc56VFJLn/8KDxCCpkjZ5d3kdPgk2R/gld7KwO0KoHs2IX5gWIxO/xTz9nxh1UkVUt52sW1TVPy6EweM12RcjZrfaGQc7n+ucrDe9IBvWH2ORnyVXlqAuFoNKL4q6t8+ujpaMBTVe1XItFoOFDj8QRY6s07Z9fU+P0KyfRKjYe2FdEKF1I0nXy1PxSNxaIqGcRhCAekVQ1zK8VmN5ezfTYPK/S80OEfYecI3ov9ujc1k56iP/DpP58NhCamQh9pnz8saxqJToz72Gdf/1S0py8wOJXKPgq36ANwu6p7HpCistlHscGhSq0jHTejNOnDfDSeGBlhN+jaySWhjjxKP55Jjk0Emnga29CTSfQnN1ydSGSfpDMZtaWjvGU4MjGdzj4y51AsVkkoQ2R+yMCT4avGzcOcM9NBJhHL8AG0kz6RLeBNsk6SfOqOLR86NnSNR+m404mRlPBzsOPKpHOojpA5nElNiAWSlZxR2obCZNEyE9W0i/MsGctCiiUL6Jyz1fEkSbTGJnvrZnjKvNI4qJL1b9x6sTkxtpFw3l/XNmypj3SOTdpV7Sb7nCQXwmTcqMHqCI7Nzo4N6V9tvkcW/0FAr8shKz8xNMQnpw5FJh4EG/muGyD7A9uaZMXyCvK5tyZZkNlEnO9jqYkJxyooOtH+AfbZbbMKsk3J9o3JSIzPCdtjByrZHLJt1ONQhW/OLVvbLQl1jOzS0wm6c17z9ZP9jc2hfhRY90OyY+tl0nLG/S0pOp9jw2xgh72uOj6ZGEyas3GRjMWbYbBWGQs66KhOezibJV8D25PZEmnrn5ZJf9mYSYonE1OnTUbODxPT5OKWzkzpJwpduqtklItGlw7/4FSMt0lwPZb1LUvsUUb4bLB163wSkzalsOzzOlrXkRFHRlgiZNYnwxOTKpnEPM48dNLCIrOZ1I7fuXfL8qZkiO/kdNvp3Z0nlGN1SavFcVeRdBqGrmoaUR5PJ+ybfmnOVNqG07tbsrVLxURCHRymy05/msuds3vIHGZHhKOJqKZIkEr00Lma/3pbSw7bCa37g6C2meiOoe/wwvxrWndgx4uLlrOvsalL15elydkPvz4svFQkt9s7nN9MwmOmK1LOZuGaR+uEQiJvjUcL12b3dKRGD9yUZNjL421C9VQFonqQzkYDlozu3BxE6qPl7IgelmmJ/JuZopPhamGQdDKZ1iaJnF1Apdjs5rI+B1lPAxANQ7SKkZybtIO/m/7Wlp0Kx1NqnFxiZ7O8eQkLKMZdMh2Fn4nIOSv7eCadeRCy/dTlGWQ/FMaTvq5xUhSvdeCT0wczgxo9e6bJqWFkXI2nwhMzRm0Ey1UzycwD0j2SeULKTJHsQuZwkP7Gl+hhGc4slp6ms1MZtWvA308G4Is258x4tQH6yYKPR7XKCTaHdAGz0yTx8LmanQ6bl3ZuOf0VeCYxOOz7aISMS1Yay/FsZT5+xMp8kHw8m9ZOzeQ6of0OzhfhCV8EtgayZEmHxrUsK8wS3RazmRQrtoOuELM7iTJOZ8+rk9LVQrzS2KV9eRWRbHmMFKXXZ1ukv8xmx4Ys1xtrZXZZ00SKrGEhu5THHzlFc0PtOqHtjXQXpUvBiiqnxWb82qySXSWTCen7AM/x0takW4RuzRi5bvJVR3fjLLnYs+50Kz/OaLeLorR2ja0ol80qyH4IIjtbKDYc7J80NqW+x9I5DI2Qq6bw25GhObdsz596FCMblBYySz4n6VFg3w/JNdVYav3ndVrOk+zUdIwsVBe5t3E5hOn+MBk0IhS5tdZ+iGijP22ZB53eeExfhJwHnTgtcmNg7OFs0dIZtuf3Z1JkFbGdpILcpI2QXZoO77cdRM4H4MWkGs+QOaSbTG5wQuOOfjunBZ0n0S76OcexLBwC2p7Gu7udxLT1kNKXkSwLP3XM82il+795xmCVrGyXm8eZh03aPFLITGqH+Tx3S3KGCX40oJDdzziROh8UuVaXZbW47CoWhWHYnq8tY7c3RkZ5MktOy/ZN/zTOVGJ2Z7vN0Ahr8T8Z6RFvGplz5ux6sqfxbefoPNcb2460ezA2HBpj1zsysNoXNHZ4ejjLJetrxu1KzcI0uSiw64u+CbSTj7nDL0YpIlOt7+/L4YX7tnGZWsp0xxaBaQcWo5MhGrFp2Kahlvyr1XOzJxID4YRBmHynBZDhq2rUKO+aSERVn55788vZjl+l+mxPTUCNRJPis5XI2QVUis1uLmvO5sc5uXwKt+b07C/kyAajpoEGFLecPTs1Lp/omcaFgXox4dOegRMuMFRLzhYrCWgj8uwIvXenJ1YWsGh3eo6bNdo10jPaSJJ+Nosl5Zjle1oGtIdC5piZejr1WEwPQ+wmREtv1gWkpyp7XQudw0yKz7ma8F5kj8XQxjnCymxLxcbojwDSdYK2jOdpj85Dymi9Uz8cE5tYaF/JtuB5Qu9Ot5FTzjZXLJfN5Lg1qeiyMh2yILGB3hWYaUazZSKptXqcsFen0V9yzeGFXYWrx2Wzi1UaASdG6Aok42anyCVE29yBMb6t2f6g5/i3PqS/xmjXHnNrknUlbE29Ilxew7QiTdgVdekWZwU6b1bR+m5PV9JYA8acWOewgxZor8K37LGzyX4tufLdg9ditpBhHoVY8qOFkPWmT8vcbWg5wrajXx0PYbpd9BsY+pnt3vTnBeGg077O+6Cje7jwEz/7qtJgYZkleoNkpB/bbsl1PwClI1qUNrTVSzMjbO5j2QxwlpztdhKj60FYRtZggC3jPI/Wt7rVjHF40s/a6hX2Vde5pZMW1hWZSa0c2j33bkkLf6TyXaW7uu0mfSLZ7aCY76nPdVcR1Hb1YWMYekTo7crobJsrX3BJz1RUrWLCOB4HyHafzT6KDY2HhuhNo3yjQnfXnDmbrt5cA8xzvZWrN/1d+m9xrE5Bv8K67vC0ZPNc53iY072arD1j6sImWLKc/fDr9y9O/sVcUfvHF50bjRB5zHTFHoFpfTWJ1smwl1ZH03pskl5piA6ylv5sBNrSQ4S2kqY51xtQeRcN3k7EPhEd92BNEb5aUjRtn60EanzVVVVV1VrrFOTsgirFZjeXOWeze2Kycc1aLu26xX4D1dRqjEispOdi15ztePbUqpDJ/frYhOVJcOMCo301T7LsdC88GLSBjM4Ch/V6YzkHmedus1hadTH7eDo2OOJvMR8bn2Nm6OXTcuokqU7LRtICkq/2UMvuWGbTEw9CwvPydGUagcm0m9w/pIeEJ+61Gi/606d1JbMrepRWV2hOkKsITWO0u1hnoxU4z5ytX56t0mjrWEXE2jU6VJjRShe9CnNsWOzL4o4ZgsVdheuWtAz1BSQ7ALk+kXljqY5eh7Tdw7o/OO0D8ioytK5h675kaGxxx81qk70nIT4eGWO/jLMRXedQ1NxjLXNlHdicQ/lAozcJbLcxy6G6H8LCDQxtZK8d9SxUjRj7GKt0ZBHHugiWFWXMIZvWFK1f59J6OFqjbJ0lyz7gtvXdD0C3zUSlqUvrJcRZ96Ks69Cas+0biKmtB2F9GvuhZauxiTodrdoRoS0ynVs