pima_diabetes.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 피마 인디언 당뇨병 데이터베이스를 이용한 기계학습\n",
    "\n",
    "- Pima Indians Diabetes Database를 이용해 당뇨병과 유전/생활양식과의 관계를 알아보자<br>\n",
    "자료 출처: https://www.kaggle.com/uciml/pima-indians-diabetes-database/data\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sb\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_dia = pd.read_csv('./dataset/diabetes.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Pregnancies</th>\n",
       "      <th>Glucose</th>\n",
       "      <th>BloodPressure</th>\n",
       "      <th>SkinThickness</th>\n",
       "      <th>Insulin</th>\n",
       "      <th>BMI</th>\n",
       "      <th>DiabetesPedigreeFunction</th>\n",
       "      <th>Age</th>\n",
       "      <th>Outcome</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6</td>\n",
       "      <td>148</td>\n",
       "      <td>72</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "      <td>33.6</td>\n",
       "      <td>0.627</td>\n",
       "      <td>50</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>85</td>\n",
       "      <td>66</td>\n",
       "      <td>29</td>\n",
       "      <td>0</td>\n",
       "      <td>26.6</td>\n",
       "      <td>0.351</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>183</td>\n",
       "      <td>64</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>23.3</td>\n",
       "      <td>0.672</td>\n",
       "      <td>32</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>66</td>\n",
       "      <td>23</td>\n",
       "      <td>94</td>\n",
       "      <td>28.1</td>\n",
       "      <td>0.167</td>\n",
       "      <td>21</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>137</td>\n",
       "      <td>40</td>\n",
       "      <td>35</td>\n",
       "      <td>168</td>\n",
       "      <td>43.1</td>\n",
       "      <td>2.288</td>\n",
       "      <td>33</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Pregnancies  Glucose  BloodPressure  SkinThickness  Insulin   BMI  \\\n",
       "0            6      148             72             35        0  33.6   \n",
       "1            1       85             66             29        0  26.6   \n",
       "2            8      183             64              0        0  23.3   \n",
       "3            1       89             66             23       94  28.1   \n",
       "4            0      137             40             35      168  43.1   \n",
       "\n",
       "   DiabetesPedigreeFunction  Age  Outcome  \n",
       "0                     0.627   50        1  \n",
       "1                     0.351   31        0  \n",
       "2                     0.672   32        1  \n",
       "3                     0.167   21        0  \n",
       "4                     2.288   33        1  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_dia.head()\n",
    "# 임신 횟수, 식후 2시간 경과 혈당 수치, 혈압 등의 8개 항목과 당뇨병 여부를 나타내고 있다"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.matrix.ClusterGrid at 0x202881f9ba8>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAloAAAK8CAYAAAA3RhF1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzsvXmYZHV1//861XvP9CwwLDPMYDPKpqgoiH6DyoTFgCCERI1rQFFCosYVGdRoS7484g8TxWhMSEAkGsWIRKKgIH4RtyGyiciAwjAMs0DPwqw9vVTV+f1xb83crq51qqpvfU6f1/PM0123PnXvqamue9/3fM7nfURVcRzHcRzHcZpPJu0AHMdxHMdxrOJCy3Ecx3Ecp0W40HIcx3Ecx2kRLrQcx3Ecx3FahAstx3Ecx3GcFuFCy3Ecx3Ecp0W40HIcx3Ecx2kRLrQcx3Ecx3FahAstx3Ecx3GcFuFCy3Ecx3Ecp0V01jN4YvgPLenX03Xg4VLpeRFZAlwPHAzkgatV9aoyY18GrAD+QlW/0+xYHcdxHMdxaqUuoaXZ8VbFUY0s8GFVvU9EBoB7ReR2VX04OUhEOoDPAj9KI0jHcRzHcZwkdQktJsZaFEZlVHUDsCH+fYeIrAQOAR4uGvo+4EbgZdMboeM4juM4zlRCyWjtQUQGgZcAdxdtPwQ4FzgZF1qO4ziO47QBbZHREpELgQsTm65W1atLjJtNlLH6gKpuL3r6C8AlqpoTqVjy5TiO4ziOMy2Iau317WMP3d6SYvieY06rqoxEpAv4PvAjVf3HEs8/ART2swAYAS5U1f9uZqxO+yAiy4HetONwzDOqqlekHYTjOGFS59RhOjVaEqWorgFWlhJZAKp6WGL8dcD3XWSZp1dVh9IOwrGNiAylHYPjOOHSFlOHNXAi8HbgtyLyQLztY8ChAKr6L2kF5jiO4ziOU476hFYu26IwKqOqP2fvtGAt489vXTSO4ziO4zi1UWdGK/1Vh47jOI7jOKEQRI2W4ziO4zhOiHhGy3Ecx3Ecp0W40HIcx3Ecx2kR9QmtNnCGdxzHcRzHCQXPaDmO4ziO47SIOoXWRIvCcBzHcRzHsYdntBzHMN6mqCkMujt8Q3gLI2dG4xktx7GNtylyUsVFqjPTqc9HKyWhJSLXAmcBw6p6TJkxy4AvAF3AJlU9afoidBzHcRzHmUooGa3rgC8B15d6UkTmAf8MnK6qa0TkwGmMzXEcx3EcpyRBCC1VvUtEBisMeQvwXVVdE48fno64HMdxHMdxKlGn0EqnqXQNHAF0icidwABwlaqWzH45juM4juNMF21RoyUiFwIXJjZdrapX17GLTuA44BSgD/iViKxQ1d83MUzHcRzHcZy6qC+jNd6ajFYsquoRVsWsJSqA3wXsEpG7gBcDLrQcx3Ecx0mNtshoNYHvAV8SkU6gG3g58Pl0Q3Icx3EcZ6YTRI2WiHwTWAYsEJG1wKeIbBxQ1X9R1ZUi8kPgQSAP/LuqPpRKsI7jOI7jODH1ZbTGc62Ko/JxVd9cw5grgSunIRzHcRzHcZyaaIsaLcdxHMdxHIvUWaOVTkbLcRzHcRwnROqcOsy3Kg7HcRzHcRxzuNByHMdxHMdpEXUJrbwLLcdxHMdxnJqpM6PVqjAcx3Ecx3HsUWdGq1VhOI7jOI7j2KM+oTUhrYrDcRzHcRzHHPUJrawLLcdxHMdxnFrJ1DM4NyEt+VcNETldRB4VkcdEZHmJ5w8Vkf8nIveLyIMi8tp63pfjOI7jOE4rqCujlZuoS5c1BRHpAL4MnAasBX4tIjer6sOJYZ8Avq2qXxGR5wO3AIPTHqzjOI7jOE6CthdawAnAY6q6CkBEvgWcAySFlgJz4t/nAuunNUJnCnHmsbfFhxkUkaEW7n9UVa9o4f4dx3Ec49QltLLZVITWIcBTicdrgZcXjRkCbhOR9wGzgFOnJzSnAr2qOpR2EI3QYhHnOI7jzADqy2i1SGiJyIXAhYlNV6vq1YWnS7xEix6/GbhOVf9BRP4P8B8icoyqusOq4ziO4zipUWdGq6MlQcSi6uoyT68FliQeL2bq1OAFwOnxvn4lIr3AAmC4yaE6juM4juPUTH1CK5fK1OGvgcNF5DBgHfAm4C1FY9YApwDXicjRRLVBG6c1SsdxHMdxnCLaXmipalZE3gv8COgArlXV34nIZcA9qnoz8GHg30Tkg0TTiueravH0ouM4juM4zrRSl9CayKeS0UJVbyGybEhu+2Ti94eBE6c7LsdxHMdxnErUl9HSdISW4ziO4zhOiNSX0XKh5TiO4ziOUzP1Ca2STguO4ziO4zhOKeoTWuJCy3Ecx3Ecp1Y8o+U4juM4jtMiPKPlOI7jOI7TIlxoOY7jOI7jtIg6hVarwnAcx3Ecx7FHXUJr3IWW4ziO4zhOzdRljDUhrflXDRE5XUQeFZHHRGR5ied7ROSG+Pm7RWSwnvflOI7jOI7TCtp+6lBEOoAvA6cBa4Ffi8jNcdudAhcAz6rq80TkTcBngb+Y/mgdx3Ecx3H20vZCCzgBeExVVwGIyLeAc4Ck0DoHGIp//w7wJRERbyztOI7jOE6a1OmjlYpuOQR4KvF4LfDycmNUNSsi24D9gU3TEqHjOI7jOE4J6hRarUFELgQuTGy6WlWvLjxd4iXFiq+WMY7jOI7jONNKnVOHrdEusai6uszTa4EliceLgfVlxqwVkU5gLrCl2XE6juM4juPUQwhTh78GDheRw4B1wJuAtxSNuRk4D/gV8HrgJ16f5TiO4zhO2rS90Iprrt4L/AjoAK5V1d+JyGXAPap6M3AN8B8i8hhRJutN0x6o4ziO4zhOEXUJrVxKZU+qegtwS9G2TyZ+HwXeMN1xOY7jOI7jVKLtM1qO4ziO4zihUqfQyrcqDsdxHMdxHHN4RstxHMdxHKdF1Ce01DNajuM4juM4teJTh47jOI7jOC2iLqGV9YyW4ziO4zhOzXhGy3Ecx3Ecp0V4jZbjOM4+ICLLgd604wiAQREZSjuINmdUVa9IOwinNdQ5dZhrVRyO4zih0auqQ2kH4YSPC1Hb1JnRaj+hJSL7ATcAg8Bq4I2q+myJcf8fcCaQAW4H3u/9EB3HcRzHaSWZegZPkG/JvwZZDtyhqocDd8SPJyEifwScCLwIOAZ4GXBSowd2HMdxHMephIWpw3OAZfHvXwPuBC4pGqNEtRTdgABdwDPTE57jOI7jODOV+oRWvjVCS0QuBC5MbLpaVa+u8eUHqeoGAFXdICIHFg9Q1V+JyP8DNhAJrS+p6spG43Ycx3Ecx6lEW9RoxaKqrLASkR8DB5d46uO17F9EngccDSyON90uIq9W1bvqjdVxHMdxHKdWgpg6VNVTyz0nIs+IyMI4m7UQGC4x7FxgharujF9zK/AKwIWW4ziO4zgto75i+HyuJf8a5GbgvPj384DvlRizBjhJRDpFpIuoEN6nDh3HcRzHaSl1ZrSyrYqjEa4Avi0iFxAJqjcAiMjxwEWq+i7gO8DJwG+JCuN/qKr/k1K8juM4juPMENqiGL4RVHUzcEqJ7fcA74p/zwF/Nc2hOY7jOI4zwwleaDmO4zhOPbRh+6R2bFPkbYGaRH2rDl1oOY7jOOHj7ZOq0IbCL1g8o+U4juM4jtMiPKPlOI7jOI7TIuoSWjkXWo7jOI7jODVTl9AaH1srrQrEcRzHcRzHGnUZljqO4ziO4zi140LLcRzHcRynRbjQchzHcRzHaREutBzHcRzHcVqECy3HcRzHcZwWUdeqQ8dxHMdx0mca2ghNR1ugGdHmx4WW4ziO44RH8G2EZkqbH586dBzHcRzHaREutBzHcRzHcVqECy3HcRzHcZwW4ULLcRzHcRynRbjQchzHcRzHaREutBzHcRzHcVqECy3HcRzHcZwW4ULLcRzHcRynRbjQchzHcRzHaRHuDO84TtszDe1G9oXpaFFSLzOipYnjhIQLLcdxQiD4diPTQRsKP8eZ8fjUoeM4juM4TotwoeU4juM4jtMiXGg5juM4juO0CBdajuM4juPMSETkWhEZFpGHEtuOFZEVIvKAiNwjIifE298qIg/G/34pIi+u6Riq2qr4HcdxHMdx2hYReTWwE7heVY+Jt90GfF5VbxWR1wIfVdVlIvJHwEpVfVZEzgCGVPXl1Y7hqw4dx3Ecx5mRqOpdIjJYvBmYE/8+F1gfj/1lYswKYHEtx3Ch1SLa0fensDz+6ANPCD6Nee//PTHtEBrmuE/8Iu0QmsJB3XOqD2pzNk3sTDuEhsmIpB1Cw3yjb0HaITSFsYmOtENomJ6uXNohNIUXrf7+XwEXJjZdrapXV3nZB4AficjniEqs/qjEmAuAW2uJoWGh1Y6Cok0YVNXzm7WzZv4/93V0N2M3qSJLDks7hIbp6/h12iE0hdF8Nu0QGqY305V2CA2TJ/j7J1btHkg7hKawpGtX2iE0zMaRvrRDaAqxqKomrIr5a+CDqnqjiLwRuAY4tfCkiPwxkdB6ZS07a7hGS0SG3EhwKi0QoIuI0pevAP6kjtc9AawBUNVlAD856I3Bn5Ef7g5fLAK8+46L0g6hYV617ONph+AY4siu/dMOoWFmS/iTRZt0LO0QmsJ/Pfm9qqneeOrw+4karW3APFVVERFgm6rOiZ97EXATcIaq/r6WGML/a2hT2qENRiz2TqXoc+6SfDoBNZG+4KUivO2H58H47rTDaJgJAxktMTDtZmFh00t6F6YdQlPoNrCgP0P434kGWA+cBNwJnAz8AUBEDgW+C7y9VpEFLrSCpyhzVsh6QZTSnA/sKH7N3N7w71QyE+HPVmtuIu0QmsLsjvA/iwk1UI9i4Lq4MT+adghNQTM9aYfQMBOEf0NeCyLyTWAZsEBE1gKfAt4NXCUincAoe2u8PgnsD/xzfHOWVdXjqx3DhVb47OkBVyS6VsU/1xe/4ODnbp+eyFrI7x83UMsxsi3tCJqChdqg3Ay5qLQ7c8VGScCYhv/3NGuGyANVfXOZp44rMfZdwLvqPcbM+J+cISSnKxPThkcUj8uNhZ/WtnBxp9PGRaXDQCql20BNjQWsCF4L72KmZLSmAz+7BEaJIvvBou2F6cNFQBY4pXgfE6PhC639Nfzl01YwIHnJGchAWGBLPvyyBoADMuGv2OsyULdYCyJyJHBDYtNSoinCdcAQcDRwgqreE48fBFYCj8bjV6hqxVVNLrTCoze5ylNEhpLbRaSwjHUp0ef70+Id7N5pI5MSOtIzO+0QHKetmGXAZgNMlMsxaqFusQZU9VHgWAAR6SASWDcB/cCfAf9a4mWPq+qxtR7DhVb4jMZiazB+vIoos7WKMnYQPb3hrxJbM2qjkNwC3RJ+dtHCe7CAlcziuIGV3SbKM+rnFCIR9WRhQzNWJLvQCpxCXZaIDJXx7loR/9vDyI7wM1r7WfjT7TDwHoBxA3e+WQP1KBaW4y+Q/rRDaApdBuwdZihvAr5Zw7jDROR+YDvwCVX9WaXBNs70DkRLUI8quNEXia5Xkvisc/nwTwIZAzdbn3nttVz6w7oXsLQd4V/eodPAhdFCBmJtdjuLO8Nv6WQBKwsTRORCamjBIyLdwNnApVV2uQE4VFU3i8hxwH+LyAtUtexyfhdabUSNbvKDpTaq6hWJeq1idgPhV2cmyBu4un/81vBFlhU6DBT+Svg6iwM7Z5m4wFtYiZs1INyhrhY8ZwD3qeozVfY3BozFv98rIo8Tre6/p9xrXGi1F5MK3UtRQUzB3notiATZI0TCrY/IvBSAHaPhTx0Svh8gdNvQvhaaGecNuKpbYJbYKIa3UEhuYSq6Tt5MDdOGInIAsEVVcyKyFDicvb6VJXGhZYgSPlpLiQrif54cN25g6vDgnIGTwISNpexi4ISsBu7ebUwdTmlkEST9mfBvZp+fMWAKXSMi0g+cBvxVYtu5wD8BBwA/EJEHVPVPgFcDl4lIFsgBF6nqlkr7d6EVIDVMMZZrPD0EcH9P+OmgdRkDd4wLlqQdQlOYbaDdiIXpKgsMGHGGD/9WFrYQ/ur0WlHVEaLWOsltNxHZPBSPvRG4sZ79u9AKk3JtdwoUVhmWElssmQj/znesO/zl+Lrbxt37SH487RAaxkSvQwPM6wxftAP0GrAL2aluodMsXGiFxyhwVOJxsYFpQXiVFVv75cP/Am2z0DLFSFNpC3SIhRxE+PQY+Rws1GhZeA+1UMEZ/vp4+yCwGnijqj4rIhcDb43HdhI5xx9QafrQwNVqZlFldSFMzXatIFpNsbswIGsisW0AIxcVx2kWFpoxg41VhzPFxLeCM/xy4I74mrs8fnyJql4JXBmPfx3wQa/Rsknx6kJgj7AaTGS1BlX1fBFZlHxxr4R/p7LLQOGvFSxkg2bGJaX9sbAoAUANCK0ZWre4xxleRM4BlsXbvwbcCVxSNL6mlYoutAKkaHXhUOKpXqIUZ6Hv4fLE80sLg3oMFJL3GfjT/buTruSyO96fdhgOdlq/hM5wbsTEir25Bor6O2bmzEfSGf4gVd0AoKobROTA5MB4peLpwHur7TT8q5VTzChwVFGRfKHBNABj+fDv3+dr+O/hgz+p+v0MAgsixYI1goVsUH+m20QmJfx8lo2Vk9ASZ/gCrwN+UW3aEFxoWWDSNGI8VZj00ALIEqVEAdhC+KaA6yX8QnLps+FTY6GWY8LAxd0CVgxLcwZEb5+FBUc05Az/jIgsjLNZC4HhovG19kV0oRU6xdOItTSWHpghq0naHiOGpRYuKhaychboNJELgnHCP8fONXBDXifF9VY3A+cBV8Q/v1d4QkTmAicBb6tlxy607LGnPou9gqtgYDoEsCUT/sc+YODiTpcNzyALzvBOe2ClLshCk/IdOnMMS0s5wxMJrG+LyAXAGuANiefOBW5T1V217D/8K+40U2Pj531lsMHX7/HYSixJXUbU8PKnhUHjBnrTZQ1M9YiB1XpWsLBy0gIWmnsD5Az0zrRQK1crZZzhN5MouSl67jrgulr370Krfqo2ft5XqvhjVaXgsZUQg68D/ofoc97zWWcMZINmafgXRvXpqrZBDVwYxYBI2WXEjdxCZq7fQO1lu+BCyx6FrNYj8eNlxQM6w7+mMCzhp7U9o9U+WMhoWZjCnWPAFgFgwsBN1M4ZMHUoIkuIHOAPBvJEKxKvEpEXA/8CzCayTHqrqm4XkS7g34GXEumn61X1M9WO40LLGIkpw6NU9bhEdmuPaemYgTtfCwXYdNj4+nUZECkWsGBRMW5AoICNz8LKwoQqZIEPq+p9IjIA3CsitxOJqY+o6k9F5J3AxcDfEdVp9ajqC+O6rodF5JuqurrSQWyc6Z1JFMRWPBVZEFh7DEuP7d2aSlzNpGN0btohNMx3jh3iT//jpLTDaJg141VtZNqereM70w7BifnzOS9IO4SGWau7qw9qc16hNuxnKhEbkhZMSXeIyErgEOBI4K542O3Aj4iElgKzRKQT6APGge3VjuNCK0DqLMhfD7ySZI2WhH+31ZcP/z2c/l9npB1CUxjo6Es7hIbZ3TGedggOcFTfQh7Lhy96D8i0ar3U9LHBQKu2ehCRQeAlwN3AQ0Tmpd8jymItiYd9BziHSJz1U0OfQ3ChFQQlhFXSmLTUN7pg51CSrSPhnwQ2dxqYrpqzf/UxATBuoJaj00Dhb8ZAScDBmfBFO9hw6R8x4AUGtTnDi8hs4EbgA3Et1juBL4rIJ4n8tAp3YicAOaKZovnAz0Tkx6q6qlIMLrTCYNJKx8TqxCkrIGPxtQb4dLmdzekL3yhze9bACTlrI4tiYcWeBZGSN/A5WLBtARsLE3oNrJyE6s7wcYH7jcA3VPW78WseAV4TP38EcGY8/C3AD1V1AhgWkV8AxwMutAxSaLszCFMyXoNEKw57mZrZGgIbU4f7W7jZGq3J667tsZDRGs+H/x4sMGqka8UsA+1r5hnoJ1sNiTxRrgFWquo/JrYfqKrDEi0N/wTRCkSIkhgni8jXiaYOXwF8odpxwv9rmIEU2u4kCt4HVfX8wjYiS4eyU4fP7go/G7TOgKm67L847RCawryO/rRDaJhuAxdGC1hwVAcYMSAYNxmw0KmBE4G3A78VkQfibR8DDheR98SPvwt8Nf79y/HvDxH1Dv+qqj5Y7SB+dgmYEoKrYm1WgRecEv4qsTt/vjDtEBomv/JXaYfQFHbmw5+KHsmF/x4sYGXqcMzA+1io9nsdqurPoew871Ulxu9kciuemnChFQavrNE1Ptk8etJKwyQ7H2tSVCkyNx9+DQTzD0w7gqYwlg/fzduCYamFGq2lEn52FGAT4X8nfq1VXQucGnGhFQad5dr+VFh5uIrI2mEK2bHw597nh5+Zh5yN1HzOiMlk6FgQWiMGMkEA4wbex/5ioD6jChWc4fcDbiCqeV4NvFFVn93X47jQai9Gy2Su5lV4TS9R252lTPXMOqLkQUbCTwn3GLio/OBtP+XMG05LO4yGsbDCysJyfAtZuR+OrOL0/qXVB7Y5AwYurRZMV2ugnDP8+cAdiU4ry4FL9vUg4f81GKJQc1WMiCwrely8ynA1UQarl73LTKcYlRbo7gk/k5LbkXYEjXP2N0+BfPh3vjasEdKOwAF4Zf9z2GnAv2k24c8ahC/bq1PBGf4c9vYJ/hpwJy60Zhx7/LMKGbCkSIuF2FJgAJhyWzIxEf5JwEKJlvTOSjuEptCTCT9DamH600JWzsJ7ANiGgZtZA7MG9VDkDH9QLMJQ1Q0i0lBBrQstgyTSnQUm1WqJAR+tzR0GlJaRptIWpqwsmK6GLxXtsEPDL4bfP2OjRmsfneGbGoONM/3MZhQ4qvCgaFpxPXAGRbVas+eHv5S9M/x2aOjOfa6tbCssiJTdeRsu/aFjod4PIGPgfXQbmTzcF2d44BkRWRhnsxYCw43E4EIrcOLs1Y8TRfSD7HWGB9hB0eecz4V/Etgd/lswM3VooU9gX6Y77RAaptl34WmwWcO/CbSCBbFYjXLO8ET9Dc8Droh/fq+R47jQCoPVVZ7vTNRsFeqz1hMZmJ5SPNiCvUNf+EkUM1jw0cpbqA0ykFk8wIilQIcBkfJU3kaLsCqUc4a/Avi2iFxA1HanbpPSJC60wmB1Lc8npg0LNVnzgZ+2LKoUCf/SjtdotRHh33rYYNRIpZkFoTUTMlpVnOGnJCn2FRtnemd1/LM3aWwqIosoYVraOyd8mTJ7Y9oRNI4aaSqdMSC0rKx2C53h3EjaITSF/TPh95Pt8/6fTcP/J21QMDp9XVGdRkFkFXogDgEYuC6auNe64c0/5i++viztMBoma6CBroWCfgs8nd/O83oWpB1Gw3QZqJfboeFbVNSKiHQA9wDrVPUsEXkv8AHgucABqropHncUUVPplwIfV9XP1bJ/F1oGSDSXHiRahVjckqeTxBRidix8pdVl4Lr4Z99YlnYITcGCSLFQo2UhK3d0z0Fph9AULEwdWlkBWiPvB1YCc+LHvwC+T2RUmmQL8LfAn9azcxdatngEOEpVz4/rtc4lMiwdIFp9CEBHV/h1EKPha0UzWBApFupRwv9WQ7+BFawAOQPfCQvvoRZEZDFwJnA58CEAVb0/fm7SWFUdBoZF5Mx6juFCyxCx1cNQ/LAX+A3R9OGkWi01YKs+J2fgJDBhw7vJgqt6+JOfNjJaWQPZUYBeA4LRQlauRr4AfJQoIdESXGjZY0+9FvA/8bYXkzAtnTCQDjJxCujtTzuCptBl4KJiwYPKwhRu1kReDnaH/1GYyWhVcoYXkbOAYVW9t7incDNxoWWMRL3WIvZmsiaZlmY6wv8CTYR/XYScjWLTnIGLo2j4f1AWMlqzjKx0C/8bATsMNLyHqs7wJwJni8hriWaB5ojI11X1bc2MwcZftX1GE1OCELm/V0RVLyxqx7OHuS8PP5OyaX34F0bWPZF2BE0hbyCTggGRkjGQlbNSgB1+jteGWKyGql4KXAoQZ7Q+0myRBS60gqCQpSpQJLqS25Ou8K9k7+c7qRhed4Xf5mKBhRqtrvDbvoCNqUMLWFiUsN1AM2awkeXtsuADtI+IyN8S1W0dDDwoIreo6rtE5GAiG4g5QF5EPgA8X1W3V9qfCy1b9AKr4p/z2VujNZlM+HeNOQt3vkZS8xbqm8bz4U/jWpg6HDHi3XRQZspEQnCMGPDHqwdVvZPYzkFVvwh8scSYp4HF9e7bhZYxEjVax8WbCmalUDAs7Qs/k2KhRutbH3mcN332OWmH0TAWWvBYwMLn8OT4ZhNeWtsMZOY6LdzMtgkutAKnqA5rMPHULyhRnwWQ3bCzxVG1ng7mpR1Cw7zhX1+adghNwUI2qDsT/qnQwqrDxd3zTWS1FmTCb449UzJaInItUFh9eEy87cXAvwCziVrcvVVVt8em4CuBR+OXr1DVi6odI7izS7kC72lkMMVjl2JPf8Nk7VbsqbU8raBazZiFmy0jU4cWarR25cKvW7RQDG/Bfwqgk/Czi10GpqJr5DrgS8D1iW3/TlQY/1MReSdwMfB38XOPq+qx9RwgOKFFUePk6aZcIXq7UCREFwFrgE8nx+RHwv8CzbNQDG/A6BNsiBQL9U05Axmt2UbsHXoMCK3thD/9WQuqelecqUpyJHBX/PvtwI/YK7TqxsZftVNgFDiKqBVPQWwtJWFWCtAxL/y7xt0GCvqt+Gj1ZrrSDqFhxg1MV1nAir3DDsL/exqfIVOHZXgIOBv4HvAGYEniucNE5H5gO/AJVf1ZtZ250AqUROZqsLAtMV1Y6Hd4K9FnPMneQQyIFBu5IBtYWMrutAdW3Mi7DAjGTgOLK6CyM3wF3gl8UUQ+CdwMFPqlbQAOVdXN8YKz/xaRF7i9g116VXWoeCqzILbi7c8QZbdOJfFZ50fDv1OxcQqwgYVieCuZlNCxYX4LGKiXG5DwM9VQ1Rm+3GseAV4DICJHEDWdRlXHgLH493tF5HGiGaN7Ku3PhZZtFgMvJJHNApjYEv7J7OBs+Bd3Omx8/ToNFDBnDUyTWBApnQYECsCogfpLCwa4+4qIHKiqwyKSAT5BtAIRETkA2KKqORFZChxO5F1ZERtnegeYNJ1Y6HG4Kv45iUz4NlqIhZOAkVWHFvybLKzYs8DG/GjaITSFeRK+vcNMQUS+CSz8Z1mOAAAgAElEQVQDFojIWuBTwGwReU885LvAV+PfXw1cJiJZIAdcpKpbqh3DhVb4JPsgDhJNFRbE1VLgOIoyWrnd4V9U1MBUz88ueoBX/dvxaYfRMBkDn8Wu3Hj1QW2OBbH42Ogwz+lZkHYYDWPgo2CRhO9uXwuq+uYyT11VYuyNwI31HsOFVuAk+yAWFcIvB37O5J6HAHQOhJ8NmjBwcT/5Ky+EbPgXeAsteCxk5SxwSM98E4srLIjeEcKfTm8XXGiFScHGYRJFhfCF6cOfM7kNDx394V9U7u0N/z2cPGtO2iE0hW4D3keaCf/mwwJ9RgqwRw3U/A0Y+F5XQ0R6ifyyeoj00HdU9VMich1wErAtHnq+qj4gIvOBa4HnEl2H36mqD1U7jv3/SYPEgmqo3HMwpV5rRfwPgLHh8O8Y5+fDv2Okpz/tCJqChaJZC+/BAl1GMotdBtZFj86MjNYYcLKq7hSRLuDnsS0SwMWq+p2i8R8DHlDVc0XkKODLwCnVDuJCywglWhMdB9xLCcPS3kPD/9ifeSJ8scju8HtOgg1X9R4DpqsWsCBQALoMTB1u0fDLGqqhUYPQwom4K/5X6YT2fOAz8WsfEZFBETlIVZ+pdJzwr7gzmOKG0qp6fuK5W4lWUkx9XXf4y/F7NPwTGRM2TmQWmhlbsHewUCtnpdfhhIHvRIeBOthaEJEOoqTE84Avq+rdIvLXwOWxYekdwPLYQ+s3wJ8RZb5OAJ5DZKPkQsswkxpKFwmvp4hqtCbVZwFkh8PvTTfHgtAyUqNlwUfLwnuwwNZ8+OcmgG4Df08WMtVQ3RleVXPAsSIyD7hJRI4BLgWeBrqJzE4vAS4DrgCuEpEHgN8C90P1fksutOwwSjRNuJ6oLis5ZfjT5MCJis0CwsCAQwXS3Zd2CE3B65ucZrFfxob/VNZARutZtSF6a3WGV9WtInIncLqqfi7ePCYiXwU+Eo/ZDrwDQKIU8hPxv4q40AqXSSsPS9g8FJhiWNo5u7WBObWh47vTDsGJyRtw8raAhSk3sHHzYWUatxKx0/tELLL6iNrVfVZEFqrqhlhM/SlRk2nirNeIqo4D7wLuqtbnEFxoBUsdKw+XsXfqcAhADXSv2SbhXxgvu+BOPnnNsrTDaJix/ETaITSMhWkSC/0afzv2NKf1PiftMBpm1IAX2HqdETeCC4GvxXVaGeDbqvp9EflJLMIEeAC4KB5/NHC9iOSAh4ELajmIC62wSbrCF6YLS32mk6YOuxaE/7H3GajRuviLL4adW9MOo2EsiJScZ7Tagpf3LWEL4Qv3BYS/itXKCtBKqOqDwEtKbD+5zPhfEfU3rIvwr7gzmDLThYUarcKU4RRneDrDFyndBu7eraw6nMiHv2LPgli00FTagkAB2GHAg8q+zJo+XGgZoYToKqw+nNJYuvfSoekLrEXoT/4l7RAa5p8+/EjaITSFrkz4tRwWRIqF1jVZA4IX4IW57rRDaJinOlweNAv/n2yAEiah+8oUC4YWMQSQ/a+qCzDanrcesCHtEBrmD0+F3zwX4PJdw2mH4AC5fPhC67jOI6oPCoAeA8L9l5nRtENoOSKyBLgeOBjIE1k/XCUiVwKvA8aBx4F3xAXzg8BK4NF4FytU9aIpOy7ChVZj7PGxaoRYsK2oOnAyBXH2BLCmruPtN7fOQ7UfuYnwhZYVejvCn+6xUKOVMTCd3mdALAL8vif8iTcLGdIayAIfVtX7RGQAuFdEbgduBy5V1ayIfJbIV+uS+DWPq+qx9RzEhVZ7sM+CLZFVm1qLVY5s+PUDB5wWfp/A//qv8AUKQG82/GkSCyJFDPQJ/FFv+IXwAEfkw/cDm5WxLw9UdQOwIf59h4isBA5R1dsSw1YAr2/kOPb/Jw1TrRaLclOSBu4adTx8sXjYePjTCwAdBvx2cgZa8Gg+fN+WHRa8Z4BNBnpnzjIiD6o5wyfGDRKtQLy76Kl3AjckHh8mIvcD24FPqOrPqsVg438yIMrUdQ1WeC5JcjVh8nHx9gIrKDElmXv0ydoDblM6jjg07RAaZmvHrrRDaApjo+GvnrTQJ9BCz8nF0oyS1/Q5OB9+dvHXYqNGqxZneBGZDdwIfCBpQCoiHyeaXvxGvGkDcKiqbhaR44D/FpEXVDMtdaE1/UyZJkx4YVWcQkwIsYKwKoirZLudquSeDf8LJBs2ph1Cw7z5rXDb18K/sIzlwp/usbDq0ILQ+rdNv+amWVNsjYLjrt7whdaYAYuKWhCRLiKR9Q1V/W5i+3nAWcApGn+54sbSY/Hv94rI40TX3nsqHcOFVntQMB4drHElY13CqhjpDv8kQD78i8oT38lyeJ8BkbIr/M+iIxP+d8LCqsN37388t6YdRBOYE/5XgoOMZBcrEbfYuQZYqar/mNh+OlHx+0mqOpLYfgCwRVVzIrKUyLx0VbXjuNBqAxItc4aoszC+SJjVZBORHzFwpzIcfmfsRcfYqEfp+Vn49SgTBuqbMgamP2dp+IIXYK2EfwM1QzgReDvwWxF5IN72MeCLQA9we1xWULBxeDVwmYhkgRxwkapuqXYQF1rtxZ5G0XV4dCXrtErWZBXT/cJDGgixPdj9q7Vph9Aw99y3MO0QmsLc7pHqg9qcx7eWKnEMCwNJFMJfVhExy8A7edFE+O+hGqr6cyi55PiWMuNvJJpmrAsXWm1E3Ch6eWEaUVXPLzWuSIQVRFbN04n54WcbDzZlug4OP4vybIeNE9m2kfCL+i0Uw2OgRqvDgM0G2BApz4b/FtoGF1otoEo2arDSa4umEcvts3i1Yblm0qXJhl/L0bl4v7RDaJidGRsXFQv1Td0GTFctMC/8UxMAT3aGL3rDfwe1IyIdRAXt61T1rLh26/8CbyCaIvyKqn4x3n4V8FpgBDhfVe+rtn8XWq2hbJ1VsYCqwGjR2EFVPT8WXPu02rBAZtH+9b6k7cit25R2CA3z3InwHfrBhtlnb6cLrXbgyQ4bSqtHw/9OrJPwbVvq4P1ErXXmxI/PB5YAR6lqXkQOjLefQVQAfzjwcuAr8c+KuNBqU5JNoiHKaMXCK+mZ9fP4Z129EjNHHdmMEFPlN18J36JifUf4juoAA4Tv0j+RD3+BiBrIQTyDjYv7szqWdggN8/cd4f891YKILAbOBC4HPhRv/mvgLapRby5VLTR0PQe4PrZ7WCEi80RkYewwXxYXWilTY9F7UlzVbExajswLTqx1aNvy1J6enuEyPxf+xR2gvyP8diPdveGfCvMGhNaEgZ6TAEsy4d98XGlgJS7A16oP+QLwUWAgse25wF+IyLnARuBvVfUPwCHAU4lxa+NtLrSaTHJKb7AJ+6tq51AkxmrvaViG/PATjby8LXjZgvANSwEeH56fdggNkzWQDbLQQNeC6er9o+s5ovfA6gPbnE4D0+njBtpSQeUWPCJyFjAcm48uS4zpAUZV9XgR+TPgWuBVlF6hWPWL50KrTpJTenXUWzX1mKWEWYli+bIF8tLd14Iop5ddO8PPoqwbmZV2CE3BQibFgkixwLG9i9IOoSl0Ev4CETFw8wFVW/CcCJwtIq8lun7OEZGvE2WqCjYONwFfjX9fS1S7VWAxpWeZJuFCa/qZUuReaXCJqcVy7vHJYvmlwHxgR6l96kT49QOj4+H/6e7fGf7nADA+YWOKIXQsGJZaaftiwUdrloR/jq2Gql4KXAoQZ7Q+oqpvE5ErgJOJMlknAb+PX3Iz8F4R+RZREfy2avVZ4EJr2ilR5D5U5SWTphZjIXVUscdWUbH8z6mA9A9UejoIRMLPQPyG2WmH0BQsXOAtZLRyRuqbLDBg4NLaY2D6swGuAL4hIh8EdgLvirffQmTt8BiRvcM7atlZ+H8NM4wiU9NXENk7rCkaVtHyQVevbFF008chg9vSDqFhnnjcxtThjonwneGd9sBKf73BXPgZrV9mwjcirgdVvRO4M/59K9FKxOIxCryn3n270Eqf4qnEYgaLNxTVbBVPI1Yvljdw957pCv/ufa6BInKw4apuwQvMQq2cFTYY8APLGPACaxdcaKVM8VRiMUkRVkJUJW0fCqxKbCvtrzURvldN94LwTwKbjbTgwUu0nCYx10BtE8Aw4TeVHpgBNVoAInItUFh9eEy87Q3AEHA0cIKq3hNv7wb+FTgeyAPvjzNhFZkZ/5N2KFWvVZxrPxR4d6Wd6ObNzY9smsn0hn9CXsrutENoCp0GTsgT+fAvjBZ4TG1MQw9K+Cu7h42Yx9bAdcCXgOsT2x4C/oxIVCV5N4CqvjB2i79VRF5WMDYtR/hnyBlMoV6LyWJrDfDpii80sEosMxC+q3rGQEE/QFcmfNFrwVXdAjvUxsV9h4R/fuo1kl2shqreJSKDRdtWQsmyiOcDd8RjhkVkK1F2638rHcOFVvszySC1lN1Dwtahlxr6H8q8OZWeDoKJDX9IO4SGee4JI6x7IPzPoiMbvmdQp8yMi0q789juYY7pD99La7cBm4oRI4alTeY3wDmxvcMS4Lj4pwutkClhkDpl+rCoB2IWOKXiTgfCb2acNzDD8OADB6UdQlOwYI1gAQv2Dkf2H8yohp9x389ARusQIytAKznD7wPXEtVt3QM8CfySGqpUXWiFxShwVJnarEIB/Aqi6cOyzvD0hW8rYKAsiLWd4Z+MATLZ8BcmqIFVhxkDWbkuA47qVrDQRgiqOsPXu68s8MHCYxH5JVB1esXA5WrmENdkDVGlP2IsxCASX1NXJnaFf4Hf+IfwxeJzc6Nph9AUOiT8i+PuXPi1QVXqcYOg18IdFJhoXmPhPTQbEekHRFV3ichpQFZVH672Oht/1U65FYhQqmZrfbG/aXjMmht++5qubhs1EDs2hD+PKwbEooWpQyteYLMNFJKPzhCpJSLfBJYBC0RkLfApYAvwT8ABwA9E5AFV/RPgQOBHIpIH1gFvr+UYLrQao5zZ6GArj0np6cPBQlueRL/D0kXxs8NvwTO+O/w/3Ynx8E/GAJ2Z8D8LCyLFwudgZboqa0AwThh4D7Wgqm8u89RNJcauBo6s9xjhfzNTpJzZaA39Cxs6Zg1F8RD1OyxtWJoNv9i000A2yIrQ6jZwgbcwdWiBjXkb3nILO8IvJNcZktGaDsI/Q85MRoGjkhsKok9EriaqySq04vnplFdbWHWYC//Od+vO8E/GAON5A8LdQCG5BS+wAQOr9QADn4SNrFwtlHGGvxJ4HTAOPA68Q1W3xn5bK4FH45evUNWLqh3DhVaAJLJapVhFNKWYbMUzmfHwi7D79ws/A7H56Z60Q3BiLEwdWhBaFqwdAGb56smQuI6pzvC3A5eqalZEPgtcClwSP/e4qh5bzwFcaIVLsj6seIrwCSKLh9I1Wlu3tDSw6eDyJw5OO4TG6ctz2Ykb046iYXbfGv7CBHUvsLbgwZ1r+Js5dV3D2pL5BjLuqztmxneijDP8bYmHK4DXN3IMF1ptTIWVhMWsiP8VKEwblmTi179rMLL0yRL+9Ofn/noW8Jy0w2gYveWZtENoGAumq5mp7UKC40MDLzEx7zY4Hn5m7u4+7/8Z807ghsTjw0TkfmA78AlV/Vm1HbjQaiEVhFJFIZRgANjR1KCAzucd0uxdTjufXLkq7RAaZu11Nmq0ejrCr6sZyYY/nT6eD3+ByG0afsN7gGf65qUdQsO8LBd+Y2xozBleRD5O5Pz+jXjTBuBQVd0sIscB/y0iL1DV7ZX240KrtZQ0FhWRoUqGo7WSEHJTTUkrkHvy6UYPnTq9h4R/9x6taQifjk3h16OM58LPQFiY/jy4oz/tEJrCDgO9Djdlwv9ew747w4vIeURF8qdo/OVS1TFgLP79XhF5nKhE555K+3KhFRAlMmQFgfViqjSSTqIT4Z8EOgf3SzuEhhl7dFvaITQFMeB9NNBt4+49dDbnbdx8zMuE71X4ODasNvYFETmdqPj9JFUdSWw/ANiiqjkRWQocTrTwrCIutNqMKnVZe0xJi8bvoI7PsmPBnH2Or13IPb017RAaxoJFBUBWwxfuGQNi0YKr+q58+KuJAeZmwrcL2WWgW0ItlHGGvxToAW6XqPaxYOPwauAyEckCOeAiVa26usyFVvtRto9hYZVhkRhbRGROCuUMSov3M2d2w0GmzeiK4bRDaJiBRdXHhEDf0+HXaFkohrdg72BB8AL8QXelHULDdBvwlquFMs7w15QZeyNwY73HcKEVFgVLh+J2OwXRVVh9WFlw9Yc/TdK1X/gn5KcfsFGPIoRv75A14N9kQSxayMqBjYbMBxH+DVS74EIrIBLu78sTHlrJQvjC78V2D5PpDP9OZfe68IXW7HnhCxSAsa2+DLwd8IxW+zBPutIOoWHGDPw91YqIdBAVtK9T1bMS2/+JyBV+dvz4UOBrwDygA1iuqrdU278LrQCp0GOxkN0ql9EaamFY08r2TeFn5QC6e8LPpIznwxdaO8ZnbuFvO3H/2E4u3u/laYfRMBsJ/3s9aiIvVzPvJ2qts6eAWUSOJxJUST4BfFtVvyIizwduAQar7dyFVsCUsHdI+nNN7XFYeN2c8D1eLBSS9/RlUQ3/fXRK+KeRvs7wp0nEgGHpm+a8kM0GrBEs2DvMFERkMXAmcDnwoXhbB3Al8Bbg3MRwZa8Ym0uNtkrhnyFnNr2qOpQQXGuBR9hbs1VTcbyTDmO7bXz9LJh9WnBVt8B8tbHSrcvA1OGqva4G1vkC8FEig/AC7wVuVtUNRTcwQ8BtIvI+YBZwai0HsHGmnyGUsH4YhMm1W8BS9qrsNcCnp+yoL/xpt2wu/BNyJmOjBmL/nvDtQjaPVTR2DgMDxfDbxMZ01ZMa/lT0oIR/nYDKzvAichYwHJuPLou3LQLeQGT5UMybgetU9R9E5P8A/yEix6hW7krvQqs17FkdWMvgIgFV6TWTrB9EZKiE+EqmMpdSysh0JPylx/MWhH+3tWF9+P0aAXbnwvc+slBIbsEZfhPh1/uBDZFiZWFCFWf4E4GzReS1RNfROcDviNzfH4uzWf0i8piqPg+4ADg93u+vRKQXWABU9BtyodUCEhmmoRpfskdAlfHKKjBY9HiUyRksqGW6sDf8k8Cqp8J3hj/6+eF7gQF0/S78VaxzumalHYJjiFEDwv3xvIEsbxVU9VIic1LijNZHkqsO4+07Y5EF0SzRKcB1InI00TV6Y7XjuNBqX6YYl5YRbklrh6VEn2nZQngA3RJ+49bejvCLTbvCX5MAwHg+/BVWVTL/QZAz8B66jGRR+gm/tGGmGJbWyYeBfxORDxIVxp+vNaSSXWiFxWiR2Bpkb/H7UmA+UTueisiSw1oR27SyI1dzD+22Zfh3NrIoHbIz7RAaJmtApFhYddhhRGhZEIyzZpg8UNU7gTtLbJ+d+P1hounGuphZ/5OBU+yfVVT8vooal5qycUPTY5tu5nWGXxc0e374q/UARjaG/z4siBQLNTUW3gOAkXUuTpNwoRUwSeFVg1kpxIal+VWrWx1ay+npCn+6asszs1hyXNUEZNuTfST8aVwLrV8sFMN/a8v9fHj+CWmH0TAW5OLoDPICK3aGF5GTgc8B3cC9wAWqmpXojuwq4LXACNHU4X3V9u9CKzDKFMkXKBiWVqzRstBUejwb/vLpJc/bynj45XIm6JDwa2osrJz89JzjTTQKfDb8PydGdeYILRLO8CKSIWqzc4qq/l5ELgPOI2o0fQZwePzv5cBX4p8VcaHVfkyxhihh/5A0JS3wSmqs0aKvnE4Lh/nzwhdaYzttFJv2dIRvzjiRD/+iYmH681cd4X+vAY7W8Fd2H5AJ/zpRCyWc4fcHxlT19/GQ24lWJl4DnANcHxfArxCReSKyUFUr1uO40GozylhDJO0flgNHMVVsFWq0kk2mSx/jmU3NCzglegfC99vJdISfgQAbIiVr4O7dwtThXCOXpM7wPwpyYuBN1EaxM/wmoEtEjlfVe4DXA0vi5w4Bnkq8dm28zYWWJVT1ioLYUtXzi5+vUqs1BKDbwzcs7V0Q/oUxH34NuRl86rA9mDDwHgDWZ8KvIe3HRsa9Xmd4VVUReRPweRHpAW6DPV3CS6WN3d7BIrHYGoKSNVvVDUsz4U8xbFvdk3YIDXPAieFf3AH67g//s9idHUs7hMYJ/2tNrwH/KbCxenKj2rgTrNcZXkS+rqpvA14FICKvYW+HlbXszW4BLKaG1f4utMJlTy2Xqp5fpUh+Evkd4X+BuvvDv2PM7wr/ZAywYyL8dkgWLowWVk4ekrdxSXo6E37GvcOI6K1EGWf4t4nIgao6HGe0LiGq3wK4GXiviHyLqAh+W7X6LHCh1WqKDUYLDDa642Qj6fgYhdqsNcAdVPhsOxaE3wQ4O7Y17RAaJj8afp0ZQKc7SLcFFi6LO43UBeUMiN6cheWf+87F8bRiBviKqv4k3n4LkbXDY0T2Du+oZWcutFpIscFogRp7ICZF2mCtx4gzW1ApndkX/lTP3c8cmHYIDXPmGeFnFgFm3xm+F5iFrJwFskaE1k4DHlSHSH/aIUwrSWd4Vb0YuLjEGAXeU+++XWi1KUVmpEMlpgaLVxcWPLQKHEEZpKe7WWGmxkA+/AvjXdd28apzw8/MjefDz8xZKIbPG1h1eM3WB/ibuS9JO4yGmSvhX1rXqQ2rjXYg/L+GmcOkJtMlhNdaSvtrTS2ON3BR6Tdwx3jcn+/EQgVzdyZ8Hy2nPThv4IWEL9tdpISEiBwJ3JDYtBT4JJGf1jlEFrrDRC7w6+PXLCOyhegCNqnqSZWO4UIrUKr0Pay48lB3hm/vsF9f+NNuMsuGIeBAZ/jmjNsmws8GWeDu3Ja0Q2gK8zLhl2fMk5lxA6WqjwLHwp5WPOuAm4BnVfXv4u1/SyS+LhKRecA/A6er6hoRqVrH4kLLCAl/rV6igvhPlx3cGX7xcnd3+Bkt3WHh3h3GDEwdWvCgUg2/eLnXyMKKAQNTh+Mzsxj+FOBxVX2yaPss9vplvQX4rqquAVDV4Wo7Df+vYYZT3J6nyOqhOLM1BEDWgEgJ/7qI9IdfKwc2nOG7M+GfCi3UaA2Ije/EqAHRu0PH0w4hDd4EfLPwQEQuB/4S2Ab8cbz5CCLn+DuJ3OSvUtXrK+00/LPLDCUhpgYLDvGx1cMPqWJYmt+0rfUBtpiurvAv7tI3K+0QmoKJ9jX58C+MYqD2cpeGnx0F6DOQ0VqamZ12CE2hkjN80bhu4GxiXy0AVf048HERuRR4L/ApIt10HFH2qw/4lYisSPRGnEL4fw0zg1Gi/oZJelV1KOGjBbW4wgOdJ76syeFNP7t/cF/aITTMjp9tTDuEptDfEX49ioXpTwtYECgAHQYWuTyjBrolUNUZPskZwH2q+kyJ5/4T+AGR0FpLVAC/C9glIncBLwZcaIVMoeVO8TRh4TnYk+FaEf+rXAy/KfwLfE9v+BfGjp7wp3oAdmTDt9qwQM7AdJWVAmwLnQa2zLypwzczedrwcFX9Q/zwbKJV/QDfA74kIp1AN5FD/Ocr7diFVljssXgo0euw4Ku1iL2CqzRj4X+BemcZEFpG/AB7MuHX1ViY/rRwcX/aSH+9gyX8FcUW/p5qRUT6gdOAv0psviK2fsgDTwIXAajqyrhE58H4uX9X1Ycq7b8ZQqtcm5lWMTiNx2pnJvU6LGyMhdcySme0hgDUgNDauil8lbJ1Ewz+SfifRVa3px1Cw+QMCC0L3LtzNS+d9Zy0w2iYfgOrJyeMuPTXgqqOEPlmJbf9eYXxVwJX1rr/hoVWuTYzrWKaRV3bkpgyLOUaX8hoFbvFAyCzwxcp49nwL+6Hn7IDCzfwFlzVLVgjWOC42eGLLLDR4HuG2ju0BJ86DJ9R9hqVFqYPCwJrAJjSiE5HwnctntUXfiZIx22cyHbnbBTNOulzsIRvfguww0Dnil4TbcprIzYqvQdYp6pnicjJwOeIarDuBS5Q1ayIXAy8NX5ZJ3A0cICqVnTadaEVDqVWHhb3RCy4w6+iQlNpyYT/Bdq1O/y6IOkNXyyCjYxW1oAXmAXC/0uKmGvg0rreQrq9dt4PrATmSOST8jXgFFX9vYhcBpwHXJOcMhSR1wEfrCaywIVWMBRWHtYwpiC2Sk4bAuhE+IXk3Z3hXxilO/w6DitYEIsWpqt2G5mu6jMgGSeMfBbVEJHFwJnA5cCHiGq1xhK+WLcTeWtdU/TSSasUK+FCyxgJsQVl+h7mh7dOe1zNZsGi8Ps16mj4J2OwsTpp3EAxvAVn+Cdy4ddeAvR1zEs7hIY5FhuGpTXwBeCjRKU2AJuInN+PV9V7gNcDS5IviFcpnk5kYloVF1oBUqL4HWo0KwXIHLRf02OabjZvCP/i3nuQjYtKxsAKq90GvMDUgNA6IGOjRqvTwM1Hj4b/HqCyM7yInAUMq+q9IrIMQFVVRN4EfF5EeoDbgGzRbl8H/KKWaUNwoRUakywdigRX0jtrEdH0YenPdzT8ufc+A8XwHbNtfP0sNGTu7gj/s7AgtKwQfn4UhjMW3kVVZ/gTgbNF5LVE19I5IvJ1VX0b8CoAEXkNUX/DJJN6IlYj/LPLDCLpAl/KQ6tA0dThFLJraxLhbU1X+H6AZPYP32YDYHZn+NmgrWM70w6hYSwIrV4D2VGAYQOF5M/HRi/WSqjqpcS9DeOM1kdU9W0icqCqDscZrUuI6reIx80FTgLeVutxXGgFSAnBVeu04VALw5pWnn56TtohNMzT34YXfnJJ9YFtTgfhC/eF/eFPp1uo0Vox8hQLugaqD2xzjuoK/+9ppYZ/A9UAF8fTihngK6r6k8Rz5wK3xb0Oa8KFVsAUm8UWTSWWFV86En5K2MKqwyMvPRQMGGVuz4bvy2Zh+lMM1AU9v39R2iE0hZyBv6cFEr6FTj2o6p3AnfHvFwMXlxl3HXBdPft2oZUOxW2LBmt5UZki+CQFw9KKGa7xTeFf3LeNhX8SkP0WpB1CU8hI+Bd4DIgUCxmtxRkb0+kWzG5ihu0AACAASURBVD5X5cOfTm8XXGilQIlM1FCNL93TVLocsRgrLoyfVKvVaWDV7kYDjYzJhy94AXIGsnIWskEWsnJzsFGjZYHwv9XtgwstAxRlupLCqrD6cNKKiZyBqfdDCL/YlHEbrWss9AkMX6LYMCxdreFPQwMcLuFn5g7KGFhxVAMiciRwQ2LTUuCTqvqF+PmPELnBH6Cqm0REgKuA1wIjwPmqel+lY7jQCpQicTUIPBI/TmavSlo89BwafjZo64NdaYfQONlia5Yw6e8M/4Q84S142oJdGn7XCoBxCV/07lIb56dqqOqjwLGwp+fhOuCm+PES4DRgTeIlZwCHx/9eDnwl/lkWF1ptTglBVdh2VMHaIdF2pyCyKtdorQ3fg2rCQMsUem2YM1oQKTsmDKR5DXDwLBvfiS2ELxhfNHOc4ZOcAjyuqk/Gjz9P5Br/vcSYc4DrNfJTWSEi80RkoapuKLdTF1rtz566rEQtVy+wOt5WEGI1iSyAroPD/9hHfht+TQ3Z8E/GAFkD7Wt6O8LPkIqBm48uA0XkVlgj4d+QQ2Vn+BLsMSIVkbOBdar6G5m84OcQ4KnE47XxNhdaBinnEr+CKM1Z0qwUQDrDP5l1hp+ZNzN12GnAZHLCQp2ZgffQbURoWViYsFGt1JBWdIbfg4h0A2cDl8a9DD8OvKbU0FKHqbRvF1qBUsklvoKf1tD0RulU4vuf2MCZnwrf4sHCqkMLheQWPocbtz/En885Ju0wGuYZAyLFQkF/nZwB3Keqz4jIC4HDgEI2azFwn4icQJTBSjpNL6ZCYgNcaIXGnixWYcO+uMRvubd1AU4Xxy3YmHYIDbPwQ8emHUJTyBuYOswYsHew0Nz71IEjGDHQKXBQwq81e1hnnI/Wm4mnDVX1t8CBhSdEZDVwfLzq8GbgvSLyLaIi+G2V6rPAhVZQJETVjxP1Wkk7h1cSfaY/rbSfuYeHXxu0ZWVP2iE0TH5dxZugYLCQDcoaKOgXA8axXQYEL0Cfhj8FOm7gBqpW4qnC04C/qmH4LUTWDo8R2Tu8o9oLXGiFSWeiQD45TbgbqJqv6pgXvr3DHzbOTzuEhlnYE/7nADaaGbvQag/mG7kk7Zbwp3EXzqCpQ1UdAfav8Pxg4ncF3lPP/m38Vc88Vhd+SbrMi8itwLIU4pl2+iX8C6PuDr+OA2zYO2QMrNizQPhSMeKAfPjTuI+wI+0QzOBCK0xWQ8neh08Bh1KlRiu3Nfxlu/nwkyjIgeEXwgP0dYY/jZvN21gBGjo9akNqrc6EX56Rs3CSrQERmQf8O3AM0erBdxJdQ98NFIqBP6aqt8TjLwUuAHLA36rqj6odw4VWmCSL4guO8BDVaq0nsngo1G4V6rb2kOkP/26rryP8LAojNkwyLdgKWJh2s8CYAUd1gDEDnQI7zOQXq3IV8ENVfX1s8dBPJLQ+r6qfSw4UkecTeW29gOga+2MROUK1ckGbC60ASa40JHaIL5HdKvQ4HIDJOeDM/PDn3jMGTsg6YqOvmwWjzIwBsWhhUcJqNdDDFOg18J04JBP+daIaIjIHeDVwPoCqjgPjFW68zgG+papjwBMi8hhwAvCrSsdxodVGlBBLkLByKEZVr0i6xRcVyP+87IHGw58mmciHX0gu/eEvAQfoMHBRyRu4excDQqvDSGbRQjZoq5G+k1VYSjQ9+FUReTHRYrL3x8+9V0T+ErgH+LCqPkvkAL8i8fqCK3xFXGi1F3vEUoGEkCrHnmnEhFArTBu+giirtWbSKzrDnzrs7QxfLD7y+WGOfHf4/cS8vslpFnfteIxzB45OO4yG2W1g6nCHEaFVpQVPJ/BS4H2qereIXAUsB74E/D1RzdbfA/9AVLtVtyt84SBOwBRPIxLVbBUMmlYQfcbmPue+3vBPAoe+aTZMhP8+LEwdWqgzs8ArZz+XjRr+Yp25En7vzH6xcdmo0oJnLbBWVe+OH38HWK6qzxQGiMi/Ad9PjK/LFR4MXoBnKvE04nKiVCjxz5Kfr/SFP+22bVfaETQBIz5a3ZnwTyPjnpVrC7oMiHaAUQPu9jkDWblqqOrTIvKUiBypqo8CpwAPi8jChNv7ucBD8e83A/8pIv9INHN0OPC/1Y4T/hnS2UNRdgtqUNqhEn4FBCZq5cCG0NqZtbEwIXS2G8hmAXQTfnnGUpmVdgjTxfuAb8QrDlcROb1/UUSOJZoWXE3sGK+qvxORbwMPA1ngPdVWHIILrSAoUyRfTNLGobDS8IhSA3VX+Ct7ns3NSTuEhpED9ks7hCbxZNoBNEyXgT6BFqZwOwj/PQDs1vBvoqwsTKiGqj4AHF+0+e0Vxl8OXF7PMVxohcGUInmYIsDmA/8T/z7FOytJx3MObnZ8086zHeG7quvmZ9MOoSl0