homework-practice-04-classification_Petyaeva.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Машинное обучение, ФКН ВШЭ\n",
    "\n",
    "## Практическое задание 4. Классификация. \n",
    "\n",
    "### Общая информация\n",
    "Дата выдачи: 29.10.2020\n",
    "\n",
    "Мягкий дедлайн: 23:59MSK 10.11.2020 (за каждый день просрочки снимается 1 балл)\n",
    "\n",
    "Жесткий дедлайн: 23:59MSK 13.11.2020"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### О задании\n",
    "\n",
    "В этом задании вы:\n",
    "- ознакомитесь с тем, что происходит \"внутри\" метода опорных векторов и логистической регрессии\n",
    "- познакомитесь с калибровкой вероятности\n",
    "- изучите методы трансформации переменных и методы отбора признаков\n",
    "- попробуете оценить экономический эффект модели\n",
    "\n",
    "\n",
    "### Оценивание и штрафы\n",
    "Каждая из задач имеет определенную «стоимость» (указана в скобках около задачи). Максимально допустимая оценка за работу — 10 баллов.\n",
    "\n",
    "Сдавать задание после указанного срока сдачи нельзя. При выставлении неполного балла за задание в связи с наличием ошибок на усмотрение проверяющего предусмотрена возможность исправить работу на указанных в ответном письме условиях.\n",
    "\n",
    "Задание выполняется самостоятельно. «Похожие» решения считаются плагиатом и все задействованные студенты (в том числе те, у кого списали) не могут получить за него больше 0 баллов (подробнее о плагиате см. на странице курса). Если вы нашли решение какого-то из заданий (или его часть) в открытом источнике, необходимо указать ссылку на этот источник в отдельном блоке в конце вашей работы (скорее всего вы будете не единственным, кто это нашел, поэтому чтобы исключить подозрение в плагиате, необходима ссылка на источник).\n",
    "\n",
    "Неэффективная реализация кода может негативно отразиться на оценке.\n",
    "\n",
    "\n",
    "### Формат сдачи\n",
    "Для сдачи задания переименуйте получившийся файл *.ipynb в соответствии со следующим форматом: homework-practice-04-classification-Username.ipynb, где Username — ваша фамилия и имя на латинице именно в таком порядке (например, homework-practice-04-classification-IvanovIvan.ipynb).\n",
    "\n",
    "Для удобства проверки самостоятельно посчитайте свою максимальную оценку (исходя из набора решенных задач) и укажите ниже.\n",
    "\n",
    "**Оценка**: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 641,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm\n",
    "%matplotlib inline\n",
    "\n",
    "pd.set_option('display.max_rows', 100)\n",
    "pd.set_option('display.max_columns', 100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Часть 0. SVM, LR и калибровка вероятностей. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Сгенерируем синтетические данные."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 642,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import make_classification\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X, y = make_classification(\n",
    "    n_samples=10000, n_features=10, n_informative=5, n_redundant=5,\n",
    "    random_state=42)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.3, random_state=42)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Задание 1.__ Обучение и применение метода опорных векторов и логистической регрессии.\n",
    "\n",
    "__(1 балл)__\n",
    "\n",
    "Обучите метод опорных векторов. На занятиях мы проходили линейный вариант без ядрового перехода, что соответствует линейному ядру (`LinearSVC` или же `SVC(kernel='linear')` в `sklearn.svm`). Подберите параметр регуляризации `C` (можете воспользоваться кросс-валидацией или отделить валидационную выборку от обучающей). __Здесь и далее, если не сказано иначе, подбирайте оптимальные гиперпараметры с точки зрения AUC-PR.__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 647,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_valid, y_train, y_valid = train_test_split(X_train, y_train, test_size=0.3 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 648,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn import metrics\n",
    "from sklearn.svm import LinearSVC, SVC\n",
    "from sklearn.metrics import precision_recall_curve\n",
    "from sklearn.metrics import auc\n",
    "\n",
    "# https://sinyi-chou.github.io/python-sklearn-precision-recall/\n",
    "\n",
    "\n",
    "\n",
    "def find_best_c(X_train, y_train, X_test, y_test, learn_model):\n",
    "    \n",
    "    auc_pr_list = []\n",
    "\n",
    "    for c in tqdm(np.linspace(0.001, 1, 10)):\n",
    "        model = learn_model(max_iter=100000, C=c).fit(X_train, y_train)\n",
    "        pred = model.predict(X_test)\n",
    "        precision, recall, thresholds = precision_recall_curve(y_test, pred)\n",
    "        auc_precision_recall = auc(recall, precision)\n",
    "        auc_pr_list.append(auc_precision_recall)\n",
    "\n",
    "    return np.linspace(0.01, 1, 10)[np.argmax(auc_pr_list)]\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 645,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 10/10 [00:04<00:00,  2.22it/s]\n"
     ]
    }
   ],
   "source": [
    "c_best_svc = find_best_c(X_train, y_train, X_valid, y_valid, LinearSVC)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 646,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.01"
      ]
     },
     "execution_count": 646,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "c_best_svc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "На тестовой части постройте ROC и PR кривые, а также посчитайте AUC-ROC, AUC-PR с точностью до 5 знаков. Сравните AUC-ROC и AUC-PR для вашей модели с этими же метриками для случайного классификатора (который на каждом объекте выдаёт один из двух классов с равными вероятностями)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from sklearn import datasets, metrics, model_selection, svm\n",
    "\n",
    "svc = LinearSVC(max_iter=100000, C=c_best_svc).fit(X_train, y_train)\n",
    "metrics.plot_roc_curve(svc, X_test, y_test) \n",
    "\n",
    "plt.show() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 660,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC_PR:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.83264"
      ]
     },
     "execution_count": 660,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision, recall, thresholds = precision_recall_curve(y_test, svc.predict(X_test))\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "print('AUC_PR:')\n",
    "np.round(auc_precision_recall, 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 659,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC_ROC:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7733419981499846"
      ]
     },
     "execution_count": 659,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('AUC_ROC:')\n",
    "metrics.roc_auc_score(y_test, svc.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 661,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.precision_recall_curve.PrecisionRecallDisplay at 0x7ff3a27c74c0>"
      ]
     },
     "execution_count": 661,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "metrics.plot_precision_recall_curve(svc, X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 653,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DummyClassifier(strategy='uniform')"
      ]
     },
     "execution_count": 653,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.dummy import DummyClassifier\n",
    "dummy_clf = DummyClassifier(strategy=\"uniform\")\n",
    "dummy_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 654,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "metrics.plot_roc_curve(dummy_clf, X_test, y_test) \n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 655,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.precision_recall_curve.PrecisionRecallDisplay at 0x7ff3a1bd3d30>"
      ]
     },
     "execution_count": 655,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAsTAAALEwEAmpwYAAAkFElEQVR4nO3de5xWVd338c/X4ZyIJlAmEoSoqCAqiqQZaJl5pEjBw22iSVqaUeatj4fEzEOeikd7EAsPHRRPKSjKbR7LNBlkRAEVRJJBvEVUDsqZ3/PH3jNeDDNzbWCua5y5vu/Xa15z7b3W3vu3h8Nv1lp7r6WIwMzMStdWjR2AmZk1LicCM7MS50RgZlbinAjMzEqcE4GZWYlr0dgBbKqOHTtGt27dGjsMM7MmZerUqe9HRKfayppcIujWrRvl5eWNHYaZWZMi6T91lblryMysxDkRmJmVOCcCM7MS50RgZlbinAjMzEpcwRKBpHGS3pP0ah3lkjRa0hxJ0yXtU6hYzMysboVsEdwOHF5P+beBnunXCOD/FTAWMzOrQ8ESQUQ8C3xQT5VjgTsj8QKwraQdChXPqIkzGDVxRqFOb2bWZDXmC2U7AvNztivTfQtrVpQ0gqTVQNeuXTfrYjPfWbpZx5mZNXdNYrA4IsZGRL+I6NepU61vSJuZ2WZqzESwANgpZ7tLus/MzIqoMRPBBOCU9OmhA4AlEbFRt5CZmRVWwcYIJN0FDAQ6SqoEfgm0BIiIMcAk4AhgDvAJMLxQsZiZWd0Klggi4oQ85QH8uFDXNzOzbJrEYLGZmRWOE4GZWYlzIjAzK3FOBGZmJc6JwMysxDkRmJmVOCcCM7MS50RgZlbinAjMzEqcE4GZWYlzIjAzK3FOBGZmJc6JwMysxDkRmJmVOCcCM7MS50RgZlbinAjMzEqcE4GZWYlzIjAzK3FOBGZmJc6JwMysxDkRmJmVOCcCM7MS50RgZlbinAjMzEqcE4GZWYlzIjAzK3FOBGZmJa5FlkqSOgMHAl8CVgCvAuURsb6AsZmZWRHUmwgkDQIuAD4PTAPeA9oAg4Eeku4Dro+IpQWO08zMCiRfi+AI4IyIeLtmgaQWwFHAN4H7CxCbmZkVQb1jBBHxi9qSQFq2NiIejIg6k4CkwyW9LmmOpAtqKf+ypCckTZf0tKQum34LZma2JTZ7sFjS8DzlZcDNwLeB3YETJO1eo9p1wJ0R0Qe4HLhqc+MxM7PNsyVPDY3KU74/MCci5kbEauBu4NgadXYHnkw/P1VLuZmZFVi+weLpdRUBX8hz7h2B+TnblUD/GnVeBr4L/A74DtBe0vYRsbhGHCOAEQBdu3bNc1kzM9sU+QaLvwB8C/iwxn4B/2qA658H3CTpVOBZYAGwrmaliBgLjAXo169fNMB1zcwslS8RPAxsHREVNQskPZ3n2AXATjnbXdJ91SLiHZIWAZK2BoZExEd5zmtmZg2o3kQQEafXU3ZinnNPAXpK6k6SAIYBGxwjqSPwQfpi2oXAuCxBm5lZwynYFBMRsRY4G5gMzALuiYgZki6XdExabSDwuqQ3SLqhfl2oeMzMrHaZppjYXBExCZhUY9+lOZ/vA+4rZAxmZlY/TzpnZlbinAjMzEpc5kQgaWx922Zm1jRtSovgljzbZmbWBGVOBBExtb5tMzNrmvJNMTERqPNN3og4pq4yMzNrGvI9PnpdUaIwM7NGk+/N4meqPktqC3SNiNcLHpWZmRVNpjECSUcDFcBj6XZfSRMKGJeZmRVJ1sHiy0jWF/gIIJ2ErntBIjIzs6LKmgjWRMSSGvs8HbSZWTOQda6hGZJOBMok9QR+QsOsR2BmZo0sa4vgHGAPYBVwF7AU+GmBYjIzsyLK1CKIiE+AiyRdk2zGssKGZWZmxZL1qaH9JL0CTAdekfSypH0LG5qZmRVD1jGCPwI/ioh/AEg6CLgN6FOowMzMrDiyjhGsq0oCABHxT2BtYUIyM7NiyjfX0D7px2ck3UIyUBzAUODpwoZmZmbFkK9r6Poa27/M+ez3CMzMmoF8cw0NKlYgZmbWODIvXi/pSJJ3CdpU7YuIywsRlJmZFU/Wx0fHkIwLnAMIOA74cgHjMjOzIsn61NBXI+IU4MOIGAUMAHYpXFhmZlYsWRPBivT7J5K+BKwBdihMSGZmVkxZxwgelrQtcC3wEskTQ38oVFBmZlY8Weca+lX68X5JDwNtapmW2szMmqB8L5R9t54yIuKBhg/JzMyKKV+L4Oh6ygJwIjAza+LyvVA2vFiBmJlZ48j61JCZmTVTTgRmZiWuoIlA0uGSXpc0R9IFtZR3lfSUpGmSpks6opDxmJnZxrJOMdFO0iWSbk23e0o6Ks8xZcDNwLeB3YETJO1eo9rFwD0RsTcwDPj9pt6AmZltmawtgttIFq4fkG4vAK7Ic8z+wJyImBsRq4G7gWNr1Algm/RzB+CdjPGYmVkDyZoIekTEb0imlqhazF55jtkRmJ+zXZnuy3UZcLKkSmASyaR2G5E0QlK5pPJFixZlDNnMzLLImghWS2pLuhiNpB4kLYQtdQJwe0R0AY4A/iRpo5giYmxE9IuIfp06dWqAy5qZWZWscw1dBjwG7CTpL8CBwKl5jlkA7JSz3SXdl+t04HCAiHheUhugI/BexrjMzGwLZZ1r6H8kTQUOIOkSOjci3s9z2BSgp6TuJAlgGHBijTpvA4cCt0vqRbLojft+zMyKKFMikDQR+CswISI+znJMRKyVdDYwGSgDxkXEDEmXA+URMQH4OXCrpJEk3U6nRoTXQjYzK6KsXUPXkaxQdrWkKSRPAD0cESvrOygiJpEMAufuuzTn80ySbiYzM2skWbuGngGeSd8NOAQ4AxjHp49+mplZE7Upi9e3JZmNdCiwD3BHoYIyM7PiyTpGcA/JC2KPATcBz0TE+kIGZmZmxZG1RfBH4ISIWFfIYMzMrPjyrVB2SEQ8CXwOOFba8GVir1BmZtb05WsRfB14ktpXKvMKZWZmzUC+Fcp+mX68PCLeyi1LXxQzM7MmLutcQ/fXsu++hgzEzMwaR74xgt2APYAOkr6bU7QNyXQQZmbWxOUbI9gVOArYlg3HCZaRvFRmZmZNXL4xgoeAhyQNiIjnixSTmZkVUb6uofPTBWlOlHRCzfKI+EnBIjMzs6LI1zU0K/1eXuhAzMysceTrGpqYfq+eVyhdQWzriFha4NjMzKwIMj0+KumvkraR9DngVWCmpF8UNjQzMyuGrO8R7J62AAYDjwLdgf8qVFBmZlY8WRNBS0ktSRLBhIhYQ7qQvZmZNW1ZE8EtwDySyeeelfRlwGMEZmbNQNYVykYDo3N2/UfSoMKEZGZmxZR1sLiDpBskladf15O0DszMrInL2jU0jmRaiePTr6XAbYUKyszMiifrCmU9ImJIzvYoSRUFiMfMzIosa4tghaSDqjYkHQisKExIZmZWTFlbBGcCd0rqkG5/CHy/MCGZmVkx5U0EkvoCOwPDgAUAnl7CzKz5qLdrSNKlwD3AEOARYKiTgJlZ85KvRTAU6BsRn0jaHngMuLXwYZmZWbHkGyxeFRGfAETE4gz1zcysicnXIviKpAnpZwE9craJiGMKFpmZmRVFvkRwbI3t6woViJmZNY58C9M8U6xAzMysceR7amiipKPTKahrln1F0uWSTqvn+MMlvS5pjqQLaim/UVJF+vWGpI826y7MzGyz5esaOgP4GfBbSR8Ai4A2QDfgTeCmiHiotgMllQE3A98EKoEpkiZExMyqOhExMqf+OcDem38rZma2OfJ1Db0LnA+cL6kbsAPJ1BJvVD1NVI/9gTkRMRdA0t0kYw4z66h/AvDL7KGbmVlDyDrFBBExj2Rxmqx2BObnbFcC/WurmC500x14so7yEcAIgK5du25CCGZmls9n5b2AYcB9EbGutsKIGBsR/SKiX6dOnYocmplZ81bIRLAA2Clnu0u6rzbDgLsKGIuZmdWhkIlgCtBTUndJrUj+s59Qs5Kk3YDtgOcLGIuZmdUh61KVB0p6PH3Ec66ktyTNre+YiFgLnA1MBmYB90TEjPSR09w3kocBd0dEbO5NmJnZ5ss6WPxHYCQwFai1H782ETEJmFRj36U1ti/Lej4zM2t4WRPBkoh4tKCRmJlZo8iaCJ6SdC3wALCqamdEvFSQqMzMrGiyJoKq5//75ewL4JCGDcfMzIotUyKIiEGFDsTMzBpH1qeGOki6QVJ5+nV9zkL2ZmbWhGV9j2AcsAw4Pv1aCtxWqKDMzKx4so4R9IiIITnboyRVFCAeMzMrsqwtghWSDqrakHQgySykZmbWxGVtEZwF3JGOCwj4ADi1UEGZmVnxZH1qqALYS9I26fbSQgZlZmbFU28ikHRyRPxZ0s9q7AcgIm4oYGxmZlYE+VoEn0u/ty90IGZm1jjyLVV5S/p9VHHCMTOzYsv6QtlvJG0jqaWkJyQtknRyoYMzM7PCy/r46GHpAPFRJOsW7wz8olBBmZlZ8WRNBFVdSEcC90bEkgLFY2ZmRZb1PYKHJb1G8hLZWZI6ASsLF5aZmRVLphZBRFwAfBXoFxFrgI+BYwsZmJmZFUe+9wgOiYgnJX03Z19ulQcKFZiZmRVHvq6hrwNPAkfXUhY4EZiZNXn53iP4Zfp9eHHCMTOzYsv6HsGVkrbN2d5O0hUFi8rMzIom6+Oj346Ij6o2IuJD4IiCRGRmZkWVNRGUSWpdtSGpLdC6nvpmZtZEZH2P4C/AE5KqlqccDtxRmJDMzKyYsq5HcI2kl4FvpLt+FRGTCxeWmZkVS9YWAcAsYG1E/F1SO0ntI2JZoQIzM7PiyPrU0BnAfcAt6a4dgQcLFJOZmRVR1sHiHwMHAksBImI20LlQQZmZWfFkTQSrImJ11YakFiRvFpuZWROXNRE8I+n/AG0lfRO4F5hYuLDMzKxYsiaC/wYWAa8APwQmARfnO0jS4ZJelzRH0gV11Dle0kxJMyT9NWvgZmbWMPI+NSSpDJgREbsBt2Y9cXrczcA3gUpgiqQJETEzp05P4ELgwIj4UJLHHczMiixviyAi1gGvS+q6iefeH5gTEXPT8YW72XgNgzOAm9MpK4iI9zbxGmZmtoWyvkewHTBD0oski9IAEBHH1HPMjsD8nO1KoH+NOrsASHoOKAMui4jHap5I0ghgBEDXrpuaj8zMrD5ZE8ElBbx+T2Ag0AV4VlLv3AnuACJiLDAWoF+/fn5aycysAeVboawNcCawM8lA8R8jYm3Gcy8AdsrZ7pLuy1UJ/Dtd/vItSW+QJIYpGa9hZmZbKN8YwR1AP5Ik8G3g+k049xSgp6TukloBw4AJNeo8SNIaQFJHkq6iuZtwDTMz20L5uoZ2j4jeAJL+CLyY9cQRsVbS2cBkkv7/cRExQ9LlQHlETEjLDpM0E1gH/CIiFm/OjZiZ2ebJlwjWVH1I/2PfpJNHxCSSdw5y912a8zmAn6VfZmbWCPIlgr0kLU0/i+TN4qXp54iIbQoanZmZFVy+xevLihWImZk1jqxTTJiZWTPlRGBmVuKcCMzMSpwTgZlZiXMiMDMrcU4EZmYlzonAzKzEORGYmZU4JwIzsxLnRGBmVuKcCMzMSpwTgZlZiXMiMDMrcU4EZmYlzonAzKzEORGYmZU4JwIzsxKXb6nKJmHNmjVUVlaycuXKOuv8eO+2AMyaNatYYZlttjZt2tClSxdatmzZ2KFYCWgWiaCyspL27dvTrVs3JNVap9Wi5QD06LR1MUMz22QRweLFi6msrKR79+6NHY6VgGbRNbRy5Uq23377OpOAWVMiie23377eFq5ZQ2oWiQBwErBmxX+frZiaTSIwM7PN40TQQMrKyujbty977LEHe+21F9dffz3r169v7LCAZDD9ggsuoGfPnuyzzz4MGDCARx99FIBu3brx/vvvN8h1JkyYwNVXXw3AokWL6N+/P3vvvTf/+Mc/OOKII/joo4+26Pw//elPefbZZ6u333//fVq2bMmYMWM2qNetWzd69+5Nnz59OOyww3j33Xe36LoAV111FTvvvDO77rorkydPrrXOqaeeSvfu3enbty99+/aloqICgNdee40BAwbQunVrrrvuuur6q1ev5uCDD2bt2rVbHJ/ZlnAiaCBt27aloqKCGTNm8Pjjj/Poo48yatSoxg4LgEsuuYSFCxfy6quv8tJLL/Hggw+ybNmyBr/OMcccwwUXXADAE088Qe/evZk2bRpf+9rXmDRpEttuu23mc61bt26D7cWLF/PCCy9w8MEHV++79957OeCAA7jrrrs2Ov6pp55i+vTp9OvXjyuvvHLzbig1c+ZM7r77bmbMmMFjjz3Gj370o43iq3LttddSUVFBRUUFffv2BeDzn/88o0eP5rzzztugbqtWrTj00EMZP378FsVntqWaxVNDuUZNnMHMd5ZutH/lmuQfbpuWZZt8zt2/tA2/PHqPzPU7d+7M2LFj2W+//bjsssu44447KC8v56abbgLgqKOO4rzzzmPgwIFsvfXWnHXWWUyaNIkddtiBK6+8kvPPP5+3336b3/72txxzzDHcfvvtPPjgg3z88cfMnj2b8847j9WrV/OnP/2J1q1bM2nSJD788EOOO+44XnrpJQBmz57N0KFD+ec//8mtt97KW2+9RevWrQH4whe+wPHHH79R3IMHD2b+/PmsXLmSc889lxEjRrBu3TpOP/10ysvLkcRpp53GyJEjGT16NGPGjKFFixbsvvvu3H333dx+++2Ul5fzgx/8gPPPP58VK1ZQXl7O888/T69evSgvL6djx478+c9/ZvTo0axevZr+/fvz+9//nrKyMrbeemt++MMf8ve//52bb76Zgw46qDq2+++/n8MPP3yDeO+66y6uv/56TjzxRCorK+nSpctG93TwwQczevTozH92tXnooYcYNmwYrVu3pnv37uy88868+OKLDBgwINPxnTt3pnPnzjzyyCMblQ0ePJgLL7yQk046aYtiNNsSbhEUyFe+8hXWrVvHe++9V2+9jz/+mEMOOYQZM2bQvn17Lr74Yh5//HH+9re/cemll1bXe/XVV3nggQeYMmUKF110Ee3atWPatGkMGDCAO++8kx49etChQ4fq7ojbbruN4cOHM2fOHLp27co222yTN+Zx48YxdepUysvLGT16NIsXL6aiooIFCxbw6quv8sorrzB8+HAArr76aqZNm8b06dM36prp27cvl19+OUOHDqWiooK2bdtWl82aNYvx48fz3HPPUVFRQVlZGX/5y1+qfxb9+/fn5Zdf3iAJADz33HPsu+++1dvz589n4cKF7L///hx//PF1/lb98MMP07t37432jxw5sroLJ/erqmsr14IFC9hpp52qt7t06cKCBQtqvd5FF11Enz59GDlyJKtWraq1Tq4999yTKVOm5K1nVkjNrkVQ12/ub35G3yNo1apV9W+6vXv3pnXr1rRs2ZLevXszb9686nqDBg2iffv2tG/fng4dOnD00UdXHzN9+nQAfvCDH3Dbbbdxww03MH78eF588cU6/8OqzejRo/nb3/4GJP/Rzp49m1133ZW5c+dyzjnncOSRR3LYYYcB0KdPH0466SQGDx7M4MGDM1/jiSeeYOrUqey3334ArFixgs6dOwPJOMuQIUNqPW7hwoV06tSpenv8+PHVrZphw4Zx2mmn8fOf/7y6fNCgQZSVldGnTx+uuOKKjc534403Zo45q6uuuoovfvGLrF69mhEjRnDNNddskMxrU1ZWRqtWrVi2bBnt27dv8JjMsmh2ieCzYu7cuZSVldG5c2datGixwcBx7vPhLVu2rH5UcKuttqruvtlqq602GESs2l9fvSFDhjBq1CgOOeQQ9t13X7bffnvatm3L22+/zdKlS+ttFTz99NP8/e9/5/nnn6ddu3YMHDiQlStXst122/Hyyy8zefJkxowZwz333MO4ceN45JFHePbZZ5k4cSK//vWveeWVVzL9XCKC73//+1x11VUblbVp04aystq77tq2bbvBz+2uu+7i3XffrW5NvPPOO8yePZuePXsCyRhBx44d64xj5MiRPPXUUxvtHzZsWPU4R5Udd9yR+fPnV29XVlay4447bnTsDjvsACR/VsOHD99gYLg+q1atok2bNpnqmhWCu4YKYNGiRZx55pmcffbZSKJbt25UVFSwfv165s+fz4svvliQ67Zp04ZvfetbnHXWWdVdOO3ateP000/n3HPPZfXq1dXx3XvvvRscu2TJErbbbjvatWvHa6+9xgsvvAAkT+asX7+eIUOGcMUVV/DSSy9V38egQYO45pprWLJkCcuXL88U46GHHsp9991X3WX2wQcf8J///Cfvcb169WLOnDkAvPHGGyxfvpwFCxYwb9485s2bx4UXXljroHFdbrzxxupB3dyvmkkAkkHwu+++m1WrVvHWW28xe/Zs9t9//43qLVy4EEiS3YMPPsiee+6ZN47FixfTsWNHTyVhjaqgLQJJhwO/A8qAP0TE1TXKTwWuBar6L26KiD8UKp6Va9ZVdxE1tBUrVrD7nn1Ys3YNLcpaMPi4YZx81jm8uWg5X9xlLzrtsBM9d92NHj13pVfvvVjw0QreXLSciE+7rT74eDWrWFW9XVX23rKVLFmxpnr/2vXBW+8vZ0m02ajs60d8h3vvf4Aee3+1et/wcy/gxqsup+euvWjdpjXt2n2Oc8+/iDcXLa8+V899D2Lp6JvZeZdd6d6jJ3vtux8LPlrB8ldm89/nnkWkLZrzLr6MN95dwslDT2DZsqVEBCed/kMWr2mxQSx1xfz5Tl055/yLGXjIN1i/fj0tWrbksquvZ2277Tf4WdTU98BB3HXHbQw6dhg3/+EOBn7ryA3q7j/o25w74vucdNbPNvj5NIQ2nb/MN44cTM9de9GiRRkXX3kd8z5YAcDpJwzhyhtv4gtf3IGTjx/GB4vfJyLotUcffnXtb3lz0XIW/e//Mviwg1m+bBlbbbUV191wI4/9cwrt22/DoxMf5cBB36z1vhctW8VltzzfIPdgzcOmPriSlSKiwU8KIKkMeAP4JlAJTAFOiIiZOXVOBfpFxNlZz9uvX78oLy/fYN+sWbPo1atXvcctXr6Kj1asyRx/U/WHm3/HsmVLGXnBJY0dSoMbetQ3ufUv97JNh20bO5QG86NTT+QXl4yie4+eG5UteGsON09b0QhR2WfVliQCSVMjol9tZYVsEewPzImIuWkQdwPHAjPrPapAtt+6Ndtv3Tp/xSbsO9/5Dm+++SZPPvkkHTt+tgbFG8LNo39L2Scf0GPnjR8TbYpWr17NSUO/xzcO2Lv28vdbM/6HfYsblJWkQiaCHYH5OduVQP9a6g2RdDBJ62FkRMyvWUHSCGAEQNeuXQsQavNQ9cRPc9W/f21/fZquVq1accoppzR2GGaNPlg8EegWEX2Ax4E7aqsUEWMjol9E9Mt9hLBGncJFaVZk/vtsxVTIRLAA2ClnuwufDgoDEBGLI6LqrZs/APuyGdq0acPixYv9j8eahar1CPxIqRVLIbuGpgA9JXUnSQDDgBNzK0jaISIWppvHAJu1fFiXLl2orKxk0aJFWxKv2WdG1QplZsVQsEQQEWslnQ1MJnl8dFxEzJB0OVAeEROAn0g6BlgLfACcujnXatmypVdyMjPbTAV7fLRQant81MzM6lff46ONPVhsZmaNzInAzKzENbmuIUmLgPyT09SuI9Awy3E1Hb7n0uB7Lg1bcs9fjohan79vcolgS0gqr6uPrLnyPZcG33NpKNQ9u2vIzKzEORGYmZW4UksEYxs7gEbgey4NvufSUJB7LqkxAjMz21iptQjMzKwGJwIzsxLXLBOBpMMlvS5pjqSNFqGV1FrS+LT835K6NUKYDSrDPf9M0kxJ0yU9IenLjRFnQ8p3zzn1hkgKSU3+UcMs9yzp+PTPeoakvxY7xoaW4e92V0lPSZqW/v0+ojHibCiSxkl6T9KrdZRL0uj05zFd0j5bfNGIaFZfJBPcvQl8BWgFvAzsXqPOj4Ax6edhwPjGjrsI9zwIaJd+PqsU7jmt1x54FniBZFnURo+9wH/OPYFpwHbpdufGjrsI9zwWOCv9vDswr7Hj3sJ7PhjYB3i1jvIjgEcBAQcA/97SazbHFkH1EpkRsRqoWiIz17F8ugjOfcChklTEGBta3nuOiKci4pN08wWS9SGasix/zgC/Aq4BVhYzuALJcs9nADdHxIcAEfFekWNsaFnuOYBt0s8dgHeKGF+Di4hnSWZjrsuxwJ2ReAHYVtIOW3LN5pgIalsic8e66kTEWmAJsH1RoiuMLPec63SS3yiasrz3nDaZd4qIR4oZWAFl+XPeBdhF0nOSXpB0eNGiK4ws93wZcLKkSmAScE5xQms0m/rvPa9CLkxjn0GSTgb6AV9v7FgKSdJWwA1s5hoXTVgLku6hgSStvmcl9Y6IjxozqAI7Abg9Iq6XNAD4k6Q9I2J9YwfWVDTHFkHeJTJz60hqQdKcXFyU6Aojyz0j6RvARcAx8ekSoU1VvntuD+wJPC1pHklf6oQmPmCc5c+5EpgQEWsi4i3gDZLE0FRluefTgXsAIuJ5oA3J5GzNVaZ/75uiOSaC6iUyJbUiGQyeUKPOBOD76efvAU9GOgrTROW9Z0l7A7eQJIGm3m8Mee45IpZERMeI6BYR3UjGRY6JiKa8qlGWv9sPkrQGkNSRpKtobhFjbGhZ7vlt4FAASb1IEkFzXrd2AnBK+vTQAcCS+HTJ383S7LqGItsSmX8kaT7OIRmUGdZ4EW+5jPd8LbA1cG86Lv52RBzTaEFvoYz33KxkvOfJwGGSZgLrgF9ERJNt7Wa8558Dt0oaSTJwfGpT/sVO0l0kybxjOu7xS6AlQESMIRkHOQKYA3wCDN/iazbhn5eZmTWA5tg1ZGZmm8CJwMysxDkRmJmVOCcCM7MS50RgZlbinAis4CStk1Qh6VVJEyVt28Dnn5c+M4+k5XXUaSvpGUllkrpJWpHGNFPSmPRN5E25Zj9Jo9PPAyV9NafsTEmnbMk9pee5TNJ5eercLul7m3DObnXNalmj3q8lza/585R0tqTTsl7PmgYnAiuGFRHRNyL2JHlv48eNEMNpwAMRsS7dfjMi+gJ9SGasHLwpJ4uI8oj4Sbo5EPhqTtmYiLhzSwNuZBNJJnyraRzNfy6fkuNEYMX2POkEWZJ6SHpM0lRJ/5C0W7r/C5L+Junl9Our6f4H07ozJI3YxOueBDxUc2c66eC/gJ3T35af1KdrNnRNr3tc2pp5WdKz6b6Bkh5WspbFmcDItIXxtarf5CXtJunFqmul538l/bxv2kKZKmmy8sweKekMSVPSGO6X1C6n+BuSyiW9IemotH6ZpGvTY6ZL+uGm/LAi4oXa3lZNZ7CdJ6m2JGFNlBOBFY2kMpKpAKre+h0LnBMR+wLnAb9P948GnomIvUjmZZ+R7j8trdsP+ImkTDPGplMTfCUi5tVS1i6N6RXg/wJ3REQf4C9pHACXAt9K49ngbez0nGOAG9NWzz9yyl4DWknqnu4aCoyX1DK91vfS+xkH/DrPbTwQEfulMcwimV+nSjeS396PBMZIapOWL4mI/YD9gDNy4qi69y9JmpTnurUpB762GcfZZ1Szm2LCPpPaSqogaQnMAh6XtDVJd0rVlBcArdPvhwCnAKRdOUvS/T+R9J30804kk6llmT6hI/BRjX090pgCeCgiHpX0J+C7afmfgN+kn58Dbpd0D/BAhuvluockAVydfh8K7EoyId7j6b2XAfnmitlT0hXAtiRThUzOvUY60+ZsSXOB3YDDgD454wcdSH5eb1QdFBHvkExVsKneS69hzYQTgRXDiojom/72PZlkjOB24KO0nz4vSQOBbwADIuITSU+TTC6W6fq11H0z67Uj4kxJ/Ul+454qad+M1wUYT5LsHkhOFbMl9QZmRMSATTjP7cDgiHhZ0qmkE8tVhVgzZJLVq86JiNyEgRpmWdY2JD9TaybcNWRFk/Yv/4RkkrBPgLckHQfV67DulVZ9gmQ5zaq+7g4kv9F+mCaB3Uimlc563Q+BsrTLpD7/4tMJCE8C/pHG0CMi/h0Rl5LMarlTjeOWkUx7Xdu13ySZ/O0SkqQA8DrQScnc+UhqKWmPPLG1Bxam3Uon1Sg7TtJWknqQLOn4OknCPSutj6RdJH0uzzWy2gXI++SRNR1OBFZUETENmE6ymMhJwOmSXiYZB6hagvBcYFA6sDqV5Kmex4AWkmaRdLO8sImX/h/goDx1zgGGS5oO/FcaB8C1kl5JH7v8F8m6ubkmAt+pGiyu5bzjgZP5dM781STTn1+T3nsFOU8d1eES4N8k3VSv1Sh7G3iRZNW5MyNiJfAHYCbwUhr3LdToAahvjEDSb5TMfNlOUqWky3KKDwQezxOvNSGefdRKgpJlK0dGxH81dixNmZJ1LX7mn2Pz4haBlYSIeAl4Kn1yyTZfR5LWiTUjbhGYmZU4twjMzEqcE4GZWYlzIjAzK3FOBGZmJc6JwMysxP1/26ma3sY7v5YAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "metrics.plot_precision_recall_curve(dummy_clf, X_test, y_test) # рандом победил!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 662,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC-PR:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.62612"
      ]
     },
     "execution_count": 662,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision, recall, thresholds = precision_recall_curve(y_test, dummy_clf.predict(X_test))\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "print('AUC-PR:')\n",
    "np.round(auc_precision_recall, 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "А теперь проделайте все то же самое для логистической регрессии (LR) — обучение, подбор параметра регуляризации (используйте L2-регуляризацию), вычисление всех метрик и построение кривых. Сравните результаты LR и SVM с точки зрения всех вычисленных критериев качества, объясните различия (если они есть)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 657,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 10/10 [00:00<00:00, 34.46it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from sklearn import datasets, metrics, model_selection, svm\n",
    "import numpy as np\n",
    "from sklearn import metrics\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.metrics import precision_recall_curve\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import auc\n",
    "# https://sinyi-chou.github.io/python-sklearn-precision-recall/\n",
    "\n",
    "c_best_lr = find_best_c(X_train, y_train, X_valid, y_valid, LogisticRegression)\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000, C=c_best_lr).fit(X_train, y_train)\n",
    "metrics.plot_roc_curve(lr, X_test, y_test) \n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 663,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC_PR:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8343"
      ]
     },
     "execution_count": 663,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision, recall, thresholds = precision_recall_curve(y_test, lr.predict(X_test))\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "print('AUC_PR:')\n",
    "np.round(auc_precision_recall, 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 664,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC_ROC:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7753951507549193"
      ]
     },
     "execution_count": 664,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('AUC_ROC:')\n",
    "metrics.roc_auc_score(y_test, lr.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 665,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.precision_recall_curve.PrecisionRecallDisplay at 0x7ff3a24ee4f0>"
      ]
     },
     "execution_count": 665,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "metrics.plot_precision_recall_curve(lr, X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# результаты очень похожи"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В названии метода опорных векторов присутствуют некоторые \"опорные векторы\". Сгенерируйте синтетические данные с помощью `make_classification` с 2 признаками, обучите на нём метод опорных векторов. Визуализируйте разделяющую прямую, все объекты и выделите опорные векторы (атрибут `support_vectors_`). В этот раз, если необходимо, вместо `LinearSVC` воспользуйтесь `SVC(kernel='linear')`, так как только в нём есть информация об опорных векторах."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import make_classification\n",
    "\n",
    "X, y = make_classification(n_samples=10000,\n",
    "                              n_features=2,\n",
    "                              n_informative=2, \n",
    "                              n_redundant=0,\n",
    "                              n_repeated=0, \n",
    "                              n_classes=2, \n",
    "                              random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_, X_test_, y_train_, y_test_ = train_test_split(X, y, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 10/10 [00:00<00:00, 13.05it/s]\n"
     ]
    }
   ],
   "source": [
    "# https://sinyi-chou.github.io/python-sklearn-precision-recall/\n",
    "c_best = find_best_c(X_train_, y_train_, X_test_, y_test_, LinearSVC)\n",
    "\n",
    "svc = SVC(kernel='linear', max_iter=100000, C=c_best).fit(X_train_, y_train_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# https://stackoverflow.com/questions/51495819/how-to-plot-svm-decision-boundary-in-sklearn-python\n",
    "# https://scikit-learn.org/stable/auto_examples/svm/plot_linearsvc_support_vectors.html\n",
    "\n",
    "def make_meshgrid(x, y, h=.02):\n",
    "    x_min, x_max = x.min() - 1, x.max() + 1\n",
    "    y_min, y_max = y.min() - 1, y.max() + 1\n",
    "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
    "    return xx, yy\n",
    "\n",
    "def plot_contours(ax, clf, xx, yy, **params):\n",
    "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
    "    Z = Z.reshape(xx.shape)\n",
    "    out = ax.contourf(xx, yy, Z, **params)\n",
    "    return out\n",
    "\n",
    "svc = SVC(kernel='linear', max_iter=100000, C=c_best).fit(X_train_, y_train_)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "# title for the plots\n",
    "title = ('Decision surface of linear SVC ')\n",
    "# Set-up grid for plotting.\n",
    "X0, X1 = X[:, 0], X[:, 1]\n",
    "xx, yy = make_meshgrid(X0, X1)\n",
    "\n",
    "plot_contours(ax, svc, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
    "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
    "\n",
    "ax.set_ylabel('y label here')\n",
    "ax.set_xlabel('x label here')\n",
    "ax.set_xticks(())\n",
    "ax.set_yticks(())\n",
    "ax.set_title(title)\n",
    "ax.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plt.figure(figsize=(10, 5))\n",
    "# for i, C in enumerate([1, 100]):\n",
    "#     # \"hinge\" is the standard SVM loss\n",
    "#     clf = LinearSVC(C=C, loss=\"hinge\", random_state=42).fit(X, y)\n",
    "#     # obtain the support vectors through the decision function\n",
    "#     decision_function = clf.decision_function(X)\n",
    "#     # we can also calculate the decision function manually\n",
    "#     # decision_function = np.dot(X, clf.coef_[0]) + clf.intercept_[0]\n",
    "#     # The support vectors are the samples that lie within the margin\n",
    "#     # boundaries, whose size is conventionally constrained to 1\n",
    "#     support_vector_indices = np.where(np.abs(decision_function) <= 1 + 1e-15)[0]\n",
    "#     support_vectors = X[support_vector_indices]\n",
    "\n",
    "#     plt.subplot(1, 2, i + 1)\n",
    "#     plt.scatter(X[:, 0], X[:, 1], c=y, s=30, cmap=plt.cm.Paired)\n",
    "#     ax = plt.gca()\n",
    "#     xlim = ax.get_xlim()\n",
    "#     ylim = ax.get_ylim()\n",
    "#     xx, yy = np.meshgrid(\n",
    "#         np.linspace(xlim[0], xlim[1], 50), np.linspace(ylim[0], ylim[1], 50)\n",
    "#     )\n",
    "#     Z = clf.decision_function(np.c_[xx.ravel(), yy.ravel()])\n",
    "#     Z = Z.reshape(xx.shape)\n",
    "#     plt.contour(\n",
    "#         xx,\n",
    "#         yy,\n",
    "#         Z,\n",
    "#         colors=\"k\",\n",
    "#         levels=[-1, 0, 1],\n",
    "#         alpha=0.5,\n",
    "#         linestyles=[\"--\", \"-\", \"--\"],\n",
    "#     )\n",
    "#     plt.scatter(\n",
    "#         support_vectors[:, 0],\n",
    "#         support_vectors[:, 1],\n",
    "#         s=100,\n",
    "#         linewidth=1,\n",
    "#         facecolors=\"none\",\n",
    "#         edgecolors=\"k\",\n",
    "#     )\n",
    "#     plt.title(\"C=\" + str(C))\n",
    "# plt.tight_layout()\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 10/10 [00:00<00:00, 22.14it/s]\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from sklearn import datasets, metrics, model_selection, svm\n",
    "\n",
    "c_best = find_best_c(X_train_, y_train_, X_test_, y_test_, LogisticRegression)\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000, C=c_best).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В отличие от метода опорных векторов, логистическая регрессия не пытается построить разделяющую гиперплоскость с максимальным отступом, а приближает в каждой точке пространства объектов вероятность положительных ответов $p(y=+1|x)$. Попробуйте нарисовать это распределение на плоскости, не забудьте отметить на ней все объекты. Вам может помочь функция `plt.pcolor`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# https://scipython.com/blog/plotting-the-decision-boundary-of-a-logistic-regression-model/\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000, C=c_best)\n",
    "lr.fit(X_train_, y_train_)\n",
    "\n",
    "# Retrieve the model parameters.\n",
    "b = lr.intercept_[0]\n",
    "w1, w2 = lr.coef_.T\n",
    "# Calculate the intercept and gradient of the decision boundary.\n",
    "c = -b/w2\n",
    "m = -w1/w2\n",
    "\n",
    "# Plot the data and the classification with the decision boundary.\n",
    "xmin, xmax = -1, 2\n",
    "ymin, ymax = -1, 2.5\n",
    "xd = np.array([xmin, xmax])\n",
    "yd = m*xd + c\n",
    "plt.plot(xd, yd, 'k', lw=1, ls='--')\n",
    "plt.fill_between(xd, yd, ymin, color='tab:blue', alpha=0.2)\n",
    "plt.fill_between(xd, yd, ymax, color='tab:orange', alpha=0.2)\n",
    "\n",
    "plt.scatter(*X_train_[y_train_==0].T, s=8, alpha=0.5)\n",