9G1n2Vee5tW4C88wz927J6rPJpon0C6/acDso5ru6XHcVfQA+jHB88V9RtIlad2zBJT1T8QbfZFW7HL/sUSWWUHXJ/FhXiCzddsZtjDY8R9tAeaw35Vp114ja/yA28YhkL33Vue7wRsnuhzldD5aaIHMTLF3Onr7h9CikxfcmW0blsQz56nPFIQLToO1XVb9ef+2vCiqK12gSQiO32TzEhA4otCcRcZgIR+rj/tUlRSdVr0vJEtKqhrmVYrOby56z5VN5tXyC69CvXtaLinDhzHGJIqNv6EpFaLs0sguJP3MLJxHzJCufXIwGBtaZtAxmnrvFYut7/PEHtLkhmazxKEzumXE817OSpQUkXx1CLW1pNxymTa7NV2rYVqZmgpzsknHhpG/+CmldyfT0TVuyJgRpA03S3VIfbGwjowvTdvk0630l3aqI6umDQdnMuHO7ZE16lRK3mrUym2irM7ZdIewOsapBeuWggYYsCLldoVuK3xBKO63DPmDbmrrWNWzb5TTFLW7frMZgVFaFRm7q6A/Q/SPqGJ8T1zkUNfdYy1xZBzbnkMwVr0WWeol7fq5DWBuS1sXSsMIrHemkaUtQcTdjbYKti2BZUcYc0mmxFreitPWtdZbYV31XdNs07gegNHWrtAKV1g6KEda9KOuxPO+cbUmTxlh01c19tJJetMKYJjkauI2MaKwi97mlH4ReZCaNw3yO3ZKoJIKDtPEuPQdOsedw8l/z1tXiuqvoA2jDWLeUsOmNYmWX8ky19iO6nqNX9Z8dHLQFU7opLfMjq21Q1kJJ0DzQ5rne2A0AuRuaSow9CPWkYmYItq1GXaNk98OcLo7l+mJugiXM2SMPs7/rfeg/O/ljKV7r7o25/M1IJtlJc+IUgWmTa4JWf81iN8EMucmwr8oX1N4nkk1GAnrszsaCHk9NSHslSTqm+ozYzZ6o9If1ltQCUp/55GzaaNsXou3H6ABB+iAke612MuyvCaI+uzBKsdnNFZezxeoHplE/Ya2oEM6eOS5R5OAv5ye+DvpSJ61ka/ASWuvS85Hx6znR8iu8OZOWc5B57jauW+RzfQc/AdWzFyHxWjGnmTHGMiuVNc1aZ2kByVeHUGsW/uFa1hqHFkVWphiI6WvLaf26UFHHNC8zdA2YK1loM2CRdhdnlW4ah1mSfg7mCVi4nOi6VBGxt2KR1W69dPn6J61P57Cp6L9CyJXZVO0yZiRvWu9oWXwnySpKj2SSZLXQokjsfhQZmiZbQdus0k7rsA/QrSmuOuPV7NY1bN2XDIUt7rRZhSG1n1yMVeeyxwpzKGrusZa5sg5sziGdKzHwGbuNuOcTXQ9hIv2pJxkfJl30Tcbu+oSDztC6CJYVZc4hSSSOTYCkWaJf9eDilvbcD0C3zaRJm+Q+zgTE347ci7IGODN25DiJ0YUVl1E7fdHnDSxbzfVopWpNBYbIbJgHoGVfdZ5b7Y5UX1fi7fQcuyWxvF4/8BVat02n63ZQuM+AdbW47iqC5o8nmuJNtfNRQFyaMxVRe7uR/JZA7fFZc3mtlzbinDlbO4nprdu5rP2edljNb71p68o4MYqz4brDmyW7HuZ0rxZvyIVNkOPV2nkoRWTTL79uaXXO2Xuv5wrZRB4zXXHK2azK2gzWLHabTx4SsolQDW2zQakmKdmIz6lIwEtGpXhqFCH0khBMR3Co77b0mU/ONkahkKnwN/uxl524/yUbaVXD3Eqx2c2Vl7NZ47z00ABrbtHlo9cYLTZpZ5bxauXDtReHo+REpJ+J3C9R9IAnFwaWjbp8I8aE6Gk0PTa0ob6jom08SX/f1LrTs+3s4ymVPY7N6iF41cW8LvnGdYt/YGdtZYDcwhp3+Q4zY4ylNQef0N4t1bEhPkWvXqx6TFpA8tUWaj8lHfWfLMm47LVQpBxpZZKJsosEa+I5E/mINRBXaGNB/VFC65XbMku8YobNKnvQk3cn24iepGyzRNReI5Nhj7drm5I9Z/PWteDIVKJfrwNzriJiw+sv4xOlSZrWD7GXXdT3BMhCCQ0fyfLa20eylsSPQi1kebUVq28+2vjE8TWxWoteI1ayRp/kvMVb68o7rcM+IG9N+rYvVhcurWHn65m+xV02qzAkezRzMsDWUnkL25nZnLjOoag5t5a5sg5sziFbCU8Scbbmxd3GLIfpeghT6Y3BY8vTn2zr8INOK1bbl6yLYFlR5hxapqVt3CnlPdss0a9CztbXmFXXA9BtM+nSXz+yj83dI0dRbO9lu6JyMzhCDxxttetbXBvdIl1YenyxZWT1pnrUs2w196OVSW+5n+gNu5mu+6qw4OwdmjFyrNV3e/vpS+vZppl7t1xHVzL/Ha+8iRzgZD7J9nU7KFxnQFotbruKqDgMmXRCOCE4HwXUJTlTaVNPaseLRZZEJ1IeMkp9t9+cnK6Zs60zIEp/x5tNT6TYcz7kQjZCa1X142V+6439FDM2pB071Wz76rsT2bXIkrILhFWzZNfDnC7d7Oy0dn2xboI+dWIqar5ffIFaI/Kj7o6Hey9Obj/r3G7ke2cmW4ay1lEcJHO8ZLj9XZsl+3s3FvKbirSqYW6l2OzmysvZNC6ktB87KI+nQvo7ocrbJmiMoExH4uaFUzqniOqj0AfmzD9oQs47/A3/tHyl36jfYlfTOPujA5QnyUF+gpvXJd+8bvE/pELzBNnN9T9P4DwzYiBoGIpk6CiM6aj+U6O0gOSrQ6g1/s4CHXfG+FsJJHvRdgUa5so0/hoCJTth/NwpXbmlWTJXiD45CrnihlyastA/EGOMbq5/moZn+R87YFc4exURXS0y+ox1+bTVq5GdMt4axi4nUmW25rXgGA0rDHPlsKkIzRxFWRWXUadFq2SEqjhpp3XaB+iqY3eDGsbWlNawZV8yNLe4y2Y11dcw+5M9U5Elbjcym3kQ0RfK/Lsb4lIz3Q5hKg0r/DEs3WuBIXODGsVaF8GyosQ5tEzLWEXSLNGves6u79MWwaHC0uUAdNtMhnQfFiMs0a0o4YhIDaVIItFWe46TGFsPkxESrzWyk0HeDs2y1ahuRyuVHXTim+at+6rLgpsHjuUwn3O31Edkf7pFeILC+aBwnQHbanHeVaxahkmN6G9QdTsKNJfgTEVnXkLfn8ubzPOneGHimjlbnAFZ655PM3e1nvXnud7ES1J24gGZJX136vCy2hOyvNKkxZJdDnOasxP9rD0bw9z62rE/12+JcypFZOrw1Putk784bUbtH5+e/IuLD1sG5qjGNuTzuuqRVjXMrRSb3Vz+nO3m2oaeDfROXe6ev/S3qjyLWtwfYNd0/tvyc8+M69/0nod0pTmN67Iy5/U3wInsbw47/B3g+W4j5VqBNqVo/n/Cl2wR68qhTYZsF5LCupitaei2WQ0LMpU5NS6xZH7ms+bnu3toOh8v83XOVTR/F7AyhebmFguyXYz7DVraPNYnHWyef7Xbar5zO/c6p38/3OFE6jah+c7AfHYVOkz+5/AlOVO5Os8NOod5/ZYC4lIAAASVSURBVJF2l0vS/E+k9CRgDcruh3lhWonYlSJyQeQxc9UjrWqYWyk2u7lyczaESy1JMM7V8NBJWyUZ/LCi6WZ1LJOW3hdRUI2cLXWH8OnaUXGRVsM7tlx/mkoRuSDymLnqkVY1zK0Um91EzoarV0/XiM949TKcy+DYE/3PPUKuf2Qm+3gmYfxB0CWwevBRlv+5RwiX0UQ484S2TlnkL72LVorIBZHHzFWPtKphbqXY7CZyNoQQQgiLQykiF0QeM1c90qqGuZVis5vI2RBCCCEsDqWIXBB5zFz1SKsa5laKzW4iZ0MIIYSwOJQickHkMXPVI61qmFspNruJnA0hhBDC4vB3X34tpeTFy2Pm6iYz/Vha1TC3Umx2EzkbQgghhMXhhVv3pZS8eHnSXN1ERzLSqoa5lWKzm8jZEEIIISwOf9wQk1Ly4uVJc3XzJ43XpVUNcyvFZjeRsyGEEEJYNP7wb6+dGxhPFK4BCU+aq5L09OOOu+l/93efSysZzqkUm91EzoYQQgghhDAPpdjsJnI2hBBCCCGEeSjFZjeRsyGEEEIIIcxDKTa7iZwNIYQQQghhHkqx2U3kbAghhBBCCPNQis1uImdDCCGEEEKYh1JsdhM5G0IIIYQQwjyUYrObyNkQQgghhBDmoRSb3UTOhhBCCCGEMA+l2OwmcjaEEEIIIYR5KMVmN5GzIYQQQgghzEMpNruJnA0hhBBCCGEeSrHZTeRsCCGEEEII81CKzW4iZ0MIIYQQQpiHUmx2EzkbQgghhBDCPJRis5vI2RBCCCGEEOahFJvddM3ZxMHkF7MAAAAAAAAABonHUmDOIXI2AAAAAAAA8wI5GwAAAAAAgMKDnA0AAAAAAEDhQc4GAAAAFs7dwd99dKzmt7WbWn69qWXvc637NrW8/YujO39x/m/+89/v/8VvA8+3BJ5vDVR/+M4vlN1bLtRuuhTcEjn4Xy4F/0vkoDdy+C8vHNn1nv/Po3/rf3j/Li8RAFAqIGcDAAAAC+c377xR++LPA7/aVPerTe/seK7uV5v3v7rpze2b/mbbn/36r/9sj/c/Bf76ubf+2rNvx3N+8uHF/xT41eY3X9q0x7c58Oqfv71ry8G/efnXr/zywO6/OHvkb3iJAIBSATkbAAAAWDjv125pendH6J3tZwN/dfjXvzpf99ehd/7qb2v/sv7VTb4/+9mfVPyb7/3zf/p//K+//7Mf/fGvqjc07vVe2P+X5+q27d+9/f1fb/u7fX91vs53oe6lC++8+EHdi7xEAECpgJwNAAAALJxw7XOt+/5zC20fsvlgzQt/v/8XoT0/f23T//PD/+33/udv/+Pyf/Rf/d5/9998/19954/+4F/90b/+Nz//2R9/sLf60rt/Xvf/vvD37/x5W90Lbe++8Pfvbrn0rpfISwQAlArI2QAAAMDCCdf+/Ld7N7fse56k7bd3PP/er/7Dz//Pf/HP/9t/tO5f/k9//C//6R/9L//kO2v+8b/4J//1D7/3r3/4b//tj77/g5/98f/+d/6fv7uT5GwastvqtpCc/eG7L3yInA1AyZFHzv6Hmf8flPEVNBW4LNcAAAAASUVORK5CYII="
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Лучшая модель - model_3, на сайте Kaggle получили значение метрики для тестового набора - 0.72370, что на 12.02.2020 является 7 результатом из 75. Разница с 3 местом составляет 3 тысячных.\n",
    "\n",
    "https://www.kaggle.com/c/telecom-clients-prediction2/leaderboard\n",
    "\n",
    "![111.png](attachment:111.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4  Анализ качества и оценка экономического эффекта от внедрения модели\n",
    "#### 4.1  Анализ качества для построенной модели"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- лучшее качество на тестовой выборке с kaggle показала model_3, обученная на X_DATA датасете (на всем объеме тренировочного датасета)\n",
    "\n",
    "- полученное значение roc_auc метрики на тестовой выборке - 0.72370 на 0.28% хуже 3-го места и на 1.32% хуже 1-го места\n",
    "\n",
    "- т.к. у нас отсутствуют реальные ответы на тестовом датасете, то более подробный анализ качества в данном случае получить сложно\n",
    "\n",
    "- возьмем model_2, которая обучалась на data_merge датасете (т.е. data_merge делился на тренировочную и отложенную выборки, и следовательно итоговое качество измерялось как на отложенной выборке, так и на тестовой с Kaggle)\n",
    "\n",
    "- итоговое качество на тесте для model_2 - 0.72246, что на 0.17% хуже нашей лучшей модели (model_3)\n",
    "\n",
    "- проведем детальный анализ качества модели model_2 по результатам на отложенной выборке"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### анализируемая модель - model_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(40000, 38) 38 - категориальных признаков\n",
      "(40000, 174) 174 - вещественных признаков\n",
      "(29532, 212)\n",
      "(29532,)\n",
      "используем 60 признаков после отбора\n",
      "Метрика auc_roc по кросс-валидации для 5 фолдов - 0.7494\n",
      "Метрика auc_roc на hold_out выборке (финальная оценка) - 0.7473\n",