    "plt.scatter(*X_train_[y_train_==1].T, s=8, alpha=0.5)\n",
    "plt.xlim(xmin, xmax)\n",
    "plt.ylim(ymin, ymax)\n",
    "plt.ylabel(r'$x_2$')\n",
    "plt.xlabel(r'$x_1$')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Задание 2.__ Калибровка вероятностей.\n",
    "\n",
    "__(1 балл)__\n",
    "\n",
    "Перейдём к оценке качества выдаваемых алгоритмами вероятностей. Начнём с калибровочных кривых. \n",
    "\n",
    "Допустим, алгоритм возвращает некоторые числа от нуля до единицы. Хорошо ли они оценивают вероятность? Для этого разобьем отрезок $[0, 1]$ на несколько маленьких отрезков одинаковой длины. Рассмотрим $i$-й отрезок с границами $[a_i, b_i]$ и предсказания $p_1, p_2, \\dots, p_k$, которые попали в него. Пусть им соответствуют истинные ответы $y_1, y_2, \\dots, y_k$. Если алгоритм выдает корректные вероятности, то среди этих истинных ответов должно быть примерно $(a_i + b_i) / 2$ единиц. Иными словами, если нарисовать кривую, у которой по оси X отложены центры отрезков, а по оси Y — доли единичных ответов этих в отрезках, то она должна оказаться диагональной. Ниже приведена функция, которая должна рисовать такие кривые. В ней допущено две ошибки — найдите и исправьте их."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 667,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_calibration_curve(y_test, preds):\n",
    "    bin_middle_points = []\n",
    "    bin_real_ratios = []\n",
    "    n_bins = 10\n",
    "    for i in range(n_bins):\n",
    "        l = 1.0 / n_bins * i\n",
    "        r = 1.0 / n_bins * (i + 1)\n",
    "        bin_middle_points.append((l + r) / 2) # поменяла - на + (центры отрезков)\n",
    "        bin_real_ratios.append(np.sum(y_test[(preds >= l) & (preds < r)] == 1) \n",
    "                               / len(y_test[(preds >= l) & (preds < r)])) # доли единичных ответов\n",
    "    plt.figure(figsize=(6,6))\n",
    "    plt.plot(bin_middle_points, bin_real_ratios)\n",
    "    plt.ylim([-0.05, 1.05])\n",
    "    plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Сгенерируйте синтетические данные аналогично использованным в самом первом задании. Постройте калибровочные кривые на тестовой части для логистической регрессии и метода опорных векторов (не забудьте перевести его предсказания в $[0;1]$). Изучите распределение ответов классификаторов (постройте гистограммы с помощью `plt.hist`). Чем они различаются? Чем вы можете объяснить это?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 668,
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = make_classification(\n",
    "    n_samples=10000, n_features=10, n_informative=5, n_redundant=5,\n",
    "    random_state=42)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 669,
   "metadata": {},
   "outputs": [],
   "source": [
    "svc = SVC(kernel='linear', max_iter=100000, C=c_best_svc).fit(X_train, y_train)\n",
    "svc_pred = svc.decision_function(X_test)\n",
    "\n",
    "# переведем предсказание как в семинаре\n",
    "svc_pred = (svc_pred - svc_pred.min()) / (svc_pred.max() - svc_pred.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 670,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_calibration_curve(y_test, svc_pred) # нуууу, почти диагональное"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 671,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='Count'>"
      ]
     },
     "execution_count": 671,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.histplot(svc_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "lr = LogisticRegression(max_iter=100000, C=c_best).fit(X_train, y_train)\n",
    "lr_pred = lr.predict_proba(X_test)[:, 1]\n",
    "plot_calibration_curve(y_test, lr_pred) # оооОООоО!!!!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='Count'>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.histplot(lr_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# мы видим, что у lr вероятсноти распределены примерно одинаково\n",
    "# у svc же они распределены нормально\n",
    "# вероятно, это происходит из-за распределяющей полосы svm, которая влияет на уверенности модели"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Воспользуйтесь `CalibratedClassifierCV` из `sklearn` для калибровки вероятностей метода опорных векторов на обучении и постройте с его помощью предсказания для тестовой выборки. Нарисуйте для них калибровочную кривую. Улучшилась ли она?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 672,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbkAAAG5CAYAAAATVEooAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAsTAAALEwEAmpwYAAB2QElEQVR4nO3ddXjV5fvA8fezMWKMLklpgY0eKd0N0h2iYvJTEANRsfUrKKkIKN0lMLpBuru7cwzGWD+/P56DDlicbSe2s/t1XefaOZ+8z0fZvaeV1hohhBDCFbk5OwAhhBDCXiTJCSGEcFmS5IQQQrgsSXJCCCFcliQ5IYQQLkuSnBBCCJclSU4IKyiltFKqqOX9OKXU55b3dZRSV+x8725KqdX2vIcQrkqSnEgxlFJdlVJ7lFKBSqnrSqkVSqka8b2O1vpNrfU3doqxoCWhpopyvxla60b2uJ8Qrk6SnEgRlFIDgBHA90AuoADwG9DawXG4O/J+jhA1IQuR1EiSEy5PKZUJ+Bp4R2u9UGv9SGsdprVeqrUeZDmmslJqu1LqvqWUN0YplTqG601WSn37zLbBSqk7SqkLSqluzxz7u1JquVLqEVBXKdVcKbVfKfVAKXVZKTU0yqU2W37et5Q4qymleiul/olyzepKqd1KqQDLz+pR9m1USn2jlNqqlHqolFqtlMoey7NprZQ6YInlrFKqiWX7BaVUgyjHDVVKTbe8f1La7KuUugSst5SK333m2geVUm0t70sopdYope4ppU4qpTpGOa6ZUuqYJd6rSqkPY4pXiPiSJCdSgmpAWmBRLMdEAB8A2S3H1wfetvL6L1jOywv0AsYrpV6Ksr8r8B2QAfgHeAT0BDIDzYG3lFJtLMfWsvzMrLX20lpvj3ojpVRWYBkwCsgG/AIsU0ple+Z+fYCcQGog2qShlKoMTAUGWWKpBVyw8jsD1AZKAo2BWUCXKNcuBbxoiS09sAaYaYmpM/Cb5RiAP4F+WusMgA+wPh4xCBErSXIiJcgG3NFah8d0gNZ6r9Z6h9Y6XGt9AfgD80vcWp9rrUO01pswSahjlH2LtdZbtdaRWutgrfVGrfVhy+dDmARh7b2aA6e11tMssc4CTgAtoxwzSWt9Smv9GJgLlIvhWn2Bv7TWayyxXNVan4jHdx5qKRU/xvwBUU4p9aJlXzdgodY6BGgBXNBaT7LEvB9YAHSwHBsGlFJKZdRa+2ut98UjBiFiJUlOpAR3geyxtR0ppYorpfyUUjeUUg8wbXcxVvM9w19r/SjK54tAniifLz9zrypKqQ1KqdtKqQDgzXjcK4/l+lFdxJQin7gR5X0Q4BXDtfIDZ628b3T+/V5a64eY5N7ZsqkLMMPy/kWgiqUq+L5S6j4mCb5g2d8OaAZcVEptUkpVS0RMQjxFkpxICbYDIUCbWI75HVMiKqa1zggMBpSV189iqZJ7ogBwLcrnZ5f6mAksAfJrrTMB46LcK65lQa5hkkZUBYCrVsYa1WWgSAz7HgGeUT6/EM0xz8Y6C+hiSVJpgQ1R7rNJa505ystLa/0WgNZ6t9a6NaYq829M6VMIm5AkJ1ye1joA+AIYq5Rqo5TyVEp5KKWaKqX+ZzksA/AACFRKlQDeiudtvlJKpVZK1cRUz82L5dgMwD2tdbClXaxrlH23gUigcAznLgeKW4ZDpFJKdQJKAX7xjBdMW1gfpVR9pZSbUiqv5bsDHAA6W56TL9DeiustxyTgr4E5WutIy3Y/S8w9LNfzUEpVUkqVtDyzbkqpTFrrMMx/g8gYri9EvEmSEymC1no4MAAYgkkkl4F3MSUHMJ0zugIPgQnAnHhc/gbgjyllzQDejKNt623ga6XUQ0zy/bfkorUOwnRS2Wqp2qv6zPe4i0miAzHVsB8BLbTWd+IR75Nr7cJ0UPkVCAA28V8p8XNMKc8f+ApT+ozreiHAQqBB1OMtVZmNMFWZ1zDP6ycgjeWQHsAFSzXxm5iqTCFsQsmiqUIIIVyVlOSEEEK4LElyQgghXJYkOSGEEC5LkpwQQgiXlewmVs2ePbsuWLCgs8MQQgiRhOzdu/eO1jrHs9uTXZIrWLAge/bscXYYQgghkhCl1LMzAQFSXSmEEMKFSZITQgjhsiTJCSGEcFnJrk0uOmFhYVy5coXg4GBnh+IS0qZNS758+fDw8HB2KEIIkSgukeSuXLlChgwZKFiwIEpZO3G8iI7Wmrt373LlyhUKFSrk7HCEECJRXKK6Mjg4mGzZskmCswGlFNmyZZNSsRDCJbhEkgMkwdmQPEshhKtwmSQnhBBCPEuSnBBCCJclSc6GvvvuO7y9vSlTpgzlypXjq6++4tNPP33qmAMHDlCyZEkAAgMD6devH0WKFKFixYrUqVOHnTt3xuuer732GseOHbPZdxBCCFfiEr0rk4Lt27fj5+fHvn37SJMmDXfu3OHYsWP07t2bH3744d/jZs+eTZcuXQCToAoVKsTp06dxc3Pj/Pnz8U5YEydOtOn3EEIIV+J6Se799+HAAdtes1w5GDEi1kOuX79O9uzZSZMmDQDZs2enVq1aZMmShZ07d1KlShUA5s6dy6pVqzh79iw7d+5kxowZuLmZAnWhQoVi7Lb/6NEjOnbsyJUrV4iIiODzzz+nU6dO1KlTh2HDhuHr68uff/7JTz/9RObMmSlbtixp0qRhzJgx9O7dm3Tp0rF//35u3brFX3/9xdSpU9m+fTtVqlRh8uTJtnpSQgiRpNitulIp9ZdS6pZS6kgM+5VSapRS6oxS6pBSqoK9YnGERo0acfnyZYoXL87bb7/Npk2bAOjSpQuzZ88GYMeOHWTNmpVixYpx9OhRypUrh7u7u1XXX7lyJXny5OHgwYMcOXKEJk2aPLX/2rVrfPPNN+zYsYOtW7dy4sSJp/b7+/uzfft2fv31V1q1asUHH3zA0aNHOXz4MAds/UeBEEIkEfYsyU0GxgBTY9jfFChmeVUBfrf8TJw4Slz24uXlxd69e9myZQsbNmygU6dO/Pjjj3Tq1Inq1aszfPjwp6oq46t06dIMHDiQjz/+mBYtWlCzZs2n9u/atYvatWuTNWtWADp06MCpU6f+3d+yZUuUUpQuXZpcuXJRunRpALy9vblw4QLlypVL2BcXQogkzG4lOa31ZuBeLIe0BqZqYweQWSmV217xOIK7uzt16tThq6++YsyYMSxYsID8+fNTqFAhNm3axIIFC+jUqRNgksvBgweJiIiw6trFixdn3759lC5dmiFDhvD111/HK7Yn1ahubm7/vn/yOTw8PF7XEkKI5MKZvSvzApejfL5i2ZYsnTx5ktOnT//7+cCBA7z44ouAqbL84IMPKFy4MPny5QOgSJEi+Pr68uWXX6K1BuDChQssW7Ys2utfu3YNT09PunfvzqBBg9i3b99T+ytVqsSmTZvw9/cnPDycBQsW2ONrCiGETRy5dYSWs1oSEBxg1/ski44nSqk3gDcAChQo4ORoohcYGMh7773H/fv3SZUqFUWLFmX8+PGAqTrs378/o0ePfuqciRMnMnDgQIoWLUq6dOnInj07P//8c7TXP3z4MIMGDcLNzQ0PDw9+//33p/bnzZuXwYMHU7lyZbJmzUqJEiXIlCmTfb6sEEIkwoEbB2gwtQFpUqXhdtBtMqW13+8q9aQUYZeLK1UQ8NNa+0Sz7w9go9Z6luXzSaCO1vp6bNf09fXVz64Mfvz48X/HnqVkgYGBeHl5ER4eziuvvMKrr77KK6+8kqBryTMVQtjL9svbeX3p6/zd+W+KZi1qk2sqpfZqrX2f3e7M6solQE9LL8uqQEBcCU7EbujQoZQrVw4fHx8KFSpEmzZtnB2SEEL860bgDQCq5a/GwTcP2izBxcZu1ZVKqVlAHSC7UuoK8CXgAaC1HgcsB5oBZ4AgoI+9YklO7t69S/369Z/bvm7dOrJlyxbrucOGDbNXWEIIkShbLm6h2cxmjGoyij7l++DuZt3wqcSyW5LTWsfaV16betJ37HX/5Cpbtmwybk0I4VLWn19Py1ktKZCpAI2LNnbovWXuSiGEEHaz6swqms9sTqHMhdjYayN5MuRx6P0lyQkhhLCLKw+u0Hp2a17K9hIbem0gl1cuh8eQLIYQCCGESH7yZczH5DaTaVSkEVnTZXVKDFKSsyFnLLUjhBBJzbyj89h0wczf29mns9MSHEhJzmactdSOEEIkJdMPTafX371oXKQxtQvWdnY4kuRsxd5L7QghRFL31/6/eG3Ja9QpWId5HeY5OxzABZOck5aTo1GjRnz99dcUL16cBg0a0KlTJ2rXrv3vUjtVqlR5aqmdJUuWxGupHSGESMrG7RnHW8veolGRRizqtAhPD09nhwRIm5zNPFlqZ/z48eTIkYNOnToxefJkOnXqxPz584mMjEzUUjtCCJFUaa3ZenkrzYs1Z3HnxUkmwYELluSctJwc8N9SO3Xq1KF06dJMmTKF3r17P7XUzvbt24Gnl9qR0pwQIrkKDA3EK7UXk1pPIlJHkto9tbNDeoqU5GzE3kvtCCFEUvPd5u+o8EcFbj+6TSq3VEkuwYEkOZsJDAykV69elCpVijJlynDs2DGGDh0KmKV2jh49+lxV5cSJE7l58yZFixbFx8eH3r17kzNnTidEL4QQ1tNa8+WGLxmyYQhV8lUhS7oszg4pRi5XXeksFStWZNu2bdHuy549O2FhYc9tz5gxIxMmTLB3aEIIYTNaaz5d9yk/bf2JPuX6MKHlBIdNtpwQUpITQghhtRE7RvDT1p94s+KbTGw1MUknOJCSnBBCiHjoUbYHYZFhDKo+CKWUs8OJk5TkhBBCxCpSR/Lb7t8IjQglu2d2Pnr5o2SR4ECSnBBCiFhEREbw6uJXeWf5O8w/Nt/Z4cSbVFcKIYSIVnhkOD0X9WTWkVl8Xedrupbu6uyQ4k2SnBBCiOeERYTRdWFX5h+bz4/1f+TjGh87O6QEkepKG/Hy8npu29ChQ8mbNy/lypWjVKlSzJo1ywmRCSFE/J31P8vac2v5pdEvyTbBgZTk7O6DDz7gww8/5PTp01SsWJH27dvj4eHh7LCEECJa4ZHhpHJLRYnsJTj17ilypM8R90laQxLtiCIlOQcpVqwYnp6e+Pv7OzsUIYSIVlBYEE1nNOXHf34EsC7BBd+BjU3h1mY7R5cwrleS2/s++B+w7TWzlIOKIxJ1iX379lGsWDGZtksIkSQFhgbSYmYLtlzaQo8yPaw76f5h2NQaHl+Dwn3sG2ACuV6SS2J+/fVXJk2axKlTp1i6dKmzwxFCJDEXL8K4cRASAsOGgZsT6tcehDyg6Yym7Lyyk+mvTKdLaSuWBLu8CLb3AI+M0GAzZK9s/0ATwPWSXCJLXLb2pE1uyZIl9O3bl7Nnz5I2bVpnhyWEcCKtYcMGGDMGFi82n7UGT0/49lvHxhIeGU7j6Y3Zc20Pc9rPoV2pdrGfoCPhyDdweChkqwI1F4JnHofEmhDSJucgrVq1wtfXlylTpjg7FCGEkwQGmlJb6dJQvz5s3gwffQQXLkDfvvDddzDfweOtU7mlom/5vizouCDuBBcWCP90MAmuUC9osDFJJzhwxZKckwQFBf27VhzAgAEDnjvmiy++oGvXrrz++uu4OaNOQgjhFGfOwNixMGkSBARAhQrmfadOkC6dOWbsWDh2DHr1guLFoUwZ+8Z069EtTtw5Qa0Xa/FahdfiPiHwPGxuAwFHoMIv8NL7SbZHZVSS5GwkMjIyzmMqVqzIyZMnHRCNEMLZIiNh1SoYPRpWrIBUqaBDB3j3XahW7fn8kCYNLFgAvr7Qpg3s3g3ZstkntusPr1N/an1uPbrF+f87T4Y0GWI/4eYGU4KLjIA6KyB3I/sEZgdSnBBCCBsKCIARI+Cll6BZM9i/H778Ei5dgpkzoXr1mAtAuXPDwoVw9Sp07Ajh4baP78qDK9SeXJtLAZdY0HFB7AlOazg1FtY3hDQ5ofGuZJXgQJKcEELYxLFj8PbbkDcvfPAB5MhhktrFizB0qElg1qhSBf74A9avh0GDbBvjxfsXqT25NjcCb7Cq+ypqF6wd88ERobCrH+x5F3I3hcY7IGOxZ46JMI2MN2/aNlAbkupKIYRIoIgIWLrUVEmuX2+qHLt0MVWSFSsm/Lq9e5sS4IgRUL489Oxpm3jH7x3P3aC7rO25lsp5Y+nyH3wLtrSD2/9AqU+hzDfw7OKou3fDW2/B3r1w/z588oltgrQxSXJCCBFPd+/CxInw22+mGjJ/fvj+e3jtNVOCs4Vhw+DwYXjjDShRAionYhia1hqlFN/U+4ZXy79KkaxFYj743n7Y3BpC7kD1WVCw89P7/f3hs89MCS5XLpg1y/SgSaKkulIIIax04IDp6p8vnym4FC5sOoucOweffmq7BAfg4QFz55pqzrZt4caNhF3n+O3j1JhUg0sBl3BTbrEnuItzYM3LgIaG/zyd4LSGqVNNY+Mff0D//nDiBHTunKR7WUqSE0KIOGht2tXKl4fZs003/0OHzIDutm1Nz0l7yJ4d/v7bFJ7atTOzosTH4ZuHqT25NmfvnSUwNDDmA3UkHPwMtnaGrBWg8R7z84kjR6B2bfPFixY1VZQjRkCmTAn5Wg4lSc5G3N3dKVeuHD4+PnTo0IGgoKB4nT9o0CC8vb0ZlICW5u+//z7e5wghrBMaCn36wFdfmbaxK1f+G9DtCGXLmjF127bBe++ZhGuN/df3U3dKXTzcPdjUexOlcpSK/sCwB2b829HvochrUG89pMtl9gUGmt4v5cvD0aOmjvaff6BcOVt8NcfQWierV8WKFfWzjh079tw2R0ufPv2/77t27aqHDx9u1XlhYWFaa60zZsyow8PDE31vW0kKz1QIZ7t/X+v69c2kW0OHah0Z6bxYPv3UxPHbb3Efu//6fp35x8y6wK8F9Jm7Z2I+MOCU1ktLaj3TXeuTY/77gpGRWs+fr3W+fOamr72m9e3btvkidgLs0dHkDCnJ2UHNmjU5c+YMjx494tVXX6Vy5cqUL1+exYsXAzB58mRatWpFvXr1qF+/Pq1atSIwMJCKFSsyZ84cbt++Tbt27ahUqRKVKlVi69atAAQGBtKnTx9Kly5NmTJlWLBgAZ988gmPHz+mXLlydOvWzZlfWwiXcvky1KgBmzaZktSXXzq36embb8y4u/79zXRgsXkx04vUL1Sfzb03x9wGd30NrKoMIbeg3hoo/o75gmfOmBu1b29Go2/bBhMmmLrTZEhpa8u+SYSvr6/es2fPU9uOHz9OyZIl//1cZ3Kd587r6N2Rtyu9TVBYEM1mNHtuf+9yveldrjd3gu7Qfm77p/Zt7L0xzri8vLwIDAwkPDycdu3a0aRJEy5fvkypUqXo3r079+/fp3Llyuzfv5958+YxZMgQDh06RNasWZ86H6Br1668/fbb1KhRg0uXLtG4cWOOHz/Oxx9/TEhICCNGjADA39+fLFmyPHWurTz7TIVISfbvh+bN4dEj07GkQQNnR2QEBJhelv7+sGcPFCjw9P591/dRKkcp0qaKZRJ4reHkCNj/IWQsBbWXgFchCA6Gn36CH36A1KlNVn3nHfs1ONqYUmqv1tr32e3JI/pk4ElpCkxJrm/fvlSvXp0lS5YwbNgwAIKDg7l06RIADRs2/DfBPWvt2rUcO3bs388PHjwgMDCQtWvXMnv27H+3Z8mSxU7fRoiUa8UKM9tI5sym+clRbW/WyJTJrFpQpQq88gps2WJWLgBYd24dLWe1pG/5voxuNjr6C4Q9hN3vwIVpkO8VqDYVPLxg5UozuO/sWdNbcvhwyJO0J162lksmudhKXp4enrHuz+6Z3aqS27PSpUvHgQMHntqmtWbBggW89NJLT23fuXMn6dOnj/FakZGR7NixQ5bkEcLBJkww45tLl4Zly5Lm7/kSJWDGDGjVCl5/HaZPh1VnV/LKnFcomrUoQ2oNif7Eu7tha1d4dA5KfwU+Q+DqNfigj1n6oHhxWLMm6RRbbUTa5OyocePGjB49midVwvv377fqvEaNGjF69H9/iT1Jng0bNmTs2LH/bvf39wfAw8ODsLAwG0UtRMqjtRnf/MYb0LChafNKignuiRYtTG3izJnQ96eltJ7dmhLZS7Ch1wZyeeV6+mAdCcf+B6urQ2QI1N8AJT6F4b+YjOnnZxaxO3TI5RIcSJKzq88//5ywsDDKlCmDt7c3n3/+uVXnjRo1ij179lCmTBlKlSrFuHHjABgyZAj+/v74+PhQtmxZNmzYAMAbb7xBmTJlpOOJEAkQEgLdu5sZS15/HZYsgQxxTMqfFAweDG06BDHpTj8Kpi3L+p7rye75TOeQoGtmcuUDH0O+NtDsIJxUZq2fQYOgbl0z6eZnn5k5yVyQS3Y8EYknz1SkBP7+pm1r0yaT5D75JElP3vGcwECo0Pgot87kY8/WTBQtGmXnlSWw81UIfwy+oyBdU/j4Y1O/WaCAmXCzVSunxW5rMXU8kZKcECJFOn/eLHuzfbtp4/r00+ST4KYdnMZXG7/CywtWTffGPTwTrVvDw4eYpLb7HTP/pGcBaLADFgeYqsm5c02p7dgxl0pwsXHJjidCCBGbPXvMEIHQUFi92sxYlVz8ue9PXl/6OnUL1SUsIoxChTyYOxcaN4Yh7x5mRMcuqICjUGIg3G8ItTqbpNasGYwcydPFPdfnMiW55FbtmpTJsxSubOlSk9Q8Pc045+SU4H7b/RuvLX2NxkUb49fFDw93DwDq19OsGTuGn+pW4tHdO+AzDX66DPWbwOPHpqHRzy/FJThwkSSXNm1a7t69K7+cbUBrzd27d2X4gnBJY8dCmzZQqpSppkxOzc4jd4zkneXv0LJ4S/7u9DfpPNKZHcG3YVNL6mZ4j5P361H4vUP8XWmpSWxffWXmnGzZMvnUxdqYS1RX5suXjytXrnD79m1nh+IS0qZNS758+ZwdhhA2Exlp+lwMG2Z+38+aBbEMVU2Ssnlmo0OpDkxvO53U7qnNxutrYHtPCPUHz7d4aexmCj04T49Uk9mx9B7ejfI6N+gkwCV6VwohREyCg83qAfPmmVmqRo4Ed/e4z0sqTt89TbFsxYD/Fj8lIhQOfQbHh0G6YrA4H0zZAMWKcfWLP/AdVBcvL9i1C1LKxEjSu1IIkeLcuQP165sEN2yY6TWfXBKc1prP13+Oz+8+HLhxAMAkuAenYHU1k+ACfKH3JZi308w5efgwebvXZcECuHgRunSxfmkeVyVJTgjhks6eNUME9u41SW7gwOTTLKW15pO1n/Dtlm/pWaYnZXKVMdnq7F+wojz4n4aZueDtPdCktVmh+5NP/h3QXb26mRFl1SqzqEBK5hJtckIIEdWOHabtTWtYtw5eftmGF58+HTZ+Dd6pIJsvlGgKZVtCGi+bXF5rzYBVAxixcwRv+b7FmGZjcAsLgF394NI8uJENvr0LufPDuplQr16013my+ehRKFbMJqElS5LkhBAuZeFC6NYN8uY1KwrY9Bf8tWvwTT/4LAg0kOo4nJkGJ4B76UEXghyVoERzKNscPOLfS3n+sfmM2DmC96u8zy+Nf0Hd/ge2doOgqzDfHdaHwJfDzTLhHh4xXudJz9EjR0yP0pRKkpwQwiXcvm2WQ/vlF7MUzZIlkCOHjW8y4APo+hjSZoEWx+HMUTi0GG5th4jTkPUIuB2BU5PgKHAvA6jCkKMKeLcAn8aQKnWst2hXqh0LOi7gleItUYe/hCPfwV03GBEJ1bvDif9B7txxhurlBYUKmSSXkkmSE0Ika3fumE4lY8aYcc99+pj36dLZ+EZr1sDduVAM8B0BnrmgTC4oE6W6MDwcDq2Fo0vh9naIPAvZDoI6CMfGw37APyO4FYVcVcG7JXg3IFzBx2s+5u1Kb1MkaxHa5q8AflXg0X7YDOwtCZN/h5o14xWyj48kOUlyQohk6e5ds7bn6NFmBe/OneGLL8wUjTYXHAyfvAHvukHOulCoR/THpUoFFZqY1xNhoXBwJRz1g7s7gXOQZR/ofXDkN8J2Qc8bHsyOCKPgub28V7ohnP8aQkJhdjpo8z8Y82aCVuj28TFVtqGhZrHvlEiSnBAiWbl3z1RJjhplZuHv2NEkt1Kl7HjTn36E2hdM78Wq4+PXTdMjNfi2Mq8nQh7D/mWEHl9KlxuLWBjxkP9lgfcybYJLm+AscLsDzBwDOXMmOGwfH1O4PH0avL0TfJlkTYYQCCGShfv34csvTTvTd99BkyZmnc/Zs+2c4E6fhlXfgS9Q7hvwKpz4a6ZJR4hvS9qn9WdhxENGNB7BoDceQsEZENIfmm+HsXMTleDgv8SWkqsspSQnhEjSAgJgxAj49Vfzvm1bk+zKlHHAzbWGD96AHhHg5QMlPrDZpUMjQrkTdIffmv3GW5XeMhurdzUvG3npJTP4/cgR6NTJZpdNViTJCSGSpAcPzBRcv/xiSnFt2pjkVq6cA4OYOxdybYSMCmpMBrfE/8p8FPoIgAxpMrC5z2ZS2eCaMUmb1gyhkJKcEEIkEQ8fmva24cPNyt2tWsHQoVC+vIMDCQiA0W/D28BL70PWiom+5MOQh7SY1YI07mlY1X2VXRPcEz4+cOCA3W+TZEmbnBAiSQgMNNMvFiwIQ4aYWUr27IHFi52Q4AC+GAxt7kGqPFD2m0RfLiA4gMbTG7P10lb6lu9r5qF0AB8fM8XZ48cOuV2SY9ckp5RqopQ6qZQ6o5T6JJr9BZRSG5RS+5VSh5RSzewZjxAi6Xn0CP73P9OhZPBgqFrVzJ6/dClUTHzhKWH27YNrv0EeoOYkSJW4dXn8H/vTcFpDdl/bzdwOc+nk47gGMm9v07R4/LjDbpmk2C3JKaXcgbFAU6AU0EUp9WwfqCHAXK11eaAz8Ju94hFCJC2PHplB3IUKmbXeKlY0c04uWwaVKjkxsIgIGNwLWgJ5OkLuRom+ZLeF3Th48yALOy6kbcm2iY8xHnx8zM+U2i5nzwrhysAZrfU5AKXUbKA1cCzKMRrIaHmfCbhmx3iEEElAZKRpc/vhB7h1Cxo2NAtYV6vm7Mgsxo+DqkcglRdUHWOTS/7c8GeuPLhC46KNbXK9+Cha1AwET6lJzp7VlXmBy1E+X7Fsi2oo0F0pdQVYDrwX3YWUUm8opfYopfbI6t9CJG/TpsEHH5gSxj//wOrVSSjB3bwJKz6E4pgElzbhk19ef3id4duGo7XGO6e3UxIcmIlSSpaUJOcsXYDJWut8QDNgmlLquZi01uO11r5aa98cNp9xVQjhKFqbMW/e3rB2rY2XwLGFz96C1sGQoToU6pngy1x5cIXak2vz5cYvuXD/gu3iSyAfH7PkTkpkzyR3Fcgf5XM+y7ao+gJzAbTW24G0QHY7xiSEcKItW0x39v79k+ACpuvWQcZFkCYV1J2W4AAv3L9ArUm1uPnoJqt7rKZQlkI2DjT+vL3h0iUz9jClsWeS2w0UU0oVUkqlxnQsWfLMMZeA+gBKqZKYJCf1kUK4qBEjIGtW6N7d2ZE8IyQERvc0U3eV/irBU3edvXeW2pNr4x/sz9oea6mev7pt40ygJ51PUmJpzm5JTmsdDrwLrAKOY3pRHlVKfa2UejJT6UDgdaXUQWAW0Ftrre0VkxDCeS5cMGPe3ngDPD2dHc0zfvkGGl4D9yJQ+qMEX+bIrSOEhIewvud6KuV1ZhfRp6XkHpZ2HW6vtV6O6VASddsXUd4fA5JarbwQwg7GjDE1gG+/7exInnHuHJz/EWoDDeckaOqux2GPSeeRjtYlWlO/cH28UnvZPs5EePFFSJ8+ZSY5Z3c8EUKkAIGBMHEitGsH+fPHfbzDaA1fd4U6EZC/X4Km7jp08xBFRxfF75QfQJJLcABubqZdTqorhRDCDqZMMVNBvv++syN5xoLZ4LMTIrJC9eHxPn3f9X3UnVIXhaJ4tuJ2CNB2Uuoq4ZLkhBB29WTwd6VKZsquJOPhQ1jez0zdVXd6vKfu2nllJ/Wm1CNDarOaQFJPct7eZhhgShtqLElOCGFXq1bBqVPwf/+XxIYN/Pgu1HkIXk0gX9N4nXre/zwNpzUkm2c2NvXeROEsNlhI1c5Sag9LSXJCCLsaORJy54YOHZwdSRQH9kPqqaDTQqOp8T69YOaCDK45mM29N/Ni5hftEKDtpdQelpLkhBB2c/y4Kcm9/baZPzFJiIyE39pBMaDir/GaumvduXWcuHMCpRSf1PiEvBmfnakw6cqdG7JkkSQnhBA2M2oUpEkD/fo5O5Io/hoGVc5DpDd4Wx/Y8tPLaT6zOR+s+sCOwdmPUilzei9JckIIu/D3h6lToWtXSDJTzt66BaeHgIcbtF5sdSPh4hOLaTO7Dd45vZn+ynQ7B2k/3t6mJJeSptyQJCeEsIuJEyEoyHQ4STKGd4KyYVDgA8hQxKpT5h+bT/t57Smfuzzreq4jm2c2OwdpPz4+cP8+XEtBi5pJkhNC2Fx4OIweDXXqQNmyzo7GYuMyKLARHueCWj9adYrWmnF7xlE5b2XW9FhD5rSZ7RqivaXEzid2ndZLCJEy/f03XL5s2uSShNBQ8OsO5YB6C6yauisiMgJ3N3cWdVqEUipJzmQSX97e5ueRI9DYOcvbOZyU5IQQNjdyJBQqBC1bRrPz9GkzS/OcOWa+L0f47T2ocB/StoE8cU+XO2HvBOpNrUdgaCAZ0mRwiQQHkD07vPBCyup8IklOCGFTe/eaFb/few/c3aM54LPPYMIE6NzZ9Ehp2xZmzrTfYmfnT0LEBHjkCW3i7jQydtdY3vB7A6/UXqRKwGTNSd2TzicphSQ5IYRNjRwJXl7w6qvR7Dx3DhYsgA8/hE2b4PXXYedO6NYNcuaE1q1h2jTTO8JWJraC3BqqjI9z6q5ft//KuyvepfVLrVnYcSFpU6W1XRxJxJNhBJGRzo7EMSTJCSFs5sYNmD0beveGTJmiOWDECFO8e/99qFXLNNpdvmyKfm+9Bfv2Qc+eJuG1aAGTJ5uxCAm1aDS8dAqCKkCZbrEe+tvu3xiwegDtS7VnXod5pEmVJuH3TcJ8fEyv1wsXnB2JY0iSE0LYzLhxEBYG/ftHs/PePfjzTzNwLm+UmULc3ODll+HXX+HiRdi+3VzgyBHo08ckvKZNzbl371ofTOBDOPEhhLpD56VxHt6gcAPeq/wes9rNwsPdw/r7JDMprYelJDkhhE2EhMDvv0Pz5lCsWDQH/P67KUIMHBjzRdzczFIFw4bB+fOwaxcMGAAnT8Jrr0GuXNCoEYwfbwZ2x+a3tvBiKOT/FLzyRHuI1pqlJ5eitaZ4tuKMajrKJdvhoipVyvyUJCeEEPEwe7bJO9EO/g4ONgPnGjeG0qWtu6BSZn2en36Cs2dNj5aPPjLJr18/Mxlj/fomed648fS5e5ZCtrVwLw80+jray2ut+WjNR7Sa3Yr5x+bH78smYxkzQoECKaeHpdLJbH4XX19fvWfPHmeHIYSIQmuoWNGU5o4ciWa2rD//NCWxNWugQYPE3+zwYZg3z7xOnjQ3rFUL2reDqqFw9CMI1dBkN+R/frVvrTXvr3yfUbtG8U6ldxjVdBRuKuX8zd+8OVy5AgcPOjsS21FK7dVa+z673bXL5UIIh/jnH9i/H/74I5oEFxkJw4dDuXKm5JVYSkGZMub19ddw7JhJdn5z4FR/yAqcAor8HG2Ci9SRvLPsHcbtHccHVT9geKPhqCS10J39+fjA2rWm/dTDdZsfAUlyQggbGDkSsmaF7t2j2blihVlzZ/p026+aqpQZ+JXjLpT7C4JSwd26EF4Weg+I9pSDNw4ycf9EPnn5E76v/32KS3BgklxoKJw5AyVLOjsa+5IkJ4RIlAsXYNEiGDQIPD2jOWDYMMiXDzp2tP3NI8Pg8Ndw7HtIXxgab4NslWI9pXzu8hx88yAls5dMkQkOnu5h6epJLuVUQgsh7GLsWFOgeuedaHbu2QMbN5pxcbauF3t4FtbUhKPfQqHe0HR/jAkuLCKMbgu7MefIHABK5SiVYhMcQIkSpiNrSuh8IiU5IUSCPXpkltRp1w7y54/mgGHDTHe+11+33U21hgvTYffboNzh5TnwYsylxNCIUDrP78yiE4uo8EIF28WRjKVLB0WKpIxhBJLkhBAJNnWqmYEr2mEDFy6YDiEDBphEZwuhAbD7Lbg4C3LUhOrTIX2BGA8PDg+m/dz2LDu9jJFNRtK/SnSj1FMmH5+UkeSkulIIkSCRkabDia8vVKsWzQEjRpg6MVutmnp7G6woB5fmQplvof6GWBNcWEQYrWe3ZtnpZYxrPk4S3DN8fMyCEMHBzo7EvqQkJ4RIkNWrzRC1aDtN+vubeswuXUynk8SIDIcj38LRbyB9QWj4D2SvGudpqdxSUTF3RTp5d+LV8tHNFp2y+fiYP1ROnDCjO1yVJDkhRIKMHGkmHenQIZqdf/xhGuxim8LLGoEXYFs3uLMNCvaASmPAI/aqz4chD7n68Colspfg+/rfJ+7+LuxJD8ujR107yUl1pRAi3k6cgJUrzcIBqVM/szMkxGTAhg2hbNmE3+TCLFhRFgKOQPWZUH1qnAnufvB9Gk1vRP2p9QkKC0r4vVOAYsVMh1dXb5eTkpwQIt5GjTLJrV+/aHbOnGnmkpwyJWEXD3sIe96F81Mhe3WoPgO8CsZ52r3H92g0rRGHbh5iTvs5eHpEN2hPPOHhAS+9JElOCCGe4u9v8teTdU6forUZNlCmjCnJxdednbCtKzy6AKWHgvdnYMWqALcf3abhtIYcv3OchZ0W0qJ4i/jfOwXy8YEdO5wdhX1JdaUQIl7+/NOsmBNtp8mVK81ckh9+GL8pvCIj4Mh3sOZl0BHQYDOU/tKqBAcwdONQTt49ydIuSyXBxYOPjxnp8fChsyOxH1mFQAhhtfBwM4i4UCEzkclz6tc3XS7PnYumsS4Gjy7B9h5wazO82AUq/QapM8crrqCwII7cOkLlvJXjdV5Kt3gxtGljSnNVqjg7msSJaRUCKckJIay2eDFcumRm6XrOvn2wfr0p4lmb4C7Ng+Vl4d4+qDbVtL9ZmeAuB1ymy4IuBAQH4OnhKQkuAby9zU9Xnt5L2uSEEFYbOdKU4lq2jGbnsGGQIQO88UbcF4oIht3vwLm/IFsVk9wyFLE6jgv3L1B3Sl3uPb7HOf9zlM9d3vovIf5VqJCZ4suVO59IkhNCWGXfPtiyxSwN5+7+zM6LF2HuXFOKy5Qp9guFBcLm1nBzvelYUvpLcLN+8uYz985Qb0o9AkMDWddznSS4RHB3h1KlJMkJIQQjR0L69PBqdJOHjBxpOppEW48ZRag/bGgG93ZDtWlQKLoF6GJ28s5J6k2tR0h4COt7rafcC+Xidb54no+Pmb3GVUmbnBAiTjdvwuzZ0KcPZM78zM7792HCBOjUKYalCCyCb8G6euC/F2rMi3eCA0jtnpo8GfKwsfdGSXA24uMD16/DvXvOjsQ+JMkJIeI0bpxZSfq996LZOX48BAbGPoVX0BVYWwsenIRaSyH/K/G6/8X7F4nUkRTKUohdr+3CJ6dP/L6AiJGrdz6RJCeEiFVICPz+OzRrBsWLP7MzNNRUVdavD+VjaBt7srjp4+tQdzXkaRyv+++9tpfyf5Tnyw1fAqToxU7tIeoq4a5I2uSEELGaM8dUV0bb3DZ7Nly7ZkaIRyfgGKxvAJGhUH89ZK0Yr3vvuLKDJtObkCVdFllJwE7y5TPL/blqkpOSnBAiRlqbglqpUtCgQTQ7hw0zRYHG0ZTO7u01VZQA9TfFO8H9c+kfGk5rSHbP7GzqvYlCWQol7EuIWCnl2guoSpITQsRo61YzdKB//2hm6Vq9Gg4fjn4Kr1v/mE4mqbygwRbI7B2v+z4MeUib2W3ImyEvm/tspkCmmBdHFYn3JMklswmwrCJJTggRo5EjIUsW6NEjmp3DhkGePGZh1Kiur4ENjSBdbrPAaTwGeT+RIU0G5naYy6bem8iTIU/CghdW8/Y2vStv3nR2JLYnSU4IEa2LF2HhQjOBieezq9YcOABr15oiXtQpvC4vgk0tIENxM8myZ/xWBV9+ejlTD04FoF6heuTyypW4LyGs4sqdTyTJCSGiNXasqYV8551odg4fDl5eTy8od346/NMBslSABhsg7bPr8MTu7xN/02Z2G8bsGkN4ZHjighfxIklOCJGiPHpkxne3bRvN+O7Ll02vytde+29k+OlxsL0n5KwN9dZA6izxut+8o/PoMK8DFXJXYHWP1aSycokdYRs5c0KOHJLkhBApxLRpZiKTaNeMGznS9FB4Mqbg2M+w+y3I0xzqLAMPr3jda+bhmXRe0Jmq+aqyusdqMqfNnMjoRUK4ag9LSXJCiKdoDaNGga8vVK/+zM6AADPDSceOUKAAHPwcDnwEBTpBrYXgnjbe9ztz7wy1X6zNim4ryJgmo22+hIg3Hx8z64mr9bCUJCeEeMqRI3D8uOlw8tywgQkTzDLSAwbAvg/g6LdQpK9ZKiceKwkA3HtsJkv8vNbnrOy+Eq/U8SsBCtvy9jazs1265OxIbCvOJKeUelkpld7yvrtS6hel1Iv2D00I4QxLl5qfLVo8syM0FEaMgHp1IGIcnBwJL70PlSeA27Nr78RuzK4xFBtdjFN3T6GUIrW7lYusCrtx1c4n1pTkfgeClFJlgYHAWWCqXaMSQjiNnx9UqgS5cz+zY+5cuHEV3oyEs3+CzxdQ4ZdoinuxG75tOO+teI9aL9aiYOaCNotbJM6TiZpTYpIL11proDUwRms9Fshg37CEEM5w6xbs2BFNKU5rGPE/+NwLwjZD+Z+hzFfxTnA/bPmBD9d8SIdSHZjbfq6U4JKQzJnNPJauluSs6af7UCn1KdADqKmUcgPiV/kuhEgWVqww+ey5JLdmKTQ+DEUVVBoHxfpFe35sZh2exeD1g+lWuhuT20yWYQJJkCv2sLSmJNcJCAFe1VrfAPIBP9s1KiGEU/j5mZm6nlo1J9QfjvWEEkDlvxKU4ABeKfkKvzb+lSltpkiCS6K8vU2no4gIZ0diO3EmOUtiWwCksWy6AyyyZ1BCCMcLDYVVq0wp7t9ayMc3wa8qZAmAu92hWO94XVNrza/bf+Xe43ukTZWW96u+j3s8O6kIx/HxMesHnj3r7Ehsx5rela8D84E/LJvyAn/bMSYhhBNs3mxGB7Rsadnw6LJZKifwHIxJCz1Hxut6kTqS/iv6M2D1ACYfmGzzeIXtuWIPS2uqK98BXgYeAGitTwPxm5ROCJHkLV0KadNCvXrAwzOwtiYEXYcfNVR/A7JmtfpakTqSN/3eZMzuMQysNpAPqn5gv8CFzZQsaUrxKS3JhWitQ598UEqlAlxsTLwQKZvWJsk1aACeoUdgTU0ID4Tjr8BJDR9Yn6QiIiPou6QvE/ZNYHCNwfzc8GdUPHthCudInx4KF055SW6TUmowkE4p1RCYByy1b1hCCEc6cQLOn4fOLS/Cujqg3KDqcvj1b+jQAQoWtPpadx/fZdOFTXxV5yu+rfetJLhkxtvbTO/lKqxJcp8At4HDQD9gOTDEmosrpZoopU4qpc4opT6J4ZiOSqljSqmjSqmZ1gYuhLAdM8uJpm2BdyAiGOpvhDn/wIMHZuVvK4RFhBERGUHO9DnZ328/X9T+QhJcMuTjA6dOmQ4orsCafrxtgKla6wnxubBSyh0YCzQErgC7lVJLtNbHohxTDPgUeFlr7a+UkrY+IZzAzw8+7jyXdPeWQYVfIV1BM4VX7dpmpuY4hISH0Gl+J7Kmy8qfrf4kU9pMdo9Z2IePD4SHm0RXurSzo0k8a0pyLYFTSqlpSqkWljY5a1QGzmitz1na9GZjZk2J6nVgrNbaH0BrfcvawIUQtnH3Lhzd78+Qpv0hqy8Ufw/mzTPrxllRigsOD6bt3LYsPrmYCrkrSOktmXO1HpbWjJPrAxTFtMV1Ac4qpSZace28wOUon69YtkVVHCiulNqqlNqhlGoS3YWUUm8opfYopfbcvn3bilsLIay1ciX82Pkj0qe6C1UmmPa4YcNMV7tmzWI9NygsiFazWrHi9Ar+aPEH71Z+10FRC3t56SVIlcp1kpxVpTKtdZhSagWmV2U6TBXmaza6fzGgDmYmlc1KqdJa6/vP3H88MB7A19dXenYKYUNntm3iy7oT0SU+gizlTClu/36YOBHcYv87uNP8Tqw9t5a/Wv9F73K9HRKvsK/UqaFYMddJctYMBm+qlJoMnAbaAROBF6y49lUgf5TP+SzboroCLNFah2mtzwOnMElPCOEAYcHBdCnSj9uPC6GKf2yGCnTsaBpjunWL8/z3q7zP9LbTJcG5mCcLqLoCa9rkemJmOHlJa91ba71cax1uxXm7gWJKqUJKqdRAZ2DJM8f8jSnFoZTKjqm+PGdd6EKIxLq6+geKv3CSU8FDoWot09nk3XfNUgRpo1/l+37wfRYcWwBA/cL16Vq6q+MCFg7h4wPnzsGjR86OJPHirK7UWndJyIW11uFKqXeBVYA78JfW+qhS6mtgj9Z6iWVfI6XUMSACGKS1vpuQ+wkh4un+UfI9+IEZ27vSesL7kC2NaaBr3DjGU+49vkejaY04cusIlfNWJn+m/DEeK5IvHx8zQcDx41Z1rk3SYkxySql/tNY1lFIPeXqGEwVorXXGuC6utV6OGVcXddsXUd5rYIDlJYRwFB0JW3rxMCg9S6e1pFvLYPjjD8iePcZTbj+6TYNpDTh55yQLOi6QBOfCovawdNkkp7WuYfkpC6QK4Uq0hpmvgdrL+9MnU7N1QZg6P9YFUG8E3qD+1Pqc8z/Hki5LaFSkkePiFQ5XpAikSeManU+s6XgyzZptQohkwN8fereBx5O4eKY4U7f0pMW3VeNc4dvvlB8X719kedflkuBSAHd3M4LEFTqfWNPxxDvqB8tg8Ir2CUcIYTfr1plek1mXQtpUDN7jR+nSihdfjPmUSB0JwGsVXuPEuyeoW6iug4IVzuYqq4THmOSUUp9a2uPKKKUeWF4PgZvAYodFKIRInOBgGDDALDFQEaikeVzyG+YsL0aLFjGfdt7/PBX+qMDuq7sByJcxn2PiFUmCjw9cuQL37zs7ksSJMclprX+wtMf9rLXOaHll0Fpn01p/6sAYhRAJdfAgVKoEv/4K/V+HXhoyl2bp6YFERERZIPUZp++eptbkWlwKuISbsqbCR7iaJ51PknuVZWwluRKWt/OUUhWefTkoPiFEQkREwM8/Q+XKcOcOrFgBPdJA8HWoPIElfh5kz252P+v47ePUnlyb4PBgNvTaQMU80jqRErnKHJaxjZMbALwBDI9mnwbq2SUiIUTiXLoEPXvCpk3Qtq0ZGsAZWD0Wir9LeOYqrFhhSnHu7k+fes7/HHWm1EGh2NhrI945vaO7g0gBChQAL6/kX5KLbQjBG5af0tIsRHKgNcycCe+8Y0pykyZBr16gw2HF6+CZF8p+x/btcO8e0bbH5c+YnzYvtWFAtQG8lP0lx38HkWQoZRZQTe4lOWuGEHRQSmWwvB+ilFqolCpv/9CEEFbz94cuXaB7d1PPdPAg9O5tflMdHwYBR8D3N/DIgJ8feHhAoygjAfZf38+tR7fwcPfgj5Z/SIITgGv0sLSmRflzrfVDpVQNoAHwJzDOvmEJIaz2ZGjAggXw3XemmrJwYbPvwWk4/BXkbw/5TC+TpUvNWqgZLXMWbb+8nTpT6vD60tedE79Isnx84PZtuJWMV/q0JslFWH42B8ZrrZcBqe0XkhDCKlGHBmTIYCZVHjz4v4Y2rWF3P3BPC76jADh71sxH+KSqcsvFLTSa3oic6XMypukYJ30RkVS5QucTa5LcVaXUH0AnYLlSKo2V5wkh7GXXrv+GBrzzDuzdCxWf6QV5fgrc3ADlfoJ0uQFYtszsatEC1p9fT5MZTciXMR+bem+SuSjFc7wt/Y5cPcl1xKwW0NiymGlWYJA9gxJCxCAgwCS1qlVN75Hly2HMGPD0fPq44FuwbyDkeBmK/lcNuXSpma6pUOFIBq4eSOEshdnYayN5MuRx8BcRycELL0DWrMm7h6U1S+0EKaXOAo2VUo2BLVrr1fYPTQjxL61hzhyzqOmtW/Dee/DNN/81rD1r3wAIfwiVx4NlMPeDB6a57oMPwE254dfFjzSp0pDdM+aVB0TKplTy73xiTe/K/wNmADktr+lKqffsHZgQwuLsWWjSxPSezJvXVFWOHBlzgru2Ci7MgFKDIVOpfzevXg1hRRZxsHBPIiIjyJsxryQ4EacnSU7ruI9NiqypruwLVNFaf2FZC64qIN2whLC3kBD49lvzW2b7dhg9GnbufL7tLarwR7D7Tcj4Eng/Pfve6A1zoGMHAtzPEBQWZOfghavw8TG1AFeuODuShImzuhKzSGpElM8Rlm1CCHvZtAnefBNOnIAOHWDECMhjRbvZ4aHw6AI02ATuaf7dPPXAdDZn70WOkJdZ3WMZGdLIMpHCOlE7n+RPhn2TrCnJTQJ2KqWGKqW+AnZgxsoJIWztzh0ziLtOHVOSW74c5s61LsHd2wcnfoEir0POWv9unnJgCr0X94QLdfhfmRWS4ES8PElyybXzSZxJTmv9C9AHuAfcAfporUfYOS4hUpbISPjrL3jpJZgxAz791Pzp3LSpleeHw643IE0OKP/TU7sKZylM8chXcJvjR5tm6e0QvHBl2bJB7tzJt/NJfMa7qWd+CiFs4ehRU3Lr29f82XzgAHz//fPDAmJzajTc2wsVR0HqLAAcvHEQgJov1iT1ogXUqpaOzJltHr1IAZJzD0treld+AUwBsgDZgUlKqSH2DkwIlxcUZGYoKVfOJLo//4SNG/+rH7JW4AU4OATytIACHQAYtm0Y5f4ox/LTy7l4EQ4fjnntOCHi4uMDx46Zeb+TG2s6nnQDymqtgwGUUj8CB4Bv7RiXEK5txQozqPv8edMG9/PPkD0B3fm1ht1vmQFNlcaCUny3+TuGbBhCJ+9ONCzckPGWmWZjWwVciNj4+MDjx+Z/16JFnR1N/FhTXXkNSBvlcxrgqn3CEcLFXbsGHTtCs2aQJo0puU2alLAEB3BxDlxfCWW+Q3vm58sNXzJkwxC6l+nO9LbT8XD3wM8PihWD4sVt+k1ECpKcO59Yk+QCgKNKqclKqUnAEeC+UmqUUmqUfcMTwkVERJhxbiVKmLm1vv3WtL3Vrp3wa4bcg33/B1l9ofi77Liyg683f02fcn2Y3HoyqdxSERgI69dLVaVInFKWOQWSY7ucNdWViyyvJzbaJxQhXNS+fdCvH+zZYxZx++03KFIk8dc98BGE3IW6q8DNnWr5q7GmxxrqFaqHm2Uqr7VrITRUqipF4mTIAAULumiS01pPcUQgQricsDAYNMiU4HLmhNmzTVWlSmAH5eDbcGujWVng5gZ4cILIEoP4ePcM2pR4xMsFXqZB4QZPneLnB5kyQY0aif86ImVLrj0srSnJCSESYsYMM8fkm2/CDz8Q7/77Iffg1qb/klqA5TdMKi/IUZPIIq/T79QRJh6YRDqPdLxc4OWnTo+MNEmuSROzErgQieHjAytXmpqB1MloRVFJckLYy7JlZqaS336zrvQWGgC3t/yX1PwPABrc00GOGlCwK+SqC1krEoEbfZf0ZcrBKXxW8zO+qvPVc5fbuxdu3pT2OGEb3t4QHg6nT8d/lIszxZjklFLTtNY9lFL/p7Ue6cighEj2wsLMtP+xVU+GBcLtrXBzvSWp7QUdCW5pIHs1KD3UJLVslZ+ahzI8Mpyei3ow68gsvq7zNZ/X/jzayy9dCm5upiQnRGI9WSX86FEXSXJARaVUHuBVpdRUnpnpRGt9z66RCZGcbd1qpm5v3vy/beGP4c62/0pqd3eBDgeVCrJXAe/PLEmtKqRKF+vlI3UkP9b/kY9rfBzjMX5+UL26mZZJiMQqUcL80XTkiPnbLbmILcmNA9YBhYG9PJ3ktGW7ECI6y5ZBulRQLi0c/sqU1u7sgMhQUO6m23/JD01Sy/EypIp7TsmQ8BACQgLImT4nM9vN/LcHZXSuXoX9++Gnn2I8RIh4SZvWjLdMbp1PYkxyWutRwCil1O9a67ccGJMQyd+JufB7JGxrCijIUh6Kv2eSWs6a4BHDgqcxeBz2mHZz23Hh/gX299tPmlRpYj3ez8/8lKEDwpZ8fODQIWdHET/WDCF4SylVFqhp2bRZa53MvqYQDnThApS4BG6ZoNYUs+yNZdLkhAgKC6L17NasO7eOP1r8EWeCA5PkCheGkiUTfFshnuPtDQsXmim+0sVeo55kWDNBc39gBpDT8pqhlHrP3oEJkWwtXwg+QJ5XIF/rRCW4wNBAms1oxvrz65nUehKvV3w9znOCgswg8BYtEj4kT4jo+PiY6VJPnHB2JNazZlqv14AqWusvtNZfAFWBuP+lCZFSHZoOHoBPr0Rf6v2V7/PPpX+Y/sp0epWz7nrr10NwsAwdELb3pIdlcmqXs2acnAKiLrAQgawpJ0T0Hj8Gj4MQltaMbUuk7+p9xyslXqF58eZxH2zh5wdeXlCrVtzHChEfRYuageDJKclZU5KbBOxUSg1VSg0FdgB/2jUqIZKr9WugTCR41QS3hM21cDfoLp+s/YSwiDByeeWKV4LT2iS5xo2T16wUInnw8DBDCVwqyWmtfwH6APcsrz5a6xF2jkuI5Gnbn5AeKJ+wGv1bj25Rd0pdRuwYwcGbB+N9/oEDZviAVFUKe/H2Tl5Jzqo/NbXW+4B9do5FiORNa3iwAcLdoID1pa8nrj+8Tv2p9blw/wJ+Xf3wzeMb72ssXWo6mzRtGu9ThbCKjw/MmmXmOsgYv5EwTmFNdaUQwhrHj0Pxh6BLQyrPeJ165cEVak+uzaWAS6zotuK51QSs5ecHVauaRQ+EsIcnnU+OHXNuHNaSJCeErawZD9mAUj3ifeqNwBs8Dn/Mqu6rqF0wYQupXr8Ou3fLAHBhX8mth6WsQiCErVz52yS50tYPHQgIDiBT2kz45vHlzHtnrBroHZPly81PSXLCngoWBE/P5JPkrBkM3lYpdVopFaCUeqCUeqiUeuCI4IRINgICIPtFCMoPabNbdcrpu6fx+d2HMbvGACQqwYGpqixQAEqXTtRlhIiVm1vy6nxiTXXl/4BWWutMWuuMWusMWutk0NwohAOtmQ75MTOcWOH47ePUmlyLkPAQar2Y+AFtwcFmZR+Z5UQ4gqsluZta6+N2j0SI5OzIFPOzetwz3h2+eZjak02728beGymTq0yib79xo5nOS4YOCEfw8TEL8t654+xI4mZNm9wepdQc4G8g5MlGrfVCewUlRLISGQnuByAgE2QuHuuh94PvU29qPdK4p2F9r/UUzxb78dZauhTSp4c6dWxyOSFiFXUB1doJ6yflMNYkuYxAENAoyjYNSJITAmD3eigUBtSJ89DMaTMzrOEwahSoQZGsRWxy+yeznDRsaNb8EsLeovawTPZJTmvdxxGBCJFsbR9r1ueoHHNV5bbL2wgJD6FuobpWT7RsrcOH4dIl+OILm15WiBjlyQOZMyePdjlrelfmU0otUkrdsrwWKKXyOSI4IZKFwI3wIDUUqRft7s0XN9NoWiMGrB5ApI60+e2fLJDarJnNLy1EtJRKPp1PrJ2geQmQx/JaatkmhLhxAfLdh8jy0XZrXHduHU2mN6FApgIs77ocN2X7+Rf8/KBSJcid2+aXFiJGPj6mTU5rZ0cSO2v+xeXQWk/SWodbXpOBHHaOS4jkYc2vkJpo145beWYlLWa1oGjWomzsvZHcGWyfhW7dgh07ZAC4cDwfH/D3NzPtJGXWJLm7SqnuSil3y6s7cNfegQmRLFxbAkEKqr/63K6FxxdSInsJ1vdaT8709plMcsUK85e0DB0QjpZcpveyJsm9CnQEbgDXgfaYpXeESNlCgyHLRbhfGKLMVhIaEQrA781/Z2OvjWT3tG4GlIRYuhTy5oVy5ex2CyGi5e1tfib7JKe1vqi1bqW1zqG1zqm1bqO1vuSI4IRI0jZPAC8NBdr8u2nOkTmU+b0M1x5ew93NnUxpM9nt9qGhsGqVzHIinCNHDrPaRVJPcjEOIVBKfaS1/p9SajRmXNxTtNb97RqZEEndsalmFGnjDwCYdnAavRf3pkaBGmRIncHut9+0CQIDpT1OOM+TzidJWWzj5J5M5bXHEYEIkaxoDR4H4Xo2yJaXv/b/xWtLXqNuobos6byE9KnT2z0EPz8z+Lte9CMXhLA7Hx/4808z6Y9bEl24LcYkp7VeankbpLWeF3WfUqqDXaMSIqk7sRoyhUFkXeYenUvfJX1pUrQJCzsuJJ1HOrvfXmvTHteggVn2RAhn8PGBR4/g4kUoVMjZ0UTPmtz7qZXbhEg5doyBSODld6lfqD4Dqg7g705/OyTBgVmE/Px5qaoUzpUceljG1ibXFGgG5FVKjYqyKyMQbu/AhEjSgjYzzz8VrTpVIZtHWoY3Hu7Q2y9ebH5KkhPOVKqU+XnkSNIdxhJbSe4apj0uGNgb5bUEaGz/0IRIou6d5JvIB3SMCOe3Pb87/PZaw/TpUL26GT4ghLNkygT58yfTkpzW+iBwUCm1CHiktY4AUEq5A1YtYayUagKMBNyBiVrrH2M4rh0wH6iktZaOLiLJ0lrzxdxufHsPemSpSv8qju9kvH8/HDsG48Y5/NZCPCep97C0pk1uNRC1oSEdsDaukyzJcCzQFCgFdFFKlYrmuAzA/wE7rQlYCGfRWvPx2o/59uZe+rrDpFdX4+7m7vA4pk2D1KmhY0eH31qI5/j4mDbi8CTaiGVNkkurtQ588sHy3pr+XJWBM1rrc1rrUGA20Dqa474BfsJUiwqRZF15cIWJ+ybwVkYY/6gw7l72Hwv3rPBwmDnTtH9kyeLw2wvxHB8fMzHBmTPOjiR61iS5R0qpCk8+KKUqAo+tOC8vcDnK5yuWbf+yXDe/1npZbBdSSr2hlNqjlNpz+/ZtK24thO1oyzTr+TPlZ3+lNxmbE9wKtHVKLKtXm0mZe/Rwyu2FeI6Pj1lf7tYtZ0cSPWtWBn8fmKeUugYo4AWgU2JvrJRyA34Besd1rNZ6PDAewNfXN4kv7CBcSURkBP38+lEwc0GG1BrCixeWQxDQtJ9T4pk6FbJlg6ZNnXJ7IZ5ToQJcversKGJmzdyVu4ESwFvAm0BJrfVeK659Fcgf5XM+y7YnMgA+wEal1AWgKrBEKeVrXehC2Fd4ZDh9Fvfhz/1/EhIeAuGPwf0InM0MRYo6PJ6AADN0oHNn0yYnhIibNSU5gJcwnUfSAhWUUmitp8Zxzm6gmFKqECa5dQa6PtmptQ4A/p2eXSm1EfhQeleKpCAsIowei3ow5+gcvqn7DUNqDYEzc8EjEjLXd0pM8+dDcDD07OmU2wuRLMWZ5JRSXwJ1MEluOaa35