      "Wall time: 2min 5s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "#кодирование категориальных признаков\n",
    "X_cat_filled=data_merge[categ_features_not_null].fillna('w/o_cat')\n",
    "X_cat=cat_to_real_encode(X_cat_filled,'freq')\n",
    "print(X_cat.shape,X_cat.shape[1],'- категориальных признаков')\n",
    "#отбор вещественных признаков\n",
    "X_real_mean=get_X_real_mean(data_merge.drop('labels',axis=1),1)\n",
    "print(X_real_mean.shape,X_real_mean.shape[1],'- вещественных признаков')\n",
    "#метки\n",
    "y=y_train_orig.labels\n",
    "\n",
    "n_folds=5\n",
    "random=0\n",
    "\n",
    "TRAIN, HOLD = hold_out_train(X_real_mean,X_cat,y)\n",
    "\n",
    "#тренировочный датасет разбиваем на матрицу объектов-признаков и вектор меток, данные используем непосредствеено в DF-ах\n",
    "X=TRAIN.drop('labels',axis=1)\n",
    "y=TRAIN['labels']\n",
    "\n",
    "X_validation=HOLD.drop('labels',axis=1)\n",
    "y_validation=HOLD['labels']\n",
    "\n",
    "#### Делаем отбор объектов c помощью ALLKNN kind_sel='mode'\n",
    "\n",
    "allknn = AllKNN(kind_sel='mode')\n",
    "X_resampled, y_resampled = allknn.fit_resample(X, y)\n",
    "print(X_resampled.shape)\n",
    "print(y_resampled.shape)\n",
    "\n",
    "#### Делаем отбор признаков с помощью ExtraTreesClassifier\n",
    "\n",
    "clf=ExtraTreesClassifier(random_state=random)\n",
    "clf.fit(X_resampled,y_resampled)\n",
    "\n",
    "feature_importances=clf.feature_importances_\n",
    "FI=pd.DataFrame({'features':list(X),'feature_importance':feature_importances})\n",
    "FI.sort_values('feature_importance',ascending=True).head()\n",
    "\n",
    "#### model_2\n",
    "\n",
    "seed=np.random.seed(0)\n",
    "param_grid={'base_score': [0.5],\n",
    "         'booster': ['gbtree'],\n",
    "         'colsample_bylevel': [1],\n",
    "         'colsample_bynode': [1],\n",
    "         'colsample_bytree': [0.5],\n",
    "         'gamma': [5],\n",
    "         'learning_rate': [0.055],\n",
    "         'max_delta_step': [0],\n",
    "         'max_depth': [4],\n",
    "         'min_child_weight': [2],\n",
    "         'missing': [None],\n",
    "         'n_estimators': [130],\n",
    "         'nthread': [1],\n",
    "         'objective': ['binary:logistic'],\n",
    "         'random_state': [0],\n",
    "         'reg_alpha': [0],\n",
    "         'reg_lambda': [1],\n",
    "         'scale_pos_weight': [1],\n",
    "         'silent': [None],\n",
    "         'seed': [0],\n",
    "         'subsample': [0.8],\n",
    "         'verbosity': [1]}\n",
    "\n",
    "stat = StratifiedKFold(n_splits=n_folds,shuffle=True,random_state=random)\n",
    "clf=xgb.XGBClassifier(random_state=random)\n",
    "\n",
    "grid_searcher=GridSearchCV(clf,param_grid,n_jobs=-1,cv=stat,scoring='roc_auc')\n",
    "\n",
    "best_features=list(FI[FI.feature_importance>= 0.006184305460633972 ].features)\n",
    "\n",
    "print('используем {} признаков после отбора'.format(len(best_features)))\n",
    "\n",
    "grid_searcher.fit(X_resampled[best_features],y_resampled)\n",
    "\n",
    "print ('Метрика auc_roc по кросс-валидации для {} фолдов - {}'.format(n_folds,grid_searcher.best_score_.round(4)))\n",
    "\n",
    "prediction=grid_searcher.best_estimator_.predict_proba(X_validation[best_features])[:,1]\n",
    "\n",
    "print ('Метрика auc_roc на hold_out выборке (финальная оценка) - {}'.format(roc_auc_score(y_validation,prediction).round(4)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "На выходе нашей модели имеем:\n",