D9ArElOax2ulHoXWIUZQvCX1vqoUuprYI/WekkiYxfCLrTWdF3YlfnH5vO/Bv9j0MuDzI59401VZY3XnRLXtGlQvLhZBVwIYR1rSnLtgbLAfq11H6VULmC6NRfXWi/HJMao276I4dg61lxTCHtTStGsaDNq5K/B/1X9P7MxMgKCtsKRVNCzrsNjunDBrDrw7beyrI4Q8WFNknustY5USoUrpTICt3i6rU0Il/A47DEHbx6kar6q9Cn/zLrAd7ZB6mCIrOyUBrEZM8zPbt0cfmshkjVrhhDsUUplBiZgpvXaB2y3Z1BCONqj0Ee0mNWC+lPrcyPwxvMH7J8IYUA5xzeIaW16VdauDQULOvz2QiRrsZbklFIK+EFrfR8Yp5RaCWTUWh9yRHBCOMLDkIc0n9mcrZe3Mrn1ZF7weuHpA7SGG35wDHjf8ePjdu+GU6fgo48cfmshkr1Yk5zWWiullgOlLZ8vOCIoIRwlIDiApjOasuvqLma0nUFnn87RHHQUPO7BrfyQO7fDY5w61SyO2r69w28tRLJnTXXlPqWU9OcSLmn83vHsvrabOe3nRJ/gAE7PMj8LOr4UFxoKs2dD69ZmxnchRPxY0/GkCtDdMmD7EWbWE621LmPPwIRwhIHVB1K/cH0q5K4Q80GnZ8NpoHEMSdCOVqyAu3dlbJwQCRVjSU4pVcDytjFQGKgHtARaWH4KkSzdDLxJi5ktuHj/Im7KLfYE9+gycA5OeDplgNq0aZAzJzRq5PBbC+ESYivJ/Q1U0FpfVEot0Fq3c1BMQtjN9YfXqTe1HpcCLnHh/gVezPxi7Cdc/tv8zFQf3B27rI6/PyxdCm+/DamsnZtICPGU2P7pRB1yWtjegQhhb1ceXKHelHpcD7zOim4rqPVirbhPOj4NrgG1u9g9vmfNnWva5GTFASESLraOJzqG90IkO5cCLlFrUi1uPrrJqu6rrEtwofchaK8ZHdq4sb1DfM7UqeDtDeXLO/zWQriM2EpyZZVSDzAlunSW9/Bfx5OMdo9OCBvJmCYjRbMWZXa92VTOW9m6k64tBxUJoWUga1b7BviMs2dh2zb48UeZxkuIxIgxyWmtHdsAIYQdnPM/xwteL5A5bWZW91gdv5NPzwZ/wLejXWKLzbRpJrnJNF5CJI414+SESJaO3T5G9T+r88bSN+J/ckQI3FptJrFr1sLmscVGa5Pk6tWDfPkcemshXI4kOeGSDt08RJ3JdVBK8VnNz+J/gZvrQYXAhWxQxrFDQrdtg3PnZGycELYgSU64nH3X91F3Sl1Su6dmU+9NlMxRMv4XubgAgoGirR3eKDZtGnh6QlvHT7AihMuR0TfCpYRHhtNlQRcypM7A+l7rKZwlAaNfdCRcXAgHgGatbB1irIKDYc4ck+C8vBx6ayFckiQ54VJSuaViXod5ZEqTKe6B3jG5sxMi/eGgO3xf37YBxmHZMrh/X8bGCWErUl0pXMKmC5v4etPXaK0pk6tMwhMcwJW/IQLIUNPhxampU81CB/Udm1uFcFmS5ESyt/bcWprOaMrsI7MJDA1M/AXPzzdrxzVonfhrxcOdO7B8uRk24OAZxIRwWZLkRNIXERHjrhWnV9BiZguKZi3Kxt4byZAmQ+LuFXACgs/BHqB588RdK55mz4bwcOlVKYQtSZITSdvKlZAlC/zyy3O7lpxcQps5bfDO6c2GXhvImT5n4u935W/z815BKFYs8deLh2nToGxZKF3aobcVwqVJkhNJ19Gj0LGjKckNHAjffGNGSls8Cn1ExdwVWddzHdk8s9nmnpcWwnkFNR1bVXnyJOzaJaU4IWxNkpxImm7fhjYtIFdak+x69IAvvoBPP+X6g2sAdCndhS19tpA5bWbb3PPxdfDfDbs1NGtmm2taado0cHODLo5f7EAIlyZDCETSExIC/RpAv4uQR8PVT2H0UEifnikrf+LNtMNY03s9NQrWwt3Nhj00riwxP4+mhdq1bXfdOERGmiTXqJHpWSmEsB0pyYmkJTwY/ucLbQ9BnixQ/D2TfJZ7M6HBUfq0gRrnIqjw7Z+xdkhJkCt/w51UUKohpElj22vHYssWuHRJxsYJYQ+S5ETScW8vTC8IRY5AYHlofxZ8R0Grs4z1eJk3jmyhiZcbS32r4Dl7qskKYWG2uXfYA7ixFnaGQ3PHTsg8daoZjtemjUNvK0SKIElOOF9EKBz6AlZWhsCbsKs29NsLqTMDsPHmCd49upnWRRqyqGpn0mbfBePSwqNZ0KWtqd5MrGsrQYeboQNNmyb+elZ6/BjmzYP27c18lUII25I2OeFc/gdgey+4fwi2ucMRXzNsIMqkyLVfrM2frf6kR5keeLh7wP1P4eBn0GkJ3PeDj8rBd9vBK3PC47jyNzz2gHQlIH/+RH4p6y1eDA8fSlWlEPYiJTnhHJFhcPgrWFkJgm7ApKywNC/M8YO0adFa8+v2Xzl99zRKKV4t/6pJcACZfaD2Ymi4DTIUhyonYHpuOD4BIhPQThcRCleXmarKpo4dAD5tmsmpdeo49LZCpBiS5ITj+R+CVVXg8FDI2xZG5IYdobB0KeTKhdaazzd8zoDVA5iwb0LM18lRDbqfAD6Ge8Gw/w3wKw1Xlj41ni5OtzZC+APYox06y8nNm7BqlZnGy03+JQphF/JPSzhOZDgc+Q5W+cLjq1BjPoyKgG2HYOZMKFMGrTUfrfmI77Z8x+sVXufHBj/Gfk2loOuPUHIB/OYOl87B5law5mW4ucm6uK78DRGp4GpmqFo1sd/SarNmmQ6iUlUphP1IkhOOcf8orK4Kh4ZAvrbQ7CiM3w8LFsDPP0PLlmiteX/l+wzbPox3Kr3DuBbjcFNW/i/6Slv4fBl8Aix7AR6ch3V1YENTuLc/5vN0pBmicMQd6jWBVI5rpp46FSpWhFKlHHZLIVIcSXLCviLD4dhPsLICPLoINeZCjdkwfyV89x307QsDBgAQHB7M3ut7+aDqB4xuOtr6BPdE48awbBUsCYQh6aDgYLi7y9z7n87w4PTz59zba0qV20IcOsvJkSOwf79M4yWEvUnvSmE/AcdhR2+TaPK3g0q/QdqcsG2bSW516sBvvxGhIwkJD8HTw5M1PdaQNlVaVJTelfFSuzasWWOGAXSeBqvWQOgiOPkrXJ4PRfqCzxfgmdccf+Vv0AoOaGjSxEZfPG7TppnldDp3dtgthUiRpCQnbC8yAo4PgxXl4eEZqD4LaswzCe7CBTPquUABWLCA8FRu9F7cm+YzmxMWEUY6j3QJT3BPVK0K69ebQWj1mkOqztDyLBR7G85NgqVFYf9HEHLXJLkrXuBdBXLksMGXj1tEBMyYYfJwThssnCCEiJkkOWFbD07B2pqwfxDkaQLNj0LBzqaDyIMH0LKlmaXEz4+wTBnovrA70w9Np2Hhhv8NEbCF8uVh0yZz39q14fhVM3tKi1NQoKNJwosLQcAx2PDQob0qN2yAq1elw4kQjiBJTthGZASc+BVWlIUHJ6DaNKi5CNK9YPZHRJgp9o8fh3nzCC1aiE7zOzHn6Bx+bvgzg2sOtn1MpUrB5s2QPj3Uqwfbt4NXQag2BZodhhfqg/aE3Ti0PW7aNMiUyeR7IYR9SZITiffwjOnJuG8A5GoAzY5Aoe5PzVrChx/C8uUwZgw0aMDby95m0YlFjGwykg+rf2i/2IoWNYkuRw5o2BA2bjTbM3tDrUWwpDmkfsGU/Bzg0SPTobRDB0iXziG3FCJFkyQnEk5HwsnRsLwM3D8MVSdD7SXgmefp48aPhxEjoH9/ePNNAAZWG8ifrf6kf5X+9o/zxRdNonvxRdMQtnKl2R4eDqtWm20OGo29aJFJdNKrUgjHkCQnEm7n67C3P+SsDc2PQOFeT5fewHQAeecdaNqURz98zbg949BaUzJHSV4t/6rjYs2d25TiSpSAVq1Mttm2DQICHNoeN3UqFCwIL7/ssFsKkaLJEAKRMJcXwrm/oORHUO7H55MbwKlT0K4dvPQSD6dOoPmclmy9vJXKeStTIXcFx8ecI4fp9dG0qakvrFTJDP5u0MAht796Fdatg88+k2m8hHAU+acm4u/xTdjVD7JUgLLfRp/g7t2DFi0gVSoCFsyg0eL2bLu8jZltZzonwT2ROTOsXg01a8KOHeZnpkwOufXMmWYVcOlVKYTjSElOxI/WsOsNCHsI1aeBWzTd/sPCzAJpFy9yb9XfNN78GgdvHGReh3m8UvIVx8f8rAwZYNkyGDzYYV0ctTZVlVWrQrFiDrmlEAJJciK+zk+Bq0ug/HDIFM2ki1qbNrgNG2DqVPbkd+f4tuMs7LSQFsUdu+J2rDw9TWcYBzl40EzlNXasw24phECSnIiPRxdhT3/IWQtKvB/9MSNGwIQJhH/6Mal69KARcP7/zpMjvWNmE0mqpk0DDw/o1MnZkQiRskibnLCOjoQdfQBthgpEN3nysmUwcCDXOjalfL5lLDy+ECDFJ7jwcDONV/PmkC2bs6MRImWRkpywzsnRcHMDVJ4AXoWe33/4MHTuzOVq3tSrfoobATfJ4Zmyk9sTa9eaBVJlbJwQjidJTsQt4AQc/ATyNDez+D/r5k1o2ZILeTyp2zaAe48DWNNjDVXzOW4B0qRs6lTIksWhM4cJISykulLELjIctvcEd0+oMuH54QLBwfDKK9x5eJNaryoCwgNZ13OdJDiLBw/g77/Nkjpp0jg7GiFSHinJidgd/QHu7TaLnabL/fQ+reG112D7drLNncurOY7SpkQbyr1QzimhJkULFpgVf2RsnBDOIUlOxOzeXjjyNbzYBQp0eH7/999zdPUM1Df9KdWhA0OJ5pgUbto0M0d0VSnYCuEUUl0pohcRbKop0+YE3zHP71+4kIOjh1CnXxp65PwHrbXjY0ziNmwwr969o58URghhf1KSE9E7OMQsKFpnJaTJ+vS+/fvZ+2E3GvZ1J32W7MxuNzvxq3m7mEePoG9fU4r74ANnRyNEyiVJTjzv5iY48QsUfRPyNH56340b7OjbmCadQ8iSLR/r+2yiUJZohhSkcJ9+ChcumMXJPT2dHY0QKZckOfG0sIewozd4FYbyPz+9LzgY2rThh1L3yJ4lL+v7/kOBTAWcEmZStmULjB5tls+rWdPZ0QiRskmSE0/bN8BM39VwC3h4/bdda/RrfVE7dzJj4HQeNqlH7gy5Y75OChUUBK++CoULw/ffOzsaIYR0PBH/uboMzk6EUh9BjqdX9Vz9fV8aus/k4Tef49WhmyS4GAwZAmfOwJ9/Qvr0zo5GCCFJThghd2Hna5C5NJT+6qldy6cMoVXwJO7kyUzIB+85KcCkb9s2Mz/1W29BnTrOjkYIAZLkBJhB3bvfgtC7UG0quP83NcffK36lzdnv8HnkyfpBR8iewidbjsnjx6aaskAB+OknZ0cjhHhC2uQEXJwNl+ZB2e8gS7l/Ny/eMZkOOwZQ8W5qVg7aS+aseZ0XYxI3dCicPAlr1pg1WYUQSYNdS3JKqSZKqZNKqTNKqU+i2T9AKXVMKXVIKbVOKfWiPeMR0Qi6CnvegWxVoeRH/20PDsZn8EjannRndc81ZC5YwnkxJnE7d8KwYfD669CggbOjEUJEZbckp5RyB8YCTYFSQBel1LNLSe8HfLXWZYD5wP/sFY+Ihtaws6+Z3aTaVHAzBfttl7aiX3+NIhsOMKfDHDJWqeXkQJOukBBTTZk3r0l0QoikxZ4lucrAGa31Oa11KDAbaB31AK31Bq11kOXjDiCfHeMRzzrzB1xfZcbDZSwGwPi943l5Ug3+ODEDvvkG2rVzcpBJ29dfw7FjMGECZMzo7GiEEM+yZ5LLC1yO8vmKZVtM+gIrotuhlHpDKbVHKbXn9u3bNgwxBXt4BvYNhBcaQrG3ABizawz9/PrR7BT0Lt4RPvvMyUEmbXv3mk4mffpA48ZxHy+EcLwk0btSKdUd8AV+jm6/1nq81tpXa+2bI4f07ku0yAgzq4mbB1T9C5Qbw7cN570V79H6tDsLz1Yk7cTJMqtwLEJDTXLLlQt++cXZ0QghYmLP3pVXgfxRPuezbHuKUqoB8BlQW2sdYsd4xBMnhsPtrVBtGnjm45z/OQavG0yHc+mYsSkLHruWQrp0zo4ySfvuOzh8GPz8IHNmZ0cjhIiJPZPcbqCYUqoQJrl1BrpGPUApVR74A2iitb5lx1jEE/cPw6HPIX9bKNgNgMLp8vDP1pcov+U0qbYshdwym0lsDhwwU3b16AHNmzs7GiFEbOyW5LTW4Uqpd4FVgDvwl9b6qFLqa2CP1noJpnrSC5hnWarlkta6lb1iSvEiQmFbD0idGe37O59v+JxS2UvSddgqKq06DPPmQYUKzo4ySQsLM9WU2bOb2U2EEEmbXQeDa62XA8uf2fZFlPcyqsiRjnwF9w+ia/7NoC3/Y/j24bzlXpWu03bAV19B+/bOjjDJ+/FHU5JbtAiyZo3zcCGEk8mMJynFnR1w7Ed0od7835F1jN41mnezN2fUe8ugc2f4/HNnR5jkHT5sRlV06QJt2jg7GiGENSTJpQThj2B7T3TavLx1W/HH/tEMLNqTn/vNR/lWgr/+kp6UcQgPN9WUWbLAqFHOjkYIYa0kMYRA2NmBT+Dhaag2mexeefi0fH9+HrwBlTkL/P239KS0ws8/m3FxY8ea9jghRPIgJTlXd30N4SfHcDF/b4q8UI9vMleH+vVQt++YJazz5HF2hEnesWNmAub27aXZUojkRkpyrizUn7Adfeh6NwNVd/px59FtVL9+qG3bYcoUqFjR2REmeRERZm7KDBlMKU4IkbxISc5VhQcRsqEFnc5dY3GgZnijoWQfOwmmTjXFkg4dnB1hsvDrr2aVgVmzIGdOZ0cjhIgvSXKuKDKM4M3taHdoG8uDYHTT0bx7owB80gY6doQvvojzEsKsDzdkiOlJ2amTs6MRQiSEJDlXoyNhx6v8cHQlK4IUf7QYxxupq0G36qZ6ctIk6UlphSfVlJ6e8Pvv8siESK4kybkSrWHvB3BhOp/U+pIqnpVplskXKlc268AsXmx+a4s4jR4N27bBtGnwwgvOjkYIkVCS5FzIgwOf8/E/o/ih6ptkLvMlzUJDoX59uHULNm+WnpRWOnMGBg+GFi2gWzdnRyOESAzpXeki7h8dTqPV3zHxgWJX9tbw+DG0bg1bt8LkyeDr6+wQk4XISOjbF1KnhnHjpJpSiOROSnIu4N6pv2i09EMOhSrmd5hHoxzVzCqe27bBn3+azibCKr/9Zgq9kyZB3tiW+BVCJAuS5JK52+fm0nBhX06EKv7uuIBm2WpAvXpmosXZs2WoQDycPw+ffAJNm0KvXs6ORghhC5LkkrM7O3i8rRehKg1LO82ioVdlqF3b/LZevNj8thZWeVJN6eYGf/wh1ZRCuApJcsnUnev/kPWfVhTImJdDr2wk1c1QqFkT7tyBlStNshNWGz8eNmwwP/Pnj/t4IUTyIB1PkqFL17ZSdXIdBt4KhXprSHXxAdSoAQEBsH69JLh4ungRBg2CBg3gtdecHY0QwpYkySUz52/spvaUOtwJj6Bz/T/g5D2oVcuMkdu0SXpRxpPW8Prr5v3EiVJNKYSrkerKZOT0zf3Um/Qyj8LDWdf2TyreLgDN60K2bLB2LRQp4uwQk52//oI1a0yvyhdfdHY0Qghbk5JcMhEW+pAmk6sTHBHGhjZjqHg5jxkmkDevWTJHEly8Xb0KAwZAnTrQr5+zoxFC2IOU5JKDyHA8dvRgXNZg8vj+D++TuaBrK/D2hlWrZHr8BNAa3nzTrPg9caLpVSmEcD2S5JK4A9f3c3j7/9EjeAsNa42G7V7QtxNUrQrLlkHmzM4OMVmaORP8/MxSOlIIFsJ1SZJLwvZc3U2jKbXIqINp12wwnqs19O8DDRvCokWQPr2zQ0yWbt6E/v2hWjV47z1nRyOEsCeppEmitl/eTv3JNcmkg9lYoweey9OZ38xt2sDSpZLgEuG99+DRI9PpxN3d2dEIIexJklwStPniZhpNrUtOFcLmqq0ouCwXDPkcevSAefMgTRpnh5hsLVhgHuHQoVCihLOjEULYmyS5JGjnkYnkcwthU4U65J+fE34eBm+/bVYTSCU1zAl19y688w5UqAAffujsaIQQjiC/MZOQoLAgPO9uZ9DDObxdtjLpZ+SE6RPNrMHffy8jlRPpgw9Molu1Sv5WECKlkJJcEuF3yo8iIwpwaHVL8CpG+r+ywvS58MMP5iUJLlGWLzerfA8eDGXLOjsaIYSjyN+zScCi44voNL8TZdNEks8zD/yRGfxWwtixpppSJEpAgBns7eMDn33m7GiEEI4kSc7J5hyZQ7eF3aiUzp2VBbKQaWI2WLsDpk41HU1Eon30EVy7BgsXmhW/hRAphyQ5J1p/fj1dF3bl5fRpWJYnFRkmZIEtx0z3v1decXZ4LmHdOrN8zqBBUKmSs6MRQjiatMk50cu5yzMkT25W5I4gw1+ZYNsVMwZOEpxNPHpkVhgoVgy++srZ0QghnEGSnBPMPjKbO4E3SbOrL195XiP9tIyw64GZDr9RI2eH5zI++wwuXDCDvtOlc3Y0QghnkOpKBxu1cxT/t/L/GFi4IsPc98KiDLAH2LgRypd3dnguY+tWGDXKjIurUcPZ0QghnEWSnAMN2zaMQWsG8Uq+snzvthf254RVj2H7RrOigLCJx4/h1VehQAEz+kIIkXJJknOQ7zZ/x5ANQ+hUuCbT3P/B404BGH4JFiyUBGdjX30Fp06Z2l8vL2dHI4RwJmmTc4CHIQ+ZdGAS3V9qxvTUe/GILAAfXYKPPpZOJja2Zw8MGwavvQYNGjg7GiGEsymttbNjiBdfX1+9Z88eZ4dhFa01kToSdzd3bt3aT7Z/muEeDvT3B++XZX4pGwsNBV9fM3XXsWOQKZOzIxJCOIpSaq/W2vfZ7fIb1k601gxYNYC7j+8yqdlIcu7pBeGP4JeMkCo7zJolCc7GfvgBDh82ozAkwQkhQKor7SJSR/Lu8ncZsXMEWdNmxu2fThBwHFaVhH23YP58yJnT2WG6lEOH4NtvoVs3aNHC2dEIIZIKSXI2FqkjedPvTX7b8xuDqn3Ir1mCUDfXwI3m8NcuGDECqlZ1dpguJTzc9KbMmhVGjnR2NEKIpETqy2zsveXvMWHfBD6r+Rnf5EyPOjQM0naGgXPMXJRvveXsEF3OL7/A3r0wdy5ky+bsaIQQSYkkORvr4N2BvBnzMjh/YdjWBbK1hC4roXRpGDdOlsyxsZMn4YsvoG1baN/e2dEIIZIaqa60gdCIUFacXgFAnYJ1GFyiNuzoDdmqw+eXIVLDggXg6encQF1MRISppvT0NKsSyd8PQohnSZJLpJDwENrPbU/zmc05eusoPDgNm1tD+gKwqBDsPmBW6yxa1NmhupyxY2HbNtMO98ILzo5GCJEUSZJLhMdhj2kzpw1LTy1ldNPReGfMBRubAQpu9YY/ZphZglu2dHaoLuf8efj0U2jaFLp3d3Y0QoikStrkEigoLIjWs1uz7tw6xrcYz+vlesD6BhB0GfKOhW7vQMOGssaLHWhtZjRxd4c//pBqSiFEzCTJJZDfKT/Wn1/PpNaT6FW2B2ztCre3QtmJ0PwbyJULZs40v4mFTU2cCOvXmwSXP7+zoxFCJGWS5BKoo3dHSucsTckcJeHAYLg0B8p8DwPnwfXr8M8/kD27s8N0OVeuwMCBUK+eWRBVCCFiI21y8XA/+D6NpjVi55WdACbBnZkAx36Aom/A3CAzH+Xo0VCpkpOjdT1aQ79+plflhAlSTSmEiJuU5Kx0N+gujaY34vDNw9x6dMtsvL4adr8FuRvDzWbwTRvo00eKGHYyYwYsX24mjSlc2NnRCCGSA0lyVrj16BYNpzXk5J2TLO68mKbFmsL9w7ClPWQqBXl/gip1zMreMmDLLm7cgP79oXp1ePddZ0cjhEgupLoyDneC7lB3Sl1O3z2NX1c/k+CCrpmhAh4ZoPJ86NjLJLb58yFdOmeHHK1792DePLh0ydmRJMy770JQEPz5p/TlEUJYT0pycciUJhMVc1fkt2a/UbtgbQgLhE0tINQf6m+Ggd+bKfD9/JJkHdq9e2Zux1Gj4OFDs83XF9q1M1NhFS/u3PisMX++mTDmhx+gRAlnRyOESE5k0dQYXAq4RBr3NOTyyvXfxshw2NwGrq+AWkvB7zK8+SZ8+SUMHWr3mOLj7l2T3EaPNsmtfXsT6t69JmHs2mWO8/H5L+GVLp10alofPIAtW2DjRvjrLyhUCHbskCX4hBDRi2nRVEly0Tjnf466U+pSIFMBNvfejFLKdO3b8y6c/g18x8J9X6hZ0/RlX7YM3JJGze+dO/8lt0ePoEMH+Pxzk8yiunwZFi2ChQtNMomMNDOPtW1rkl6lSo5NeIGBZtTFhg0mse3da3pRpk5t2uF+/11KcUKImEmSs9Kpu6eoN6Uej8Mfs6bHGirkrmB2HP8F9g+EEgMh/8dQsaJpHNq71yxk5mS3b8Pw4TBmjGm76tjRJDdv77jPvXULFi82Jbx168z6bPnymYTXti3UqGH7drCgINi69b+ktnu3ua+HB1SuDHXrmle1akm2mVMIkYRIkrPCsdvHqD+1PhGREaztuZYyucqYHVrDlldApYJqs6BpM1P82bYNKlSwSyzWun0bhg0znTqDgqBTJ5PcSpVK2PX8/U3z4sKFsHIlBAdDjhzQpo0p4dWta0pX8fX4MWzf/l9S27kTwsJM8qxU6b+kVr06pE+fsNiFECmXJDkr1Jlch5N3T7Ku5zpK5XgmS0RGgA6DL76B77833fxefdUucVjj1q3/ktvjx9ClCwwZAiVL2u4egYGwYoVJeH5+5nPmzGa+6XbtoFGjmEtZISGmDW3DBvPasQNCQ02tbsWK/yW1l1+GDBlsF7MQImWSJGeF6w+v8zD0IcWzxdDlcPFiU6R57TUz5YYT3LwJP/9s2qiCg/9LbvZurwoOhrVrTZXm4sWmxJc+PTRrZqo0GzaE48f/S2rbt5tzlDLDB58ktZo1IWNG+8YqhEh5JMnFYNfVXUzYO4HfW/xOKrdYuu6dPm363hcrZnpIpE1rsxiscePGf8ktJAS6djXJ7aWXHBoGYKoZN20yJbxFi0xsUZUt+19Sq1XLlP6EEMKeYkpydu2QrZRqAowE3IGJWusfn9mfBpgKVATuAp201hfsGVNU2y5vo8n0JmT3zM6doDu84BXLypv9+5v+6wsWODTB3bgB//sfjBtnklv37maJOmeOb/PwgAYNzGv0aFMVuXGjaQesVQuyZXNebEIIEZXdkpxSyh0YCzQErgC7lVJLtNbHohzWF/DXWhdVSnUGfgI62SumqDZd2ETzmc3JkyEP63utjz3BAUyeDOfOwYsvOiI8rl+Hn34yy8mEhf2X3IoVc8jtrebubtrVXn7Z2ZEIIcTz7FmSqwyc0VqfA1BKzQZaA1GTXGtgqOX9fGCMUkppO9ehrju3jpazWlIwc0HW9VxH7gy5n9p/7pwZR/a0XOa1yZ6RmfFqixf/l9x69DDJrWhR+95XCCFckT2TXF4gaqq4AlSJ6RitdbhSKgDIBtyxY1yk80hH+dzlWdRpETnT53xu/7hxpv3LWdzdoVcvGDwYihRxXhxCCJHcJYtJkpRSbwBvABQoUCDR16uevzr/9PnHzGQSjTfegKZNE32bBCtaVFa8FkIIW7BnkrsKRP1Vnc+yLbpjriilUgGZMB1QnqK1Hg+MB9O70hbBxZTgwCQZqR4UQojkz54TLu4GiimlCimlUgOdgSXPHLME6GV53x5Yb+/2OCGEECmH3Upylja2d4FVmCEEf2mtjyqlvgb2aK2XAH8C05RSZ4B7mEQohBBC2IRd2+S01suB5c9s+yLK+2Cggz1jEEIIkXIljfVhhBBCCDuQJCeEEMJlSZITQgjhsiTJCSGEcFmS5IQQQrgsSXJCCCFcliQ5IYQQLkuSnBBCCJclSU4IIYTLkiQnhBDCZUmSE0II4bIkyQkhhHBZKrmtbKOUug1ctMGlsmPnFciTMXk2MZNnEzN5NjGTZxMzWz2bF7XWOZ7dmOySnK0opfZorX2dHUdSJM8mZvJsYibPJmbybGJm72cj1ZVCCCFcliQ5IYQQLislJ7nxzg4gCZNnEzN5NjGTZxMzeTYxs+uzSbFtckIIIVxfSi7JCSGEcHGS5IQQQrgsl09ySqkmSqmTSqkzSqlPotmfRik1x7J/p1KqoBPCdAorns0ApdQxpdQhpdQ6pdSLzojTGeJ6NlGOa6eU0kqpFNM93Jpno5TqaPl/56hSaqajY3QWK/5NFVBKbVBK7bf8u2rmjDgdTSn1l1LqllLqSAz7lVJqlOW5HVJKVbDZzbXWLvsC3IGzQGEgNXAQKPXMMW8D4yzvOwNznB13Eno2dQFPy/u35Nk8d1wGYDOwA/B1dtxJ5dkAxYD9QBbL55zOjjsJPZvxwFuW96WAC86O20HPphZQATgSw/5mwApAAVWBnba6t6uX5CoDZ7TW57TWocBsoPUzx7QGpljezwfqK6WUA2N0ljifjdZ6g9Y6yPJxB5DPwTE6izX/3wB8A/wEBDsyOCez5tm8DozVWvsDaK1vOThGZ7Hm2Wggo+V9JuCaA+NzGq31ZuBeLIe0BqZqYweQWSmV2xb3dvUklxe4HOXzFcu2aI/RWocDAUA2h0TnXNY8m6j6Yv7SSgnifDaW6pT8WutljgwsCbDm/5viQHGl1Fal1A6lVBOHRedc1jyboUB3pdQVYDnwnmNCS/Li+/vIaqlscRHh2pRS3QFfoLazY0kKlFJuwC9AbyeHklSlwlRZ1sGU/jcrpUprre87M6gkogswWWs9XClVDZimlPLRWkc6OzBX5eoluatA/iif81m2RXuMUioVpgrhrkOicy5rng1KqQbAZ0ArrXWIg2JztrieTQbAB9iolLqAaUNYkkI6n1jz/80VYInWOkxrfR44hUl6rs6aZ9MXmAugtd4OpMVMUJzSWfX7KCFcPcntBooppQoppVJjOpYseeaYJUAvy/v2wHptaQl1cXE+G6VUeeAPTIJLKe0qEMez0VoHaK2za60Laq0LYtorW2mt9zgnXIey5t/U35hSHEqp7Jjqy3MOjNFZrHk2l4D6AEqpkpgkd9uhUSZNS4Cell6WVYEArfV1W1zYpasrtdbhSql3gVWYnk9/aa2PKqW+BvZorZcAf2KqDM5gGkY7Oy9ix7Hy2fwMeAHzLH1xLmmtWzktaAex8tmkSFY+m1VAI6XUMSACGKS1dvnaESufzUBgglLqA0wnlN4p4Y9qpdQszB8+2S3tkV8CHgBa63GY9slmwBkgCOhjs3ungOcrhBAihXL16kohhBApmCQ5IYQQLkuSnBBCCJclSU4IIYTLkiQnhBDCZUmSE8mOZdb/6VE+p1JK3VZK+TkzrvhSSl2wjCNDKbUtjmN7K6XyxPP6BWOa9T0xEnJdpdTG6AbLK6VaPZmtXyk1VCn1oeX915aJCFBKva+U8rRF7CLlkSQnkqNHgI9SKp3lc0NsNDtCYllmzYk3rXX1OA7pDcQrySWWUsrd3vfQWi/RWv8YzfYvtNZrLR/fByTJiQSRJCeSq+VAc8v7LsCsJzuUUukt61ftsqzb1dqyvaBSaotSap/lVd2yvY6lpDFfKXVCKTUjupUoLMeMVEodUEodUUpVtmwfqpSappTaiplYIIdSaoFSarfl9bLluGxKqdXKrLE2EbOsyJNrB0Z5/7FS6rBS6qBS6kelVHvM3KEzLPdOp5SqqJTapJTaq5RapSwztlu2H1RKHQTeie7BWb7vZqXUMmXWPhunzHycKKUClVLDLedXU2ZNwSOW1/tRLpPK8pyOW56bp+X8Lyzf+YhSavwzz7FHNM+ut1JqTDQxTlZKtVdK9cck9w3KrMP2qlJqRJTjXldK/Rrd9xQCcO315OTlmi8gECiDWRopLXAAM5uCn2X/90B3y/vMmLkT02NKA2kt24thZqHAcm4AZr48N2A7UCOa+24EJlje18KyNhZmZvm9QDrL55lPzgcKAMct70cBX1jeN8fMeJH9yXey/GwKbOO/dfyyRrm3r+W9h+WYHJbPnTCzawAcAmpZ3v9MNOt3Wb5vMGbdM3dgDdDesk8DHS3vKwKHLc/OCzgKlAcKWo572XLcX8CHUeO1vJ8GtIzj2fUGxkR5jk+uMzlKTBeiPCcvzJptHpbP24DSzv5/Ul5J9yUlOZEsaa0PYX7ZdsGU6qJqBHyilDqA+eWaFpNsPDBTKh0G5mEWrXxil9b6ijazwR+wXDs6syz33wxkVEpltmxforV+bHnfABhjuf8Sy3FemF/u0y3nLwP8o7l+A2CStqzjp7WObg2ulzATRK+x3GMIkM8SS2ZLbGCSTEx2abPuWYTlO9WwbI8AFlje1wAWaa0faa0DgYVATcu+y1rrrZb306OcX1cptdPyjOsB3lHuGdOzs5oljvVAC6VUCUyyOxzf64iUw6XnrhQubwkwDFMyiboGoALaaa1PRj1YKTUUuAmUxZTYoi52GnWFhQhi/rfx7Dx4Tz4/irLNDaiqtX5qMdVoakATSgFHtdbVnrl+5nhcI6bvEWxJfPE+XymVFvgNU+K8bHneaa24Z3xNBAYDJ4BJCbyGSCGkJCeSs7+Ar6L5S34V8N6T9iBlVlMAs4zSdUtprQemqi6+OlmuWQMzU3pANMesJspimEqpcpa3m4Gulm1NgSzRnLsG6BOljSurZftDzBI/ACeBHMqsR4ZSykMp5a3Nem33LbEBdIvle1RWZrZ8N8t3+ieaY7YAbZRSnkqp9MArlm0ABZ7c3/Kd/uG/hHbHUnJt/8z1rHl20Yn63dFa78Qsy9KVKG2xQkRHkpxItizVi6Oi2fUNpmrykFLqqOUzmFJGL0unihI8XfqyVrBSaj8wDrM2WHT6A75KqUPKzMT/pmX7V0AtS0xtMcuuPPudVmJKqHssVZEfWnZNBsZZtrljEshPlu9yAHjSO7MPMNZyXGxFx93AGOA4cB5YFE0s+yz33QXsBCZqrfdbdp8E3lFKHcck698tSXYCcATzh8buZy5pzbOLznhgpVJqQ5Rtc4GtWuvoqnyF+JesQiCElZRSGzEdI5L1unFKqTqY79HCyaEkmDJjIn/VWq9zdiwiaZOSnBAi2VBKZVZKnQIeS4IT1pCSnBBCCJclJTkhhBAuS5KcEEIIlyVJTgghhMuSJCeEEMJlSZITQgjhsv4fZgv/t1QMIaoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# код из шестого семинара\n",
    "from sklearn.calibration import CalibratedClassifierCV, calibration_curve\n",
    "\n",
    "plt.figure(figsize=(7, 7))\n",
    "\n",
    "sigmoid_svc = CalibratedClassifierCV(svc, cv=3, method='sigmoid').fit(X_train, y_train)\n",
    "sigmoid_svc_pred = sigmoid_svc.predict_proba(X_test)[:, 1]\n",
    "\n",
    "svc_true_prob, svc_pred_prob = calibration_curve(y_test, svc_pred, n_bins=15)\n",
    "sigm_svc_true_prob, sigm_svc_pred_prob = calibration_curve(y_test, sigmoid_svc_pred, n_bins=15)\n",
    "lr_true_prob, lr_pred_prob = calibration_curve(y_test, lr_pred, n_bins=15)\n",
    "\n",
    "plt.plot(sigm_svc_pred_prob, sigm_svc_true_prob, label='SVC_sigm', color='red') # мы видим, что стало лучше!!\n",
    "plt.plot(svc_pred_prob, svc_true_prob, label='SVC', color='blue')\n",
    "plt.plot(lr_pred_prob, lr_true_prob, label='LR', color='orange')\n",
    "plt.plot([0, 1], [0, 1], label='Perfect', linestyle='--', color='green')\n",
    "\n",
    "plt.xlabel('Mean predicted probability')\n",
    "plt.ylabel('Fraction of positives')\n",
    "plt.title('Calibration curves')\n",
    "plt.legend()\n",
    "plt.show() # да, стало значительнос лучше!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# воспользуемся определением из семинара"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Бонусное задание (0.5 балла).__ Реализуйте свою функцию для калибровки вероятностей. Опишите ваш подход и продемонстрируйте результаты. Ключевые слова для вдохновения: `Platt`, `Isotonic`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([0.00858369, 0.08189655, 0.16336634, 0.235     , 0.31472081,\n",
       "        0.26011561, 0.52469136, 0.53333333, 0.59893048, 0.75294118,\n",
       "        0.72395833, 0.83561644, 0.82553191, 0.86821705, 0.89142857]),\n",
       " array([0.03792749, 0.09835124, 0.16476404, 0.23288661, 0.30311015,\n",
       "        0.36556707, 0.43666038, 0.50158776, 0.56427544, 0.63068835,\n",
       "        0.69917531, 0.76721351, 0.83369974, 0.90142586, 0.9571743 ]))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sigm_svc_true_prob,sigm_svc_pred_prob"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Калибровка Платта**\n",
    "\n",
    "Допустим, у нас есть обученный класификатор $b(x)$, который выдает уверенность (скор) в том, что $x$ относится к положительному классу. Метод приближает вероятность положительного класса с помощью сигмоидной функции:\n",
    "\n",
    "$$\n",
    "p\\big(y = +1 \\big| b(x)\\big) = \\frac{1}{1 + \\exp (A \\cdot b(x) + C)}\n",
    "$$\n",
    "\n",
    "Здесь обучаемыми параметрами являются $A, C \\in \\mathbb{R}$, их подбирают с помощью метода максимального правдоподобия (точно так же, как в логистической регрессии). Сделать это можно по кросс-валидации или на отложенной выборке. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Часть 1. Работа с переменными."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В этой части мы научимся обрабатывать категориальные переменные, так как закодировать их в виде чисел недостаточно (это задаёт некоторый порядок, которого на категориальных переменных может и не быть). Существует два основных способа обработки категориальных значений:\n",
    "- One-hot-кодирование\n",
    "- Счётчики (CTR, mean-target кодирование, ...) — каждый категориальный признак заменяется на среднее значение целевой переменной по всем объектам, имеющим одинаковое значение в этом признаке.\n",
    "\n",
    "Начнём с one-hot-кодирования. Допустим наш категориальный признак $f_j(x)$ принимает значения из множества $C=\\{c_1, \\dots, c_m\\}$. Заменим его на $m$ бинарных признаков $b_1(x), \\dots, b_m(x)$, каждый из которых является индикатором одного из возможных категориальных значений:\n",
    "$$\n",
    "b_i(x) = [f_j(x) = c_i]\n",
    "$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Подготовка данных.__\n",
    "\n",
    "Загрузим данные [UCI Bank Marketing Dataset](https://archive.ics.uci.edu/ml/datasets/bank+marketing). Этот датасет содержит информацию о банковском телефонном маркетинге, объектом в нем является телефонный звонок потенциальному клиенту с предложением некоторой услуги (утверждается, что это краткосрочный депозит), целевой переменной - ответ клиента (согласился ли он открыть депозит?). В качестве признакового описания используются характеристики клиента (образование, брак и т.д.), данные о звонке и различные экономические индикаторы - более подробная информация представлена в файле `bank-additional-names.txt`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2021-11-09 18:12:29--  https://archive.ics.uci.edu/ml/machine-learning-databases/00222/bank-additional.zip\n",
      "Распознаётся archive.ics.uci.edu (archive.ics.uci.edu)… 128.195.10.252\n",
      "Подключение к archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... соединение установлено.\n",
      "HTTP-запрос отправлен. Ожидание ответа… 200 OK\n",
      "Длина: 444572 (434K) [application/x-httpd-php]\n",
      "Сохранение в: «bank-additional.zip.14»\n",
      "\n",
      "bank-additional.zip 100%[===================>] 434,15K   459KB/s    за 0,9s    \n",
      "\n",
      "2021-11-09 18:12:31 (459 KB/s) - «bank-additional.zip.14» сохранён [444572/444572]\n",
      "\n",
      "Archive:  bank-additional.zip\n",
      "replace bank-additional/.DS_Store? [y]es, [n]o, [A]ll, [N]one, [r]ename: ^C\n"
     ]
    }
   ],
   "source": [
    "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/00222/bank-additional.zip\n",
    "!unzip bank-additional.zip\n",
    "df = pd.read_csv('bank-additional/bank-additional-full.csv', sep=';')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>age</th>\n",
       "      <th>job</th>\n",
       "      <th>marital</th>\n",
       "      <th>education</th>\n",
       "      <th>default</th>\n",
       "      <th>housing</th>\n",
       "      <th>loan</th>\n",
       "      <th>contact</th>\n",
       "      <th>month</th>\n",
       "      <th>day_of_week</th>\n",
       "      <th>duration</th>\n",
       "      <th>campaign</th>\n",
       "      <th>pdays</th>\n",
       "      <th>previous</th>\n",
       "      <th>poutcome</th>\n",
       "      <th>emp.var.rate</th>\n",
       "      <th>cons.price.idx</th>\n",
       "      <th>cons.conf.idx</th>\n",
       "      <th>euribor3m</th>\n",
       "      <th>nr.employed</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>56</td>\n",
       "      <td>housemaid</td>\n",
       "      <td>married</td>\n",
       "      <td>basic.4y</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>telephone</td>\n",
       "      <td>may</td>\n",
       "      <td>mon</td>\n",
       "      <td>261</td>\n",
       "      <td>1</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>1.1</td>\n",
       "      <td>93.994</td>\n",
       "      <td>-36.4</td>\n",
       "      <td>4.857</td>\n",
       "      <td>5191.0</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>57</td>\n",
       "      <td>services</td>\n",
       "      <td>married</td>\n",
       "      <td>high.school</td>\n",
       "      <td>unknown</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>telephone</td>\n",
       "      <td>may</td>\n",
       "      <td>mon</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>1.1</td>\n",
       "      <td>93.994</td>\n",
       "      <td>-36.4</td>\n",
       "      <td>4.857</td>\n",
       "      <td>5191.0</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>37</td>\n",
       "      <td>services</td>\n",
       "      <td>married</td>\n",
       "      <td>high.school</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>no</td>\n",
       "      <td>telephone</td>\n",
       "      <td>may</td>\n",
       "      <td>mon</td>\n",
       "      <td>226</td>\n",
       "      <td>1</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>1.1</td>\n",
       "      <td>93.994</td>\n",
       "      <td>-36.4</td>\n",
       "      <td>4.857</td>\n",
       "      <td>5191.0</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40</td>\n",
       "      <td>admin.</td>\n",
       "      <td>married</td>\n",
       "      <td>basic.6y</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>telephone</td>\n",
       "      <td>may</td>\n",
       "      <td>mon</td>\n",
       "      <td>151</td>\n",
       "      <td>1</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>1.1</td>\n",
       "      <td>93.994</td>\n",
       "      <td>-36.4</td>\n",
       "      <td>4.857</td>\n",
       "      <td>5191.0</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>56</td>\n",
       "      <td>services</td>\n",
       "      <td>married</td>\n",
       "      <td>high.school</td>\n",
       "      <td>no</td>\n",
       "      <td>no</td>\n",
       "      <td>yes</td>\n",
       "      <td>telephone</td>\n",
       "      <td>may</td>\n",
       "      <td>mon</td>\n",
       "      <td>307</td>\n",
       "      <td>1</td>\n",
       "      <td>999</td>\n",
       "      <td>0</td>\n",
       "      <td>nonexistent</td>\n",
       "      <td>1.1</td>\n",
       "      <td>93.994</td>\n",
       "      <td>-36.4</td>\n",
       "      <td>4.857</td>\n",
       "      <td>5191.0</td>\n",
       "      <td>no</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age        job  marital    education  default housing loan    contact  \\\n",
       "0   56  housemaid  married     basic.4y       no      no   no  telephone   \n",
       "1   57   services  married  high.school  unknown      no   no  telephone   \n",
       "2   37   services  married  high.school       no     yes   no  telephone   \n",
       "3   40     admin.  married     basic.6y       no      no   no  telephone   \n",
       "4   56   services  married  high.school       no      no  yes  telephone   \n",
       "\n",
       "  month day_of_week  duration  campaign  pdays  previous     poutcome  \\\n",
       "0   may         mon       261         1    999         0  nonexistent   \n",
       "1   may         mon       149         1    999         0  nonexistent   \n",
       "2   may         mon       226         1    999         0  nonexistent   \n",
       "3   may         mon       151         1    999         0  nonexistent   \n",
       "4   may         mon       307         1    999         0  nonexistent   \n",
       "\n",
       "   emp.var.rate  cons.price.idx  cons.conf.idx  euribor3m  nr.employed   y  \n",
       "0           1.1          93.994          -36.4      4.857       5191.0  no  \n",
       "1           1.1          93.994          -36.4      4.857       5191.0  no  \n",
       "2           1.1          93.994          -36.4      4.857       5191.0  no  \n",
       "3           1.1          93.994          -36.4      4.857       5191.0  no  \n",
       "4           1.1          93.994          -36.4      4.857       5191.0  no  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "no     36548\n",
       "yes     4640\n",
       "Name: y, dtype: int64"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.y.value_counts() # норм такой дисбаланс"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Задание 3 (0.5 балла).__ Разделите выборку на обучающую и тестовую в соотношении 3:1. Зафиксируйте `random_state=777`, также используйте `stratify`. Один из столбцов в данных не несет информации для решения задачи — какой? Отбросьте его. \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Из описания данных:\n",
    "# 1 - duration: last contact duration, in seconds (numeric). \n",
    "#     Important note: this attribute highly affects the output target (e.g., \n",
    "#     if duration=0 then y='no'). Yet, the duration is not known before a call is performed. \n",
    "#     Also, after the end of the call y is obviously known. Thus, this input should only be included \n",
    "#     for benchmark purposes and should be discarded if the intention is to have a realistic predictive model.\n",
    "\n",
    "# уберем этот признак\n",
    "\n",
    "df = df.drop(columns=['duration'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Закодируйте категориальные признаки с помощью `OrdinalEncoder`. Посчитайте качество (в этом задании будем работать c `AUC-PR`) при применении логистической регрессии. Замерьте время, потребовавшееся на обучение модели (с учетом кодирования признаков).\n",
    "\n",
    "__Вопрос__: почему в данном задании мы выбрали метрикой именно `AUC-PR`, а не, к примеру, `AUC-ROC`?\n",
    "\n",
    "__Ваш ответ__:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "# AUC-PR точнее работает на несбалансированных выборках (см 5 лекцию)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "age                78\n",
       "job                12\n",
       "marital             4\n",
       "education           8\n",
       "default             3\n",
       "housing             3\n",
       "loan                3\n",
       "contact             2\n",
       "month              10\n",
       "day_of_week         5\n",
       "campaign           42\n",
       "pdays              27\n",
       "previous            8\n",
       "poutcome            3\n",
       "emp.var.rate       10\n",
       "cons.price.idx     26\n",
       "cons.conf.idx      26\n",
       "euribor3m         316\n",
       "nr.employed        11\n",
       "y                   2\n",
       "dtype: int64"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "# категориальные признаки (тоже взяла из описания): \n",
    "# job, marital, education, default, housing, loan, contact, month, day_of_week, poutcome\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import OrdinalEncoder \n",
    "enc = OrdinalEncoder()\n",
    "\n",
    "categorical = ['job', 'marital', 'education', 'default', 'housing', 'loan', 'contact', 'month', 'day_of_week', 'poutcome', 'y']\n",
    "\n",
    "df[categorical] = enc.fit_transform(df[categorical])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df[df.columns[:-1]]\n",
    "y = df[df.columns[-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import average_precision_score\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "y_score = lr.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.501124576890935"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Задание 4 (0.5 балла).__ Закодируйте все категориальные признаки с помощью one-hot-кодирования. Обучите логистическую регрессию и посмотрите, как изменилось качество модели (с тем, что было до кодирования). Измерьте время, потребовавшееся на кодирование категориальных признаков и обучение модели.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "df[categorical] = enc.inverse_transform(df[categorical])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "\n",