    "\n",
    "- вероятность отнесения к классу\n",
    "- предсказанный класс (с базовым порогом 0.5)\n",
    "- реальный класс, на котором проводили валидацию модели (метки из hold_out выборки)\n",
    "\n",
    "Создаем DF с реузльтатами классификации model_2 (предсказанные класс, соответствующая вероятность и реальный класс) и рассмотрим динамику изменения различных ошибок классификации"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8000,)\n",
      "(8000,)\n",
      "(8000,)\n",
      "TP (верно определенные клиенты из класса \"отток\") =  523\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  72\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  4399\n",
      "Precision =  0.106\n",
      "Recall =  0.879\n",
      "roc_auc =  0.747\n",
      "общее число клиентов в валидационной выборке =  8000\n",
      "TP (верно определенные клиенты из класса \"отток\") =  331\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  264\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  1604\n",
      "Precision =  0.171\n",
      "Recall =  0.556\n",
      "roc_auc =  0.747\n",
      "общее число клиентов в валидационной выборке =  8000\n",
      "TP (верно определенные клиенты из класса \"отток\") =  135\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  460\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  355\n",
      "Precision =  0.276\n",
      "Recall =  0.227\n",
      "roc_auc =  0.747\n",
      "общее число клиентов в валидационной выборке =  8000\n",
      "TP (верно определенные клиенты из класса \"отток\") =  74\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  521\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  99\n",
      "Precision =  0.428\n",
      "Recall =  0.124\n",
      "roc_auc =  0.747\n",
      "общее число клиентов в валидационной выборке =  8000\n",
      "TP (верно определенные клиенты из класса \"отток\") =  34\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  561\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  21\n",
      "Precision =  0.618\n",
      "Recall =  0.057\n",
      "roc_auc =  0.747\n",
      "общее число клиентов в валидационной выборке =  8000\n",
      "TP (верно определенные клиенты из класса \"отток\") =  14\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  581\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  6\n",
      "Precision =  0.7\n",
      "Recall =  0.024\n",
      "roc_auc =  0.747\n",
      "общее число клиентов в валидационной выборке =  8000\n"
     ]
    }
   ],
   "source": [
    "#вероятность отнесения к классу\n",
    "predicted_prob=grid_searcher.best_estimator_.predict_proba(X_validation[best_features])[:,1]\n",
    "print(predicted_prob.shape)\n",
    "\n",
    "#вероятностный класс (с базовым порогом 0.5)\n",
    "predicted_class=grid_searcher.best_estimator_.predict(X_validation[best_features])\n",
    "print(predicted_class.shape)\n",
    "\n",
    "#реальный класс, на котором проводили валидацию модели (метки из hold_out выборки)\n",
    "print(y_validation.shape)\n",
    "\n",
    "DF_model_res=pd.DataFrame({'predicted_prob':predicted_prob,\n",
    "                                 'predicted_class':predicted_class,\n",
    "                                 'true_class':y_validation})\n",
    "\n",
    "#перебираем различные значения порога классификации и смотрим динамику изменения различных ошибок\n",
    "TP_list=[]\n",
    "FN_list=[]\n",
    "FP_list=[]\n",
    "threshold_list=[0.05,0.1,0.2,0.3,0.4,0.5]\n",
    "\n",
    "for threshold in threshold_list:\n",
    "    TP,FN,FP, _ = mistakes_balance_tuning(DF_model_res,threshold)\n",
    "    TP_list.append(TP)\n",
    "    FN_list.append(FN)\n",
    "    FP_list.append(FP)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Визуализируем результаты"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(threshold_list,TP_list)\n",
    "plt.plot(threshold_list,FN_list)\n",
    "plt.plot(threshold_list,FP_list)\n",
    "plt.plot(threshold_list,[y_validation.sum()]*6,linestyle = '--')\n",
    "plt.legend(['TP','FN','FP','class_1'])\n",
    "plt.xlabel('threshold')\n",
    "plt.ylabel('customers (samples in validated dataset)')\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Рассмотрим значения в точке (с порогом = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TP (верно определенные клиенты из класса \"отток\") =  135\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  460\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  355\n",
      "Precision =  0.276\n",
      "Recall =  0.227\n",
      "roc_auc =  0.747\n",
      "общее число клиентов в валидационной выборке =  8000\n"
     ]
    }
   ],
   "source": [
    "TP,FN,FP, DF_res = mistakes_balance_tuning(DF_model_res,0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Можно наблюдать, что в завимисоти от выбранного порога классификации, наблюдается различное соотношение ошибок нашей модели.\n",
    "\n",
    "Для порога равного 0.2 наша модель выдает следующие результаты:\n",
    "\n",
    "- всего клиентов из класса отток - 595\n",
    "- верно определенные клиенты из класса \"отток\" -  135\n",
    "- ложные пропуски клиентов из класса \"отток\" - 460\n",
    "- ложные срабатывания - неверное отнесение клиента к классу \"отток\" - 355\n",
    "- объем класса отток , предсказанный моделью - 490\n",
    "\n",
    "Насколько данная модель может быть полезной в бизнесе, и насколько высока/низка ее предсказательная способность для бизнеса можно определить только задав некторые вводные параметры, для которых можно подобрать оптимальный порог классификации (баланс между видами ошибок), при котором наша модель будет демонстрировать наилучшую эффективность."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.2  Оценка ожидаемого экономического эффекта от внедрения построенной модели в бизнесе"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Пусть:\n",
    "\n",
    "- один пользователь в среднем в месяц приносит - 100 у.е.\n",
    "- для удержания одного пользователя в среднем в месяц будем выделять - 10 у.е.\n",
    "- вероятность того, что пользователь лоялен к спец. предложениям - 50 %\n",
    "- из клиентов склонных к оттоку будем работать с топ - 25 %\n",
    "\n",
    "У нас имеется всего 8000 пользователей в текущем месяце\n",
    "\n",
    "#### Сценарий 1 - не применяем модель по прогнозированию оттока\n",
    "\n",
    "- Потенциальная прибыль = 8000 * 100 = 800 000 у.е.\n",
    "- Число пользователей попавших на конец месяца в группу отток (число 1 в векторе целевых признаков y_validation) - 595 человек\n",
    "- Потерянная прибыль из-за оттока = 595 * 100 = 59 500 у.е.\n",
    "\n",
    "#### Результат:\n",
    "\n",
    "- Реальная прибыль = Потенциальная прибыль - Потерянная прибыль из-за оттока = 740 500 у.е\n",
    "\n",
    "#### Сценарий 2 - применяем модель\n",
    "\n",
    "- в качестве порога классификации использем 0.2\n",
    "- из 595 человек мы определим верно 135 - 22 % (округляем вниз)\n",
    "- также мы неверно определим в группу склонных к оттоку 355 человек \n",