    "ohe = OneHotEncoder()\n",
    "\n",
    "x = df[df.columns[:-1]]\n",
    "y = df[df.columns[-1]]\n",
    "\n",
    "categorical_ = categorical[:-1]\n",
    "\n",
    "x = ohe.fit_transform(x[categorical_])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "lab = LabelEncoder()\n",
    "y = lab.fit_transform(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 10/10 [00:06<00:00,  1.52it/s]\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)\n",
    "\n",
    "c_best = find_best_c(X_train, y_train, X_test, y_test, LogisticRegression)\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000, C=c_best)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 285 ms, sys: 5.34 ms, total: 290 ms\n",
      "Wall time: 305 ms\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.48793188490577577"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Как можно было заменить, one-hot-кодирование может сильно увеличивать количество признаков, что сказывается на памяти, особенно, если некоторый признак имеет большое количество значений. Эту проблему решает другой способ кодирования категориальных признаков — mean-target encoding (для простоты будем называть это счётчиками). Основная идея в том, что нам важны не сами категории, а значения целевой переменной, которые имеют объекты этой категории. Каждый категориальный признак мы заменим средним значением целевой переменной по всем объектам этой же категории:\n",
    "$$\n",
    "g_j(x, X) = \\frac{\\sum_{i=1}^{l} [f_j(x) = f_j(x_i)][y_i = +1]}{\\sum_{i=1}^{l} [f_j(x) = f_j(x_i)]}\n",
    "$$\n",
    "\n",
    "__Задание 5 (1 балл).__ Закодируйте категориальные переменные с помощью счётчиков (ровно так, как описано выше без каких-либо хитростей). Обучите логистическую регрессию и посмотрите на качество модели на тестовом множестве. Сравните время обучения с предыдущими экспериментами (с учетом кодирования признаков). Заметили ли вы что-то интересное?\n",
    "\n",
    "\n",
    "__Бонус (1 балл)__: Здесь и далее реализуйте вычисление счетчиков с помощью трансформера (наследуйтесь от классов `BaseEstimator, TransformerMixin` из `sklearn.base`). Обратите внимание, что все вычисления должны быть векторизованными, трансформер не должен модифицировать передаваемую ему выборку inplace, а все необходимые статистики нужно считать только по обучающей выборке в методе `fit`. Ваш трансформер должен принимать при инициализации список из категориальных признаков и изменять только их. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "#получается, мы делим количество определенного значения признака при котором y=1 на количество значения этого признака"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 706,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_ = df.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 707,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 12.2 s, sys: 84.4 ms, total: 12.3 s\n",
      "Wall time: 12.9 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "for feature in categorical_:\n",
    "    unique = df_[feature].unique()\n",
    "    \n",
    "    for value in unique:\n",
    "        \n",
    "        value_with_one = df_[\n",
    "            (df_[feature] == value) &\n",
    "            (df_['y'] == 'yes')]['job'].count()       \n",
    "        \n",
    "        all_value_count = df_[\n",
    "            df_[feature] == value]['job'].count()\n",
    "        \n",
    "        df_.loc[df_[feature] == value, feature] = value_with_one / all_value_count\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 708,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df_[df_.columns[:-1]]\n",
    "y = df_[df_.columns[-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 709,
   "metadata": {},
   "outputs": [],
   "source": [
    "lab = LabelEncoder()\n",
    "y = lab.fit_transform(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 710,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5561082211115029"
      ]
     },
     "execution_count": 710,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)\n",
    "\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Отметим, что такие признаки сами по себе являются классификаторами и, обучаясь на них, мы допускаем \"утечку\" целевой переменной в признаки. Это ведёт к переобучению, поэтому считать такие признаки необходимо таким образом, чтобы при вычислении для конкретного объекта его целевая метка не использовалась. Это можно делать следующими способами:\n",
    "1. Вычислять значение счётчика по всем объектам расположенным выше в датасете (например, если у нас выборка отсортирована по времени).\n",
    "2. Вычислять по фолдам, то есть делить выборку на некоторое количество частей и подсчитывать значение признаков по всем фолдам кроме текущего (как делается в кросс-валидации).\n",
    "3. Внесение некоторого шума в посчитанные признаки. \n",
    "\n",
    "__Задание 6.__ Реализуйте корректное вычисление счётчиков самым простым способом — добавление шума к значениям (постарайтесь найти баланс между борьбой с переобучением и сохранением полезности признаков). Снова обучите логистическую регрессию, оцените качество. Сделайте выводы.\n",
    "\n",
    "__(0.5 балла)__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 711,
   "metadata": {},
   "outputs": [],
   "source": [
    "for feature in categorical_:\n",
    "    noise = np.random.normal(0, 10, df_.shape[0])\n",
    "    df_[feature] = df_[feature] + noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 712,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df_[df_.columns[:-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 713,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5582717758108648"
      ]
     },
     "execution_count": 713,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000) \n",
    "lr.fit(X_train, y_train)\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall # шум действительно борется с переобучением!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__(Бонусная часть, 0.5 балла)__ Посчитайте корректные счётчики первым или вторым способов из описанных выше (не забудьте добавить и шум). \n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "А теперь ответьте на следующий вопрос: что будет, если некоторая категория встречается в выборке всего несколько раз? По этой причине производится сглаживание счётчиков. Например, на практике хорошие результаты показывает использование сглаживания средним по всей выборке:\n",
    "$$\n",
    "g_j(x, X) = \\frac{\\sum_{i=1}^{\\ell} [f_j(x) = f_j(x_i)][y_i = +1] + C \\times global\\_mean}{\\sum_{i=1}^{\\ell} [f_j(x) = f_j(x_i)] + C}\n",
    "$$\n",
    "где $global\\_mean$ — доля объектов положительного класса в выборке, $C$ — параметр, определяющий степень сглаживания (например, можно использовать 10 или подобрать для каждого признака свой). Основная идея в том, что мы \"разбавляем\" среднее значение по некоторой категории глобальным средним значением. И тем меньше, чем большее количество объектов этой категории встречается в выборке. \n",
    "\n",
    "Однако для сглаживания вместо среднего значения целевой переменной можно использовать любое другое значение от 0 до 1 (этот параметр иногда называют $prior$). Можно сделать несколько признаков с разными значениями параметра. На практике в задачах бинарной классификации полезными бывают даже отрицательные значения!\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Задание 7 (1 балл).__ Добавьте сглаживание, описанное выше и повторите эксперименты."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 714,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = df.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 715,
   "metadata": {},
   "outputs": [],
   "source": [
    "global_mean = df1['y'].value_counts()['yes'] / len(df1['y'])\n",
    "c = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 716,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 12.3 s, sys: 80.5 ms, total: 12.4 s\n",
      "Wall time: 13.2 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "for feature in categorical_:\n",
    "    unique = df1[feature].unique()\n",
    "    \n",
    "    for value in unique:\n",
    "        \n",
    "        value_with_one = df1[\n",
    "            (df1[feature] == value) &\n",
    "            (df1['y'] == 'yes')]['job'].count() + c * global_mean\n",
    "        \n",
    "        all_value_count = df1[\n",
    "            df1[feature] == value]['job'].count() + c\n",
    "        \n",
    "        df1.loc[df1[feature] == value, feature] = value_with_one / all_value_count\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 717,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df1[df1.columns[:-1]]\n",
    "y = df1['y']\n",
    "\n",
    "lb = LabelEncoder()\n",
    "y = lb.fit_transform(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 718,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5584465104124989"
      ]
     },
     "execution_count": 718,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)\n",
    "\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall # пока шум выигрывает"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Задание 8 (0.5 балла).__ В данных имеется признак \"возраст клиента\". Сейчас мы интерпретируем его как числовой, что в общем случае для линейной модели может быть неверной гипотезой. Тем не менее, у этого признака есть довольно много уникальных значений (сколько?), поэтому применять к нему one-hot кодирование может оказаться излишним. Попробуйте закодировать возраст с помощью счетчиков. Стало ли лучше?\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 719,
   "metadata": {},
   "outputs": [],
   "source": [
    "unique = df['age'].unique()\n",
    "\n",
    "for value in unique:\n",
    "\n",
    "    value_with_one = df[\n",
    "        (df['age'] == value) &\n",
    "        (df['y'] == 'yes')]['job'].count() + c * global_mean\n",
    "\n",
    "    all_value_count = df[\n",
    "        df['age'] == value]['job'].count() + c\n",
    "\n",
    "    df1.loc[df_['age'] == value, 'age'] = value_with_one / all_value_count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 720,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df1[df1.columns[:-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 721,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.555026443761822"
      ]
     },
     "execution_count": 721,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)\n",
    "\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall # стало хуже"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 722,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/5b/j4k5qtvx2j7b9nzkp2clls4c0000gp/T/ipykernel_50103/240578827.py:10: RuntimeWarning: invalid value encountered in long_scalars\n",
      "  df_.loc[df_[feature] == value, feature] = value_with_one / all_value_count\n"
     ]
    }
   ],
   "source": [
    "for value in unique:\n",
    "\n",
    "    value_with_one = df_[\n",
    "        (df_[feature] == value) &\n",
    "        (df_['y'] == 'yes')]['job'].count()       \n",
    "\n",
    "    all_value_count = df_[\n",
    "        df_[feature] == value]['job'].count()\n",
    "\n",
    "    df_.loc[df_[feature] == value, feature] = value_with_one / all_value_count \n",
    "\n",
    "\n",
    "noise = np.random.normal(0, 10, df_.shape[0])\n",
    "df_['age'] = df_['age'] + noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 723,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df_[df_.columns[:-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 724,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.555026443761822"
      ]
     },
     "execution_count": 724,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000) \n",
    "lr.fit(X_train, y_train)\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Можно пойти и в обратную сторону. У нас есть признаки \"месяц и день недели\" для звонка. Попробуйте интерпретировать их как числовые (месяц от 0 до 12, дни недели от 0 до 4). Стало ли лучше в этот раз?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 725,
   "metadata": {},
   "outputs": [],
   "source": [
    "lab = LabelEncoder()\n",
    "df_['day_of_week'] = lab.fit_transform(df['day_of_week'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 726,
   "metadata": {},
   "outputs": [],
   "source": [
    "lab = LabelEncoder()\n",
    "df_['month'] = lab.fit_transform(df['month'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 727,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df_[df_.columns[:-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 728,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x,y, \n",
    "                                                  train_size=0.75, \n",
    "                                                  random_state=777, \n",
    "                                                  stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 729,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5597825712927184"
      ]
     },
     "execution_count": 729,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Замечание.__ Усложнение методов вычисления счётчиков не делают результаты модели гарантированно лучше. Особенно с учётом того, что логистическая регрессия не такая сложная модель, чтобы переобучаться. Поэтому вы необязательно должны были получать на каждом шаге всё лучшие и лучшие результаты (но необходимые результаты у вас должны были получиться).\n",
    "\n",
    "Как вы должны были заметить, счётчики являются хорошей альтернативой one-hot-кодированию. Напишите, какие плюсы и минусы использования счётчиков по сравнению с one-hot-кодированием вы заметили.\n",
    "\n",
    "__Ответ:__ \n",
    "OHE - достаточно простой способ, но дает много признаков\n",
    "Счетчики - более запарный вариант, там надо следить за утечкой данных, там много эксперементов, которые тоже требуют времени, но они могут дать кач-во получше, чем OHE (но не факт)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Часть 2. Отбор признаков"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 730,
   "metadata": {},
   "outputs": [],
   "source": [
    "numeric = ['age', \n",
    "           'campaign', \n",
    "           'pdays', \n",
    "           'previous', \n",
    "           'emp.var.rate', \n",
    "           'cons.price.idx', \n",
    "           'cons.conf.idx', \n",
    "           'euribor3m', \n",
    "           'nr.employed']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 731,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('bank-additional/bank-additional-full.csv', sep=';')\n",
    "\n",
    "X = pd.get_dummies(df[categorical_])\n",
    "X[numeric] = df[numeric]\n",
    "\n",
    "y = (df.y == 'yes')\n",
    "lb = LabelEncoder()\n",
    "y = lb.fit_transform(y)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=777, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 732,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(41188, 62)"
      ]
     },
     "execution_count": 732,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape # 62 признака"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Как вы помните, в данных много категориальных признаков (сейчас давайте интерпретировать возраст как числовой). Давайте закодируем их с помощью one-hot кодирования. Исходные колонки с категориальными признаками можно удалить. Сколько признаков мы получили?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "В качестве основной модели будем использовать логистическую регрессию, а целевой метрики — `AUC-PR`. Обучите модель и посчитайте качество на тестовой выборке. Давайте запомним полученное значение."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 733,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.510598442653082"
      ]
     },
     "execution_count": 733,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Встроенные методы"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Допустим, мы хотим оставить только 40 лучших признаков. Попробуем сделать это несколькими способами.\n",
    "\n",
    "Начнём с отборам признаков с помощью линейной модели. Как известно, веса линейной модели означают вклад каждого признака в предсказание модели, а значит, модуль этого вклада можно интерпретировать как важность признаков. Такой метод отбора называются встроенным или embedded method, так как он заложен в особенности модели.\n",
    "\n",
    "__Задание 8 (0.5 балла).__ Оставьте 40 признаков с наибольшим модулем соответствующего параметра линейной модели. Обучите модели заново и оцените её качество. Замерьте скорость такого отбора признаков.\n",
    "\n",
    "Изменилось ли качество? Как?\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 734,
   "metadata": {},
   "outputs": [],
   "source": [
    "weights = lr.coef_[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 735,
   "metadata": {},
   "outputs": [],
   "source": [
    "number = np.sort(np.abs(weights))[-40]# 2.00523338e-02"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 736,
   "metadata": {},
   "outputs": [],
   "source": [
    "weights = np.abs(weights)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 737,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40"
      ]
     },
     "execution_count": 737,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "weights_mask = weights >= number\n",
    "len(X.columns[weights_mask])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 738,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_new = X[X.columns[weights_mask]]\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_new, y, test_size=0.2, random_state=777, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 739,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.49945244152139134"
      ]
     },
     "execution_count": 739,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "А теперь давайте подумаем, что мы не учли. Мы предположили, что признаки вносят вклад равномерно, но не учли их масштаба. Если мы умножим один из признаков в 100 раз, то без учёта регуляризации его вес уменьшится в эти же 100 раз. А мы на основе этого отбираем признаки! Давайте сначала отмасштабируем признаки одним из способов, а только потом будем удалять признаки. \n",
    "\n",
    "Кстати, в таком случае надо пересчитать качество на всех признаках (сделайте это ниже). Если вы сделали нормирование признаков в самом начале, то попробуйте отобрать признаки на неотмасштабированных данных.\n",
    "\n",
    "Что получилось?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 740,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "ss = StandardScaler()\n",
    "X[numeric] = ss.fit_transform(X[numeric])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 741,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40"
      ]
     },
     "execution_count": 741,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=777, stratify=y)\n",
    "\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "\n",
    "weights = lr.coef_[0]\n",
    "\n",
    "number = np.sort(np.abs(weights))[-40]# 0.05978735905958852\n",
    "\n",
    "weights = np.abs(weights)\n",
    "weights_mask = weights >= number\n",
    "len(X.columns[weights_mask])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 742,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.49194210630288226"
      ]
     },
     "execution_count": 742,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_new = X[X.columns[weights_mask]]\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_new, y, test_size=0.2, random_state=777, stratify=y)\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Вопрос на засыпку: one-hot кодирование возвращает нам единичные признаки-индикаторы. Попробуйте также отскалировать их, как и обычные числовые, и снова выбрать 40 главных по вкладу признаков. Изменился ли их список? Изменится ли качество?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 743,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 743,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(numeric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 744,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_feature = X.columns[:-9]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 745,
   "metadata": {},
   "outputs": [],
   "source": [
    "ss = StandardScaler()\n",
    "X[new_feature] = ss.fit_transform(X[new_feature])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 746,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40"
      ]
     },
     "execution_count": 746,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=777, stratify=y)\n",
    "\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "\n",
    "weights = lr.coef_[0]\n",
    "number = np.sort(np.abs(weights))[-40]# 0.012913628932186972\n",
    "\n",
    "weights = np.abs(weights)\n",
    "weights_mask = weights >= number\n",
    "len(X.columns[weights_mask])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 747,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.49194210630288226"
      ]
     },
     "execution_count": 747,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_new = X[X.columns[weights_mask]]\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_new, y, test_size=0.2, random_state=777, stratify=y)\n",
    "\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall # не поменялось"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Методы фильтрации\n",
    "\n",
    "\n",
    "Давайте отбирать признаки умнее, а именно через подсчёт некоторой функции для каждого признака. На основании значений этой функции будем оставлять наиболее важные признаки. Методы этого семейства называют фильтрующими или  filter methods. \n",
    "\n",
    "В качестве такой функции будем считать t-статистику:\n",
    "\n",
    "$$t(j) = \\frac{|\\mu_+ - \\mu_-|}{\\sqrt{\\frac{n_+ \\sigma^2_+ + n_- \\sigma^2_-}{n_+ + n_-}}},$$\n",
    "\n",
    "где $\\mu$, $\\sigma$, $n$ соответственно среднее, стандартное отклонение и количество объектов каждого из классов.\n",
    "\n",
    "__Задание 9 (0.5 балла).__ Оставьте 40 признаков с наибольшим значением $t$ и замерьте качество. Не забудьте замерить скорость отбора признаков в этом случае.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 749,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.get_dummies(df[categorical_])\n",
    "X[numeric] = df[numeric]\n",
    "\n",
    "y = (df.y == 'yes')\n",
    "lb = LabelEncoder()\n",
    "y = lb.fit_transform(y)\n",
    "\n",
    "X['y'] = y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 750,
   "metadata": {},
   "outputs": [],
   "source": [
    "t = []\n",
    "\n",
    "for column in X.columns[:-1]:\n",
    "    plus = X[f'{column}'].loc[np.where(y==1)]\n",
    "    minus = X[f'{column}'].loc[np.where(y==0)]\n",
    "    mu_plus = np.mean(plus)\n",
    "    mu_min = np.mean(minus)\n",
    "    delta_plus = np.std(plus) ** 2\n",
    "    delta_min = np.std(minus) ** 2\n",
    "    n_plus = len(plus)\n",
    "    n_min = len(minus)\n",
    "    \n",
    "    t.append(np.abs(mu_plus - mu_min) / (np.sqrt((n_plus * delta_plus + n_min * delta_min) / (n_plus + n_min))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 751,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "39"
      ]
     },
     "execution_count": 751,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "number = np.sort(t)[-56]\n",
    "weights_mask = weights >= number\n",
    "len(X.columns[:-1][weights_mask]) # там дальше 10 признаков с одинаковым значением. Остановимся ан 39"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 752,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5027519932935146"
      ]
     },
     "execution_count": 752,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_new = X[X.columns[:-1][weights_mask]]\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_new, y, test_size=0.2, random_state=777, stratify=y)\n",
    "\n",
    "\n",
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train, y_train)\n",
    "\n",
    "y_score = lr.predict(X_test)\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc_precision_recall = auc(recall, precision)\n",
    "auc_precision_recall "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Методы-обёртки\n",
    "\n",
    "__Задание 10 (1 балл).__ \n",
    "\n",
    "Третий из рассматриваемых нами методов работает следующим образом: мы исключаем по очереди один из признаков и смотрим, как это влияет на качество. Удаляем признаки таким жадным способом, пока не окажется выполненым некоторое условие (количество признаков или ухудшение качества).\n",
    "\n",
    "Заметим, что нельзя оценивать качество по тестовой выборке, иначе мы можем переобучиться, как, например, при настройке гиперпараметров. Разделите обучающую выборку на 2 части, на одной из них обучайте модель без одного из признаков,  на второй части оценивайте качество. Исходную тестовую выборку стоит использовать только для финальной оценки качества.\n",
    "\n",
    "Снова оставьте только 40 признаков и оцените качество на тестовой выборке. Сколько времени занял такой отбор признаков? __МНОГО__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('bank-additional/bank-additional-full.csv', sep=';')\n",
    "\n",
    "X = pd.get_dummies(df[categorical_])\n",
    "X[numeric] = df[numeric]\n",
    "\n",
    "\n",
    "#X = X.drop(columns=['duration', 'y'])\n",
    "y = (df.y == 'yes')\n",
    "lb = LabelEncoder()\n",
    "y = lb.fit_transform(y)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, \n",
    "                                                    y, \n",
    "                                                    test_size=0.2, \n",
    "                                                    random_state=777, \n",
    "                                                    stratify=y)\n",
    "\n",
    "X_train1, X_test1, y_train1, y_test1 = train_test_split(X_train, \n",
    "                                                        y_train, \n",
    "                                                        test_size=0.3, \n",
    "                                                        random_state=777, \n",
    "                                                        stratify=y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "metadata": {},
   "outputs": [],
   "source": [
    "# не буду удалять, поскольку я это написала сама и ждала 20 минут пока оно обучится!!!\n",
    "\n",
    "# def find_useless(X_train, X_test, y_train, y_test):\n",
    "#     score_list = np.array([]) \n",
    "#     for column in X_train.columns:\n",
    "#         lr = LogisticRegression(max_iter=100000)\n",
    "#         lr.fit(X_train.drop(columns=[column]), y_train)\n",
    "\n",
    "#         y_score = lr.predict(X_test.drop(columns=[column]))\n",
    "#         precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "#         score_list = np.append(score_list, auc(recall, precision))\n",
    "        \n",
    "#     index_loser = np.argmin(score_list)\n",
    "#     winners = np.delete(X_train.columns, index_loser)\n",
    "#     return winners # возвращаем нужные колонки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [],
   "source": [
    "# %%time\n",
    "# useful_columns = X_train1.columns # это очень долго\n",
    "\n",
    "# for i in tqdm(range(len(X.columns) - 40)):\n",
    "#     useful_columns = find_useless(X_train1[useful_columns], \n",
    "#                                   X_test1[useful_columns], \n",
    "#                                   y_train1, \n",
    "#                                   y_test1) # Wall time: 19min 52s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#https://towardsdatascience.com/feature-selection-for-machine-learning-in-python-wrapper-methods-2b5e27d2db31"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Backward elimination"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "0it [00:09, ?it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 37.6 s, sys: 2.13 s, total: 39.8 s\n",
      "Wall time: 23.5 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "from sklearn.feature_selection import RFE\n",
    "\n",
    "model = LogisticRegression(max_iter=100000)\n",
    "#Define RFE \n",
    "rfe = RFE(model, 40)\n",
    "#Use RFE to select the top 5 features \n",
    "fit = rfe.fit(X_train1, y_train1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "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>Feature_names</th>\n",
       "      <th>Selected</th>\n",
       "      <th>RFE_ranking</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Columns</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>job_admin.</td>\n",
       "      <td>False</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>job_blue-collar</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>job_entrepreneur</td>\n",
       "      <td>False</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>job_housemaid</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>job_management</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>job_retired</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>job_self-employed</td>\n",
       "      <td>False</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>job_services</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>job_student</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>job_technician</td>\n",
       "      <td>False</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>job_unemployed</td>\n",
       "      <td>False</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>job_unknown</td>\n",
       "      <td>False</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>marital_divorced</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>marital_married</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>marital_single</td>\n",
       "      <td>False</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>marital_unknown</td>\n",
       "      <td>False</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>education_basic.4y</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>education_basic.6y</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>education_basic.9y</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>education_high.school</td>\n",
       "      <td>False</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>education_illiterate</td>\n",
       "      <td>False</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>education_professional.course</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>education_university.degree</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>education_unknown</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>default_no</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>default_unknown</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>default_yes</td>\n",
       "      <td>False</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>housing_no</td>\n",
       "      <td>False</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>housing_unknown</td>\n",
       "      <td>False</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>housing_yes</td>\n",
       "      <td>False</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>loan_no</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>loan_unknown</td>\n",
       "      <td>False</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>loan_yes</td>\n",
       "      <td>False</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>contact_cellular</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>contact_telephone</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>month_apr</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>month_aug</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>month_dec</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>month_jul</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>month_jun</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>month_mar</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>month_may</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>month_nov</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>month_oct</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>month_sep</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>day_of_week_fri</td>\n",
       "      <td>False</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>day_of_week_mon</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>day_of_week_thu</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>day_of_week_tue</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>day_of_week_wed</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>poutcome_failure</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>poutcome_nonexistent</td>\n",
       "      <td>False</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>poutcome_success</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>age</td>\n",
       "      <td>False</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>campaign</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>pdays</td>\n",
       "      <td>False</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>previous</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>emp.var.rate</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>cons.price.idx</td>\n",
       "      <td>False</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>cons.conf.idx</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>euribor3m</td>\n",
       "      <td>True</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>nr.employed</td>\n",
       "      <td>False</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         Feature_names  Selected  RFE_ranking\n",
       "Columns                                                      \n",
       "0                           job_admin.     False            9\n",
       "1                      job_blue-collar      True            1\n",
       "2                     job_entrepreneur     False           11\n",
       "3                        job_housemaid      True            1\n",
       "4                       job_management      True            1\n",
       "5                          job_retired      True            1\n",
       "6                    job_self-employed     False           12\n",
       "7                         job_services      True            1\n",
       "8                          job_student      True            1\n",
       "9                       job_technician     False           15\n",
       "10                      job_unemployed     False           21\n",
       "11                         job_unknown     False           17\n",
       "12                    marital_divorced      True            1\n",
       "13                     marital_married      True            1\n",
       "14                      marital_single     False            5\n",
       "15                     marital_unknown     False           14\n",
       "16                  education_basic.4y      True            1\n",
       "17                  education_basic.6y      True            1\n",
       "18                  education_basic.9y      True            1\n",
       "19               education_high.school     False            4\n",
       "20                education_illiterate     False           22\n",
       "21       education_professional.course      True            1\n",
       "22         education_university.degree      True            1\n",
       "23                   education_unknown      True            1\n",
       "24                          default_no      True            1\n",
       "25                     default_unknown      True            1\n",
       "26                         default_yes     False           23\n",
       "27                          housing_no     False           19\n",
       "28                     housing_unknown     False            8\n",
       "29                         housing_yes     False            6\n",
       "30                             loan_no      True            1\n",
       "31                        loan_unknown     False            7\n",
       "32                            loan_yes     False           13\n",
       "33                    contact_cellular      True            1\n",
       "34                   contact_telephone      True            1\n",
       "35                           month_apr      True            1\n",
       "36                           month_aug      True            1\n",
       "37                           month_dec      True            1\n",
       "38                           month_jul      True            1\n",
       "39                           month_jun      True            1\n",
       "40                           month_mar      True            1\n",
       "41                           month_may      True            1\n",
       "42                           month_nov      True            1\n",
       "43                           month_oct      True            1\n",
       "44                           month_sep      True            1\n",
       "45                     day_of_week_fri     False           16\n",
       "46                     day_of_week_mon      True            1\n",
       "47                     day_of_week_thu      True            1\n",
       "48                     day_of_week_tue      True            1\n",
       "49                     day_of_week_wed      True            1\n",
       "50                    poutcome_failure      True            1\n",
       "51                poutcome_nonexistent     False            2\n",
       "52                    poutcome_success      True            1\n",
       "53                                 age     False           20\n",
       "54                            campaign      True            1\n",
       "55                               pdays     False           18\n",
       "56                            previous      True            1\n",
       "57                        emp.var.rate      True            1\n",
       "58                      cons.price.idx     False            3\n",
       "59                       cons.conf.idx      True            1\n",
       "60                           euribor3m      True            1\n",
       "61                         nr.employed     False           10"
      ]
     },
     "execution_count": 252,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Create a dataframe for the results \n",
    "df_RFE_results = []\n",
    "for i in range(X_train1.shape[1]):\n",
    "    df_RFE_results.append(\n",
    "        {      \n",
    "            'Feature_names': X_train1.columns[i],\n",
    "            'Selected':  rfe.support_[i],\n",
    "            'RFE_ranking':  rfe.ranking_[i],\n",
    "        }\n",
    "    )\n",
    "\n",
    "df_RFE_results = pd.DataFrame(df_RFE_results)\n",
    "df_RFE_results.index.name='Columns'\n",
    "df_RFE_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_columns = list(df_RFE_results[df_RFE_results['Selected'] == True]['Feature_names'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5004502790563247"
      ]
     },
     "execution_count": 259,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train[best_columns], y_train)\n",
    "\n",
    "y_score = lr.predict(X_test[best_columns])\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc(recall, precision) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Стоит отметить, что с помощью такого метода можно пойти и в обратную сторону. Попробуйте _добавлять_ по одному самому полезному признаку в выборку до тех пор, пока не наберется 40 штук. Найдется ли порог, при котором добавление следующих признаков будет только ухудшать качество модели?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "metadata": {},
   "outputs": [],
   "source": [
    "# !pip install mlxtend"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [],
   "source": [
    "# та же ссылка"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Forward Selection "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "metadata": {},
   "outputs": [],
   "source": [
    "from mlxtend.feature_selection import SequentialFeatureSelector as SFS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 14min 15s, sys: 58.7 s, total: 15min 14s\n",
      "Wall time: 8min 57s\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature_idx</th>\n",
       "      <th>cv_scores</th>\n",
       "      <th>avg_score</th>\n",
       "      <th>feature_names</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>(52,)</td>\n",
       "      <td>[-0.026301261991317437]</td>\n",
       "      <td>-0.026301</td>\n",
       "      <td>(poutcome_success,)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>(52, 57)</td>\n",
       "      <td>[-0.01805962041150555]</td>\n",
       "      <td>-0.01806</td>\n",
       "      <td>(poutcome_success, emp.var.rate)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>(52, 55, 57)</td>\n",
       "      <td>[-0.015456996754722896]</td>\n",
       "      <td>-0.015457</td>\n",
       "      <td>(poutcome_success, pdays, emp.var.rate)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>(41, 52, 55, 57)</td>\n",
       "      <td>[-0.012420602488476318]</td>\n",
       "      <td>-0.012421</td>\n",
       "      <td>(month_may, poutcome_success, pdays, emp.var.r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>(41, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.00938420822222974]</td>\n",
       "      <td>-0.009384</td>\n",
       "      <td>(month_may, poutcome_failure, poutcome_success...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>(5, 41, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.008082896393838412]</td>\n",
       "      <td>-0.008083</td>\n",
       "      <td>(job_retired, month_may, poutcome_failure, pou...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>(5, 17, 41, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.007649125784374711]</td>\n",
       "      <td>-0.007649</td>\n",
       "      <td>(job_retired, education_basic.6y, month_may, p...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>(0, 5, 17, 41, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.007649125784374711]</td>\n",
       "      <td>-0.007649</td>\n",
       "      <td>(job_admin., job_retired, education_basic.6y, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>(0, 5, 17, 40, 41, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.007215355174910787]</td>\n",
       "      <td>-0.007215</td>\n",
       "      <td>(job_admin., job_retired, education_basic.6y, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>(0, 1, 5, 17, 40, 41, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.006781584565447085]</td>\n",
       "      <td>-0.006782</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, edu...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>(0, 1, 5, 8, 17, 40, 41, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.0046127315181281325]</td>\n",
       "      <td>-0.004613</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>(0, 1, 5, 8, 17, 40, 41, 48, 50, 52, 55, 57)</td>\n",
       "      <td>[-0.0033114196897368053]</td>\n",
       "      <td>-0.003311</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>(0, 1, 5, 8, 17, 40, 41, 48, 50, 52, 54, 55, 57)</td>\n",
       "      <td>[-0.0028776490802731036]</td>\n",
       "      <td>-0.002878</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>(0, 1, 5, 8, 17, 37, 40, 41, 48, 50, 52, 54, 5...</td>\n",
       "      <td>[-0.002010107861345478]</td>\n",
       "      <td>-0.00201</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>(0, 1, 5, 8, 17, 37, 40, 41, 48, 50, 52, 54, 5...</td>\n",
       "      <td>[-0.0007087960329541509]</td>\n",
       "      <td>-0.000709</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>(0, 1, 5, 8, 12, 17, 37, 40, 41, 48, 50, 52, 5...</td>\n",
       "      <td>[0.0014600570143648017]</td>\n",
       "      <td>0.00146</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>(0, 1, 5, 8, 12, 17, 33, 37, 40, 41, 48, 50, 5...</td>\n",
       "      <td>[0.002327598233292427]</td>\n",
       "      <td>0.002328</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>(0, 1, 5, 8, 10, 12, 17, 33, 37, 40, 41, 48, 5...</td>\n",
       "      <td>[0.004062680671147567]</td>\n",
       "      <td>0.004063</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>(0, 1, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41, 4...</td>\n",
       "      <td>[0.0036289100616837544]</td>\n",
       "      <td>0.003629</td>\n",
       "      <td>(job_admin., job_blue-collar, job_retired, job...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>(0, 1, 2, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41...</td>\n",
       "      <td>[0.00276136884275624]</td>\n",
       "      <td>0.002761</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>(0, 1, 2, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41...</td>\n",
       "      <td>[0.005363992499538894]</td>\n",
       "      <td>0.005364</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>(0, 1, 2, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41...</td>\n",
       "      <td>[0.0031951394522199417]</td>\n",
       "      <td>0.003195</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>(0, 1, 2, 5, 8, 10, 11, 12, 17, 25, 33, 37, 40...</td>\n",
       "      <td>[0.00449645128061138]</td>\n",
       "      <td>0.004496</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>(0, 1, 2, 5, 8, 10, 11, 12, 17, 24, 25, 33, 37...</td>\n",
       "      <td>[0.0031951394522199417]</td>\n",
       "      <td>0.003195</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>(0, 1, 2, 3, 5, 8, 10, 11, 12, 17, 24, 25, 33,...</td>\n",
       "      <td>[0.0031951394522199417]</td>\n",
       "      <td>0.003195</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>(0, 1, 2, 3, 5, 8, 10, 11, 12, 17, 24, 25, 30,...</td>\n",
       "      <td>[0.0036289100616837544]</td>\n",
       "      <td>0.003629</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 8, 10, 11, 12, 17, 24, 25, ...</td>\n",
       "      <td>[0.0031951394522199417]</td>\n",
       "      <td>0.003195</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 17, 24, 2...</td>\n",
       "      <td>[0.0031951394522199417]</td>\n",
       "      <td>0.003195</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 17, 24, 2...</td>\n",
       "      <td>[0.004062680671147567]</td>\n",
       "      <td>0.004063</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 17, 24, 2...</td>\n",
       "      <td>[0.00449645128061138]</td>\n",
       "      <td>0.004496</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...</td>\n",
       "      <td>[0.00449645128061138]</td>\n",
       "      <td>0.004496</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...</td>\n",
       "      <td>[0.0036289100616837544]</td>\n",
       "      <td>0.003629</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...</td>\n",
       "      <td>[0.0014600570143648017]</td>\n",
       "      <td>0.00146</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...</td>\n",
       "      <td>[0.0036289100616837544]</td>\n",
       "      <td>0.003629</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...</td>\n",
       "      <td>[0.0018938276238286145]</td>\n",
       "      <td>0.001894</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 16, 1...</td>\n",
       "      <td>[0.0036289100616837544]</td>\n",
       "      <td>0.003629</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>(0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 16, 1...</td>\n",
       "      <td>[0.00449645128061138]</td>\n",
       "      <td>0.004496</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>(0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 15, 16...</td>\n",
       "      <td>[0.0014600570143648017]</td>\n",
       "      <td>0.00146</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>(0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 15, 16...</td>\n",
       "      <td>[0.0018938276238286145]</td>\n",
       "      <td>0.001894</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>(0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 15, 16...</td>\n",
       "      <td>[0.0014600570143648017]</td>\n",
       "      <td>0.00146</td>\n",
       "      <td>(job_admin., job_blue-collar, job_entrepreneur...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                          feature_idx  \\\n",
       "1                                               (52,)   \n",
       "2                                            (52, 57)   \n",
       "3                                        (52, 55, 57)   \n",
       "4                                    (41, 52, 55, 57)   \n",
       "5                                (41, 50, 52, 55, 57)   \n",
       "6                             (5, 41, 50, 52, 55, 57)   \n",
       "7                         (5, 17, 41, 50, 52, 55, 57)   \n",
       "8                      (0, 5, 17, 41, 50, 52, 55, 57)   \n",
       "9                  (0, 5, 17, 40, 41, 50, 52, 55, 57)   \n",
       "10              (0, 1, 5, 17, 40, 41, 50, 52, 55, 57)   \n",
       "11           (0, 1, 5, 8, 17, 40, 41, 50, 52, 55, 57)   \n",
       "12       (0, 1, 5, 8, 17, 40, 41, 48, 50, 52, 55, 57)   \n",
       "13   (0, 1, 5, 8, 17, 40, 41, 48, 50, 52, 54, 55, 57)   \n",
       "14  (0, 1, 5, 8, 17, 37, 40, 41, 48, 50, 52, 54, 5...   \n",
       "15  (0, 1, 5, 8, 17, 37, 40, 41, 48, 50, 52, 54, 5...   \n",
       "16  (0, 1, 5, 8, 12, 17, 37, 40, 41, 48, 50, 52, 5...   \n",
       "17  (0, 1, 5, 8, 12, 17, 33, 37, 40, 41, 48, 50, 5...   \n",
       "18  (0, 1, 5, 8, 10, 12, 17, 33, 37, 40, 41, 48, 5...   \n",
       "19  (0, 1, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41, 4...   \n",
       "20  (0, 1, 2, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41...   \n",
       "21  (0, 1, 2, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41...   \n",
       "22  (0, 1, 2, 5, 8, 10, 12, 17, 25, 33, 37, 40, 41...   \n",
       "23  (0, 1, 2, 5, 8, 10, 11, 12, 17, 25, 33, 37, 40...   \n",
       "24  (0, 1, 2, 5, 8, 10, 11, 12, 17, 24, 25, 33, 37...   \n",
       "25  (0, 1, 2, 3, 5, 8, 10, 11, 12, 17, 24, 25, 33,...   \n",
       "26  (0, 1, 2, 3, 5, 8, 10, 11, 12, 17, 24, 25, 30,...   \n",
       "27  (0, 1, 2, 3, 5, 6, 8, 10, 11, 12, 17, 24, 25, ...   \n",
       "28  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 17, 24, 2...   \n",
       "29  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 17, 24, 2...   \n",
       "30  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 17, 24, 2...   \n",
       "31  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...   \n",
       "32  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...   \n",
       "33  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...   \n",
       "34  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...   \n",
       "35  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 17, 2...   \n",
       "36  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 16, 1...   \n",
       "37  (0, 1, 2, 3, 5, 6, 7, 8, 10, 11, 12, 15, 16, 1...   \n",
       "38  (0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 15, 16...   \n",
       "39  (0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 15, 16...   \n",
       "40  (0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 15, 16...   \n",
       "\n",
       "                   cv_scores avg_score  \\\n",
       "1    [-0.026301261991317437] -0.026301   \n",
       "2     [-0.01805962041150555]  -0.01806   \n",
       "3    [-0.015456996754722896] -0.015457   \n",
       "4    [-0.012420602488476318] -0.012421   \n",
       "5     [-0.00938420822222974] -0.009384   \n",
       "6    [-0.008082896393838412] -0.008083   \n",
       "7    [-0.007649125784374711] -0.007649   \n",
       "8    [-0.007649125784374711] -0.007649   \n",
       "9    [-0.007215355174910787] -0.007215   \n",
       "10   [-0.006781584565447085] -0.006782   \n",
       "11  [-0.0046127315181281325] -0.004613   \n",
       "12  [-0.0033114196897368053] -0.003311   \n",
       "13  [-0.0028776490802731036] -0.002878   \n",
       "14   [-0.002010107861345478]  -0.00201   \n",
       "15  [-0.0007087960329541509] -0.000709   \n",
       "16   [0.0014600570143648017]   0.00146   \n",
       "17    [0.002327598233292427]  0.002328   \n",
       "18    [0.004062680671147567]  0.004063   \n",
       "19   [0.0036289100616837544]  0.003629   \n",
       "20     [0.00276136884275624]  0.002761   \n",
       "21    [0.005363992499538894]  0.005364   \n",
       "22   [0.0031951394522199417]  0.003195   \n",
       "23     [0.00449645128061138]  0.004496   \n",
       "24   [0.0031951394522199417]  0.003195   \n",
       "25   [0.0031951394522199417]  0.003195   \n",
       "26   [0.0036289100616837544]  0.003629   \n",
       "27   [0.0031951394522199417]  0.003195   \n",
       "28   [0.0031951394522199417]  0.003195   \n",
       "29    [0.004062680671147567]  0.004063   \n",
       "30     [0.00449645128061138]  0.004496   \n",
       "31     [0.00449645128061138]  0.004496   \n",
       "32   [0.0036289100616837544]  0.003629   \n",
       "33   [0.0014600570143648017]   0.00146   \n",
       "34   [0.0036289100616837544]  0.003629   \n",
       "35   [0.0018938276238286145]  0.001894   \n",
       "36   [0.0036289100616837544]  0.003629   \n",
       "37     [0.00449645128061138]  0.004496   \n",
       "38   [0.0014600570143648017]   0.00146   \n",
       "39   [0.0018938276238286145]  0.001894   \n",
       "40   [0.0014600570143648017]   0.00146   \n",
       "\n",
       "                                        feature_names  \n",
       "1                                 (poutcome_success,)  \n",
       "2                    (poutcome_success, emp.var.rate)  \n",
       "3             (poutcome_success, pdays, emp.var.rate)  \n",
       "4   (month_may, poutcome_success, pdays, emp.var.r...  \n",
       "5   (month_may, poutcome_failure, poutcome_success...  \n",
       "6   (job_retired, month_may, poutcome_failure, pou...  \n",
       "7   (job_retired, education_basic.6y, month_may, p...  \n",
       "8   (job_admin., job_retired, education_basic.6y, ...  \n",
       "9   (job_admin., job_retired, education_basic.6y, ...  \n",
       "10  (job_admin., job_blue-collar, job_retired, edu...  \n",
       "11  (job_admin., job_blue-collar, job_retired, job...  \n",
       "12  (job_admin., job_blue-collar, job_retired, job...  \n",
       "13  (job_admin., job_blue-collar, job_retired, job...  \n",
       "14  (job_admin., job_blue-collar, job_retired, job...  \n",
       "15  (job_admin., job_blue-collar, job_retired, job...  \n",
       "16  (job_admin., job_blue-collar, job_retired, job...  \n",
       "17  (job_admin., job_blue-collar, job_retired, job...  \n",
       "18  (job_admin., job_blue-collar, job_retired, job...  \n",
       "19  (job_admin., job_blue-collar, job_retired, job...  \n",
       "20  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "21  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "22  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "23  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "24  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "25  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "26  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "27  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "28  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "29  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "30  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "31  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "32  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "33  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "34  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "35  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "36  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "37  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "38  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "39  (job_admin., job_blue-collar, job_entrepreneur...  \n",
       "40  (job_admin., job_blue-collar, job_entrepreneur...  "
      ]
     },
     "execution_count": 266,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "sfs = SFS(LogisticRegression(max_iter=100000),\n",
    "           k_features=40,\n",
    "           forward=True,\n",
    "           floating=False,\n",
    "           scoring = 'r2',\n",
    "           cv = 0)\n",
    "#Use SFS to select the top 40 features \n",
    "sfs.fit(X_train1, y_train1)\n",
    "\n",
    "#Create a dataframe for the SFS results \n",
    "df_SFS_results = pd.DataFrame(sfs.subsets_).transpose()\n",
    "df_SFS_results #оставлю эту попытку, потому что она тоже долго работала"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "metadata": {},
   "outputs": [],
   "source": [
    "best_columns_1 = list(df_SFS_results['feature_names'][40])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.4869238510164684"
      ]
     },
     "execution_count": 274,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression(max_iter=100000)\n",
    "lr.fit(X_train[best_columns_1], y_train)\n",
    "\n",
    "y_score = lr.predict(X_test[best_columns_1])\n",
    "precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "auc(recall, precision) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Давайте подведём итоги по отбору признаков. Назовите преимущества и недостатки каждого из методов. Какой метод привёл к наилучшему качеству?\n",
    "\n",
    "**Ответ:** К наилучшему качеству привели счетчики\n",
    "ОГРОМНЫМ МИНУСОМ методов-оберток является количество времени, которое они сжирают, в данной задаче они еще и особо не помогли.\n",
    "\n",
    "Встроенные методы оказались быстрыми, но не эффективынми,метод фильтрации тоже быстрый и чуть лучше\n",
    "Но все же не дотянул"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Часть 3. Оценка экономического эффекта модели\n",
    "\n",
    "В данной части мы займемся тем, что от вас скорее всего потребуется на реальной работе (помимо перекладки `json`, разумеется). А именно: мы соберем несколько специализированных метрик качества, попытаемся настроить модель на максимизацию _прибыли_ и оценим, сколько вообще получится заработать на этом. Разумеется, здесь будет сделано множество упрощающих жизнь допущений, но обо всем по порядку. \n",
    "\n",
    "__Задание 11 (1 балл).__ Допустим, работники вашего колл-центра получают за один звонок клиенту 1 доллар. При согласии клиента на предлагаемые условия он принесет в банк 10 долларов.\n",
    "\n",
    "Если вы всё прослушали на экономике, то напомним, что выручка — это сколько денег нам принесли клиенты, а прибыль — выручка за вычетом расходов на зарплату и прочее.\n",
    "\n",
    "Загрузите данные о телемаркетинге из предыдущего блока заданий. В этой части не нужно делить выборку - мы будем использовать кросс-валидацию. Используйте 5 фолдов, сделайте `shuffle=True, random_state=500`. По кросс-валидации у вас получится 5 вариантов обучающей и тестовой выборки. Обучите логистическую регрессию на каждой обучающей выборке (воспользуйтесь one-hot для категориальных признаков, гиперпараметры оставьте со значениями по умолчанию) и сделайте предсказания для соответствующих тестовых выборок. Допустим, всем положительным прогнозам ваши сотрудники решили позвонить. Посчитайте на всех тестовых выборках выручку и усредните. Сколько денег вы в среднем заработаете? Также вычислите стандартное отклонение.\n",
    "\n",
    "Сколько из заработанных денег придётся отдать операторам вашего колл-центра?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 753,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.compose import ColumnTransformer\n",
    "\n",
    "\n",
    "column_transformer = ColumnTransformer([\n",
    "    ('ohe', OneHotEncoder(handle_unknown='ignore'), categorical_),\n",
    "    ('scaling', StandardScaler(), numeric)\n",
    "   ])\n",
    "\n",
    "x = column_transformer.fit_transform(df) \n",
    "\n",
    "y = df[df.columns[-1]]\n",
    "lb = LabelEncoder()\n",
    "y = lb.fit_transform(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 755,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TRAIN: [    0     1     2 ... 41185 41186 41187] TEST: [    5     6    10 ... 41160 41176 41177]\n",
      "310\n",
      "TRAIN: [    1     2     3 ... 41184 41185 41186] TEST: [    0     8    21 ... 41164 41183 41187]\n",
      "330\n",
      "TRAIN: [    0     1     3 ... 41183 41184 41187] TEST: [    2     4     7 ... 41181 41185 41186]\n",
      "330\n",
      "TRAIN: [    0     1     2 ... 41185 41186 41187] TEST: [    3    11    14 ... 41178 41182 41184]\n",
      "330\n",
      "TRAIN: [    0     2     3 ... 41185 41186 41187] TEST: [    1     9    12 ... 41168 41170 41179]\n",
      "320\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import KFold\n",
    "# https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.KFold.html\n",
    "\n",
    "kf = KFold(n_splits=5, random_state=500, shuffle=True)\n",
    "\n",
    "scores = []\n",
    "money = []\n",
    "\n",
    "for train_index, test_index in kf.split(x):\n",
    "    print(\"TRAIN:\", train_index, \"TEST:\", test_index)\n",
    "    X_train, X_test = x[train_index], x[test_index]\n",
    "    y_train, y_test = y[train_index], y[test_index]\n",
    "    lr = LogisticRegression(max_iter=100000)\n",
    "    lr.fit(X_train, y_train)\n",
    "\n",
    "    y_score = lr.predict(X_test)\n",
    "    print(sum(y_score))\n",
    "    precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "    scores.append(auc(recall, precision))\n",
    "    \n",
    "    money.append(sum((y_score == y_test) & (y_test == 1)) * 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 756,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2128.0"
      ]
     },
     "execution_count": 756,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(money) # примерно столько мы заработали"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 757,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "110.34491379306978"
      ]
     },
     "execution_count": 757,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.std(money) # стандартное отклонение"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 758,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "212.8"
      ]
     },
     "execution_count": 758,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(money) / 10 # столько мы отдадим сотрудникам"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Внесем некоторую долю случайности. Пусть теперь согласный на условия клиент будет приносить не 10 долларов, а случайную величину, равномерно распределенную в интервале $[0;20)$. Проделайте все те же самые действия. Для имитации реальной ситуации **НЕ** фиксируйте `random_seed` при подсчете выручки с клиента. Что получилось?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 759,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TRAIN: [    0     1     2 ... 41185 41186 41187] TEST: [    5     6    10 ... 41160 41176 41177]\n",
      "TRAIN: [    1     2     3 ... 41184 41185 41186] TEST: [    0     8    21 ... 41164 41183 41187]\n",
      "TRAIN: [    0     1     3 ... 41183 41184 41187] TEST: [    2     4     7 ... 41181 41185 41186]\n",
      "TRAIN: [    0     1     2 ... 41185 41186 41187] TEST: [    3    11    14 ... 41178 41182 41184]\n",
      "TRAIN: [    0     2     3 ... 41185 41186 41187] TEST: [    1     9    12 ... 41168 41170 41179]\n"
     ]
    }
   ],
   "source": [
    "kf = KFold(n_splits=5, random_state=500, shuffle=True)\n",
    "\n",
    "scores = []\n",
    "money = []\n",
    "\n",
    "for train_index, test_index in kf.split(x):\n",
    "    print(\"TRAIN:\", train_index, \"TEST:\", test_index)\n",
    "    X_train, X_test = x[train_index], x[test_index]\n",
    "    y_train, y_test = y[train_index], y[test_index]\n",
    "    lr = LogisticRegression(max_iter=100000)\n",
    "    lr.fit(X_train, y_train)\n",
    "\n",
    "    y_score = lr.predict(X_test)\n",
    "    precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "    scores.append(auc(recall, precision))\n",
    "    money.append(sum((y_score == y_test) & (y_test == 1)) * int(np.random.uniform(0, 20)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 760,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[3136, 1266, 1045, 4275, 3510]"
      ]
     },
     "execution_count": 760,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "money"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 761,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2646.4"
      ]
     },
     "execution_count": 761,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(money) # примерно столько мы заработали"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 762,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1273.4078058501134"
      ]
     },
     "execution_count": 762,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.std(money) # стандартное отклонение"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 763,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "264.64"
      ]
     },
     "execution_count": 763,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(money) / 10 # столько мы отдадим сотрудникам"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Настройте по кросс-валидации коэффициент регуляризации модели для максимизации прибыли (считайте как случайную величину выше). Удалось ли получить какой-то выигрыш? При каком коэффициенте регуляризациии прибыль максимальна? Постройте график зависимости ожидаемой прибыли от коэффициента, также укажите стандартные отклонения (вам поможет `plt.errorbar`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 764,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 16s, sys: 3.9 s, total: 1min 19s\n",
      "Wall time: 45.2 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "kf = KFold(n_splits=5, random_state=500, shuffle=True)\n",
    "\n",
    "\n",
    "c_best_money = []\n",
    "c_best_std = []\n",
    "c_best_score = []\n",
    "\n",
    "for c in np.linspace(0.0001, 1, 10):\n",
    "    scores = []\n",
    "    money = []\n",
    "    for train_index, test_index in kf.split(x):\n",
    "        X_train, X_test = x[train_index], x[test_index]\n",
    "        y_train, y_test = y[train_index], y[test_index]\n",
    "        lr = LogisticRegression(max_iter=100000, C=c)\n",
    "        lr.fit(X_train, y_train)\n",
    "\n",
    "        y_score = lr.predict(X_test)\n",
    "        precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "        scores.append(auc(recall, precision))\n",
    "        money.append(sum((y_score == y_test) & (y_test == 1))* int(np.random.uniform(0, 20)))\n",
    "        \n",
    "    c_best_money.append(np.mean(money))\n",
    "    c_best_std.append(np.std(money)) \n",
    "    c_best_score.append(np.mean(scores))\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 765,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<ErrorbarContainer object of 3 artists>"
      ]
     },
     "execution_count": 765,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "yerr = c_best_std\n",
    "plt.errorbar(np.linspace(0.0001, 1, 10), c_best_money, yerr=yerr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Попробуйте запустить перебор несколько раз. Находится ли каждый раз один и тот же \"лучший\" коэффициент? Присутствует ли какая-то закономерность? Какие вы можете сделать из этого выводы?\n",
    "\n",
    "Каждый раз по новому(\n",
    "Видимо, порог не так сильно влияет на результат и выигрывает вероятность оплаты"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 766,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD8CAYAAACb4nSYAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAsTAAALEwEAmpwYAAAvBUlEQVR4nO3deXhV5bX48e/KzJgwhAAZGCSAYYYQoQqiKOKItfYKVsWp2BZ7tdre6rX3p9V6byf11ltbi0JFK4OzqLSCSkSrzJMmEIhMSRgSCAmEQEKS9fvj7NCIgZyQM+ScvT7PkyfnvHufvdeGZJ2d913nfUVVMcYY4w4RwQ7AGGNM4FjSN8YYF7Gkb4wxLmJJ3xhjXMSSvjHGuIglfWOMcRGvk76IRIrIehF513neR0RWiki+iCwUkRinPdZ5nu9s793gGA867XkicpnPr8YYY8wZNedO/x5gc4PnvwGeUtV+wCHgDqf9DuCQ0/6Usx8ikgFMBQYBk4E/iUhky8I3xhjTHF4lfRFJAa4EnneeC3Ax8Jqzy1zgWufxFOc5zvaJzv5TgAWqWqWqO4B8IMsH12CMMcZLUV7u97/AfwAdnOddgDJVrXGeFwLJzuNkoABAVWtEpNzZPxlY0eCYDV/TqK5du2rv3r29DNEYYwzA2rVrD6hqYmPbmkz6InIVUKyqa0Vkgo9ja+x8M4AZAGlpaaxZs8bfpzTGmLAiIrtOt82b7p3zgWtEZCewAE+3zh+ABBGpf9NIAYqcx0VAqnPiKCAeONiwvZHXnKSqs1Q1U1UzExMbfaMyxhhzlppM+qr6oKqmqGpvPAOxH6nq94BlwPXObtOBt53Hi5znONs/Us+sbouAqU51Tx8gHVjlsysxxhjTJG/79Bvzc2CBiPwKWA/MdtpnAy+JSD5QiueNAlXNEZFXgFygBpipqrUtOL8xxphmktY8tXJmZqZan74xxjSPiKxV1czGttknco0xxkUs6RtjjItY0jfGGBexpG+MMS5iSd8YE3Q3/OVzbvjL58EOwxUs6RtjjItY0jfGGBexpG+MMS5iSd8YY1zEkr4xxriIJX1jjHERS/rGGOMilvSNMcZFLOkbY4yLWNI3xhgXsaRvjDEuYknfj2w+EWNMa9Nk0heROBFZJSIbRSRHRH7ptL8gIjtEZIPzNdxpFxF5WkTyRWSTiIxscKzpIrLN+Zp+mlMaY4zxE2/WyK0CLlbVChGJBj4Vkb87236mqq+dsv/leBY9TwfOA/4MnCcinYGHgUxAgbUiskhVD/niQszp1f+1sfCusUGOxBgTbE3e6atHhfM02vk608K6U4AXndetABJEpAdwGbBUVUudRL8UmNyy8I0xxjSHV336IhIpIhuAYjyJe6Wz6XGnC+cpEYl12pKBggYvL3TaTtdujDEmQLxK+qpaq6rDgRQgS0QGAw8CA4HRQGfg574ISERmiMgaEVlTUlLii0MaY4xxNKt6R1XLgGXAZFXd63ThVAF/BbKc3YqA1AYvS3HaTtd+6jlmqWqmqmYmJiY2JzxjjDFN8KZ6J1FEEpzHbYBLgS1OPz0iIsC1wJfOSxYBtzhVPGOAclXdC7wPTBKRTiLSCZjktBljjAkQb6p3egBzRSQSz5vEK6r6roh8JCKJgAAbgB84+y8GrgDygUrgNgBVLRWRx4DVzn6Pqmqpz67EGGNMk5pM+qq6CRjRSPvFp9lfgZmn2TYHmNPMGI0xxviIfSLXGGNcxJK+Mca4iCV9Y4xxEUv6xhjjIpb0jTHGRSzpG2OMi1jSN8YYF7Gkb4wxrYw/F2CypG+MMS5iSd8YY1zEkr4xxriIJX1jXM6f/cem9bGkb4wxLmJJ3xhjXMSSvjHGuIglfWOMcRFvlkuME5FVIrJRRHJE5JdOex8RWSki+SKyUERinPZY53m+s713g2M96LTnichlfrsqY4wxjfLmTr8KuFhVhwHDgcnO2re/AZ5S1X7AIeAOZ/87gENO+1POfohIBjAVGARMBv7kLMFojDEmQJpM+upR4TyNdr4UuBh4zWmfi2dxdIApznOc7ROdxdOnAAtUtUpVd+BZQzfLFxdhjDHGO1716YtIpIhsAIqBpcBXQJmq1ji7FALJzuNkoADA2V4OdGnY3shrjDHGBIBXSV9Va1V1OJCC5+58oL8CEpEZIrJGRNaUlJT46zTGGONKzareUdUyYBkwFkgQkShnUwpQ5DwuAlIBnO3xwMGG7Y28puE5ZqlqpqpmJiYmNic8Y4wxTfCmeidRRBKcx22AS4HNeJL/9c5u04G3nceLnOc42z9SVXXapzrVPX2AdGCVj67DGGOMF6Ka3oUewFyn0iYCeEVV3xWRXGCBiPwKWA/MdvafDbwkIvlAKZ6KHVQ1R0ReAXKBGmCmqtb69nKMMcacSZNJX1U3ASMaad9OI9U3qnoc+O5pjvU48HjzwzTGhLOa2jpKK09QV6dEREiwwwlr9olcY0zQ7SqtZMeBoyzLKw52KGHPkr4xJqjW7T7EgYpqAOav2h3kaMKfJX1jTNDU1SmPLMohOlJI6hjLR1uK2Vt+LNhhhTVL+saYoHllTQGbCstJ69yW7h3jqFN4ZXVhsMMKa5b0jTFBUV55gt++n0dW7850aRdDXHQk49K7snD1bmrrNNjhhS1L+sZ1bHnA1uGpD7ZSVlnNI9cMwjM9F9yYlcae8uMs32qfxvcXS/rGmIDbsu8wL63YxffO60VGz44n2y/JSKJr+1jm2YCu31jSN8YElKry8Ns5dIyL4v5J/b+2LToygu9mpvDRlmL2lR8PUoThzZK+MSag3t20l5U7SvnpZQNIaBvzje1TR6dSW6e8uqagkVeblrKkb4wJmMrqGv578WYGJ3dk6ui0Rvfp1aUd5/frwoLVBdTZgK7PWdI3xgTMM8vy2Vt+nF9eM4jIM0y3MC0rjaKyY3ySfyCA0bmDJX1jTEDsPHCU55bv4LoRyYzq1fmM+07K6E6XdjHMX2kDur5mSd8YExCPvZtLTFQED1ze9BpMMVERXD8qhQ8276f4sA3o+pIlfWOM3y3bUsyHW4r594n96NYxzqvX3DA6lZo65dW19gldX7Kkb4zxq6qaWn75Tg59E9tx67f6eP26vontGdu3CwtW77YBXR+ypG+M8avZn+5g58FKHrl6EDFRzUs5085Lo6D0GP/8ygZ0fcWSfhg7WlXD37/YS35xBZv3Hg5636hNf+A+e8uP8X8f5jMpI4nx/Zu/5vVlg5Lo1Dbaplz2IW/WyE0VkWUikisiOSJyj9P+iIgUicgG5+uKBq95UETyRSRPRC5r0D7ZacsXkQf8c0nuVnzkOPNX7ea2v65ixGNL+eHL6yg/doKKqhq+9/xKSo9WBztE4yL/vXgLdar811UZZ/X62KhIvjMyhSU5+yk5UuXj6NzJmzVya4D7VXWdiHQA1orIUmfbU6r6+4Y7i0gGnnVxBwE9gQ9EpP6z1s/gWVi9EFgtIotUNdcXF+Jm+cUVLM3dz5LcfWwoKEMVUju34abzejFpUBJPLsnjyPEath84ys2zVzLv+2OIbxMd7LBdr/6vnoV3jQ1yJP6xYvtB3tm4h3+fmE5q57ZnfZypWWk8/+kOXl9XyA8uPMeHEbqTN2vk7gX2Oo+PiMhmIPkML5kCLFDVKmCHs0B6/Vq6+c7auojIAmdfS/rNVFenrC84xJLc/SzN3c/2kqMADEmO5yeX9GfSoCQGJHU4OXOhiNCxTTTP3jSKGS+t4fYXVvPi7Vm0i/XmPd+Y5qupreORRTkkJ7Thhy1M1P26tSerT2cWrNrNjHF9bQ3dFmrWb72I9MazSPpK4HzgbhG5BViD56+BQ3jeEFY0eFkh/3qTKDil/byzC9t9jp+o5Z/5B1iau58PNhdzoKKKqAhh7DlduPVbvbnk3CR6JrQ54zEuGtiNp6eOYOa8dXz/xTXMuXU0cdGRAboC4yYvr9zNln1H+PP3RtImpuU/YzdmpXHvwg2s2H6Qb/Xr6oMIGxfuf31BM5K+iLQHXgfuVdXDIvJn4DFAne9PALe3NCARmQHMAEhLa3xuDrcoq6zmoy3FLMnZz/JtJVRW19I+NooJAxK5NCOJCQO6Nbub5vIhPfj9d4dx3ysbmfnyOp69eRTRkTaeb3znYEUVTyzJ4/x+XZg8uLtPjjl5cHfiF0Uzb9VuvyZ9N/Aq6YtINJ6E/7KqvgGgqvsbbH8OeNd5WgSkNnh5itPGGdpPUtVZwCyAzMxM1xXnFpRWnuyfX73zELV1SlLHWK4bmcylGd0Z07czsVEtu3O6bmQKldW1/OKtL7l34QaenjrijPOgGNMcv1+SR2V1LY9c/a/FUVoqLtozoPvSip0crKiiS/tYnxzXjZpM+uL5X5sNbFbVJxu093D6+wG+DXzpPF4EzBORJ/EM5KYDqwAB0kWkD55kPxW40VcXEqpUlZw9h1mSu58lOfvYsu8IAP2T2vPDC8/h0owkhiTH+7wf86YxvThWXcvjizfTJjqS335nqPWVmhbbVFjGgtUF3H5+H9KTOvj02NOyUpnzT8+A7ozxNqB7try50z8fuBn4QkQ2OG3/CUwTkeF4und2AncBqGqOiLyCZ4C2BpipqrUAInI38D4QCcxR1RyfXUkIOVFbx8rtpSzN3cfS3P3sKT9OhEBmr848dMW5XJqRRO+u7fwex/fH96WiqoY/fLiNdjGRX1u2zpjmqqtTHl6UQ5d2sdxzSbrPj5+e1IHRvTsxf1UB3x/X135Wz5I31Tuf4rlLP9XiM7zmceDxRtoXn+l14ezI8RN8vLWEpbn7WbalmMPHa4iLjmBceiL3XtqfiQO7BeVP1nsvSaeyuobnPtlBm5gofj55gP0ymbPyxvoi1u8u43fXD6VjnH9KgqeOTuP+VzeyYnspY8/p4pdzhDur2fMjVeVARTXT56zi868OUl1bR+d2MVw2qDuXZiQxLj3RJ5UNLSEi/OcV51JZXcuzH39F+9hI7r7Y93dpJrwdPn6CX/99CyPSEvjOyBS/nefKoT345Ts5zF+125L+WbKk70eHKk+w/cBRalWZ/q1eXJrRnVG9OrW6QVMR4bEpgzlWXcvvl2ylTUwUd1zg/cRYxjz9wTYOHq1izq2Zfh0biouO5LqRKcxbuZtDR6vp1O6byy2aM7Ok70dlldVERggf3nchUa28LDIiQvjt9UOprK7lsXdzaRcTydQsd5fMGu/kFx/hhc92ckNmKkNTEvx+vqlZqbzw2U5eX1fIneP6+v184aZ1Z6IQpqqUHTtBfJvoVp/w60VFRvD0tBFMGJDIg29+wdsbvlFRa8zXqCqPLMqlbUwkP7tsQEDOObB7R0amJTB/1W5UXVfV3WKhkY1CUO7ew5yoVRJCbI6bmKgInr1pFFm9O3PfKxtZkrMv2CGZVuz9nH18mn+A+y7tH9BChGlZaXxVcpTVOw8F7JzhwpK+n2TnlQCE5MRmcdGRzL51NEOS47l73nqWby0JdkimFTpWXctj725mYPcO3DSmV0DPfdXQnnSIi7Ipl8+CJX0/+TivhLYxkc1eNKK1aB8bxdzbsjinW3tmvLSGVTtKgx2SaWWe/fgrisqO8cg1gwLehdkmJpJvj0jmvS/2UlZp04U3R2hmpFau/NgJ1u4+FHJdO6eKbxvNS3dk0TOhDbe/sJqNBWXBDsm0EgWllTz78VdcNbQHY/oGp3Ry6ug0qmvqeGOdjT01hyV9P/h02wFq65SEtqGd9AG6to9l3p1j6NQumlvmrGLz3sPBDsm0Ar96L5cIER668tygxZDRsyPDUm1At7ks6ftBdl4xHeOiaB8m89V3j49j3p1jiIuO4ObZK9leUhHskEwQfbKthPdz9nP3xf3oEX/m6bz97casVLYVV7B2lw3oesuSvo+pKh9vLWFc/8Swms4gtXNbXr5zDKrwvedXUlBaGeyQTBBU13gWR+nVpS13jgv+B/iuGtqT9rFRzF9V0PTOBrCk73O5ew9TfKSKCWexCHRr169be168I4ujVTXcNHsl+4O80LoJvLmf7eSrkqP8v6syWjzFty+0i41iyvCevLtpD+WVJ4IdTkiwpO9j9aWaFw4Iv6QPMKhnPC/cnsWBI1Xc9PxKDlbYYtVuUXz4OH/4cBsXDUhk4rlJwQ7npGlZaVTV1PGWfZjQK5b0fSw7r5hBPTvSrUNcsEPxm5FpnXh++mh2l1Zyy5xVlB+zOyw3+PU/tlBdU8f/u3pQsEP5msHJ8QxNibcBXS9Z0veh8mMnWLe7jIsGdAt2KH439pwuPHvzKLbuP8Jtf13F0aqaYIdk/GjtrlLeWFfEHeP60CcAaz0017SsNLbsO8J6KytukiV9H6ov1ZwQpl07p7pogGeh9Q0FZdw5dw3HT9QGOyTjB7XO4ijdO8Zx90X9gh1Oo64e1pN2MZHMX2mf0G2KJX0fqi/VHJ6aEOxQAqZ+ofXPtx/kRy+vo7qmLtghGR9buLqAL4sO8+AVA2nXSsuQ28dGcc3wnryzaQ+Hj1t345k0mfRFJFVElolIrojkiMg9TntnEVkqItuc752cdhGRp0UkX0Q2icjIBsea7uy/TUSm+++yAq+uTsl2SjVDZVZNX7luZAq/unYwH20p5icLN1BbZ/2q4aKssprfvb+FrD6duWZYz2CHc0bTstI4fqKOt9fbgO6ZeJOdaoD7VTUDGAPMFJEM4AHgQ1VNBz50ngNcjmcx9HRgBvBn8LxJAA8D5wFZwMP1bxThIHfvYUrCtFTTGzeN6cVDV5zLe1/s5eevb6LOEn9YeHLpVsqPneCRq1v/+slDkuMZ1LMj81YV2IDuGTSZ9FV1r6qucx4fATYDycAUYK6z21zgWufxFOBF9VgBJIhID+AyYKmqlqrqIWApMNmXFxNMH28N71JNb3x/fF/uvSSd19YW8sg7OfaLF+Jy9xzmbyt2cdOYXmT07BjscJokIkzLSmPz3sNsKiwPdjitVrP6IUSkNzACWAkkqepeZ9M+oL5wNxlo+PG4QqftdO2nnmOGiKwRkTUlJaEzpW92XjGDk8O7VNMb90xMZ8b4vrz4+S5+/Y8tlvhDlGdxlBzi20Rz36X9gx2O16YM70mb6MiQn3JZVf32u+N10heR9sDrwL2q+rVZt9QTnU8iVNVZqpqpqpmJiaFx11xe6SnVnNA//Es1myIiPHj5QL53Xhp/+Xg7f/woP9ghmbOwaOMeVu0s5WeXDSShrf/XoV1411gW3jW2xcfpEBfNNcN6smjjHo6E6ICuqrK7tJIdByr90k3qVdIXkWg8Cf9lVX3Dad7vdNvgfC922ouA1AYvT3HaTtce8j7JL3FVqWZT6hdav25EMk8s3crzn2wPdkimGY5W1fDfizczOLkjN4xObfoFrcy089KorK5l0cY9wQ7lrMxavp19h6uIiAB/DKN4U70jwGxgs6o+2WDTIqC+Amc68HaD9lucKp4xQLnTDfQ+MElEOjkDuJOctpCXnVfiulLNptQvtD55UHd+9d5m5ln9dMj447J89h+u4pfXDCYyonUP3jZmWEo85/boGJJdPK+tLeR//r6FLu1i6NW5rV8Gz7250z8fuBm4WEQ2OF9XAL8GLhWRbcAlznOAxcB2IB94DvgRgKqWAo8Bq52vR522kFZX55lVc7wLSzWb0nCh9Yfe+oIDNk9Pq7fjwFGe/2Q7141MZlSv0CyuExFuzErly6LDfBFCA7ofbdnPz1/fxAX9utI3sZ3fqqW8qd75VFVFVYeq6nDna7GqHlTViaqarqqX1Cdwp2pnpqqeo6pDVHVNg2PNUdV+ztdf/XJFAXayVNMFUy+cjfqF1s/r05mvSo5y8KgtbdeaPfpODrFRkTxw+cBgh9IiU0YkExcdwbwQudtfu6uUH728joweHXn25lFE+LE81m5NW+hkqaZL6/O9ERcdyfPTR9MuNpL84gpmzltHUdmxYIdlTvHh5v0syyvhnonpIV+F1jEumquG9mTRhiIqWvm8UNv2H+H2F9bQI74Nf71ttN8XX7Kk30LLtnhKNRM7xAY7lFatfWwUGd07kpzQhg8372fiE9n87wdbOVZt8/W0BnV1yqPv5nJOYjumf6t3sMPxiWlZaRytruWdVjygu6fsGLfMWUVMVAQv3p5F1/b+zyOW9FvAU6p5yBWzavpCRISQ0qkNH94/gYnnJvG/H2zjkic/5r1Ne62eP8j2Hj7OroOVPHLNIGKiwiMtjExLYEBSBxa00i6eQ0eruWXOKiqO1zD3tixSO7cNyHnD4383SD7JL6FOsVLNZkpOaMMzN45kwYwxdGwTzcx565g6a4Utuh4kVTW17Ck7xmWDkhiXHj4/y55P6KaysbCcL4ta14BuZXUNt89dze7SSp6bnhnQTzxb0m+B7LwS4ttEMzw1NKscgm1M3y68++ML+NW1g9m6/whXPv0Jv3jrCw7ZYG9A7S49hgK/uDIj2KH43LdHpBAbFcGC1a3nbv9EbR0zX17HxoIynp46gjF9uwT0/Jb0z1JdnZKdV8K49K4hWcvcWkRGCDeN6cWyn07glrG9mb+qgAm/z2buZzupqbVpmv2prLKan726kdKj1fSMjwtY90IgxbeN5sqhPXhr/R4qq4M/oFtXp/z89U0syyvhV9cOYfLg7gGPwZL+Wcrde5gDFVaq6SsJbWN45JpBLP73cQzq2ZGHF+Vw5dOf8ln+gWCHFnZUlXc27uGSJz/mjfVF9IiPo2dCm2CH5Tc3ZqVRUVXDuxv3Nr2zn/3mH1t4Y10R913anxvPSwtKDJb0z1J2nmfWCSvV9K0B3Tvw8p3n8exNozhaXcONz6/kh39bS0FpZbBDCwtFZce4Y+4afjx/PT0T2vDO3ReQ1rmtX+vCg21Ur07069Y+6DX7zy3fzl+Wb+eWsb348cXBW4GsdS6DEwKy80oYkhxvpZp+ICJMHtydCQMSeW75dv6U/RUfbSnmrvF9+cGEc2gbYz+2zVVbp/xtxS5++48t1Cn84spzue38Pq7omqyfcvmxd3PJ3XM4KNNEv7GukMcXb+bKIT14OMhrE9id/lmoL9W0qh3/iouO5McT0/nopxdy2aDuPP1RPhOf+JhFG/dYiWczbN1/hOuf/YyHF+UwqndnlvxkPHeO6+uKhF/vuhHJxARpQHdZXjH/8domvnVOF568YVjQ/90t6Z+F5dusVDOQesS34elpI3j1B2Pp3C6Gf5+/nhv+sqLVleG1NlU1tTy5JI8rn/6EnQeO8tQNw5h72+iwHLBtSqd2MVwxuDtvri8K6AcC1+8+xI/+to4B3Tvwl5tHERsVGbBzn44l/bOQnVdCQlsr1Qy00b07s+juC/if64aQX1LB1X/8lP988wtKrcTzG1bvLOWKP3zC0x/lc9XQnnxw34V8e0RKq1/y0J+mZaVx5HgN730RmAHd/OIj3PbCarp1jOWF27LoEBcdkPM2xZJ+M9XPqjkuPTHof6a5UWSEp3922f0TuPVbvVm4uoAJv1vGX/+5gxNW4snh4yd46M0v+O6zn3P8RB1zb8/iqRuG0yUAH+9v7bL6dKZvYruATLm8t/wYt8xeRVSEZ3qF1jT2Z0m/mU6WalrVTlDFt43m4asH8Y97xjEsNYFfvpPLFX/4hE+3ubfEc0nOPi598mPmr9rNHRf0YclPxlt1WQOeKZfTWLvrEHn7jvjtPGWV1dwyexWHj9fwwm2j6dWlnd/OdTYs6TfTsi2eUs3x9svUKqQndeDF27OYdfMoqmrquGn2Sma8uIbdB91T4ll8+Dg//NtaZry0lk5tY3jzR+fzX1dl0M7PszWGoutGphATGeG3u/1j1bXcMXcNuw5WMuuWUQxOjvfLeVrCfiqaKXtr6JVq+mLt0dZMRJg0qDvj+ycy+9MdPLMsn0ue+pjvj+vDjyb0C9vkV1enLFxTwH8v3kxVTR0/u2wAM8b3JdoW8zmtzu1imDy4O2+sK+SBywcSF+27gdWa2jrunreOdbsP8cyNI/nWOV19dmxf8ma5xDkiUiwiXzZoe0REik5ZSat+24Miki8ieSJyWYP2yU5bvog84PtL8b+yymrW7z7ERVa10yrFRUcy86J+fHT/BK4c0oNnln3FxCc+5q31RWFX4rm9pIJpz63gwTe+YFDPjvzjnnHMvKifJXwvTM1K5fDxGhb7cEBXVXnwjS/4cEsxj04ZzBVDevjs2L7mzU/IC8DkRtqfariSFoCIZABTgUHOa/4kIpEiEgk8A1wOZADTnH1DyifbDlCncKFNvdCqdY+P46kbhvP6D8eS2CGWexdu4PpnPw+ppfNO50RtHc8sy2fyHz4hd+9hfvOdIcz//hj6JrYPdmghY2zfLvTu0tanXTy/fT+PV9cWcs/EdG4e08tnx/WHJv/uVdXlItLby+NNARaoahWwQ0TygSxnW76qbgcQkQXOvrnNDzl4/lWqmRDsUIwXRvXqzNszz+e1tYX89v0tXPPMp9yQmcqJ2rqQvCPeUFDGA69vYsu+I1wxpDuPXD2Ibh1De4WrYKj/hO7//H0L2/YfIT2pQ4uON/vTHfw5+ytuPC+Ney9J91GU/tOSn/y7RWST0/1TX7CeDBQ02KfQaTtde8jwlGoWW6lmiImIEP5tdCof/XQCd5zfh9fWFrKxsJxdB4/yzsY9FJRWtvqun6NVNTz6Ti7X/emfHKqsZtbNo/jT90ZZwm+B74xKITpSWLC6oOmdz+Ct9UU89m4ukwd157Epg0PicxBnO8L1Z+AxQJ3vTwC3+yIgEZkBzABISwvOLHSNydlzmAMV1dafH6I6xkXzi6symJqVxvXPfsb+I1X8eP56wDO4NywlnqEpCQxPTWBoSnyrqWvPzivmoTe/pKjsGDeNSeM/Jg+kYyv5kE8o69o+lkmDuvP6ukJ+dtmAsxrQXb61hJ++upHz+nTmf6cOD5mbwbNK+qq6v/6xiDwHvOs8LQJSG+ya4rRxhvZTjz0LmAWQmZnZam7B6mfVtFLN0NavW3sGJHWgTpWHrx7EhoIyNhWWsbGgnOyt26i/6U/t3MbzJpCSwLDUBAYndwzoRG8HK6p47N1c3tqwh3MS2/HqD8YyunfngJ3fDW7MSuO9TXt5P2cfU4Y3r+NhQ0EZP/jbWtKTOvDc9EyfVgH521n9FItID1WtH/r+NlBf2bMImCciTwI9gXRgFSBAuoj0wZPspwI3tiTwQMveWsLQlPiALFxs/C9ChMHJ8U4dtWfgraKqhi+Lyk++CWzYXcZ7m/Y6+0P/pA4Mc94EhqbEM6B7B5+PDagqbzpdBhVVNdwzMZ0fXXROq5izJdyM7duFXl3aMm/l7mYl/a9KKrj9hdV0aR/D3NtGh9xfXk0mfRGZD0wAuopIIfAwMEFEhuPp3tkJ3AWgqjki8gqeAdoaYKaq1jrHuRt4H4gE5qhqjq8vxl/qSzXvvih4c2Ab/2sfG8WYvl2+tnzdgYoqNhWWsaGgnI0FZSzJ3cfCNZ5+4NioCAYnxzM0JZ7hqQkMS0mgV5e2Z92vW1BayX+++QWfbDvAyLQEfv2dofRv4SCjOb2ICGHq6DR+848tfFVSwTleVEDtP3ycW2avQoAXbz8vJMdVvKnemdZI8+wz7P848Hgj7YuBxc2KrpVYbqWartW1fSwXD0zi4oFJgOdOvKD0GBsKy9hUUMbGwjLmr9rNX/+5E4D4NtFfexMYmhpPtw5nTgw1tXW88NlOnliylQiBR6cM4qbzehERIn3Eoez6USk8sSSPBat281ATawSXHzvBLbNXUVZZzYIZY+nTtXVNr+Ct8Pyooo9l5xVbqaYBPOV+aV3aktalLdcM6wl4kvbW/RWebiHnr4I/ZX9FbZ1ngKBnfJzTJZTAsNR4hiTHn5xx8WhVDd/+02d8UVTOxIHdeOzawWG9dGFrk9ghlkszknhtbSE/vWzAafc7fqKWO+euZvuBCv56axZDUlrf9AresqTfhLo6ZfnWEsZbqaY5jajICDJ6diSjZ0emZnkqzo5V15Kzp5wNBWVsLPSME/z9y30AiEC/xPYcqKjiUOUJuraP4Y83juDKIT1CouQv3EzLSuPvX+7j/Zz9jW73TK+wnjW7DvF/00ZwQXrrnF7BW5b0m1BfqmkLppjmaBMTSWbvzmQ2qLg5dLSajc4g8abCMnYdrCSxfQxL77uQhLYxQYzW3S7o15WUTm2Yv/Kbn9BVVR5680s+2LyfR67O4KqhPYMQoW9Z0m/CMivVND7SqV0MEwZ0Y4IzNnTDXz4HsIQfZBHOGg2/ez+PYSnxXyu/fGLJVhauKeDui/px6/l9ghil74TeZ9EDLDuvmGFWqmlMWPvuqBQiI4TiI1Un21745w7+uCyfqaNTuX9S/yBG51uW9M+grLKaDQVlVrVjTJjr1jGOS87tRsmRKupUWbRxD798N5dJGUn86trQmF7BW5b0z6C+VNP6840Jf9Oy0qipU3YdrOT+VzYwuldnnp42gqgQnJzvTMLranwsO6+YTm2jGZaSEOxQjDF+Ni49kZioCIqPVHFOYvuQm17BW5b0T6OuTvk4zxZAN8YtIiOE5IQ42sVEMvf2LOLbhNb0Ct6y6p3T+HJPOQePVnPRQOvaMcYtunWIo1uHOJJCcHoFb9md/mlk55UgAuPTLekbY8KHJf3TyM4rZmhy65lX3RhjfMG6dxpx6Gg16wvK+PHFrX/ps1Cy8K6xwQ7BGNezpN+I5dtKUKXFq2RZkjPGtDbWvdOIj/NK6NQ2mqFWqmmMCTOW9E/hWQC9hPH9rVTTGBN+LOmfor5U0z6Fa4wJR00mfRGZIyLFIvJlg7bOIrJURLY53zs57SIiT4tIvohsEpGRDV4z3dl/m4hM98/ltNyyLVaqaYwJX97c6b8ATD6l7QHgQ1VNBz50ngNcjmcx9HRgBvBn8LxJ4Flb9zwgC3i4/o2itcneWszQlAQr1TTGhKUmk76qLgdKT2meAsx1Hs8Frm3Q/qJ6rAASRKQHcBmwVFVLVfUQsJRvvpEE3aGjnlk1J9jc+caYMHW2ffpJqrrXebwPSHIeJwMFDfYrdNpO196q1JdqWn++MSZctXggV1UVUB/EAoCIzBCRNSKypqSkxFeH9YqVahpjwt3ZJv39TrcNzvdip70ISG2wX4rTdrr2b1DVWaqaqaqZiYmBu+O2Uk1jjBucbdJfBNRX4EwH3m7QfotTxTMGKHe6gd4HJolIJ2cAd5LT1mp8UeTMqmmrZBljwliT0zCIyHxgAtBVRArxVOH8GnhFRO4AdgH/5uy+GLgCyAcqgdsAVLVURB4DVjv7Paqqpw4OB9XJWTVtENcYE8aaTPqqOu00myY2sq8CM09znDnAnGZFF0D1pZqd28UEOxRjjPEb+0QuUGqlmsYYl7CkD3xSP6vmQOvPN8aEN0v6ePrzO7eLYWhyfLBDMcYYv3J90q+rU5ZvLWF8elcirFTTGBPmXJ/060s1J1ippjHGBVyf9JflFVuppjHGNVyf9LPzShhmpZrGGJdwddIvPVrNxsIym2DNGOMark76n5ycVdP6840x7uDqpL9sS7GVahpjXMW1Sb+uTlm+7QAX9k+0Uk1jjGs0OfdOuNpUVE6pLYBujGmFFt411m/Hdu2dfrZTqjnOFkA3xriIi5O+lWoaY9zHlUn/YEWVlWoaY1zJlUn/k20HPLNqWqmmMcZlXJn0s/OK6dIuhiFWqmmMcZkWJX0R2SkiX4jIBhFZ47R1FpGlIrLN+d7JaRcReVpE8kVkk4iM9MUFNFd9qeZ4K9U0xriQL+70L1LV4aqa6Tx/APhQVdOBD53nAJcD6c7XDODPPjh3s1mppjHGzfzRvTMFmOs8ngtc26D9RfVYASSISA8/nP+Mlm1xZtW0Uk1jjAu1NOkrsERE1orIDKctSVX3Oo/3AUnO42SgoMFrC522rxGRGSKyRkTWlJSUtDC8b8reWsLw1AQ6WammMcaFWpr0L1DVkXi6bmaKyPiGG1VV8bwxeE1VZ6lqpqpmJib69m78YEUVmwrLmNDfqnaMMe7UoqSvqkXO92LgTSAL2F/fbeN8L3Z2LwJSG7w8xWkLmPpSTevPN8a41VknfRFpJyId6h8Dk4AvgUXAdGe36cDbzuNFwC1OFc8YoLxBN1BALLNSTWOMy7VkwrUk4E0RqT/OPFX9h4isBl4RkTuAXcC/OfsvBq4A8oFK4LYWnLvZap0F0C8a0M1KNY0xrnXWSV9VtwPDGmk/CExspF2BmWd7vpbaVFjGocoTXGhdO8YYF3PN1MrZeSVEWKmmMd/gz2l8TevjmmkYsreWMMxKNY0xLueKpF9fqmkTrBlj3M4VSX/5yQXQrWvHGONurkj62XkldG0fw+CeVqppjHG3sE/69aWa49NtVk1jjAn7pL/RSjWNMeaksE/6VqppjDH/EvZJ/+O8YptV0xhjHGGd9A9WVLGpqJwJVqppjDFAmH8i10o1TWPsE6jGzcL6Tn/ZFivVNMaYhsL2Tr+2Tlm+rYSLB9qsmqZ1sr84TDCE7Z3+1f/3KWWVJ6w/3xhjGgjbpF9+rBqA8eldgxyJMca0HmGb9MsqT9A+NoqEtlaqaYwx9QKe9EVksojkiUi+iDzgj3McqKjiaHUtCW2i/XF4Y4wJWQFN+iISCTwDXA5kANNEJMPX52kTHck5ie3o0t7u8o0xpqFA3+lnAfmqul1Vq4EFwBRfn6RdbBRd28cSFx3p60MbY0xIC3TSTwYKGjwvdNqMMcYEQKur0xeRGcAMgLS0tLM+jtVAG2PMNwX6Tr8ISG3wPMVpO0lVZ6lqpqpmJiba9AnGGONLgb7TXw2ki0gfPMl+KnBjgGMwxphGuaGHIKBJX1VrRORu4H0gEpijqjmBjMEYY9ws4H36qroYWBzo8xpjjAnjT+QaY4z5Jkv6xhjjIpb0jTHGRSzpG2OMi1jSN8YYF7Gkb4wxLmJJ3xhjXERUNdgxnJaIlAC7WnCIrsABH4UTKtx2zW67XrBrdouWXHMvVW10HptWnfRbSkTWqGpmsOMIJLdds9uuF+ya3cJf12zdO8YY4yKW9I0xxkXCPenPCnYAQeC2a3bb9YJds1v45ZrDuk/fGGPM14X7nb4xxpgGQj7pi8hkEckTkXwReaCR7bEistDZvlJEegchTJ/y4prvE5FcEdkkIh+KSK9gxOlLTV1zg/2+IyIqIiFf6eHNNYvIvzn/1zkiMi/QMfqaFz/baSKyTETWOz/fVwQjTl8RkTkiUiwiX55mu4jI086/xyYRGdnik6pqyH7hWYjlK6AvEANsBDJO2edHwLPO46nAwmDHHYBrvgho6zz+oRuu2dmvA7AcWAFkBjvuAPw/pwPrgU7O827BjjsA1zwL+KHzOAPYGey4W3jN44GRwJen2X4F8HdAgDHAypaeM9Tv9LOAfFXdrqrVwAJgyin7TAHmOo9fAyaKiAQwRl9r8ppVdZmqVjpPV+BZiziUefP/DPAY8BvgeCCD8xNvrvn7wDOqeghAVYsDHKOveXPNCnR0HscDewIYn8+p6nKg9Ay7TAFeVI8VQIKI9GjJOUM96ScDBQ2eFzptje6jqjVAOdAlINH5hzfX3NAdeO4UQlmT1+z82Zuqqu8FMjA/8ub/uT/QX0T+KSIrRGRywKLzD2+u+RHgJhEpxLMC348DE1rQNPf3vUkBXy7RBI6I3ARkAhcGOxZ/EpEI4Eng1iCHEmhReLp4JuD5a265iAxR1bJgBuVn04AXVPUJERkLvCQig1W1LtiBhYpQv9MvAlIbPE9x2hrdR0Si8PxJeDAg0fmHN9eMiFwCPARco6pVAYrNX5q65g7AYCBbRHbi6ftcFOKDud78PxcCi1T1hKruALbieRMIVd5c8x3AKwCq+jkQh2eOmnDl1e97c4R60l8NpItIHxGJwTNQu+iUfRYB053H1wMfqTNCEqKavGYRGQH8BU/CD/V+XmjimlW1XFW7qmpvVe2NZxzjGlVdE5xwfcKbn+238NzlIyJd8XT3bA9gjL7mzTXvBiYCiMi5eJJ+SUCjDKxFwC1OFc8YoFxV97bkgCHdvaOqNSJyN/A+npH/OaqaIyKPAmtUdREwG8+fgPl4BkymBi/ilvPymn8HtAdedcasd6vqNUELuoW8vOaw4uU1vw9MEpFcoBb4maqG7F+xXl7z/cBzIvITPIO6t4byTZyIzMfzxt3VGad4GIgGUNVn8YxbXAHkA5XAbS0+Zwj/exljjGmmUO/eMcYY0wyW9I0xxkUs6RtjjItY0jfGGBexpG+MMS5iSd8YY1zEkr4xxriIJX1jjHGR/w8UwGfXOyP+KgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 6min 23s, sys: 19.3 s, total: 6min 42s\n",
      "Wall time: 3min 39s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "kf = KFold(n_splits=5, random_state=500, shuffle=True)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "for _ in range(5):\n",
    "    c_best_money = []\n",
    "    c_best_std = []\n",
    "    c_best_score = []\n",
    "    for c in np.linspace(0.0001, 1, 10):\n",
    "        scores = []\n",
    "        money = []\n",
    "        for train_index, test_index in kf.split(x):\n",
    "            X_train, X_test = x[train_index], x[test_index]\n",
    "            y_train, y_test = y[train_index], y[test_index]\n",
    "            lr = LogisticRegression(max_iter=100000, C=c)\n",
    "            lr.fit(X_train, y_train)\n",
    "\n",
    "            y_score = lr.predict(X_test)\n",
    "            precision, recall, thresholds = precision_recall_curve(y_test, y_score)\n",
    "            scores.append(auc(recall, precision))\n",
    "            money.append(sum((y_score == y_test) & (y_test == 1)) * int(np.random.uniform(0, 20)))\n",
    "\n",
    "        c_best_money.append(np.mean(money))\n",
    "        c_best_std.append(np.std(money)) \n",
    "        c_best_score.append(np.mean(scores))\n",
    "        \n",
    "    yerr = c_best_std\n",
    "    \n",
    "    ax = plt.errorbar(np.linspace(0.0001, 1, 10), c_best_money, yerr=yerr)\n",
    "    plt.show()\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Задание 12 (1 балл).__ Выше мы уже описали примерную экономическую модель вашей задачи. Как вы считаете, что для вашего бизнеса важнее — хороший precision или recall модели? Почему?\n",
    "\n",
    "__Ответ:__ корее всего recall, потому что мы можем себе позволтиь если что ошибиться при звонке, главное набрать как можно больше клиентов\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Вспомним, что на самом деле логистическая регрессия предсказывает нам вероятности положительного класса для объекта. Возможно, путем настройки порога бинаризации этих вероятностей мы сможем получить какой-то выигрыш? Проверьте ваши рассуждения выше с помощью настройки порога бинаризации на кросс-валидации для максимизации прибыли. Воспользуйтесь сеткой от 0 до 1 с шагом 0.01. Напомним, что снижение порога дает нам более высокий recall и более низкий precision, и наоборот."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 767,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 100/100 [12:26<00:00,  7.46s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 18min 20s, sys: 56.1 s, total: 19min 16s\n",
      "Wall time: 12min 26s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "kf = KFold(n_splits=5, random_state=500, shuffle=True)\n",
    "\n",
    "t_money = []\n",
    "t_std = []\n",
    "t_score = []\n",
    "t_precision = []\n",
    "t_recall = []\n",
    "for t in tqdm(np.linspace(0, 1, 100)):\n",
    "    scores = []\n",
    "    money = []\n",
    "    precision = []\n",
    "    recall = []\n",
    "    for train_index, test_index in kf.split(x):\n",
    "        \n",
    "        X_train, X_test = x[train_index], x[test_index]\n",
    "        y_train, y_test = y[train_index], y[test_index]\n",
    "        \n",
    "        lr = LogisticRegression(max_iter=1000000)\n",
    "        lr.fit(X_train, y_train)\n",
    "        y_score = lr.predict(X_test)\n",
    "        \n",
    "        probability_one = pd.DataFrame(lr.predict_proba(X_test))[1]\n",
    "        y_score_new = np.where(probability_one < t, 0, y_score)\n",
    "        \n",
    "        precision, recall, thresholds = precision_recall_curve(y_test, y_score_new)\n",
    "#         print('------')\n",
    "#         print(metrics.precision_score(y_test, y_score_new))\n",
    "#         print('------')\n",
    "# #         print(precision, recall, thresholds)\n",
    "#         print(metrics.recall_score(y_test, y_score_new))\n",
    "#         print('------')\n",
    "        t_precision.append(precision)\n",
    "        t_recall.append(recall)\n",
    "        \n",
    "        scores.append(auc(recall, precision))\n",
    "        money.append((sum((y_score_new == y_test) & (y_test == 1)) * int(np.random.uniform(0, 20))))\n",
    "\n",
    "    t_money.append(np.mean(money))\n",
    "    t_std.append(np.std(money)) \n",
    "    t_score.append(np.mean(scores)) \n",
    "    t_precision.append(np.mean(precision))\n",
    "    t_recall.append(np.mean(recall))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 768,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "yerr = t_std\n",
    "plt.figure(figsize=(8, 6), dpi=80)\n",
    "ax = plt.errorbar(np.linspace(0, 1, 100), t_score)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 769,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.888888888888889"
      ]