    "\n",
    "т.е. наша модель выделит 490 человек из которых только 27% входят в класс отток\n",
    "\n",
    "- работаем с Топ 25% из класса отток (определенного моделью)\n",
    "- из 122 человек (топ 25%) действительно к 1ому классу относятся только 58 человек - 47%\n",
    "- учтем, что лояльность пользователей к спец предложениям составляет 50%\n",
    "\n",
    "т.е. из 122 человек - 64 не из класса отток, предполагаем, что неверно сделанное спец. предложение для них не переведет их в группу отток =)\n",
    "следовательно, из 58 человек, только половина решит остаться - 29 человек.\n",
    "\n",
    "\n",
    "- Потенциальная прибыль = 8000 * 100 = 800 000 у.е.\n",
    "- Модель определяет 490 человек 1ого класса из которых мы работаес с Топ25 - 122 человека\n",
    "- Расходы на удержание 10*122=1220 у.е\n",
    "- Потерянная прибыль из-за оттока = (595-29) * 100  = 56 600 у.е.\n",
    "\n",
    "#### Реузльтат:\n",
    "\n",
    "- Реальная прибыль = Потенциальная прибыль - Потерянная прибыль из-за оттока - Расходы на удержание = 742 180 у.е\n",
    "\n",
    "#### Итог:\n",
    "\n",
    "- прибыль без модели - 740 500 у.е.\n",
    "- прибыль с моделью - 742 180 у.е\n",
    "- прирост прибыли с моделью +0.227% "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.227"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#ранее определили функцию, которая реализует данный расчет\n",
    "mistakes_balance_tuning_and_final_estimation(DF_model_res,0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Рассмотрим зависимость разницы в месечной доходности при различных уровнях порога."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#максимальное значение порога, выше которого отсутствуют истинные значения класса отток \n",
    "# (TP=0, следовательно Recall и Precision = 0)\n",
    "\n",
    "max_threshold=DF_model_res[DF_model_res.true_class==1].predicted_prob.max()\n",
    "\n",
    "#минимальное значение порога, ниже которого отсутствуют истинные значения класса отток \n",
    "# (TP=595 - все истинные значения определены, следовательно Recall =  1)\n",
    "\n",
    "min_threshold=DF_model_res[DF_model_res.true_class==1].predicted_prob.min()\n",
    "\n",
    "result_list=[]\n",
    "\n",
    "threshold_list=np.linspace(min_threshold,max_threshold,100)\n",
    "\n",
    "for threshold in threshold_list:\n",
    "    #print('\\nпорог = ',threshold)\n",
    "    diff = mistakes_balance_tuning_and_final_estimation(DF_model_res,threshold)\n",
    "    result_list.append(diff)\n",
    "\n",
    "plt.figure(figsize=(10,6))\n",
    "plt.plot(threshold_list,result_list)\n",
    "plt.xlabel('threshold')\n",
    "plt.ylabel('Difference_in_total_income_per_mnth')\n",
    "plt.title('Difference_in_total_income_per_mnth = f ( threshold )')\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Наилучший итог от применения модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TP (верно определенные клиенты из класса \"отток\") =  278\n",
      "FN (ложные пропуски клиентов из класса \"отток\") =  317\n",
      "FP (ложные срабатывания - неверное отнесение клиента к классу \"отток\") =  1083\n",
      "Precision =  0.204\n",
      "Recall =  0.467\n",
      "roc_auc =  0.747\n",
      "реальный объем класса отток  595\n",
      "класс отток, определяемый моделью  1361\n",
      "% дучших от класса отток  340\n",
      "число людей действительно относящихся к классу отток из исследуемых  115\n",
      "число людей, лояльных к предложению  58.0\n",
      "прибыль без модели =  740500\n",