     },
     "execution_count": 769,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.linspace(0, 1, 100)[np.argmax(t_score)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 770,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "yerr = t_std\n",
    "plt.figure(figsize=(10, 8), dpi=80)\n",
    "ax = plt.errorbar(np.linspace(0, 1, 100), t_money, yerr)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 771,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.4747474747474748"
      ]
     },
     "execution_count": 771,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.linspace(0, 1, 100)[np.argmax(t_money)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 772,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2907.6"
      ]
     },
     "execution_count": 772,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t_money[np.argmax(t_money)] # неплохо!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Постройте график зависимости прибыли от порога бинаризации, также отметьте на нем стандартные отклонения. Выделите наилучший порог. \n",
    "\n",
    "\n",
    "__Вопрос:__ Замечаете ли вы какую-то закономерность? Для правильного ответа на этот вопрос попробуйте запустить несколько раз и задумайтесь, почему порог получается в какой-то конкретной области?\n",
    "\n",
    "__Ответ__: Да, как и было сказано ранне - для нас важнее recall, поэтому чем больше порог, тем меньше единичек будут попадать\n",
    "До какого-то момента это некритично и значение порога может меняться из-за вероятностей, а потом резко идет на спад"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 778,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 779,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 100/100 [10:06<00:00,  6.06s/it]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                   | 0/100 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Лучший порог: 0.2828282828282829\n",
      "Заработаем: 3438.2\n",
      "PR: 0.6573002199413489\n",
      "Recall: 0.22983381409204187\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 100/100 [09:38<00:00,  5.78s/it]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp8AAAIECAYAAABfSUWIAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAxOAAAMTgF/d4wjAACTHElEQVR4nO39eZQk13keeD83InKpyqx96b3RaOyLSIIEKFIWCZEUZC00RxKphTM0fTiyRXmZ8RHGI2n02bLmzNjH0szAHo392SRlG6L4SbZIakSZkihDGySKAgkSAEHsaDQavXd11165R8T9/oi4kZGZEZERmZFZmVnP75w+QFdWZUVXVlW++W5XSClBRERERDQM2n5fABEREREdHAw+iYiIiGhoGHwSERER0dAw+CQiIiKioWHwSURERERDw+CTiIiIiIbG2O8LSCqXy8mVlZX9vgwiIiIiCnHp0qW6lDIXdNvYBZ8rKyu4ePHifl8GEREREYUQQlwPu41ldyIiIiIaGgafRERERDQ0DD6JiIiIaGgYfBIRERHR0DD4JCIiIqKhYfBJREREREPD4JOIiIiIhobBJxERERENDYNPIiIiIhoaBp9ERERENDQMPomIiIhoaBh8EhEREdHQMPgkIiIioqFh8ElEREREQ8Pgk4iIiIiGhsEnEREREQ0Ng08iIiIiGhoGn0REREQ0NAw+iYiIiGhoGHwSERER0dAw+CTaZw898jgeeuTx/b4MIiKioWDwSURERERDw+CTiIiIiIaGwSfFxvIwERER9YvBJxHRBOGLRCIadQw+iYioBQNYIhokBp9ENHEYPBERjS4Gn0REREQ0NAw+iSgSs4hERJQmBp9ERERENDQMPomIiIhoaBh80khjyZeIiGiyMPgkIiIioqFh8ElEREREQ8Pgk4iIiIiGhsEnEREREQ0Ng08iIiIiGhoGn0REREQ0NAw+iYgCcM0XEdFgMPgkIiIioqFh8ElEREREQ8Pgk4iIiIiGhsEnEREREQ0Ng08iIiIiGhoGn0REREQ0NAw+iYiIiGhoGHwSERER0dAw+CQiIiKioYkdfAoh/qsQ4lkhxDNCiL8QQtznvv2cEOJl9+3PCCF+zPcxtwkhviKEeEUI8aQQ4p44txERERHRZDISvO+PSim3AEAI8UMAHgXwZve2H5NSPhPwMZ8A8Ekp5aNCiA+5H/NAjNuIiIiIaALFznyqwNM1B0BGvb8QYhXA/QA+477p8wBOCCFujbot7vUQERER0fhJkvmEEOLTAN7j/vX7fTd9WgghAHwNwM9JKa8DOAHgipTSBAAppRRCnAdwEsB2xG1n2j7nwwAeVn+fm5tLcsn77qFHHgcAPPbwg2N130RERESDkGjgSEr5USnlCQD/GMAvuW9+t5TyTQDeCuAGgF9L8wKllI9IKY+rP8ViMc27JyIiIqIh6mnaXUr5awDeI4RYklKed9/WAPCvALzLfbcLAI4IIQwAcDOjJwGc73IbEVGHhx553Mv2ExHR+IoVfAoh5oUQR31//0EA6wCqQoh537t+GMDTACClXAPwFICPuLd9EMBFKeWZqNt6/6cQ0ThhMElEdDDF7fmcA/BZIcQUABvAdQDvB3AIwOeFEDoAAeAsgI/6Pu7jAB4VQvw8gB0AH4t5GxERERFNoFjBp5TyDQBvD7n5voiPexnAO5PeRkRERESTiSccEREREdHQMPgkIiIioqFh8ElEREREQ8Pgk4iIiIiGhsEnEREREQ0Ng08iIiIiGhoGn0REREQ0NAw+iYiIiGhoGHwSERER0dAw+CQiIiKioWHwSTQBHnrkcTz0yOP7fRlERERdMfgkIiIioqFh8ElEREREQ8Pgk4iIiIiGhsEnDQz7EImIiKgdg08iOvD4QomIaHgYfBIRERHR0DD4JCIiIqKhYfBJREREREPD4JOIiIiIhobBJ40EDnwQEREdDAw+iYiIiGhoGHwSERER0dAw+CQiGhFsPyGig4DBJxERERENDYPPHqSRnWCGg8YZv3+JiKhXDD6JiKhnfCFCREkx+CQiIiKioWHwSURERERDw+CTiIiIiIaGwScRERERDQ2DTyIiIiIaGgafRERERDQ0DD6JiIiIaGgYfBIRERHR0DD4JCIaMi5mJ6KDjMEnUcoYWBAREYVj8ElEREREQ8Pgk4iIiIiGhsFnSlhqJSIiIuqOwScRERERDQ2DTyIiIiIaGgafRD5snyAiIhosBp9ERERENDQMPomIiIhoaBh8EhGlgC0bRETxMPgkIiIioqFh8EkUA7NaRERE6YgdfAoh/qsQ4lkhxDNCiL8QQtznvv02IcRXhBCvCCGeFELc4/uYnm4jIiIiosmUJPP5o1LKN0kp3wLgEQCPum//BIBPSilvB/BLvrf3cxsRERERTaDYwaeUcsv31zkAUgixCuB+AJ9x3/55ACeEELf2elvP/xIiIiIiGnlGkncWQnwawHvcv34/gBMArkgpTQCQUkohxHkAJwFs93jbmbbP+TCAh9Xf5+bmEv8jiYgovht7NWR0jgQQ0WAk+u0ipfyolPIEgH8Mp1Q+cFLKR6SUx9WfYrE4jE9LNBE4KEW92Ko0sFNt7PdlENGE6umlrZTy1+BkQC8COCKEMABACCHgZC/PA7jQ421ERLRPLFtCSkDK/b4SIppUsYJPIcS8EOKo7+8/CGAdwBqApwB8xL3pgwAuSinPSCl7uq3Pf8/YY6aKiPZTqW4CYPBJRIMTt+dzDsBnhRBTAGwA1wG83+3V/DiAR4UQPw9gB8DHfB/X621ERLQPSjU3+ASjTyIajFjBp5TyDQBvD7ntZQDvTPM2IiLaH17wydiTiAaE44wRWAInooNmt8rgk4gGK9GqJaJ+qWD+sYcf3OcrIZo8Ukpc2CyjmI33qz3o57FUs5z7YtmdiAaEwScR0YSomTaqDRu6sHq+jz237G4z9iSiAWHwSUQjT0qJi5tlFHP8lRWl2lBZy96p4BNw1i7pmujzqoiIWrHnk4hGXrVho9KwUW70ntE7CKoNG0B//ZolX/BZN+1+L4mIqAODTyIaeSqjZzMWitTMfPYeffozn1UG+0Q0AAw+iWjk1UyV0WMjYpSq6QaffXyZ/MFnjZlPIhoABp9ENPJqblDFIZholXr/PZ+lluCTmU8iSh+DTyIaeSoDZzPzGanZ89lH2b3KzCcRDRaDTyIaeV7PJ4PPSKmX3RsMPokofQw+iWjkNXs+9/lCRlwthVVLpTrL7kQ0WAw+aeLxmNT91+9joDJwElz/E4VldyIaBww+iWjk+TNwaqiGOlVSXjLPzCcRDQKDT6IDZhwzwf4M3J6vLEytvD2ffS2Zbwac7PkkokFg8EmBxjFAocnlX3ZerjH4DFP1BYt2j3upuOeTiAaNwScRjbyWzCeDz1D+IL1uJQ8cbVuiVDch3OPcWXYnokFg8HlAMJNJ46zmz3yy5zNU1ewv+Cw3LEgJ6G70WR1A2Z2/i4iIwScRjTx/5rPEzGeoqi8w72UrgPra6poTfDLzSUSDwOCTiEaePwNX4sBRKP/XqdFD5nOvPfjkwBERDQCDTyIaef4MnH8am1q1lN17yHyqHZ+Gl/lk8ElE6WPwSUQjj2X3eFoGjlh2J6IRxeCTiEZeS+aTA0ehKr4yeS9Zy46yOzOfRDQADD6JaOT5exnb93xyerqp31VLXvAp2PNJRIPD4JOIRl5L2Z0DR6FqLLsT0Rhg8ElEI88fVHHgKJw/Q9zTwJH7tWXZnYgGicEnEY28mmlDCEAAKDPzGco/7d7bqqUGAEATAgIMPoloMBh8EtHIqzYsaACE4PGaUSp9L5l3Pl4TzteaZXciGgRjvy+AiKgbJ/Mp3MwnA6Iw/Q4c7bp7PjUhIITgwBERDQSDz32yVWkgpzPxTBRHs+wumPmMUDX7W7VUqpnO19ltcagy80lEA8Dgcx9c2Cjj+m4N01l9IPev1s489vCDA7l/6g0fl97VTAsCApoAyn0MHE3yY2DbEnXThiYAW/ZYdq+bKGYNJ8ssuGqJiAaDqbd98MVnrwBwniyIqLtaw818CsFVSyFUllLt6Owl+NytmijknJyEJgQHjohoIBh87oMvPnsZAGBJBp9EcdRMC0IAmuaUhiV/djqoNUuauyapl57PUs1EMe8En860O8vuRJQ+Bp9D9vqNEp6/vAMAsJlUoBA8tadVrWFDg4AmBGy5vyuARvWxUcNGmpv5bPTY86kyn4KZTyIaEAafQ/bFbzpZT004mc84GZxRfbIjGhY1cOQm9Th0FMALPt3f6j1Nu9dMFHNOLzp7PoloUBh8DtnvfesKFqYzKGSd7EKlwbIWkV/7iy3LlqhbzqolldXrZ+hoUqmye689n1JKp+zuZT6dsjtbHIgobQw+h+jM2i5eurqL7733sHd83Xalsc9XRTTaVBClBo6A8TnffZhVi0pb2T1pybzasGFLNMvucFocTA5GElHKGHwO0X/5pjPl/v43HfWGAhh8EkVTQy8CzbJ7iWX3DjWv7N7bwNGue7Rm0Zt2d++XfZ9ElDIGn0MipcQXn72MpUIW337zInT3F/t2mcEnURQV/Gi+snuJpxx1aK5acv6etOyujtb0l92BZlBLRJQWBp9D8tLVXbx2vYTv+7bDMHTNexJl5pMomhqk8Q8cMfPZyVu11GPPp/qa+svuADOfRJQ+Bp9D8nvPNkvuAFh2J4pJBT9O2d3NfDL47FBtK7s3kpbd3XPd2zOfVWY+iShlPF5zCFTJfXUmhwdOLQJoTqQy+CSKptb9OEc+utPuQyi7Nyx7KJ8nLRVfhlig98xn0bfnE2Dmk4jSx8znENRMG+fWy/j+bzviTbmr7MQOg0+iSN7A0ZD3fG6WG1jbreHSVmXgnysNXtkdzrnsSQeO9jrK7g4Gn0SUNgafQ6B+qb//TUe8t3kDRww+iSL5exm9PZ9DWLVkusHbuAwFVtsyn0mDxr22zKfGgSMiGhAGnwMmpcRu1cSRuTzeenLBezt7PoniCV61NPiASO23HJfTlGq+PZ9CiN7L7nmW3YlosBh8DljNtGHaEj/wbUe8gBNwnyDA4JOom5pvyfwwB44sL/gczZ/R9gX2/fZ8NjOf7vGa7tsZfBJR2hh8DpiaIP0BX8ld0TTB4JOoi2bPp/AmsAc9CGTZ0st8qp/hUVdt+E+C6r3ns5jLuPejMp8suxNRujjtPmB1y4YA8JYT8x236YLBp7JXM1GpO+dIqyc9IsAXVMEJiKYy+sBL4et7Ne//x6Xs7vV8wok+k65aau75dDOfXs8nM59ElC5mPgfMsiV0TQQGVJoGbFfG44kNcJ6czlzfG8j0706lga1KA9+8uJ36fdN4q7WdWV7IGQMfOLq24ws+xyXz6Z0E1V/ZvZBtO+GIZXciSlms4FMIkRdC/I4Q4hUhxDeFEI8JIW51b/szIcTrQohn3D8/7fu4VSHEl4QQrwohnhNCvDvObZPEtmVLr6efLgR2Kg1IKYd8Vb2pmTakBM7dKKV+326FE1945lLq903jzd/zCTiZuUEPHF3bqXr/P06ZT8N9oStEL8GnhUJW935faWDZnYgGI0nm85MA7pBSvhnAFwD8qu+2n5ZSvsX98y99b/8XAJ6QUt4G4GMAfkMIkYlx28SwpPQWyrfTNIG6ZXtlxVFnuUFy0ie1OGz3vr/47BVv0IMI6Aw+p7MGSgPOfF71BZ/j0/NpIZ9Rw0K9TburHZ8AM59ENDixgk8pZVVK+fuymaJ7AsCpGB/6owD+nXsfTwK4DODBGLdNBNOyYUt4i+XbqaB0q1If5mX1zHaDwkFkQtR31vXdGp44u576/Q9K+8Qxpa+llxHONPagM59rY5r5zGecX+lCALWkA0dV09vxqe4DYM8nEaWv157Pfwgn+6n8CyHEt4QQ/1kIcRoAhBBLADJSyqu+9zsH4GTUbe2fSAjxsBDiovqzt7fX4yUP35Y7TKSHfJVVTDouQ0cqOzmITIgtpff1+N1nLqd+/zS+AjOfAw4Ix7Lns2EjZzSHheqmnailZ69mejs+gWawX2XZnYhSljj4FEL8PIBbAfwv7pv+ppTyTgBvAvAXAL6Y3uUBUspHpJTH1Z9isZjm3Q/UZsnJaGoRZXdgfE5QUYmUQQWfOUPH3Udm8QfPXWGfGXn8q5YA5wSeSsMaaHvG1Z2qt9R+nDKfU9lm2R1oLsqPY69mesNGADOfRDQ4iYJPIcQ/AvDDAL5PSlkGACnlBfe/Ukr5rwGcFkIsSSnXAZhCiMO+uzgF4HzUbT3/S0bQZlllPqPL7gc98ymlhC2dJ/oPvOUodqomHn/5eqqfY9LV3cMMJlHNO7PcMe0GWJUBHvt4bacKQ3OO89wdk+Cz0lZ2B+L3Z0spA3o+OXBERIMRO/gUQjwM4MMAHpJSbrlvM4QQh3zv80EA19zgEgA+C+Cn3NseAHAMwOMxbpsIm2Un8xk1cASMUfCpej5TftKveStiBP7Gm48CAH73myy9J3Fxq9LSpzhJqh3T7k6ANMjS+9puDbruBJ971fH4+aw2bOSN1tOJ4gaf6iS2GV/ZXfPdRkSUplhL5oUQxwH8XwDOAvhT9xVxDcB7AfyeECIHwAZwA8AHfB/6swB+XQjxKoA6gI9IKRsxbpsIWyr4nJDMpzftnnCQoRsVRGgacGx+Cg+cWsAfvXhtKEcoTgLblrBsObFBQq1htezKVUvQB/X9UTMtbJTqKOYMmLDHpuxec6fdtysN72sV92d1r23BPMBpdyIanFjBp5TyIpovptvdH/Fx1wB8T9LbJkW3srvmphZ2xiD4tG3p7eJMuwdMTS6r3tgPvOUYnjy3icdeuJbq55lUZTcTbdoSlXqz729S1EwbOaNZpJnOqsznYMrBa+6wkaEJ2FKMz8CR2XvZvXm6UWvZXddE6pUOIiKecDRAquweNnA0TplP/17FtDMhKuuisjXff+9h6Jpg6T2msi8z9/oADgDYbzXTagk+C25wPahdn2u7TvuC4ZbdS/XBDjelwbRsNCzp2/PpiPuzqn4GZ3Kt+YicoTHzSUSpY/A5QFulbpnP8Qk+/deY9pJ5FUSoL9NSMYd33baMP3/l+sg/6Y8Cf1n4tevjs4osrmrD9oIqoJmdG9QRm1e3m5lP9T056KX2/VJ9sV7wqcrucYPPamfmE1DBJzOfRJQuBp8DtOFlPoNv14RAztDGIvjcqfgzn+k+GXk9n74M8QfefBSmLcem324/levNx+Ps9UnMfLaW3VWAtDegsrs6WlPXNO97ctRL72oRv1d2d9/eiNnzqYLrYkfwqTPzSUSpY/A5QFvlOnTRHJQIMjeVGY/g0zfxm/aTUbPns/m277nnMHKGht0BThpPyulE/gD97I30M5+2lImWlafNKbsHZD4H9MLkmiq7a8KrToz6EZte8OlbMg8kGThyPr4j+Mxo3PNJRKlj8DlAm+WGN1QUZlyCz4GW3QMyn8Wcge++6xAqDTt29uag8pef0858ru1WcfZ6CTv7GHzVGjZymaCezwFlPrf9PZ/O2/Zqo/0zWnUDxPYl8yy7E9EoYvA5QFvlemi/p+IEn6OdVQFaJ/LTfjLaCwg+AXg7P1l6j+af+j57fS/VLOWVrSokmpm1/VAzLS+jB/in3QeU+dypYSZvQBPC+54cm8xnpi3zmXDa3X+8JuCU3au+zOekVAuIaH8x+BwQKSW2yo3QBfPK3FQGO5XGvpY149iuDLLsrqbdW9/+XXesAAAqA8pwTQr19ctozmS2/1zyfqmgZj+zz+2ZT1UaHtQQ0LXdKg7N5gE0XxCN+gsg9Tip3tikOzrVKU6dPZ+cdiei9DH4HJDdmgnTlrEyn3XLbskujCJ/2TX9affWPZ9KPqNDEwAH3qOpr1/OzXqdTXHiXU1RN6z97Pls2/M54CXz17arOOwFn87bRn/gqG3aXZXd4w4cBez5BNyeT5bdiShlDD4HRK1Z0roEn7NTGQCjv25pZwiZz6D+WCEEJBh9RlFfPxWgvZbirs9a2wL7YbNtibpltwwcFb2Bo/SvZ69molS3sDqbA9D8+R2XzGffZfe24DPPaXciGgAGnwOy0eVcd2VuzIJPXRNDWbWkCAAj3pGw71T5Wa3ZeW0t/cwnAJzfKKd2v3GpwMdfds8ZGjQxmLK7WrN0qC3zOfI9n+7P5FTbkvm47RKq7F5oOx0rl9FQN+2RbwsiovHC4HNANruc666MTfBZbUDAeTJOu+zunXAUcJsmGHz6BQ18qAxgVteQz2g4m2Lm0z9odG59+DtE1Qsdf+ZTCIFC1hjI8Zpq0v3wmPV8qqx083jNZNPupZqJfEaDobc+Jaive1T2U0oJkxspiCiBWGe7U3JbExZ8blca0NydpamX3esmNIHAfahCCNiMPiM12xYETi0VUu359D/Wb+xL8OlmPg0Njz38oPf2Qs4YTOZzV2U+3bL7uCyZ7zjhyHl7kuCzmMt0vF21ctTM1lOm/LYqDazv1bFRqmOxkE166UR0ADHzOSCb6mjNiSm7m9A05xsm7aXTpZoVWHIHWHaPQwVhAsAtK0Vc2qqkthqp1pL53Ieye6Oz7A44Q0eDGDhSmwJW2weORjzzWQs54SjuwNFu1UQx1xlcNoPP8O+nhmVDotmyQETUDYPPAVGZz65L5qfHJPisNtzTmuI/ocVVqpkda5YUZj67c4J352t1y0oBUgKvp1R69wex+5H5VL2M/j2fAAZWdr/aVnZ3Svy61xM5qpqrllozn3GrFKW62THpDjQ3KES94LTdmwa1fYCIJg+DzwHZLE9W5nO70oCmuWX3lBeOl2pmeOZTgLPuXZTqpteycHqlCCC9k45U8CIAnLsxOpnPQk5vOdkpLWtu2X1lJue9rZg3sDfAY17T0O8JR6Wa1THpDrSW3cOoF4ejHqAT0ehg8DkgSQeOdkY4+GxYNsp1pzQukP6qpb2o4BMsu3dTdjOfAHB6pQAgvV2fKqOWNTRc3q4Mfedj0MARMLjM57WdGpaLWWT01qX2o152r/S5ammvanYJPsO/1ir4ZOaTiOJi8DkgW+UGZnJG4BCN3zhkPtWaGV1zypCmLWGltPldSolSvRk8tVNfP07ThvMH7zcvu8FnamX35sCPlMDFzUoq9xuXf+DIr5AzULfs1DcvXN1unm6kFPOZ0R84UsFn2wlHcVYt2dLZpdp+tCYQb9pd/SoY9a8REY0OBp8DslGqY77QOT3aLp/RkTW0kQ4+1bVpbs8nkN66pZppw7JlZNldvR8FK7vbAgBgJp/BodkcXksp86kyXlk3qBl232f78nSl4A7HpFl6l1Jibbcz+JzJGSNfUg494SjGz42qLAT3fGru/XfPfI56dpiIRgeDzwHZKtexMB1v7cjcVGakg88df/Dpvi2t8mvU6UbO53T+m3aGa5K0bws4vVzE2eulVBaDq6BGBZ/D7vsMy3xOZ9X57umV3m3pHCOq1iwpquw+yovWvcGs9rJ7zMwn0Hm6ERCz59O9icEnEcXF4HNANssNzE9I8LntO91IlcHTykSqvr3wns90P9+kqZs26pbd0t5xeqWAvZqJ67u1vu/f6/nU9yfz2ez57Cy7A0A5xYDHdKOojsxn3oCUQHkfjheNq1q3kNU1r8fcW7UU4+cmOviMMe3Onk8iSojB5wBUGxYqDQsL093L7sDoB5871cGV3b3TjbqW3Uf3iX8/qbKzv2dWTby/lsLEe820IQRgaAI5Qxv6rs/mtHv7wJHz9zSzbablBFGdPZ9G6p8rbVXTatkIIISAocU7EEJlLoPK7mpvaNjPn9rxCYz214eIRguDzwHYctcsTU7ZvVkaV98wqZXdA4Inv2aZn5nPIKrs7M8c3+JOvKfR91ltWNDgBDM3LU3vX89ne9ldZT5TzEaatgo+W8vuM+7n2h3hdUvVRucJRFlDS1h2D1oyHz1w5M927g1g+wARTSYerzkAas3SfILMZ920UzuVJm1e2V00y+7VlE458no+QzOfbtk95VOVJkW51hm835Lirs+q2Szp37RUwJ++tBZrgjotXs9nW2ClAqU0S73N4DM487k7wtPc1YblZSmVrKGhHuNFYjP4bP6+UkeZfuXMDQDhwac/2znqu1CJaHQw8zkAKviMe87xqK9bCiy7pxSANHs+g29n2T3aXkDwfnR+CllDw9kb/Wc+aw3LewxOLU3DtCUubw1v3VL3gaM0y+7BPZ8qKBvlsnK1YWGqPfOpa2hY3YekbG/aPSDzqcruIS+M/V+TQexdJaLJxOBzANS57nEHjmb3Mfh86JHH8dAjj0e+z45/4Mh9W1qZyNiZT5bdA6mys79nVtcEbl4qpJL5rLVlPoHhnvEeNnCkhmPSDHgsW8LQBBbbfm69ns+RznyGlN3TGjgKy3z6viajvo6KiEYHg88BUJnPJANHwOhmPpt7Pv3BYDpP+kGZO7+0e0wnzV7IqqpbVgu4sFn2AoteVRuW94LjlBt8ptX3aUuJnUoDdsSBBe37K5Xp7GDK7qszOWhtaXiv53OEg6tqw0Le6DX4dP4bvGQ+etVSa+ZzdL8+RDRaGHwOwJYXfCYsu5dHM/jcqZooZHWIAUy7l7xp9+Db0/58o2i70ug5uG5Ou7d+AU8vFyFlvBNuotRM22uJuGlpGkB6uz53qiau7dbw1dc3Ij5/9KqlNPd8mrbEobl8x9vHI/PZOu0OOGX3WANHbvRZyEZlPruX3Ue5LYGIRguDzwHYLKuy++RkPlVrQNp7N4Omtf0mvey+W21gbbeGjVJvj31Yz6w6471uppD5dB+Do/NTyOgitcyn6rHcKNVD3yd01VLKez6ldI6MPTQTEHzmxmHVkt3Z85lG2d3r+exedu8nOI/T/kNEk4PB5wBs9pr5HNHgc7fS8K4x7eMuu51wlHaP6ahZ33O+V6yI0nOUsJ5Zteuzn8ynlLKl7K5rAicWpnEureDT/TdHrTAKPdtdld1TGjiyQtYsAaMffDYs54jajp5PXYu35zPqeE0jes+n+poYmkDdstkeQ0SxMPgcgK1yA1lD8/rSuhn14HOn2sBsvjX4TLvs3v1s98l8UlsvOacQ9dqbWfIGjlrf7mU++wg+TVvClq3DTDctTePCRiWVoyYtL/gMD+qqDQu6JpDRQ6bdUxo48tYsBZTdZ0Z81VJF7ULtY9VS1tC8I1T9up1w5AWfuvM9wol3IoqDwecAbJbrWJjOhJ7a026Ug08ppVt2d56ANaQ/cKSJZoaz3aSX3QeV+ZzNZ7Ayk+vrRYLaO+sv6d+0VEDdsr1grR/qRKFumc/2rCfgBFZZXUttyMULPgPK7oURz3x6i/gDyu5xVi1JKb2hqnZdB46qKvPpvB+HjogoDgafA7BZqscuuQPN4HNnBIPPasNGw5LNns+Uy+7luoVC1gg/XhPpfr5Ro/odrR4ziWEDRwBwerngHH/Y432rSXP/Y3PKHTpKY9G8Okt9JyKjWDOtwOATAKZzempl97CjNQEgo2vIZ7SRXaJeC9kIEHvgSAaX3AFA04Rbvg9+sam+/ob7CmVUs8NENFoYfA7AZrkRe9gIcMplWV0bycynuqZBld33amboEx/QDKrCllyPu3U3+JSyt69p1JL+0ytF2LL3rKoKOPx3fdOyU86Pk1Hrdt/qsqICFifzGdy+UsgaqR2vqQLhw3OdPZ+As2h+HDOfli27Pv62lJE/gzkjvHdUPXZe2T3Fpf9ENLl4vGbKLFtip9romvlUx9cBTmZpdkTPd1enG3VOu6d0tnvNDDxZRUk70zpqVNkdALYqdawGlH2jlGomsroWmDm+xev77C/z6Q9s1a7PfjOf13dr3v9Hld2d5enBr5ELOR17NRN6zPaWKKrsvhqQ+QScvs9RzeqF9ny6fbJ108ZURP+5bSO07A44E+9RPZ8CzReJo7yOiohGBzOfKduuNCBl/NONlLkpY6DBp21Lr0SbhGoF6Jh2T/GEo6AVL8rkl92bQdhWD3teS3UT0yHB+y19TryrjJo/sD02PwXdnWzux1pL8Nmt7B7875vOGiinNOBi2RIC4UFYMWeMcObTLbsHLJkHumfUncxneHCaM/TwsnvNhKaJZvA5ol8jIhotDD5TlvR0I2VuwJnPjXIdl7aqiXc0NsvuzpNy2me779VMb3I5yMQPHPl2XG5G7LsMU6pZgcvBAeDWVSf47LVFwiu7+xKLWUPDsfkpNPrcH7q2U/X+f7cWMXDUsDuWpyvFnJHewJElYegitPd4tIPP8LI7ANSs8ABdSgmJ8J5PoHvZXRPN7Piofo2IaLQw+EyZOt1osZA08znY4FNlKq9uV7u8ZytVdp9rL7unkPmUUjoDR1GZz0lftdRSdu8t8xmWtTo2PwUheg/c1WMs2nYR3LQ0HTrIFHdZePzMZ/C0O+AcsVmqm6msfTJt2xuaCVLMG9irpvO50laNUXYPo9pBZwKO1vTuJyL4dLZVCO9IUk67E1EcDD5TtllSpxslDz5rpt33Wdxh1JPHZsLS7k7FeTLpnHbvHgxKKSOfrGums7KnGNXz6XvfSeQ/3Ue9cEmiFJE5VpPKvWapqwGZT8Dp+5TofZAJANZ2nOBTFyI6+GyEl92LOQO2BPr9iak2nOEnPeykAzjleNOWI/l9WHWvSZ1w9NjDD+Kxhx/0Mp9Rw2Hq901Y9hxwTpcKG/grqeDT/R4Z1b5YIhotDD5TttFH2R1onrOcprXdqrfKZzNhgNMx7e6+vVtA07BsnFsvRwa7KksSnflML9M6aqSU2CjVvYGZpC8MAKBcsyJ7ZnPuxHPUEZZhggaOgOYZ7/1MvF9zy+4ZQ2C32gh9kVIzwweOVK9rvy/Y1LWoie0gxRFeNF+tR5fdozOfbvDZQ9ldSunt6VU9n8x8ElEcDD5TprJXvWQ+gd73PUZ58cqu9/9JgxBv4GhaZT7dvX9dgsHdqgnTll5JMIhaExQVPAFO8DNOZfe6aePqdrVr/9tezUTdspExnCfupANHUkpn4ChiklkFIK9c2w19nzDNns/2snv/pyet7dYgAGQ0ZxF6UHBj2xJ1K3rVEuCsqeqHagGILLuP8KJ5laFu742NF3w6/40qu4cFnzXT2QHsZD45cERE8TH4TJnKXiXNfM56mc/ULwkvXtnx/j9pabd94AiIHkBQ1PqXqNKseqKKGjgCnJ7DUSx3htmrmditmfjq2fXI91P9nqo3L+ljU23YsGV08K7u+9Uegk9vyXzb29NYNH9tpwpDb/YK7gSsW1LBbVjPp8rW9VstUEF/1MqmGTfzH7UWar+EDhypns+IgSP1tYvOfOqBLyK907W0wQ0cxe0hJqLxwuAzZVte2X10Mp8vXG4GnxulhD2f1QY00doTlstoXSeoK24pMOoYRrX6KWrNC4C+hmb2g1pYfm2nFvl+atJdBQlJWyLUQu+wVUtAM3B7uafgUy2wbw3KTiz2X3a/vluDrkX3CqrPHzbtrjK+/XaqqOy+1mXgCBjNPZbqRcJU2LR7jMxnZPCZ0WDaEmbbiw0VaGrC2RKQz2jMfBJRLAw+U7ZZcoI1lcmMa5A9ny9e2fFKikmzazsVE7NTmZYn5qjj9hQVfMbJfHYruwuR3olKw6CCsms70ZsFVAuECsKSlt29ntmIzLG671eu7SW6b6AZtLQnBPMZHYYmes58Niwb66U6DE3zAtug4FN9/qiBI6D/nk+VdQ06olRR+z93RzC4qnTLfMbo+YxaMq/2h7a3WajHTH3dirnMSAbnRDR6GHymbLNcx9xUBnpEFiXIoDKf1YaFszdKyBkadE14A1FxbVca3rCRksvoscvuEghdbq96PqOyLoAqu49Pz6cZO/h0J741AV0TPQSf3b9+QghkDQ2vXttNvCaouWS+87aMrvV8bvx1X4+lelETVM5WfcWhmU8v+Ex8CS3URoeIeaNmz+cIBlehq5bSGjhy77e9z9sru7tft6J74hQRUTcMPlO2Ve5+tGYQNdCTdubz1Wt7sGzpBJ8ieYCzU214gbGSM2KU3X09Yv5dln6lBJnPcZl2l1Ki4ZXdo4PPG3vNzKcuROKyuwrqowaOACcDtllu4PpedBtAu2rInk8AyOgCtmxdkh+XGvBx/t3O24Izn27ZPXTgyHl7v7s3vcxnnLL7CAZX3U44imqPUF+6bhsTgM7yvb/sDri7UEfw60NEo4fBZ8o2y3XMJxw2Anxl95Sr7mrYKJfRncxnD9Pus1OtT0xRS6eVii/bGRb0NAeOJqfnc6dqek/o3Xo+W8rumsBWJXzlUJC4bQsqeHg1YeldBX9BMVnGLele2qwkuk/At9rIdyxjUOZTBVVdB476DT4r3cvuozztXutn4Mj92kUHn879tlcf2oPPQja9E6eIaLIx+EyRlBKb5Xpvmc8Bld1fUMGnoUHXnCxP++BAGCkldmtmZ9ndiNHzGSPz2Rw4mpyy++WtZjC2thu/51PXBOqm3fJ166bs9tVOd/n69bpuyct8BgRlqq2kl/2h3mojXXTp+QwOqhTV69p32d3r+Qx/n5lR3vPpZYj7Kbt3H1oLzXy6n3aGmU8iionBZ4qkdEpcSXd8As6kakYXqZfdX7yyg5mc4WWZpETsYzxt6fybOsvuMXo+683b10Mzn0n2fI5H5tMffN7Yq0cO5dzYq2Eqo0MToqdF883MZ5eye4/BZy2i51MFn72U3a+7mU8n4+u8bSdy4GiwS+Z3Ks6i9LBz3QF/5nP0Vi1V6hZyhtbRNhBr4Mi9qZeez722gaNCznB3to7HC0Ui2j+xgk8hRF4I8TtCiFeEEN8UQjwmhLjVvW1VCPElIcSrQojnhBDv9n1cT7eNK5W1TLrjE3Ce+OamMqlmPqWUeOHKDu48MgMhhBcwxA1w1JN6++R+rLK7L4N3IyT4jHPCEdCcdh/Fc7XbqeBTPfFf3w0vvW+U6lgqOi9UvMcmQTBXjrknVRcCC9OZxBPv3rR70H2q4DNhHynQbEdonXYPGDgKyegpxbQGjqqNyJI7MPqrloKyw+2rloJ2Zqqf8bDsMhC/7D7KQ1lENFqSZD4/CeAOKeWbAXwBwK+6b/8XAJ6QUt4G4GMAfkMIkenztrGk1gotFJJnPgEnyEtzyfylrQp2qybuOjILwBfgxBxsUf+e2bbTT9TAUVQw6O/5vNHvwJEb/oxD9vPSlpPVU5PHVyOGjjZKdSy53yvqCTxuVhoASm7ZPWrVEuC8sLn90AxeuZps4r3asJA1tOCyu+in7F5F1tBajmUM3vOppt2DA6PpFAeOum2nyBk6svpo7rGsmlbgEaRe2T0i+27L6HYDIM7AkfN3FaCrLQxERGFiBZ9SyqqU8vdl87f8EwBOuf//owD+nft+TwK4DODBPm8bS17w2UPZHUDqmU91rKYXfCYMGFRWJGjaHYh+Umvp+Qz5fOpc6LCzuxUV+/RznOOwqMynyiSthQSfUkqsl+pYLKjMp/P2JBPvzcxxdNkdAG4/NIPdmhkZDLermlZo1rGfsvu1nRpWZ3IQQrjl7ujMZz6s7J5Wz2fF7Jr5BJzgaiR7Pht2x4J5IH7PZ7d/ezP4bMt8ul8L9eKkmFW7UEevNYGIRkuvPZ//EMAXhBBLADJSyqu+284BONnrbe2fSAjxsBDiovqzt5d8Wfaw2H2U3QEnyEuz51NNurdnPuMumldZ2KCyOxCdifT3fN4IKT2X6xYKWSOy1w5oPrmNw7qly1sV6EIg4+4QCpt436uZqJs2loo5AEjcEgE0B466tS0AwO2HZwAkWzZfCynnAk62S6D3gaNDs3kAzmNbzAUHdbUumU9dE5jK6H31fNq2xG614fWeRinmRnOgptawgsvuulq11Gfw6d53tb3nk5lPIupR4uBTCPHzAG4F8L+kfzmdpJSPSCmPqz/FYnEYn7YnKvPZy8AR4ASfEv0PUCgvXtmBJoA7DjmBR3NCOV6AY4X0fHo9YBHBoMp86gJYL4UHYHECJ/XcOA6DDJe3KjB0AcONZsJ2faqgraPsniDzuRezZxYAbl91fm5euRp/6CisnAvA6yFO2vNpWjbWS07mU5nNZ7qccBT+a6qQ6y/4LNVN2DL6XHdlVIPPSsMKDNDjZT7RNfAOzXzWTBRzzRePhREeyiKi0ZIo+BRC/CMAPwzg+6SUZSnlOgBTCHHY926nAJzv9bbk/4TR0ez57D3zCaS3aP7FKzu4ebmAKbc3TgU4sTOfKvgMWLUEdCm7uz2fhq5FLpmPUzJWYcGo93yalo2rO1U3+IzOfKpydbPs3kvm0/kaB5Vc291+SGU+EwSfDTt0wTvgXHPSsvuNvTqkhJf5BJwVPcF7PqMHjgAn4Onnx2U7xo5PpZg3RnKYptqwAlsTUi+7B2Q+/T+/M17wOfovEolof8UOPoUQDwP4MICHpJRbvps+C+Cn3Pd5AMAxAI/3edtYUgeJ9NPz6b+ffti2xLn1sldyB5LvZlRB8Fzbkvnm6pXwJxmV+TR050jPoN2ipZiZT21Myu7XdmuwJZDRnLU3hawenvncCws+k2Q+LUy5hwd0s1DIYmUmlyj4rEVkPgH0dGiB2n264st8zoT0UqoXG1GT2NNZo6+BI3W0ZtTpRspMzhjJs91Dp9316BeJti0hYwwcqfvuGDiqmi3DggVOuxNRTHFXLR0H8H8BmAfwp0KIZ4QQX3Vv/lkA3yGEeBXAowA+IqVs9HnbWGqW3fc/81lzn3D8wacmnFNl4mbXVBDc0fOpBz8Z+VUaNoT7+aQMzuiV3J7Pbsal7K6GjQy33/PQXD40+FStCMtuz6eAEyxsJ8l8xgzeldsPFfHq2l7sYK3asDuObPTThUC5bqFSj/+4qEzwakvwGVZ2j5H5zOp9ZT7Vgvmoc92VYt5A3bRTa4tJg5QSVdOKHDgK+zkt1VtXJYUJK7uXaiaKvqpIs+eTwScRRYv1zCWlvIjgdX+QUl4D8D1p3jaubClRyOqRpcoosymecqSyhHf7gk8hBOans7Gza7YdUnbPdC/nVesWNDSXp6+Xai3ZLill7MznuJTdveDTbaI7NJPH85e3A9+3vezuPDaZZNPudStW24Jy2+oM/vLMOkxbegNRUaoNy3msQ07QbE6813A8Ox3rGlTms73sXrdsp3zsC6K6DRwBbtk9IPqUUsYKSuMcramoLJ+UCPltOBhqN+djD3cuA5Hu9QSuWuqyZF4NBnUPPoN7vHdrJm7O6d6+WfX1GcXsMBGNFp5wlCLL7u10IyXVzKebpfBnPgFgsRA/wLGkRNbQOkp6YXv//MoNE0IAuhuI3dht/Zw104Zpy66n8wC+afchBJ9SSi+DndQlN/hUgd2h2Rx2qmZgZrC97A447RpbCTKfpZoZK3Os3OFOvMf9OtbMLpnPHo7Y9DKfs61ld6Bz12fYsZF+hZzuBmDNx6xu2ri0VcW59VLkCySgebJSnLK7yuyNVubT+W9Q2d3QnV2qYdPuajCo657PTOfPu3rx6C+7c8k8EcXF4DNFli17HjYCgFNLBQDp7LOsmzYWC1kc8j3JA84kftxTdGwpO3Z8Av5yXkTPZ91yjo30Zcf84p5uBPj2fA4h+Nypmnj9RglXt+Pvw1Q6yu5udi/ojHeV+VQnHAHA3HQGWwmWzJfr8Qa2lNsPORPvcb6Oti1RN8NXLQG97fq8rjKfM/7Mp/M91j50pDJtUdfQfr67lBL/9HefR6VhwZbdA2OV+Ywz7a4GatI+ArcfKugO+xpl3QMhgqjBoG6rzoLK7urrW8w1fz+on2WW3YmoGwafKbKl7HnYCABuWSlAoP8Mn5QSNcvGXe6xmn6L01lsVxqxsnu23Xm6ERBv1VLV1/MJdJ5ypEp+8cruKvM5+J7PumlDAqHl8iiXNivIGpoXyKy6wWfQxPt6qY6pjN5yNObCdAZb5Xrs4GavZnY9WtPvNnfiPc6LmzhrjppHbMYPPtd2asjqWktfdFjmM96qJVUKd75mj37lHH7za+e9qnjX4NMNeOPu+QTSGQhMi/pWyYUMhmV1LfTxLtXi9nx29nir7Ka/cuFlPhl8ElEXDD5TonrM+im7G7rmnJve6O8c84blTLHedXi247aFQga2bGZ8olhSdgwbAfFWLZXrzulFuhd8tmU+3WGHRANHXabdg86uTkr1275+o5T4Yy9vVXF0Lu8F/Ifd4DPoVKGNUq2l5A44ZXdbBh812XGdtkS1YXc9mtRvNp/Bkbl8rMynN+wTI/O5EbLHNci13SpW3NON/NcFBAWfljckF0YdsWlL4PFXruN/++ILOL1c8L623VpMvGn3WKuW3LaYUSq7w7mWsHVbWUOPyHyqloPoz9HcbtG8H9XXWfS9OFVL/xl8ElE3DD4jrO/VYh9H2Dxas7/j6XOGBktKrIWcChSHylC093sCzTVQGzH6Pm27j7J7w/IWkQPoWEae5GjI5rT74MvuKuvYW/BZwdH5Ke/vquUh6IjNjb16S8kdcMruQLx1S2rHpwq+4rrt0IyT3e0SQFW9kndE5lMkL7uv7dRa+j0Bf+az9QWR2jMaVRZWmc+qaeEf/MZTKOYM/Orfuh8Zd9gmbuYzVtl9zHo+Aef3SWjwWe192r3kDRm1/n4o5kdzET8RjRYGnxFqlo3dqukNkkRRGbN+yu5A80nkuUvJy75K2LAR0Ly+bovmbenkVNon3YHwpdPex7pZOU04wwzZgEXze20TslG0IZbd1eN4bj1Z8LlTbWC3ZrYFn6rs3hp8Silxo1T3TjdSvMcmRlY6SduC3x2HipBwsuNRmgveY2Q+Y5bdpZS4sdd6uhEQnfmMCn4BZ9US4AS15bqF/+9/9zacXinCjT1jZD4b3vny3TR7Pru/b6+klImqHqrsHrRkHnCG37qvWor+HGpqPrDs3taWU8wZ7Pkkoq4YfEZQT4qf/8bFru+rKtBpZD4B4PnLOz3fh8p03LraeRTpghvwdDtiU2V3gjKfKiAJK7urJykhBIQQWCpmO8vuSXo+h5r5dP577kY50cepYaNjvuBTrZZq7/ks1S13IKw1CFtIkPn02hYSDBwB8fs+mwvew39FaMIJbuJmPi3baU3xr1kCmhnFnfaBIzP6hCXA1/MJ4J/+jbvxnbctA4g/ib9TbWB2KtN16AaInnbfrjRwYbPcV+C1W23g9RvlRENnsQaOQqfd42U+hRDIGZr3ggTwld3bvv+KueADA8ZJGu07RBSNwWeEQtY5Peaz37jQdQhE3b5Q6C/zmfWCz34ynzayhubdl1/cAEcFYbNTncFhtkvms9yWUXGCz/aBox7K7kM44UhlPi9vV1qebLsJCj7zGR3z05mOzKfKFHaU3afiZaWB5tcvycARANyhgs8ugbz6t0dNmgshsFjIxg4+TfdnpD3zOROW+WzYoYM0igrw56YMfPSdp7y3qzJ6t80OOxUzMLsfRGXpg4LPr5y5gWrDRinBwv1237q4DUvKRC+y1JVMhbRfZA0tdNVS3IEjwHlR7L+usLJ7Iad7L4wOAgaqRL1h8BlBCIHZvIELGxV89fWNyPfdc3/hBmUbk9CEQFbXes58bpXrMG0ZOiGsguNuT8ph57oD4SeeKOpoTTWlvlTIYb1UayknJho4GmLZXb2IkBI4vxE/+3lpywkw/WV3wFkp1N6/e8Md0OkcOHK+1nF2farMcZKBI6D5/dktwIlzrjoALBZysQeOmsFncOYzaM9nt8//rttWcHxhCivF1oDWy3x2+Vo6mc94X8Nm5rPztq+dc34/hAV6cTznvuBM0lOq3jcsQ5zV4/R8dv88+YzeWnYPaZsp5jLYq5p9DUwS0eRj8NmFCr4++40Loe9zYaOM3aqJqYyOe47O9f05cxkNFzcrsTJg7f7vP34VQHjGanFaTQFHPymrAaqgsnu+ywlHKnBRT2rLxVxHVqiXPZ+DLrtXGxb8T5lJho5U5vPofGtgdWguj6vb1ZYnYy/z2R58FuI9NkDvA0eFnAFDE10zn3HOVQeA5WI2ds+n+p6KO3BUCzmz3E9NWLeXzYUQ0ET3SfydSiN25nMmFz7t/vVzmwAAs489TM9dcl5wJukpbQ4chaxairHnM1bmM6O1vPjbrYYFnzpMO1n2logOHgafXWQNDfednMfvf+tKx5Oj8qm/OAvAOT0oDSrb80LC7OcXnrmE//iX5zCV0TAXsJ8TaA61xM58Bk27dznbvVJv9nwCToACtE687yXI3Kn7GfSS+W231059/c/1FHy2Zz5zqDSsliMHNwIWzAPA/JTKfHYP5pIMbLXLuaXYqCxds+zeLfOZRaluxWpRCMt8FrIGhAje89kt8xlFFyKyt9m2JXZrZuALrCD5jAZdEx0tOHs102uTadi9r0nrJfPZveczfNWSegEYZ9gqZ+gtbS+lgFVL/r+P0sQ7S+NEo4fBZww/ev8JVBs2fu/ZKx23Xd+t4T8/eQE5QwvdtZdUL0NHL13dwc99/ltYncnh8Gw+dIBiJm9A10TXnk+VwAksuwcct+cX1PMJtC6aT5L5VN+kgy67q+BTBVxJM59LhWxHEOCdcuTr+2ye6955+hQQr+xedrPI0z0EnxldgwQi13lVvQXv0d/TqnUgTt+n6Qa77aduaZpwBlVq7QNHVtfPH0XTROSLrN2aCSmDv8eDCOFcZ3vZ/enzm74TluI9fu32aqb3/dbTtHtY8Bm1ZN7dxRtn2Kq95zOq7A7wlCMiisbgM4b3v+kI8hkNnw2Yev8Pf/k6au5RlnF+icehnnCfizl0tF1p4Kd+/RtoWDb+7UfeCkOPmFDWBOanup/vbkeU3bud7e71fHqZTyfY8E+8j+KeTxV8ZnQNM3kjYfBZ7ch6As1Ayz/xrjLA7WX3rKGhkNXjTbuHTBvHoc4xjzqDuxYz86n+DXFK75YtYWgicB3ZbD4TPHDUT+ZTE9go10ODOXXQQtyeTwBu8Nl6f0+6/eBq7dPFze6r2dq9eGXHK6EnOb3TGzgKzXwKNCwZODC5VzNj/85ygs/mi7+9iLI7EO+gBCI6uBh8xjCTz+D77z2Cb7yxiTNre97btysN/PpfvYFbV4veE08adE3g5OJ0rMynlBL/0289g3PrZfyT99+Nt9202PVjFgrZrn2FzbJ7xLR7SCayvedzyQ0+/bs+VdYlbrZYiMFPu2+XmwvHb14uxN71KaXE1Z1qR78n4D9is5n5DCu7A072M8nAUdJpd6D5uOzVwj+Pynzmu2Qevcc2xtCRaUuszOS84NdvJt+6okdKibrVvecziu72tpZDJtDVaqe4mU91ne3B59fObWA6q3uB2MXNZGu6gOZeX4Fey+7hx2sCwau19qpmrGEjwC27t2U+dU10fN4iz3cnohgYfMb0ofuPAwA+58t+fuaJN7BXM/F3H7wltayncs/RWZy9vueVsMNslhv4oxfX8EP3HcNH33lTrPtemM507fmMGjgKWjrt1575VNmx9ZbMp+X2+nX/uqk9g4Muu6v9ipomcGqpgGs7ta5ff8AJqixbBmY+Dwec775eqiOf0QIDx/npDLYqCfZ89hR8Ol/zqOyUynx2W3Xkld1jZj5XZzsDdEAFn81gWIVf/fZ8AuG7PtXRmkF9zWHay+5108YzF7Zw38l571SlXjKf33KDz3xGgy3jl967lt3dr19Qf2+pZsYaNgLcgaNGa/BZzHX+/BZ4vjsRxcDgM6Z33LyE4wtT+O2nLrrnuEv8+y+/jmPzU/jAW46m/vnuPTYHWwIvXtkNfZ9SzcR6qY47D8/gn//Qt8UOgBems9iqNCJ3l6qbggZaDF2LnJhWmSaVVVG7GG+0DByZiU7nac+8DIIqu+tC4NRyAUC8ZfNqwvlYYNk9OPO51NbvqSxMZ7HV5QAAIFnbQjsVcEQFCHFOOAJ8ZfcuL2aklDBt2bHjU5nJZ7DTkvl0/tst+I2i1i2FtTF4mc8kZfe80fJz89zlbVQbNh44tYiM7ny+XjKfz1/awbH5Ka9lJixb287LfIatWjLCN1PsJQk+3Rd/6vOp4LPdzAgOHBHR6GHwGZOmCfzI205gbdc5xm+n0sBGqY6PP3jay3ik6e6jztGYL4T0fdZNG9d2a9AE8Im/+bbQJdNBFqazsGwZmfmybQlNILR/NNs2gOBXqbdmPlWP341S68BRksCpfeBhELa9zCdwWgWfMUrvprsbJyjzuVzMQojW4HN9rxZYcgeczKczCBOd+VLBSdLjNQFf2T0q8xlz1VLcgSO1vL992EiZyRuom7aX3ZZd9lfG0e2UI6/nM0HZvZgzIH3X93V3v+cDpxa9z5c081mpW3h1bRf3HpuN9cLAz1u1lA0ruwefRialRKluJSq7+1+rhgWfzHwSURwMPhP44NuOQQgnSNksN7BczOJH7z8xkM91jxt8hvV9/uHzV2HZEgvTWdy0VEh0394RmxGDLZaUkVmRnKF5pdl2Xs+n+/esoWFuKtNWdm/NfD728IN47OEHwz9fZvBl952AzGecoSN1TnpQ8GnoGpaLOS/4lFJivVTvWDCvzLuL5q0uwafqueulLO0NHMXKfHYZOHJ7Prvt07Ss4DVLSvuiea/sPtDMZ/Ky+0zbovmvvb4JXRO47+Q8hBDI6AKXtpIFny9d3YEtgXuPzvn6ceMFbzYkhGi2wrQLy3xWGzYsWwb23wZR3wfq371XNTvWLAHNSknUC5tRwRVMRPuHwWcCxxem8R23LKFUt2DaEv/9d97c10BElNWZPFZmcqHB52eeeAMCyZ44FbWPNGqq2pbRT0w5Qw9d4VJuy3wCTgawZdVS3UrUr5iL2FeYFrVfU9MEbl5SZfckmc/gwOrQbM7r+ZQS3naEICpL3G3ReLluYjrbuVw9DvWwRmW+q414mc/ZvOGc796l5zPsaE2l/YhNFeT0lfn0ej6D2xh6zXw61+dMkH/jjQ3ce3TW6981NOeAiCTrkp5zf8bvPTbXzHzGDN6kdEruYd8HYcFn3HPdFfUiQKJZdg/KunPgiIjiYPCZkMp0agL4yDviDfj06t6js3j56m7HE9kr13bx1dc3UHRPq0lqPsaieduW0CPuOmtoodPnlbZpd8DJkKnMp5SyI/PZzbDK7gLOE/LcdAYL05nYmc+srmE5pI/TOWLTOeVIZTTb1ywp6rGxuuzb2aslC9794pR2VZa526olIZzVSd3K7ir4PBQxcAQ0Tzlqlt1TyHyGld176fn0nXL02vU9bJYbeOBUc8NERhfYq5leC0ccz7vDRvf0VHaXkY9RVldH07b+7DTPdY93jaqnVLrDUOW6hZmg4FM9jjWTmUUiCsXgM6G/fs9h5AwNi4VsooxJL+45Ooe6ZXdkLf5/T7wBIHgSPY5uR2w6QVJ0ViRq+rzaCM58bpYbkNLJnZi2TLSjMirYTct2pdGS7T0Vc92SaUkcmc+HZooPzeXRsJyvqQoql4rBgao65ahb2b2csGfWL052TWU+42QeFwvZrgNHKvhciZn59AaOUgg+w9pLvGn3JJlPVXa3gSfdIzXvbwk+k0+8P3d5G6szOazO5KG5/9zYZXcZnZ32Mp9WSplPGT2MWGDmk4hiYPCZUD6j4+TidOCi7LSpvk9/1qJUM/H5py7hzsMzXbNSYRZU2T0kYPCm1aPK7pnwk1OaA0fNt6lF85YtvSev5JnPwa9a0n0XffNSATf26qHHqiqmbePoXGe/p3LI7XO0LNsLwkLL7u5j0y3zWa5bPQ0bAYjVV1htOMMomaj0t2u5mOsafFq2Ot0oOPM52575RPSxkXE4p/dEZz51TWA6wbDejK/s/qQ3bLTg3W4knHivmzZevrqLe4/NudecvOwetStX9YI2QsvusT6N9yJEQnp7SIO+/1Q2ngNHRBSFwecIU09IVd8TxxeeuYy9momPvOOmnneLeue7h2SE1Nv1yMynHpqJLLtTtP6PVquFLFt6JdVRW7W0U2l4mScAsdYtqWD62EJE8OlOeKt9oED3snu3ReN7NbPnsrsQwjlLPbLs7ix4j/M9tljIYq9mRp7vrtZRhf27Vdl9J8XMpxDOaV5R0+5zU5lEP0de5tMNPk+vFFqy2BktWebzlWu7aFgS97ovNFUwWIqxXxZwgsFcZOYzeNq9lDTzaQRkPgMGjlQwv1cb7AvFICzzE40PBp8j7PjCFGbzRsv6mV9/4g0UcwZ+8L5jPd9vt+BTneIUtUIqq0esWmpYmG5bIK9WC5nSlzlJNHA02J5PKSW22zKf3sR7ROk9as2SorJ9/uAzdNp9Km7ms/eyO+AEHVGl0WrDih34LcbY9WnaEromQrPpoQNHfQ70Oad5hU+7zwYEUFFUqblu2ri4WcHbfSV3wJ/5jBd8qpON2jOfcY+ndMruET+nXQaOkhyvCajg03lwgno+AedrtNelWkBEBxuDzxEmhMDdR2dRM21IKfHU+S28eGUHP3TfscB+q7hmpzLQBLAZMgX8whVn+jZqzU0uo4VOn1cbVke5tKXs7n5Yoj2fGR2WLWGGlPr7VWlYaFitE/5xJt6bC+aDy8kAsBqQ+VwO6flciDFwZEuJhiV7OlpT0USXnk8z/tGWcRbNq3Pdw3QMHKH/gSPA6W+OmnZPui1CXWfJbS25vz341AQMTcQPPi8HB59xeyallJFHoIYFn+p41thld/d7QU35A+GVi2LO8O4/iO2rfhDRwcTgc8Tdc3QOUjoT1Z9xB436nbLXNYH56WzoIIZa75SLyHy2n3jiV2lYmGpber1cbAZVKnOSJIBWQcigsp9bvnPdlVPL0wCig8+GHb7jU1FHbFq27Q0ShWU+Z6cyECK67K5u6rXnE3CCnMhp94AXEGGa57sHfz/ZtnO6UXTwGTxw1O8qM5X5DPo+3ak0Eg8NquBTfR+2Zz6FEDg6PxW75/O5SztYLGRxZM75Hkm651NKRB4wEXa2u1d2T7jnUyK67K7eHnb9aztVnL1RwvYY7AElosFh8Dni7j3m9IKV6yZ+79krePupRdxxeKbv+52fznh7Ldu9eHkHGT28RAo4vWS2bE4x+5XrVscQxFIKA0dA8DGBadj2neuuzOQzWC7mosvuVvey+8J0FhldwLSczGfO0EKHXHRNYDafQVSCt9m20F/ZPWqQqmbaicvu/kME/FRQqifJfKawaglwMp+WLb3JdsWyJXZrZqI1S0Bz1RLg7Cw9sdj5uB9fmMKlGLs+pZR48coO7jk665W/4xwA4P94iW5l9+BVS7u9Dhz52mbCyu6FbHjw+dzlbUgg9IAKIjoYGHyOuHuOOuW49VIddcvGR96Zzm7RsHJkqWbi9fVS1xU7UcFgtW5hqq0krDKfpu2flk22agkYXOazea5769tvXp6O3PWpyu5R0+6aJrA6k/fK7svFXGSv3cJ0JnLVUtS0cVya5gQ4YQFStWHF7rdU/bxhZfdXru0CaD6GQYpZwxmCaj/hqM/gM+w0L9VykDTz6c/2PXDzYuDjeGx+Crs1syPgbVe3bNRM2yu5A80hvTjT7uprFFl2V8drhu757L3nM7Ts7mY+g763zl53fpaCXrQS0cHB4HPEnV4uQMApdS0Xs/jeew6ncr/z01lsBZQjX7q6Cym7P+lHlcErDQtTbdmYYs5A1tBap90TnnDkfL7BZExU2b0923tqqYCtciO0B7Nh29BFdOkTcPo+VfAZVnJX5qezXl9dkGbmuL/Mpy2bBwK0q5k28kkznyHB5wuqjSPi/jRNoJg1AvZ89ld2V6d5tQfGzQXzyYLPaV9A/sBNC4Hvc3zBade4uBVdelfbIu496gs+hYAm4k27e1+jGHs+G/1Ou2eaZXf1ecPaZoo5w/k5D7jttevOMGO3gToimmwMPkecoWvek/aPPXAiMnuUxGIh42YhW9/uDRt1+TzNTGRn8OIEn61PiEIILBeyfZfdB5X59J/r7qcm3tufvBXTkjAiemOVQzN5WLGDz0zXgSMAfQ4cRe+TTJL5VCc7bYQcsfm8O1TTLZCcyRte2V39G3vdZasshJzmte0drZnsa6hpwitVP3DzYuD7HHfXbnUbOlLfy6q1xvscQsTKfMb5GnU/XrPrpwHQfOxsf9k9rOfTtwu13WtrbubTYvBJdJAx+BwD+awOAeDDbz+Z2n2GTVW/4AUK3TKfweU825aoNuzAwGh5Jtf7wJH7BDuoU46Cej4B4GY3+AxaqG/FGKRR1K5PifBdl8rCdNYd7Ah+gu6WeYrDO989oDdPSulsLIj5Qmd2yjnmdb0U3PP5wpUdGJqI7PkEnB7bjsxnnwNHiyFl914zn4ATHGoCuPPwbODtcYPPqmlhJm/g5OJ0x/1H7WBV1NcozpL5oOAzZ2ixVy3lfSccqd8ZUWV3AN5WCz8v8ynlwPq3iWj0MfgcA0uFLG5amvbKeWlQvXDtAc4Ll3ewXMx2DRTCMpFV70zwzifEpbbM53SSVUsDLrtvh2U+3XVL7SfEAMA33nCOV4zah6ocmmuuYlI9kmHmp6N3faqSfJKTedpFZT5VRjzupLkQAguF4PPdqw0Lr10vRa7tUmbyRnPJvPu2tHo+2zOfvRytqcxOZbAwHf4zctwNJqMm3qWUTr/n0bmOAFAT8VYtSS/zmfx4zVLNDM1cBmmecOTfthD8ecMyn1vlesv3yPWQATUimnwMPseAJkSsACeJhYAAx7RsvHR1F3cdme2aEfHK7g275WQRdbRm+6olwJl4l2hOiI/UqqWK86SotV22WrdUbysTmpaNX/jCcxAA5mJkz9QRmwCwWAje8amorHRY32fU2dpxRU1Vq5OKkgR+SyHnu798dded8O8eyPrL7mlNuy91zXwm/xouFbKRrROHZnLQu+z6bFgSUnaW3AHnsYlVdnf/G1V2DxsM3KslO561OXDkVC6yhhb6mIYFn6+5w0bqBd7V7Wrsz09Ek4XB5wEVVHZ//UYJNdP2JuyjeE9qVmsmUp0LH1QKVIvVG5aEJqLLhWGfb3CZT+fJvj3zOZ01cGg219Hz+em/egMvXd3F/HQmVh+u/0zzbmV3L/MZUnb3ej7TKLsHBDlVt7UhyY7NpWI2sOczbg8x4JTd1YEK6tsyjT2fQFDm0wk+47xwSMrQNRyZy+NSRPCpvo/9k+6KM3BkRQ6dAckyn+0v2koJj2fN+crutpSRL3wKXvDZ+nZVclfDeWs7DD6JDioGnweUytz4AxwVKNx9NLiXzU/14rX3YKqsWfuqJaC5bqlh2Si0Hb8Z9/MNcs/nVMhZ5qeWCmhYtvdkv7ZTxSOPvYJj81Ndh4cU1fMJhC+YV+a9FwbBt0uvZzaFsrsv86ky2DWvdSL+r4fFQg67NbMj2/V8zB5ioDnAYknnhCNNIFY/beR95px+1Pa1Yjs9rlqK6/hC9KL5qvt9HPRCzzvlqMvEu7eIPyKrnAlZMr9XMxNWHlTZ3XlhEPWx4ZlPJ/hU7SJXGXwSHVgMPg+o+YDMp1qJc/eRGMGnHpxRUat7grKaqtdRIvmOymzI50vLdrnuZRzbnV4pwJbNQP2f/f6L2KuZ+MUP3BN7Vc3qbPyeT9USETZw5PXM9jHtrns9n527XnvKfBbCBth2MD+diRVEqlOOnOMXnYAnyQuUIKoftf18d5X57GXgKI7jC9PYqZqhfbu1hgUhmgNtfs0jNqOz/F52OOqEI7VqKWDgKOyEoiDqxYMdI/NZDMt8rpVgaML7vrq2w55PooOKwecBtRgQLDx/eQf5jBb4hNjOmz5vDz69snvnt5b/PPMkw0Ytn2+A0+5hJdjm0JHEV87cwBeeuYz33bmKh+4+FPv+Z/OGt0B8qUvP5/xU9PnuzROO+i+7p9XzGfT9ZNkSL13dbTnBJ4rKfNrSDT77XLPkXdt0trPsrno+B5j5BJr9zX7rezVUGjamM3rg0FLzsQk/gQrwld0jHqeg4zUblo26affY8+lMsUcGn960e+v379nrezi5NI2Me5IDy+5EBxeDzwNqru0McSklXriygzsPz3addAfCezDLXtk9aNq9GXQlHZYZfM9nIzQLdsq3bumffOE55AwNv/iBexLdvxAChvuku9jvtHsP2wLaqYGjoJU+6gVF0p5PoPWaz62XUK5bsTLpQHPnphN8yshychILhUznwFGlgYwu+t4jGuaYe9xqI+Ax/NLzVwGEn43ebImI/l73TjiKeJxUoOdvV1GT9EnaNgxdg64Jb+AoKmuq7td/gEXDsnF+o4xbVoreqiqW3YkOLgafB5SuCcxNNc8Qv7ZTw0apjnti9HsC4YMM1Xr3nk8gedauuWop/cynbcvIzKfKBK+X6njtegl//z234sRi8rVXhqZBoPuZ7AsB/bgt1+tOG/ezASFq1ZI6dzvJjs2loB7iy/F7iAFf2V06gVVqmc9CFtuVRksWcqdiYjaf6busH0atRQs6nOD3nr3ifh8E/wwILfyx8bNjDBwJIZA1tJbgU2W7k/8Mat5jE112bz6OyhvrZZi2xC0rRQBOL+81Bp9EBxaDzwNsYTrrZapeuOIMhsQNFMKWzEf1fPoHbZL2fA5y1dJe3YQtgfmQ4FMtAbdsiZuXC/jJd5/u6fMsF7M4PJfvGvCo4DRs2ll26bmLQ0SV3c1eyu5OVru9jQMIHqoJMuMr19pS9r1mSVmYzkLK5i5XwCm7D6rfE/CX3Vsfw+u7NTxxdh3T2eCSOxDdEuEXZ8k84PRn+8vuXvDZw89gtwXzzm3N05AUNWx0y4rzQk7XNKyx55PowGLweYAtTGe8TFWSYSMgPBiMCj4NXfOeWJNOaue9nk/n/v27Rfu1XY5eu5PP6N7AzP/6gXt6Xv+Tz+ixgkYhnNOAwjOf/S2YB5zMZ9bQQjKf6ZTdX7iyg5yh4XSMHmKgmfm0ZHPgKA3qRY9/6Gin0kh8tGYSR+by0DWBRtsxP196/qozLR7xuYM2EQRprlrqfhRuLaDsnmTJPOA8Hup7MupjVUY1MPhcdTOfunOKU5xl+kQ0eQb325cAAI89/OB+X0IolfmUUuL5yzuRRwa2CzvbvblkPjhwMDSBuiUT76gcZNl9O8bOxwX3vPV3376S+ucPogsROXDUz7CRMpMzAns+vVOqEi6ZBzq3J9x5eAZGzPaA5sCRU9pNqx9T7bT1r1vaqZpeIDQIhq7h8GweN9pO8fm9Zy8ja2goRjx+WsQmAr/mkvnoID2jt5fdncc3ceYz08x8Rr2I0jSBQlZvKbufdRfM37LcLLsDwLWdKk6vDO5xIKLRxODzAFO9hVI6WaqblwuhQWM7LxhshE27B9+PrmmAZfUxcDTA4DNk1RLQXE01LJoGNEKSQrbsPNqwlxc5xbwRmPlUq5aS9HzO5jNOttZu7kK9sVfDQ3evxr4Pf9ldyninIsXhne/uTrxLKbFXMwc26a4cX5jCle3movm13Sq++voGvufuQ14wFkRVB0r1LgNHbnDXrTc2a2gtvaelHsvu/gGwbh9bzBteRQFwMp/Lxaz3M9YMPms9BZ+q6jHKL+6JKBzL7geYelJu2DbeWC/H7s0D/CcchZTdQ4JY1eeWNHMXNuCUhjiZz2HThXDLz53ZTyll4sAhSDFnBJZ2VWtDksynpomWHuLnryRr4wD8A0eq7J5S5rOt7K4ycr0crZnEsYUpZz+s+wm/9NxVSAn8wJuORn6cynwGnT7lF+eEIwCdA0fV5NPuQGuQO9Pl+6+QM7yvs5QSr63ttQSZhnuOLYeOiA4mBp8HmFrpo47EjDtsBLSe7e4X1fMJ+ILPpE98Xtk9/VVLIxl8hqxCUkdPplF2Dws+qz2sWgKcgaqOHuIEL2j8J+OkOu0+3Zr5VL2I3TKfjz38YF+ZNTXxbrp9n1989gpyhob33RmdDXbjsq79kCq46zZwlA0ZOFJT6XH5Xwx0W1A/kzO8r/ONvbrT5uALPnW9WXYnooOHwecBpp6UveAzQZYqbO9mnJ5PIPmez4OW+VR7OLfajoVUedB+dnwqM6Fld7VqKdmvh8WCb3vC5R0IAdx5eCb2x+uaQDFnwLTV8vT09nwCzeBTXeMgp92B5sR7w5K4tlPFk+c28N47V7tmrZMOHHVbudWe+WyW3Xt7Aeh8bJzMp3N97ZPuQPP3AHd9Eh1MDD4PMNXHqLKVdyUJPkPOWo+b+Uw6cKRrAhldDOSEoy23N23YfZ1R1PGX7cdCqvVLaWU+65bd8QLCWzKfMPhbLGS94xdVD3HS9oCZvOEFh2nu+QTgnXLkZT6HFHyalo0/+NYVt+R+pOvHCTjf792CT1tKxDgPorPsXleZz976ruN8bNEtu0spm8NGLWV3dcoR1y0RHUQMPg+wRd/A0epMDisz0cc++oWdtV6pW16gGGQ6q6OQ1fHO00uJrzdn6Aeu7N4RfLqpzzR6PtV9tJ8hrjKfScvuauLdtCRev1FKlElXWoLPlDKf01kD+YzmnXKkth8NctUSAJzwFs1L/N63riCf0fDeLiV3wFm1VcwFZ6X9LFvGPo0sqOezl2l3pduaJlWWl/BnPpvBpxACy8Usy+5EB1Ss4FMI8StCiHNCCCmEeIvv7eeEEC8LIZ5x//yY77bbhBBfEUK8IoR4UghxT5zbaHgWfNPdcU82UjK6gBABZfeGhamMHrpIPaNrODo/lSjQVXJt+wrTslNR53yPzvIHFdz/7jcvt7xdlVq7nZIUhwoQ2oMcb9o94cDPUtF5TMtuZi1JD7Eyk894AXZaA0dA6/nu1pAyn4fn8gCcYP7Jc5t4352HMB0zY13MGSjV0wk+M7qGWsC0e/LMZ4Jp91xzc8Fr1/eQNTQcczPByupMvmvZvW7a3oshIpoccX+7fw7AdwJ4I+C2H5NSvsX98599b/8EgE9KKW8H8EsAHo15Gw3Jgu/EoaSBghCiI6MCOJnPuOuakhpU8LlVqaOYM2LvoxyGfEbDdFbHbz91CU+f3/TenmbmU00s79Za+0rVC4qkmU+VSVc9xEm2J3jX5HsBkGTVUzcLhWwz8xlz4KhfGV2DoQlvgCtOyV3plvm0bAlLNsvXUbK6s2pJvXDZq1kwNJE4uE9adgec79fXru/h9HKhI1A+PJfH2k4tcKODcm2niivbzI4STZpYv32klH8upbwY906FEKsA7gfwGfdNnwdwQghxa9Rt8S+b0uA/TvLuI8kDBacM3tnz2W36tle5jO6tAUpT1Lnu+0UIgZViDoYm8Iu/+7zX66kCp6TDIkFUgNAe5NR6zXwWWnuIeyu7Nx+HVDOfhSw23eEt9bWcG/CqJcA5yQdweqDfc0f8nafFfPABAIpqx4iT+cwaGvzxXalmopAzEp9rnyT4VC+OLFvi4malpeSuHJrNoW7ZXs91O9OyUTNtmLYMPXCBiMZTGr/dPy2E+JYQ4t8LIdTxLycAXJFSmgAgnZe25wGc7HJbByHEw0KIi+rP3t5eCpdMQOtxl72USNuP7QMGG3y2r4xJyygGn4Dz9f1b33EK37y4jc895bz2U8Fn3PJtlKIb6LUPtlQbFrK65k3cx6Uyn3YPPcRKS+YzxeBzYTqLvZoJWzoZQ2DwmU+gOYn+vrtWE1UECjkjctWSmtzXYwSQalOECkBLdTNxyR1oZqLVQFQU9TjWTAtSAqdXOo9YPTTrtCWEld7PrZe87Q5bbb3PRDTe+v3t/m4p5ZsAvBXADQC/1v8ltZJSPiKlPK7+FIs8ii1Nuub0bt60OJ34Y8PK7vlBld0z2kCm3bfLoxl8AsA//O7bsFzM4pe/9DJ2qw2v7N5L8NDOy3y2B5+m1dOkuer5BHp7MQMMruzuBca29DKfg+75BICMG6S9P0HJHXBaIsp1KzTjt76XLPMJNNd07VXNnjLn6sVAnBclahuD6h8Oznw6wWfY0NGLV3a9/18vMfgkmiR9BZ9SyvPufxsA/hWAd7k3XQBwRAhhAIBw6jsn4WQ4o26jIZubymBxOps4ywWoHszOgaPpQZXdB9DzKaXETtX0Fu6Pmtl8Bv/zX78DN/Zq+H/+5Iwv85nOnk+g8ySdWsNO3O8JNMvuQG8ld6A1G5nkhKVu1PnulpSwpURW11LNrIaZm8pguZjFd991KNHHqeAwbOhoveSsKIoVfOoq86l6PnvLfKrviTi/KtQwW9X9/RBWdgfC1y29dHXH+38VbBPRZOg5fSKEKADISCm33Dd9GMDTACClXBNCPAXgI3CGiT4I4KKU8oz7saG30XAt9LHbMmvoKNdN78kNGPTAUW+rli5tVUIHM1RiaVQznwDwI287gc88cR7/4cuve1/btI7XBEIynz0EZv6vYS/DRsAgM5/OtTmDOhKzU8l7Hnth6BoWprOJh9nU6UNhQ0de2T3mqiWgmflUPZ9JeZnPtq9b0ClQ6nur4fY49FJ2f6kl88l9oESTJNZvICHEJwD8AIDDAP5QCLEL4HsAfF4IocNpAzoL4KO+D/s4gEeFED8PYAfAx2LeRgPQzzGBYXKGhq2y7QWfti1RM+3BDRz1kPncKtdRrlvQhJP5aQ84LG/4ZHSDT00T+MUP3I0P/tu/8rKUqQSfEauWesl8aprwzqRPpeyeZubTzcpatoRtD6ffsx/q3PWwvs8kZfeM3uz5lFKiVLd66/kMCT6D+O//yFw+8Pu1W9n9pau+4HPMMp8PPfI4gMH83iWaBLF+A0kpPx5y030RH/MygHcmvY3GhwoGVaufd7rRAHs+66YduZql3dPntwA4Gc6rO1UcmWvdNTis02769babFvGDbzmK33nG2fuZxrT7TEjms2ZaPQfjuiZg27KnHmLnmgY07a7K7rbzmI/6461eGIRNvCcqu3uZTwlbOu/fW+Yzftndf/9BJXfAeUwyuggMPrfLDVzaqni/Y9b3mPkkmiSjs9iQxk7W0FpWH3U7WrNf6skvSfbzKd+OzBev7HTcrjKfo9rz6fdz33cXVNIpleM1Q3o+qw2759OFlopZrBZzPfUQA62Zz16yr2G8zKdUZffRfrybp091KbsnnHZXL9x6m3aPn/n0P45BJXfAyZSvzuRxLaDnU/V7qt5mDhwRTRYGn9SznKG3rD6q1Aec+TSCj/SM8o03/MHnbsftKvM5ymV35fBcHqszOczmjVQGjqYyOjQB7LUtma82LOR7PFe9mDP6CuwGuecTACzbhpSjdZpVkLAdrMr6Xh2aQKy+1ayv7N7PtoRE0+4xMp8AsDqbC8x8qpK7eiE7bmV3IorG4JN6lstoaFjSy6b0eiZ4XCqD077eKYxlS3zzwpb3pBmY+RyDgSO/2XwGh2bzqQzLeGeId5TdbeRTOlc9qdaez/SuQWW21QDMqGc+w4bBlPVSPVbJHWgvu6tDCgZbdvdvvIgKPg/N5HFjr/OUI5X5zGU06EJw4IhowjD4pJ7l9NYpWnWsYhpZucDP52U+4028v3x1F6W6hemsjowuAoNPtfNxfqr3qf9xNpPPtGTXpJSom3ZPez7T4B8ESvMacoaOYs5Aw33hMvoDR9HB50YvwadsZvqL/ez5jPHCR9OEF6TeshpcdgecbL4t0bHP9MUru1guZmFoGnRN7Hvm86FHHveGiIiofww+qWcqOFBJi1Hr+VT9nvmMjqyh4fUbJS87q1hjVHYfhGKu9RhHFQLsV+azOKCeT8ApvTe8BfMjXnYP2UQAOIHaZjl+8OlftaRivF4yn6eWCxAifjuEJpwDLA67U+1BVt1dn6Yv+LRtiZev7uLOw87GBF0T7PkkmjAMPqlnKhhUJbPBB59u5jPmKUdPvdEMPnOGDlsCr1xr7fu0x2DV0iAV80Zb5tP5b5o7NpNQJ24B6fZ8As2hI2CMMp8BS+a3ynVIidDdte38S+bV93svPZ9H56dw60oxduCaNTRMZfTIFpFDM05g6g8+39goo9KwcNeRGQDO98R2pRG73YaIRh+DT+pZ+7F9VbfsPsjjNYH4Zfenzm/i1NI0DE2E9n1aUkKI1l7Dg6S951O9kBjG6T9h1AR32tew6NtoMDY9nwGZz/UEk+5A657PNI9n7eboXB5H58KznoBTdgcA02oGny+5P6P+zCcAbPJ8d6KJweCTepYzWsvuXs/nCJTd1/dqOLdexltPLrgfq4LP9synkwXrdTXQuCvmW88QV8HJoIbG4tC84DPda2jNfI72i42oVUtJFswDrS8S+xk4SkoI0XUw7pBXdm/+TL/oTrrf6ct8Apx4J5okDD6pZ83gs63sPgKrltRy+ftucoJPQxOYyRmBmc+DWnIHOhfNSzeP3euqpTSomCqjp/uCYNF3lOyoZz5zhoaMLgIHjpIcrQn4B476m3YfBHXKkb/s/tKVHeiawK2rzpS8F3xy4p1oYjD4pJ6pJzUVCg561ZJXdm90L7urYaO3npwH4GRh7jwygxev7LSsdbHtgx18tk9Vez2f+zRwBDjnoTu9n+kGn+PU8ymEQCFgDRaQ7HQjoD3z6bxtGGX3OIo5Z2etf9r9pau7uGWl4H0PGuJgZj45YU+TjMEn9aw5cOT8fdCrlrK6c7/+xfZhvvHGJgpZHXccmvHedufhWexUTVzebi61tqQci9ONBqV9qlp6Zff9+9WwUszi+PxU93dMaNEffI74tDvQ2Y+rJC67t/R8ugNHI9J2IITAodm81/O5W23g/EbZ6/cEmv/OGzxik2hiMPiknoWW3fd52t20bDx7cRtvPjEPQ29+i991xHlCUwMNUkrntBtmPr1TjlRwsp89n4auedm6NC1Mj0/mE3Aem1KtM8uf5GhNoPXnVP2sDqovuxeHZnNez6faRqH6PYFm8LnBdUtEE4PBJ/Wsfdp94MdrZuL1fL50dReVhuUNGylqdYvq+xy3040GQQWf6nx39Vju57T7oKjMp8D+BtdxFXOG97j49VV2t4FCVh+pAbtDs86ieVtKbyBQvVAEOHBENIkm7xmGhqZ9z+fAez69affonk+v3/Om+Za333F4BkI0J94P+o5PoFl+VRk2OQKZz0FZLDiP8ygFXlEKOSN02n1+OhO7JzbTVnYflWEjxRs6sqR3rOZdvrK7JpyBQQ4cEU0OBp/Us9BVS/s87a6Wy993ojXzOZ01cGqpgBevqsynOlrzAAefbWX3Uej5HBRVdh+T2BPFvIFKw+o493yjVG/pX+2mfeBoVPo9FRV8WrbES1d2MT+d8VYwAU5f6FIxy1OOiCbI5D3D0NB0lN0bFgxNeJmWtDWn3aODz2+c38TplULLdLNy15EZnLtRQqVuMfOJ5nJ9Vd61R2DafVDU4xy3V3K/qTVYbceeY71Ux3IhF/ARwdpXLY3KpLvi3/X50tVd3Hl4piOru1TIsexONEEYfFJsjz38IB57+EHv7+0DR9WGNbBhI+fzdS+7m7aNCxuVjn5P5a7Ds7Al8PK13QN/rjsAFHPOv32U9nwOiqFryBmDGWYahIIXfDajT3Wue6LMZ3vZPTtqwaeT+aw0bOzVzJZJd2WpmMU6p92JJsZ4/BamkaTO/1bPjZW6NbCjNYF4ZfeqmxUNCz7v9E28e5lPrlrqWLU0iZlPADixMIXVmfhZw/1UDAg+1bnui8Uegk84WdRR6/k87AafZfcc+7t8k+7KUiGLUt3y+sqJaLyN1m8hGiu5tiXz5bo1sH5PoFk+rEcGn86TU/uwkeKfeD8I0+7+THWQziXzk5v5BJD64vpBKgaU3dW6oeUEmU9NE8joornjMzdaLyxW3BcDDfcH0j/priwVnfdZL9VxbAA7YBUpJcp1c6AVHCJi5pP6kB162V1lPsOzH5WGhZmcgdtWO7MnAHBsfgozeQMvXtllzyd8q5Zqk9/zOW5UVtr2RZ833L7HJGV3wMl+qlOERm3gKJ/RvSEwTSDwZ1f9ewddev+zl6/j0lY1cLk/EaWHwSf1rH3avdKwBrqip9nzGZz5lFKi1rDxlpPzoTsQhRC46/AsXry6w55PODsUp7N6s+zuvn0SVy2Nm0JE5nOxmKx1IGM0g89RK7sDgKE5v0tOLRcC9wQvF1XwOdiho+cvbwOIrq4QUf8YfFLPvD2f7t8rA858ZnQBIcKn3WumDQngvpB+T+WuIzPYrZpeEDtq07/D5j/GUWWxcxNadh8nMwE9nxvursskZXegLfM5YgNHAGDozovFuwKGjQBn2h0Y/BGbr67tAQAa7SsGiChVfIahnrWX3Qfd8ymEQM7QQsvuXr/nyfnI+1E9ZXXThi7Gqw9wEIp5w8t8qufcPMvu+y5o2t0ruycYOAKcn1V1L6OZ+XR+Bu88HNwus+T+ewd9xOYZN/g0LWY+iQaJwSf1zF92l1KibtoDnXZ3PqceWnZXb/+2Y3OR9+EfaBiX024Gacaf+YSEJpwsM+2vqIGjxD2fvvVSo9bzCfiCz4BhI6CZ+Rzkonnblnjtupv5tIaX+Xzokcfx0COPD+3zEY0CBp/UMy/4hPT6Pgc9JepkPoODT9OWEOj+xHz7oRlvwGFcFo4PUjFvYLfaPOEon9EPfDZ4FAStWvKCz+nkZff2+x0lxbyBYs7AO29ZCrxdZT4HWXa/tFXxVrWZtmT2k2iAGHxSz4QQyOqas7zaLeoNsuwOOL2IYWV305IwdNE1cJrK6ji1XADAzCfQ7PmUUkJK6b2ooP0VPO1ew/x0BkbCU8T8j+kolt1zho4jc/nQwHg6qyOf0QZadj/jZj3Vr4SrO9VU758ZTqImPstQX3KG5pbdnb8POvOZ1bXQSVTTtr3yXTeq9K4xw4diLgNbOoNjKvNJ+6/g7uNsL7snLbkDbWX3EdvzGYcQYuBHbL7m9nuq32EXNysD+1xEBx2DT+pLLqPBhvRKg4MOXMJ6PqsNC7YEdC3et/Rd7mBD2Eqmg2TGl2GzweBzVOQM3a0stJbdl3oIPjMtZffxXC026CM21bCRygwz+CQaHAaf1BdVdvcyn8MouwesWrrmlsiMmIMyKvPJuZrWwRaW3UdLMW94L+wsW2KjXPeGb5LItpTdx/PFxVIhixulekswnqYza3uYn84g736tLjH4JBoYPstQX3IZ3T0zekg9nyGrlq7tOBmRuGX3d5xeQiGnj2T/27B5vYXSGRzLMfM5Mgo53Su793KuuzLqA0dxLBVzqJs2SvX0z3eXUuLVtT3culL0+mkvbpa7fpxp2d7+VCKKj8En9cXp+WxOu+9X2V0NB8QNPgs5A0fnplhiRutUtZTSy/zQ/nP6cZ0fLjVs00vZPTviA0dxLA3wiM0be3VsVxq4dbUIXRPQRLyy+8XNSuqDSUQHAZ9lqC9Zd+DI3udVS9e2VfDJb+mkvJ5PCdhg5nOUFHM6bPfbfT2F4FOgtf9znDTXLaU/dKT6PW9dLQIADF3Dpa3o4HN9r4aGLUNPXCOicOP5W4hGRs49OUX1YQ08+MzosGzZ0fd1NWHPJzV5mU/byWAz8zk6irlmz+f6Xm/nugPNVUvjvN1B9boOYt2SWrN0ixt8ZjSBy1uVyJK6ClgtKbFdaaR+TUSTjM8y1JecoUNK6WU+h9HzCTQHnJSkZXdqUsGnNaSNBRRfIWd4L+7Uue49ZT51FXymeXXDpTKfgyi7qzVLt640M5+mLb1BxiDqHHgAOHt9L/T9iKgTg0/qiyq7S3fJ/KCP1/TOk297+7XtKnTRfcE8dVIDRyrLw2n30eFvifDK7j0MHKlSuxjj6HOQR2yeWdvDVEbHsfkpAM3jZaNK72d8wedr10upXxPRJOOzDPVFld2H2fMJILDszpJ7b1Tm07SZ+Rw1hWxzGMwru/fR8znGsedAj9g8s7aHW1YL3olnGa37xPsZZj6JejaeY480MlQwqI4AHHzZvfPUFykl1nZqDD575JXdveCTr0lHhX8Nlup1XEh4rjvgDz5bf0Yee/jBPq9weFTQnXbP5261gas7Vbzj9KL3NvW75OJGeObz1bVd58Q1y8ZZZj6JEuGzDPVFPalZwxo48srurae+1K34R2tSq86yOzOfo6I5DAasl5xz3XuZVg8LPsdJPqNjJmekfsRm+6Q70GxTCCu771QbuLZTQ87QkNEFzt5g5pMoCWY+qS9eJtIeTs9nLtM5cNScdOdrqV6oYxzDMp/jlB2bNP4drOt7vZ3rDkzGwBHgLNhPu+weFHxqAihk9dBdn+pjsoYGS0qcu1GGZUse10sUE4NP6ktuwJnP9sBHBbv+4PMaJ937Vswb2HTLmez5HB3tZffTK4We7mcSVi0BzqT/hc0K5qc6z6eXUqJctyClTDR4qNYs+YNPIQSOLUyF9nyeueYLPm2Jcr2Bi5tl3LTU2+NDdNAwVUR9yfp6PjO6GPgC66Cy+9XtZEdrUqeiu9IH4LT7KCnkmtPumz2e6w5MxsAR4ByxuRlyvvt2pYHL21X84fPXEt3na2t7MDTRETgeX5jG5a2qV9Xxe3VtF4CTUVZfW/Z9EsXHZxnqi8pEWnI4GbOgPZ9cMN8//3nfPOFodMy4j0vDsmH3eK474As+xzz6XCpkYdoSQbvf92rOme9fee1Govs8s7aHU8uFjhfOxxemULdsXA8o859Z20NW11pecL/GiXei2Bh8Ul9UD6Zly4EPGwG+TKvvyWdth0dr9kuVdwGW3UdJwQs+nW/4XhbMA80BmrEvu7vBd/vJQ7vVBioNJ/j86tmN2PdnS4nzG2Vvubyf2vkZVHp/dW0Pp1cKEEIg677o5a5Povj4bE19yfqyBVMDHjYCfD2f/rL7ThVZQxv7kuJ+mvFnPll2HxlFX+YT6G3HJwDcvOyUlLNjPpSn2g7ag88/f8XJdgoAL1/bjb2OSWWU/f2eyvGFaQDoGDoq101c3Kx4H6NrAjN5g7s+iRLgwBH1JeebjB5G5jNw2n27isOz+ZE53Wgcp8OZ+RxN7cHnUg/nugPAPUfncMtKYWIyn2Zb8PnHLzl9nvPTGWyWG/ja6xv43nsPd72/uul8XYODT5X5bA0+VW/nratFvHx1F0IInF4p4uwNZj6J4hrvl8G07/w7IYeT+ewMPq/tOMEn9c7f85ln5nNk+AeOgN7L7sD4l9yB4MynZUv82cvXkTU0zOadKfivvr4e6/7qbjtDUPB5LCT4VMNGt63OeG+7ZbmA67s17FQbcf8pRAcan2WoL/4S7VAyn21l92rDwma5gUNzDD774c98cuBodGQNDf6Qsdey+6Twej59rz6fubCJjVIdhayOjC6wXMzG7vtUmc+gFVZLhSzyGa1j0fyr7pql2w41A9Zb3OCVE+9E8TD4nECPPfzg0Eq/WV/wOeijNYHOzOfajjOJeni2t3IkOfw9nzxec7T4M5ZLPU67T4qggaM/fnENgJO9F0Lg7Tcv4sWrO9gud89C1k0bx+anMJ3t7EATQuDYfOeuzzNre9A1gVO+1Uyn3Z5a9n0SxRPrWUYI8StCiHNCCCmEeIvv7bcJIb4ihHhFCPGkEOKefm+j8eLPfA6jVzDv9Xw6Tz5qzdIhlt370lp2Z+ZzlPgH6Xo5132SLE4HB59Lhaz3u+gdp5cgJfDkuejsp5QSDctuyWC2O74wjUublZa9omfW9nDT0nTLC+/TK8x8EiURN8XxOQDfCeCNtrd/AsAnpZS3A/glAI+mcBuNkZaez6GW3R0q+DzMsntfivnmiTE5Zj5HitrNOTfV27nu+2UQFRhD1zA/nYFlO+XyCxtlvHxtF++5c9UbOPz2m5cAdO/7bFhO807QmiXl+MIUaqbtlflrpoVz6yXc1tYjetPSNIQAz3gniinWbzIp5Z9LKS/63yaEWAVwP4DPuG/6PIATQohbe72tv38K7Qf/q/9hDBy17/m8tu0Gn8x89mWcM5/DbDPZDyrzedBL7spSIetlPv/0Zafk/t13rXq337ZaxMJ0Bl99PTrzWbfCJ90VNXRkuoNJ526UYcvWYSPAqfqcWJjGa2vMfBLF0c/L6BMArkgpTQCQTl3iPICTfdzWQQjxsBDiovqzt8dXlqMkN+Tgs9nzybJ7mma4amlkqYxeP5Puk2SpmPOCzz96cQ0ZXeA7b1vxbtc0p+/zuUvb2I2YPo9as6SoXZ9q1ZU36R5Qqj+9UsDr66WOHaRE1GnkazhSykeklMfVn2Ix/BcFDV9+2Hs+Q8ruqxw46kuRS+ZHlho4OuiT7spSIQtLOn2fT7y2jnecXmr5/gWc0rstga+/sRl6P3Eyn2rXpzphSk263xJQqj+9XETdtHG5bTqeiDr1s2T+AoAjQghDSmkK5+X5SThZzJ0eb6Mxk9WH2/OZbZt2v7ZdxWIh29J7SsmpVUsC43/+96Rplt35Agtoth/s1UzULRvvu3O1432+/fQigOijNhumDV0IzEcMcR13j9g03R7TM2t7ECIk+HTXNSU9490ZfJL405fXcGGjjAsbZZx3/7vY1mrx0COPAxjPgyyI/HoOPqWUa0KIpwB8BM7A0AcBXJRSngGAXm+j8dJywtEQyu66JpDRRUvZnSX3/qlVSxOwh3ziqBcDLLs71KL57YpTUn/fXYc63ufOw7OYzRuhQ0fluhO4dnvRulzMIWtoXubzzNoeTixMB/6uUwFp0jPeL29XUa5b+Nh/fNJ7m64JWLbETsVMdF9E4yJW8CmE+ASAHwBwGMAfCiF2pZS3Avg4gEeFED8PJ6P5Md+H9XobjZFhL5l3PqcOW0pIKbG2U+uYPKXkvMwno8+Rw7J7q2U3G1gzbdx+qIgTi9Md76O7fZ9/9vJ13LQ03XG60//5h6/AlkAxF/07S9MEjs9P4aK7bunsegnv9vWX+t2yknzXZ6VuoVy3kDU0/K8fuAcnF6dxYmEaR+bzuOcX/hDVhhX7vojGSazgU0r58ZC3vwzgnWneRuNl2NPugBPwlusmbOn0bXHNUv/UC4dhhp4sHcbDsnurxULz6/DeOzuznsq337yEP3pxDZWGhYJvifzXz23gP37ldeQNDXNTmdCPV44tTOH1GyU0LKc8HtYjujKTQzFnJNr1+ezFLQBO5eHDb2+duc1nNOxUTVxjdYcmECcLqC9ZfT8ynxpsCZjuwAB/MfdPCAFNtC40p9Ggqgt3HZ7p8p4Hg3/l1Pvu6uz3VFTfZ6XezB5WGxZ+5nPPIqNrODSbj5XpP74wBQmg4mYhw4JPIQROrxQS7fp86vwWgOANE+ptT0UMTRGNKwaf1BdjH4LPrKFBAjDdlSbc8ZkOXRMd5Unaf9NZA7esFHDbIQafQLPsrgngrScXQt/v7iOzKOYML2gEgH/5R6/g7I0Sfvq7b2+p2kRR65ZKdaf/MupxuGWliGs7tdjrlp467wSWQUfaesHneQafNHn6mXYnAuAMqUg5zLK7DimlF3weYtk9FYdm80Mtu1N8B/VFQVBrxqr7fVrMGdAjUvWGruH+Uwv4s5evw5YSz1zYwqf+/CzedHwOf+ddN+O3n7oY+rF+x9yJ97KbQVW9nUHUGe8Ny4auRf8+lFLi6fNbyBla4OOb1Z1qhMqOEk0SZj6pb+rX5tCCz4wG6Su7M/OZjqmMzgXzNPJm8xmcWJjCcowe2Hecdo7arNQt/MznvgldE/g/PvTmlopNN2rXp5TAkbk8ZvLhfaLqjHe1QzTKxc0KbuzVQn/mhBDIZ3R869K2txA/qYceedxbz0Q0Spj5HDHjOIShCQFbyqH2fLLsTnRw5WL+rvn2m52+z2u7NVzeruLhh27HHQl7Z1XZHYheSA80d302YgSLUSV3JZ/RsVGq4/nL27gvosWAaNww80l9G3rm01d2zxoa5qe7T6wS0cFz77E5COGchnTXkVn83e+6JfF9rM40M6ztZ7q3u3m5ACGAutW951MNEuUjdo3m3b5Ult5p0jDzSX1zJkY7M5+DyuLmDLfsbkscjjmxSkQHT0bXMJXRUa5b+D8+9CZkEpTbFc092CJqzZKSz+g4Nj+Ftd1a1/t9+sIWlgpZZPTw31/5jA4hnCzpT+DmxNdONKoYfFLfVOw3rH7BXMYtu1s2DvFMd0rBOLa7UDyHZnIwbYl7j831fB+GpqFhWbjtUPcDLU6vFHHJXUofptqw8MLlHXzXHat4Yz18L6iuCdy2WsTTXLdEE4Zld+qbet3eS1ahF+pIPFtyxycRRTN0re8XxllDgwBwe5eyO+BMvPt70oM8e3Ebpi3x1pvmu97fW08u4PJ2FVe3q/EvmGjEMfikvmmaiFx5kjb/YnsOG+2Pxx5+kNlCOjCWClmcWJzCXIz+8lvc0nzUhLoaNrrvRPchIrXLlPs+aZKw7B6BT67xrBRzsCJKTGnL+aZDebRmcvy+JkpG10TXvZ2K2gNaiwg+nz6/CV0TePOJ7q0AKjvKk45okjD4pL7FPSkkLTnf52PZnYhGydtuWoCuCWxXGjAtu2OnqJQST53fwp2HZzCd7f4UfHq5iNm8wcwnTRSW3Wns5HyrSZj5JKJRkjN0zE1lYNoSf/Dc1Y7bL25WcH23Fnk0qJ+mCdx3cgHPXdqBPcQKE9EgMfiksePPfO53zyd7H4mo3dxUBgLAr/7F2Y6pd6/f8+R87Pt768kF1C07spRPNE4YfNLY8fd8rnLVEhGNGEMTmMkb+ObFbXy9rVfzaXdhfNzMJ9Ds+6w2rLQukWhfMfiksaPK7rpoLcETEY2KheksACf76ff0+U0sFrK4aWk66MMCveXEPIQAqg1mPmkyMPiksaPK7vqQ9ooSESWVNTS8985V/NcXrnmL5G0p8fzlHbz15Hyik9lm8hncvjrDzCdNDD5709hR0/XGEHeLEhEl9bffdTOkBP7Dl18HANQaNkxb4r4EJXflrTfNw7QlGhaznzT+uGqJxo4qtQ8i+OTwEBGl5Z2nl3D3kVn81tcv4shcHlXTyVwmGTZS7ju5gN/82gWW3mkiMPNJYyfHzCcRjQEhBP72u25GpWFhu9JApWFBE8Cbj88nvi81oMTSO00CBp80do4tTAEAcn2e10xENGjvf9NRHJrNYavSQLVh487DsyjkkhcdTy8XoIn+g8+HHnkcDz3yeF/3QdQvlt1p7Nx1ZBanlqaZ+RxBbFsgapU1NPyt7ziFX/7SywB6K7kDzrL5fEZHpW6hZlrc9EFjjZlPGksZXUs0LUpEtF/+27efhPptlWS/Z7t8RocE8OKV3VSui2i/MPgkIiIaoPnpLOamMgCAt9+82PP9qH73M2t7qVwX0X5h2Z2IiGjAlotOAHpiMf5y+XZZd7fxa9cZfNJ4Y/B5wLFHj4ho8IQQyBr9tQpldOfjzzL4pDHH4JOIKAV8IUeDJoRARhd47Xppvy+FqC/s+SQiIhoTWV3DG+slmDzpiMYYM59jgBkVIiICnNVNpbqFC5sV3Lxc2O/LIeoJM59ERERjIqOGjjjxTmOMmU8iogRYiaA0Jf1+8k+8fzcODeKSiAaOwScF4hMsEdHoybq7Ps9y6IjGGMvuREREY0LXBBamMy27PnleO40bZj6JiIhc41D1uWWlyEXzNNYYfKZkHH5h0Wji9w4RJXHLShFff2MTG6U6FgvZ/b4cosQYfBIR0dDwxVb/Tq84K5bOXt/DYqH3s+KJ9gt7PomIiMbILStFADzjncYXg08iIqIxcsuqCj4HP/GeZJiJg08UF8vuREREY+TEwhQyusDZiMxnw7JxeauCYp5P8zR6+F1JREQ0Rgxdw01LhcjM59Pnt1CqW6iZNkzLhqGz0Emjg9+NREREY+aWlQLOb5RRN+3A2x9/ZQ0AYNoSf/zS2jAvjagrBp9ERERj5vRKEZYtcX4jOPv5+CvXoQnn/3/9r94Y4pURdceyOxER0YjptpJKTbyfWesMPm/s1fDcpR0Uc85T/JfP3MBr1/e8jyHabww+aWJwf+Do4mNDlK5b3F2fQeuW/uLV6wCAQlaHoWvYq5n4zBNv4J/+jXuGeo1EYVh2JyIiGjOn3Szm2YCho8dfdoLP6ayOqYyG2w8V8blvXES5bg71GonCMPgkIiIaM3NTGSwXcx2ZT9uW+ItXb+CuI7MwdA1CCPzNd57CbtXE7zx9eZ+ulqgVg08iIqIxdMtKAa9d34OU0nvb85d3sF6q48HbV7y3/dB9x1DMGfj0X51reV+i/cLgk4hoRDz28IPsj6XYTq8UsVs1YfkCyj93+z3fffuy97ZizsAH33oML13dRbURvJopDaZto1K3Bnb/NDlSCT6FEOeEEC8LIZ5x//yY+/bbhBBfEUK8IoR4Ughxj+9jQm8jIqJoDFRJDR3VzWbw+fjL1zGd1XH/TYst7/uRd9wEANiqNAZ2PRulOi5uVXBxszywz0GTIc1p9x+TUj7T9rZPAPiklPJRIcSHADwK4IEYtxFNnDQCBQYbRAdX+8+/OuO9YdkAdOxUG/jG+U28545VZI3W3NJth2bwztNL+Kuz6zDtwWQ/VRD8xNkNfOht0wP5HDQZBlZ2F0KsArgfwGfcN30ewAkhxK1Rtw3qeoiIiCbJLctO8KlOOfrKmXVYtsSDvpK730ff6WQ/dyqDmXp3gmDgibPrA7l/mhxpBp+fFkJ8Swjx74UQKwBOALgipTQBQDpdzucBnOxyWwshxMNCiIvqz95e504zIiKig+bYwhSyhoa6G/Q9/orT7/ng7auB7//Q3YegawLblQZMqzX7+dAjj+OhRx7v+Voalg3TVplPBp8ULa3g891SyjcBeCuAGwB+LaX7hZTyESnlcfWnWOQJDURERLomcHq5gLppQ0qJP3/lOk4tTePkUnDJ29A1zE1lYNoSf5Lyee+Xtyre/1/crODCBvs+KVwqwaeU8rz73waAfwXgXQAuADgihDAAQAgh4GQ2z3e5jYiIiGI4vVKAaUvUTBuXtiotK5aCzOWdUY/f+vrFVK/jwoYTfOYzTljx1dc3Ur1/mix9B59CiIIQYt73pg8DeFpKuQbgKQAfcd/+QQAXpZRnom7r93qIiIiGYRQ2Dqjz2rfdKfYH74gOPg1dw3RWx5++vIa13Wpq13HezXTO5jMAWHqnaGlkPg8B+FMhxLNCiG8BeBDAR93bPg7g40KIVwD8HICP+T4u6jYiIiLqQgWfu1UTWV3DO04vdf2Y2XwGli3xO09fSu06LrjrlaYyOk4vFxh8UqS+Vy1JKc8CuC/ktpcBvDPpbURERNTdaXfXpwTwwM0LmM52f1ov5HQsTGfwW1+/iL/zrtNwOt/6ozKfhi7w9puX8JtfO48LG2WcWOTKJerEE46IiIjG1OmV5hBut35PRRMC/81bjuHM2h6evrCVynVc3ChD1wQ0IfCO086Ce/Z9UhgGn0RERGOqmDNgaE7m8t0xg08A+NH7TwAAPvv1C6lcx/mNMjK6cx2q9M/SO4VJ84QjIiKifbHfgz/7KZ/RUTdt3HFoJvbH3H10Fvcem8V/+eYV/ML7+zvdeq9mYrPcwA/fdwyP/NhbAKCl71PtDz3IjxG1YuaTiIhojB2ezeHE4lTi3s0fedsJ7NVM/MFzV/r6/Gqn53Fff+e3n17ivk8KxeCTiIhiGYXVQtRJCKfXMqn/5i1HkdU1/FZE6b1u2rDck4vCqGGjk77gk32fFIXBJxER0QE0P53F99xzCE+c3fDOZff7ra9fwBsbZazt1iLvR2U3TyxMeW9j3ydFYfBJRER0QKnBox13ST0ASCnxLx97BT/zuWcBANWGFXkfKvj0H+t5aDbPfZ8UisEn0QhieZOIhuGv3bqMI3N57FRNSClRN238z597Fv/3H7+Ke4/NYjqrw7Ql1vfCs58XNivI6hoOzeRb3q76PoOyqnSwMfgkIiI6oHRN4ENvOw7TltirWfjvH30Sn/vGRbznjhX85598JwpZHQDw/OWd0Ps4v1HG8YUpaFpr36nq+6zUozOndPAw+CQiIjrAPvS24wCAqztVfPnMDXz47SfxqY/ej0LOQM6IDj6llLiwUW6ZdFdU32e5S9meDh4Gn0RERAfYTUsFTGWcIPNnvvcO/PMfuheG7oQHOcP573OXtwM/9vpuDTXTxsnFqY7bVN8nM5/UjsEnERHRAXd4Lo8TC1P4e991a8u+UE0TyOgCL4RkPi9sqkn34DPcv/30Ekxbsu+TWjD4JCIiOuAMTSDvZj/b5Qwdr98oYbfa6LgtaMenH/s+KQiDTyIiIgqVd0vvQdnPCxsVAMCJ0OCTfZ/UiWe7ExENCNdlTYaD/jjmMk7w+fzlHXy7G0wq3oL5kODz0GweWV2gXLNg2RK6lvwkJpo8zHwSERFRKDXxHjR0dH6jjNm8gbmpTOjHF3IGLCnx9XM8apMczHz24KC/CiYiooND1wSOzU8Flt0vblZaTjYKUswZ2Cw38AfPXe3InNLBxMwnERERRbrn6CxeXdtrOWqzbtq4vF0JnXRXcoYGQxP4w+evwrbloC+VxgCDTyIiogmU5jG99x6bg2VLvHR113vb5a0KpAyfdFeEECjmDVzZruKbF7dSuR4abww+iYiIKNI9R2cBAM/7+j7VmqWg043azeScLr8vPXfVe9tDjzyOhx55PM3LpDHBnk8iIhor7LsfvnuPzQEAnrvU7PtUC+a7ZT4Bp/R+ZC6PP3juKn7u++5sWWRPBw+DTyIi2lcMJkff6kwOy8UsXgjIfJ5Y6Dxas50QAn/9nsN49Cvn8PzlHS+YbacyofyemGwsuxMREQ1Ymv2X+0EIgXuOzuHFq7veUZkXNyoQAjgWI/gEgO+79zCA1tI7HUwMPomIiKire47Oom7aOLO2B8DJfB6ezXt7QLu5/9QilotZ/MFzVwZ5mTQGWHYnIiJK0ThnOKOoUvnzl3fwP/7m03jt+h7uv2kx9sfrmsD33HMYv/HV83j12m73D6CJxcwnERERdXXvUTV0tA3LlrBl+LGaYb7/3iMAgD9IWHrnZPxkYeaTiABMbraGiNJxYnEKM3kDL1zegen2fZ5Y7N7v6f/d8u2nFzE/nUkcfNJkYfBJRER0wMV58ekMHc3iWxe3MZN3znKPs2bJL6NreOiuQ/jsNy7ipsVpZA0WYA8iPupEREQHRL9T9/ccnUOpbqFcNwEkL7sDwPd9mzP1vlcze74OGm8MPomIiCiWe485Jx2Vas4Z70kznwDw125dxkzOYPB5gLHsTkRENMaG2a99jzt0ZEkJAWClmEt8HzlDx/vuWsXvPHMZpZqJP3rhGtZLNdzYq+P6bg1CANuVBuamMilfPY0KBp9EREQUqD2wPb1cQD6jodqwYegCmtbbMZnfe+8R/M4zl3F5u4q//emvd9z+Q//mL/Gpv3U/blkp9nT/NNoYfBLRxOHkPtFgGLqGu47M4unzW8jovXfuffddq1guZiEl8HPfdyeWizksF3P4H37zKZRqJt7YKOMH/81f4lc+fB/ec8dqiv8CGgXs+SQiIqLY7jnq9H32E3wauoaF6SwWC1n8yP0n8J47V/Ftx+eQ0TXMT2fxax97OzQh8BOPPolP/flZSCnTunwaAQw+iYiIKDa1bD6j91Zyj+M7b1vGF/7+X8PplSL+2e+/iGu7NdgBASiXz48nBp9EMfS7noSIaFI8dPchFHI6irnBdu6dWi7g//1734H33rmK3aqJi5sVvHhlZ6Cfk4aDPZ80VAzgiGgU8HdR75aKORyd636yURpm8hl86qP344H//TFslBv4G//Pl/H3vusW/P333oqcoQ/lGih9DD6JiIaMgQ9RfLomsFTMoZAzUMgZ+JU/OYM/eO4qfvlDb9rvS6MesexOREREIy+f0fG7/+A78fBDt+Pcegk//G+/gusBvaDsAx19zHyOMWZPiCgu/r6gSZA1NPyP77sN33vvYfzM557FMxe2sFcz8ccvXsP77jq035dHMTHzSURERGPl9kMz+Pzf/Q4sF7OwpcRP/NrX8ZOf/joubVX2+9IoBmY+aeIx40NENHl0TWBhOouZnIG7j87h9751BV8+cwPTWR3zPJpzpDH4JKKRwhcLRAdXLz//hq7h3/x3b8WPvLyGX/jC8zi/UcZO1cSFjTJOLE4P4CqpXyy7ExER0dj7rjtW8V9/+t1YnM6gbtr48KeewMXN8n5fFgVg5pP6xkwVERGNgnxGx1IxB0PXcGmrgh//5BP4Tz/5DhxfYAZ0lDD4JCIioqFLI3ERdh9zUxn8o++5Az/728/iw596Av/pJ9+JY/PDWYxP3TH4JCIiokTGoeL1ow+cAAD87G8/ix//5F8xAB0hDD6JiIhopPUa7KoA9Gc+/yw+/Mkn8Js/+Q4GoCNgXweOhBC3CSG+IoR4RQjxpBDinv28HiIiIposP/rACfzSB78N5zfKeNcv/Qn+wW88hd95+hK2yvX9vrQDa78zn58A8Ekp5aNCiA8BeBTAA/t7SURERDSugrKkP/bASfzLx17BZrmBLz57BV989gp0TeBtNy3gfXeu4vRKEVMZHVNZ3ftvPqNBE2If/gWDMZXVMZsfjf2n+xZ8CiFWAdwP4HvcN30ewL8WQtwqpTyzX9dFREREk+eJn/9uAMDaThV/8tIa/ujFNXz5zHV87fWNfb6y4fiJ77wZ/+T9d+/3ZQDY38znCQBXpJQmAEgppRDiPICTALzgUwjxMICH1d/n5uaGfZ1EREQ0IVZn8/jxt5/Ej7/9JKoNC197fQM39moo1y1UGxYqdQvlhvP/Uu731abngVOL+30Jnv0uu3clpXwEwCPq78ePH5+gbwUiIiLaL/mMjnffvrLfl3Hg7OfA0QUAR4QQBgAIIQScrOf5fbwmIiIiIhqgfQs+pZRrAJ4C8BH3TR8EcJH9nkRERESTa7/L7h8H8KgQ4ucB7AD42D5fDx0g47AkmYiIaNLsa/AppXwZwDv38xqIiIiIaHj2dck8ERERER0s+112J+oJS+ZERETjicEnEdEI4wstIpo0Qo7ZBtXjx4/Lixcv7vdlEBEREVEIIcQlKeXxoNvY80lEREREQ8Pgk4iIiIiGhsEnEREREQ0Ng08iIiIiGhoGn0REREQ0NAw+iYiIiGhoGHwSERER0dAw+CQiIiKioWHwSURERERDw+CTiIiIiIaGwScRERERDQ2DTyIiIiIaGgafRERERDQ0DD6JiIiIaGgYfBIRERHR0DD4JCIiIqKhYfBJREREREPD4JOIiIiIhobBJxERERENDYNPIiIiIhoaBp9ERERENDRCSrnf15CIEKIG4PoQP2URwN4QPx+lj4/h+ONjON74+I0/Pobjb9iP4YqUMhd0w9gFn8MmhLgopTy+39dBveNjOP74GI43Pn7jj4/h+Bulx5BldyIiIiIaGgafRERERDQ0DD67e2S/L4D6xsdw/PExHG98/MYfH8PxNzKPIXs+iYiIiGhomPkkIiIioqFh8ElEREREQ8Pgk4iIiIiGhsEnACHEbUKIrwghXhFCPCmEuCfk/X5CCPGqEOI1IcSnhBCZYV8rBYvzGAoh3iuE+JoQ4gUhxPNCiF8WQvBnYATE/Rl031cIIf5ECLE1xEukLhL8Hv02IcSfCSFedP/88LCvlYLF/D2qCSEecX+PPiuE+FMhxK37cb3USgjxK0KIc0IIKYR4S8T77XsswydexycAfFJKeTuAXwLwaPs7CCFuBvC/AXgXgFsBHALwk0O8RorW9TEEsAngx6WUdwN4G4DvAPDRoV0hRYnz+Ck/DeC1YVwUJRLn9+g0gC8A+MdSyrsA3AvgL4Z5kRQpzs/hBwD8NQBvllK+CcAfA/jnQ7tCivI5AN8J4I2wdxiVWObAB59CiFUA9wP4jPumzwM4EfBK7kMAfldKeVU6KwL+HYAPD+9KKUzcx1BK+bSU8qz7/1UAzwA4NbwrpSAJfgbhZmJ+EMC/GNoFUlcJHsP/FsATUsovA4CU0pJSDvO4ZAqR4DGUAHIA8kIIAWAWwMWhXSiFklL+uZSy22MxErHMgQ8+AZwAcEVKaQKA+2CcB3Cy7f1OovXVxLmA96H9Efcx9AghDsP5IfziUK6QosR6/NzS0KcAfByANeyLpEhxfwbvBlATQnxRCPGMEOLTQoiVIV8rBYv7GP4XAH8G4CqAKwDeB+AXhneZ1KeRiGUYfNKBI4SYhfML9JellF/f7+uh2P4pgN+WUr643xdCPTMAfDecFxD3AbgE4N/u6xVRUvfDaZc4BuAonLL7v9vXK6Kxw+ATuADgiBDCAJxhBjivAs63vd95ADf5/n4q4H1of8R9DCGEmAHwJQBfkFKOzGkPB1zcx+9BAP+DEOIcgC8DmHWb65k5239Jfo/+qZTykptZ+wyAdwz1SilM3MfwowD+REq5JaW0AfwagPcM9UqpHyMRyxz44FNKuQbgKQAfcd/0QQAXpZRn2t718wA+IIQ47P5Q/hSA/zS8K6UwcR9DIUQRTuD5JSnl/z7cq6QwcR8/KeW7pJQ3SSlPwWmq35FSnmLP4P5L8Hv0twA84FYfAOD7AXxzOFdJURI8hmcBvFcIkXX//n4Azw3nKikFIxHL8HhNAEKIO+BM9S0B2AHwMSnlt4QQvwqnMfd33ff7OwB+zv2wPwPwU1LKxvCvmNrFeQyFEP8fAL8I4Hnfh35WSvnPhn291Cruz6Dv/U8BeEZKOT/kS6UQCX6P/k0APwvAhlN2/0kp5YX9uWryi/l7NAfgX8N5AdiA0/v5U2qYk/aPEOITAH4AwGEA6wB2pZS3jmIsw+CTiIiIiIbmwJfdiYiIiGh4GHwSERER0dAw+CQiIiKioWHwSURERERDw+CTiIiIiIaGwScRERERDQ2DTyIiIiIaGgafRERERDQ0/3/nhK+TJbB7PAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                   | 0/100 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Лучший порог: 0.5656565656565657\n",
      "Заработаем: 2994.0\n",
      "PR: 0.7006723316524114\n",
      "Recall: 0.19682973527832778\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 100/100 [09:34<00:00,  5.74s/it]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                   | 0/100 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Лучший порог: 0.23232323232323235\n",
      "Заработаем: 3310.6\n",
      "PR: 0.6573002199413489\n",
      "Recall: 0.22983381409204187\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 100/100 [09:46<00:00,  5.87s/it]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                   | 0/100 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Лучший порог: 0.05050505050505051\n",
      "Заработаем: 3357.8\n",
      "PR: 0.6573002199413489\n",
      "Recall: 0.22983381409204187\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 100/100 [10:06<00:00,  6.06s/it]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x640 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Лучший порог: 0.17171717171717174\n",
      "Заработаем: 3175.8\n",
      "PR: 0.6573002199413489\n",
      "Recall: 0.22983381409204187\n",
      "CPU times: user 1h 26min 43s, sys: 4min 12s, total: 1h 30min 56s\n",
      "Wall time: 49min 12s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "kf = KFold(n_splits=5, random_state=500, shuffle=True)\n",
    "\n",
    "for _ in range(5):\n",
    "    t_money = []\n",
    "    t_std = []\n",
    "    t_score = []\n",
    "    t_precision = []\n",
    "    t_recall = []\n",
    "    for t in tqdm(np.linspace(0, 1, 100)):\n",
    "        scores = []\n",
    "        money = []\n",
    "        precision_ = []\n",
    "        recall_ = []\n",
    "        for train_index, test_index in kf.split(x):\n",
    "\n",
    "            X_train, X_test = x[train_index], x[test_index]\n",
    "            y_train, y_test = y[train_index], y[test_index]\n",
    "\n",
    "            lr = LogisticRegression(max_iter=100000)\n",
    "            lr.fit(X_train, y_train)\n",
    "            y_score = lr.predict(X_test)\n",
    "\n",
    "            probability_one = pd.DataFrame(lr.predict_proba(X_test))[1]\n",
    "            y_score_new = np.where(probability_one < t, 0, y_score)\n",
    "\n",
    "            precision, recall, thresholds = precision_recall_curve(y_test, y_score_new)\n",
    "            precision_.append(metrics.precision_score(y_test, y_score_new))\n",
    "            recall_.append(metrics.recall_score(y_test, y_score_new))\n",
    "            scores.append(auc(recall, precision))\n",
    "            money.append((sum((y_score_new == y_test) & (y_test == 1)) * int(np.random.uniform(0, 20))))\n",
    "\n",
    "        t_money.append(np.mean(money))\n",
    "        t_std.append(np.std(money)) \n",
    "        t_score.append(np.mean(scores))\n",
    "        t_precision.append(np.mean(precision_))\n",
    "        t_recall.append(np.mean(recall_))\n",
    "    \n",
    "    yerr = t_std\n",
    "    plt.figure(figsize=(10, 8), dpi=80)\n",
    "    ax = plt.errorbar(np.linspace(0, 1, 100), t_money, yerr)\n",
    "    plt.show()\n",
    "    \n",
    "    print(f'Лучший порог: {np.linspace(0, 1, 100)[np.argmax(t_money)]}')\n",
    "    print(f'Заработаем: {t_money[np.argmax(t_money)]}')\n",
    "    print(f'PR: {t_precision[np.argmax(t_money)]}')\n",
    "    print(f'Recall: {t_recall[np.argmax(t_money)]}')\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Каковы значения precision и recall на выбранном пороге? Оцените по кросс-валидации. Также вычислите стандартное отклонение."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 781,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.26"
      ]
     },
     "execution_count": 781,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# для всех случаев порог меньше 0.5, иногда значительно меньше, иногда нет - в зависимости от \"удачи\"\n",
    "# возьмем средний порог\n",
    "(0.28 + 0.57 + 0.23 + 0.05 + 0.17) / 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 785,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "recall:0.22983381409204187, precision:0.6573002199413489, t_std:1074.3179071379961\n"
     ]
    }
   ],
   "source": [
    "t_precision = []\n",
    "t_recall = []\n",
    "t_std = []\n",
    "\n",
    "for train_index, test_index in kf.split(x):\n",
    "\n",
    "    X_train, X_test = x[train_index], x[test_index]\n",
    "    y_train, y_test = y[train_index], y[test_index]\n",
    "\n",
    "    lr = LogisticRegression(max_iter=100000)\n",
    "    lr.fit(X_train, y_train)\n",
    "    y_score = lr.predict(X_test)\n",
    "\n",
    "    probability_one = pd.DataFrame(lr.predict_proba(X_test))[1]\n",
    "    y_score_new = np.where(probability_one < 0.26, 0, y_score)\n",
    "\n",
    "    precision, recall, thresholds = precision_recall_curve(y_test, y_score_new)\n",
    "    scores.append(auc(recall, precision))\n",
    "    money.append((sum((y_score_new == y_test) & (y_test == 1)) * int(np.random.uniform(0, 20))))\n",
    "    \n",
    "    t_std.append(np.std(money))\n",
    "\n",
    "   # print(recall)\n",
    "    t_precision.append(metrics.precision_score(y_test, y_score))\n",
    "    t_recall.append(metrics.recall_score(y_test, y_score))\n",
    "\n",
    "print(f'recall:{np.mean(t_recall)}, precision:{np.mean(t_precision)}, t_std:{np.mean(t_std)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Как вы, вероятно, уже поняли, ваша модель склоняется к более высокому recall. Попробуйте оценить качество модели с помощью `PR-AUC` в зоне recall $\\geq$ 0.5. Сделайте это следующим образом - выберите только те пороги, на которых достигается необходимый recall, затем интерпретируйте отсеченный в единичном квадрате прямоугольник как новый единичный квадрат и посчитайте площадь под отсеченной кривой. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 777,
   "metadata": {},
   "outputs": [],
   "source": [
    "# к сожалению, модель не добила до 0.5, поэтому эту часть задания я не смогла доделать(\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Бонус (0.01 балла):__ что вы кушали в день сдачи данного ДЗ на завтрак?\n",
    "\n",
    "__Ответ:__ бутеры с авокадо и семгой"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Социализационный бонус (0.1 балл).__\n",
    "Введены новые ограничения, занятия теперь дистанционные. Но это не значит, что не надо общаться с друзьями! Созвонитесь в зуме (или любым другим способом) хотя бы с двумя студентами курса МО-1 этого года, приложите скриншот созвона с включенными камерами и напишите очень кратко, на какие темы были беседы.\n",
    "\n",
    "__Ответ:__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Мы обсуждали участие в новом хакатоне :)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.display import Image\n",
    "Image('hw_4.jpg')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}