      "затраты на удержание =  3400\n",
      "прибыль от сохраненных клиентов  5800.0\n",
      "прибыль с моделью =  742900.0\n",
      "разница в итоговой прибыли после применения модели  0.32\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.324"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#максимальное значение\n",
    "#print(max(result_list))\n",
    "\n",
    "#индекс максимального значения\n",
    "max_index=np.where(result_list==(max(result_list)))[0][0]\n",
    "\n",
    "#соответствующий порог\n",
    "#print(round(threshold_list[max_index],3))\n",
    "\n",
    "mistakes_balance_tuning_and_final_estimation(DF_model_res,threshold_list[max_index])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Вывод:\n",
    "    \n",
    "При ранее определенных начальных условиях в зависимости от выбираемого порога наша модель может как увеличивать, так и уменьшать месячную доходность.\n",
    "\n",
    "Максимальное значение достигается при пороге 0.124, которому соответствует увеличение доходности на 0.32%, относительно номинальной доходности без использования модели.\n",
    "\n",
    "Минимальное значение соответствует порогу классификации 0.012, для которого эффект от применения модели начинает носить отицательный характер. (наблюдается  снижение доходности на 0.39%)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### +0.32% это много или мало ???"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В нашем эксперименте для оценки экономическо эффективности размер выборки , для которой делался прогноз - 8000 человек (отложенная выборка , полученная при разбиении 40 000 шт на 5 фолдов)\n",
    "\n",
    "Если опираться на введные параметры:\n",
    "\n",
    "- один пользователь в среднем в месяц приносит - 100 у.е.\n",
    "- для удержания одного пользователя в среднем в месяц будем выделять - 10 у.е.\n",
    "- вероятность того, что пользователь лоялен к спец. предложениям - 50 %\n",
    "- из клиентов склонных к оттоку будем работать с топ - 25 %\n",
    "\n",
    "То мы получаем прирост +0.32% или ≈ 2 400 у.е.\n",
    "\n",
    "если увеличим объем выборки на порядок с сохранением баласна между классами, т.е. делаем прогноз для 80 000 клиентов, то прирост в месячной доходности увелчится на порядок и будет составлять ≈ 24 000 у.е.\n",
    "\n",
    "если увеличим объем выборки еще на порядок с сохранением баласна между классами, т.е. делаем прогноз для 800 000 клиентов, то прирост в месячной доходности увелчится на порядок и будет составлять ≈ 240 000 у.е.\n",
    "\n",
    "В реальности объем клиентской базы телеком оператора - 10-ки или 100-и миллионов человек, следовательно, даже небольшой процент прироста прибыли может быть значимым, и, как следствие, применение модели машинного обучения может быть экономически целесообразно."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5. Итог работы"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В результате выполнения финального проекта удалось:\n",
    "    \n",
    "- на практике применить знания и навыки, полученные в рамках специализации\n",
    "- построить модель машинного обучения, которая предсказывает вероятность отнесения клиента мобильного оператора к классу 'отток'\n",
    "- получить точность предсказания по метрике roc_auc - 0.72370, которая находится на 7 месте в общем рейтинге (из 75) и отстает от 3 места на 0.28% и на 1.32% хуже 1-го места\n",
    "- оценить потенциальный экономический эффект от применения модели в бизнесе"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}