ukhxmsSoAYECduqbnObgfw1hk7R0GCyyd1hGGYd4C61fZuft1w+EQnYGNZ9tZywU9GdsFAUwauA70W/ks6hGGWf4R4EbiOyVVgNvVNVnReQcolWIeaKpww+oankrpRgXWnZ4JGEFsajSwMySqrYfbc9cnkg7hIYZ/93G6oMCYHcu/LoaC1k5NZCVyxrwAgN4fmYg7RAaZqXa6FxRA6Wc4T8G3JFYZLYcuAS4A7hZVVVEXgR8m2i1f0VcaAVG0XThYNHjgn9WshXPVAxkg7YZ+NPN9IZ/cQeY3xW+F9iakeG0Q2gYMZCBGOgP3xYBYAvhT4EukmplweFTwRn+HKJZIYCvAXcCl6jqzsTLZ1GDhxa40AqR5HThEHAqe13glwLHQeW267plSwvDmx4GDNRy3H/bfhz/l+G/jx3ZdWmH0DBzusJfYZUzMHV4964nOaLvoLTDaJj9OsJfrLOZ8DPVNVDOGf6ggr2Dqm4QkQMLLxCRc4HPAAcCZ9ZyEBda4dOZLJQvynBVLIoPmREDy6dPujAPhH8Hb6EFj5qYsgr/cziyL/yFOgATBjyosmrhO7HPzvBlUdWbgJtE5NVE9VqnVovB5EXYGKOUnwMeBeYlNyTmlHuJPEEKrXj2FsYPhF8/kJPwfY/yO8J/D1awMO2WCf8tGJCKERZ6HY4bEIuwb87wwDMF01IRWQhMqS1Q1btE5LkiskBVN1WKwYVWm1PUOLrUc6cntxW5wydb8eyt1xoN30fLAtJp47Iylg+/jVDeyN176PQasUXoMSAZ+w0sEKlGOWf4+N95wBXxz+8BiMjzgMfjYviXAt3A5mrHsfFXPbNZXfS4V1UvLGPxMARANvylxz1+YWwbZneEXzS7cXxb2iE0jAVrBAtTbgBbCP/m4yBstAirgVLO8Bng2yJyAbAGeEM89s+BvxSRCWA38Beq1S9GLrTCZzVMqs16nYicQfRHAHAfRcXxaqCRcbeBXmIYqG0CyJuobwofC5+DjW8EdBt4JyNGRG81yjjDQ5TdKh77WeCz9R7DhZYdelV1SESWAatV9fxYfJ1K0ecsXeEXYOcN3L2TMyAWsVEMbyEbZEFo9Rm5JFlYXLHLgEVFu2Djr3pmMxrXcA3Gj1cDB4nInezteTiZrvA/9i4J/25Ls+G/B7BhWNrfGf70p4U6s6yRLErOwGdh4eajVkRkNdG1MgdkVfX4ePv7gPcSucD/QFU/KiJdRE7yLyXSUNer6mcq7T/8K+4MR1WvgD2eWgCPAONEhfAFA9NJyOzwDSY3sj3tEBonE34rJICuTPhFsxZWHVrIomzN2/hOLMj0pR1Cw4T/11Q3f5xcPSgifwycA7xIVccSXlpvAHpU9YUi0g88LCLfVNXV5XbsQisMRhNCarDcoESdVkFclTQwVQOrDsOfrII7vjXAqW/ZWX1gmzORD38K1ILZpwUez2004aVlwRrhORK+WGyQvwauUNUxAFUtWDwoMEtEOoE+osRGxTt/F1rtQfH03yQKWSuYlLmasg/gKFU9Px63nMgxfoppqfT3Nxpv6uQJ393+1NdtIl/Rwz8MLNRoeWPs9uD4/iVph9AUOg3cCu6wsOCodhS4TUQU+NfYe+sI4FUicjnR9fUjqvprIq+tc4ANRH0RP6iqFS9ILrTagBLTf/u0DxFZXtRY+lBKrJzIP7FmXw/TNuQNeLx0vuh5aYfQFLr+5zdph9AwXR3h/z1ZwMrU4f4Gpg43avgzH1DVGb7Aiaq6Pp4evF1EHiHSR/OJDL9fRmT3sBQ4gaiWa1H8/M9E5MequqpcDC60DJEQbMvZ21h6DUV1WrLf3OkPrsk827Er7RAaZ8SGM3yngRqt8Xz4K6wsFC+PG8ksWlgBOiDhr06Hqs7whTHr45/DInITkZhaC3w39sn6XxHJAwuAtwA/VNUJYFhEfkFkD+FCyypFvQ0LLIp/rieq0zoi+aTuDF+kbA8/Mw9zwl+UAFCDX5/j1ERfxsbFvcvAdHr4VWa1ISKzgIyq7oh/fw1wGbATOBm4U0SOIHKB30SUvDhZRL5ONHX4CuALlY7hQit8epNNpQuIyK0UCaw95MO/MO4I/+bdhJ8ZgEj4H0YuH/5lxUIuaJ70pB1CU7CQmZtjJKNVAwcRNYmGSBP9p6r+MHaKv1ZEHiIqeD8vbr3zZeCrwEOAAF9V1QcrHcCFVqAkMlmDZbY/FW9aSvHnbOCikjFQbEqHja+fhYxWl4GaPwuCd5uG78kGME/Cb18zYWD6sxbi2qoXl9g+DrytxPad7G3JUxM2zvQzk4IT/FBhQyyyjipyhZ9Pkb1Dfmv4lgK9hF9nhpGMlgVGDTTGtuCj1dVt4AYKGx5U++Hnp2bhQit8ij22HokfLyKyd5hCZl74tUEHhl+7jG6u2vQ9CCwU/lrAgs3GzryNjFavgQUiMyWjBSAiHcA9wDpVPUtErgNOAgrd5s+PeyIWxr8MWEHUVPo71fbvQitwijy2lhNNFRZWHL4C+JPE8CEAMVCEvTMT/oq9az+3k3f8TfgnZAseVBayQRbavqwZ28xL+xanHUbDWFgBOjFjyuEBeD+wEpiT2HZxKREVi7LPAj+qdecutMKjrLlpwUuLqEZrEdHn+9Mp48Zs3DWGzjvfH35/PYAOC/VyTltwZO9B7NLwp3EHDNRo9RP+TWAtiMhi4EzgcuBDNbzkfcCNRN5aNeFCKzCqmZsmnr+VEvVZALo9/Bqtg7PhixSdCP+CArA9G352Mavhz0VbWJRwQOb/Z+/c4+Qoq7z/PT09l0wyk4SEIOHiGBSyiMAiICuokZvLRRQXFF5REUV3V1lQQUEXjLru666uguu7rq6A64p4Be+gqBsQXWAJgiJEkSRAEpIQcs/cu8/7R1XP1PT0dPdM90z1Of18P598ZqaquvtUqqvqV+c5z+/YP68BFmN/9uSftDftEGaKa4H3A11Fyz8uItcAPweujPsd7gecTWT7EIRWs1HCT+sp4N6SGzuwdxiyn5lH5s9LO4S6ML91dtoh1IwHsZhxMOuw30nPydXY/z55oZwzvIicCWxW1ZUisiyxzVXARiLvrC8CHyDy1roW+ICq5iYzyzcILeMUDRXCqAv8gcDFRZsvB5BZ9p+25jiwqNBt29MOoS70OShg9tBUOmf/+clBZVNEl9i/tT6jPtohVXCGPx44S0ROJ7qPdovIV1W1YOswICI3ApfHfx8NfD0WWQuB00VkWFW/Wy4G+9+GwIhhaSy6lhEVwK+hRH0WADn7xct5B19dabdfx+GFtoz975OHgv7dDuqzAOY4EFpZN7J3YlT1KqLsFXFG63JVvUBE9lXVpyVSVK8lMidFVZ9XeG08M/GHlUQWBKFlmXFF8XF9VrLfYQdwAkXHWRYtnLEgp4unW3emHULNeKiVAxAXF2T7IsWDvcNCJzVaAw5m7A07OCdq4CYR2Zsoyfog8Ne1vFkQWkZJFsUX1WctZnT48ARKFMTrs9tmKsxpoz1vf0aMLJifdgh1ISNr0w6hZtSFWLTPsAOBAjDgYCi62VDVFcCK+PcTq9j+wmrfOwgt+/QTu8HDuKL41YyKrhFkXnfxInN0qP1skG58Ju0Q6kJ/zv5wj4dCcg/0OvBkA+iRWWmHUDOP6Z60Q3BDEFrGKXhnxcOIBYPSNUQdxqFEY2nZa6+ZC3CaWIJ9kfLwjXDwX2xNO4yamdViv9bMQ6/DIQciZdXAZo7pWFx5wwbn9v4n0g6hZo7s2DftEGYEEbkBKMw+PCxetpxoMlnhRvNBVf1x3Gj6C0RF8Xng0jgTVpYgtBqYEpYNlbgn/jeuLmsMDvybcmo/A3HI0VtwMGGPWRn7QmvTgP3hdA+tkJ4/6zmsGbZff7m41X4v1qPy9rNyVfJl4HPAV4qWf0ZVP1W07GIAVX2RiCwCbhORY1TLjxUHodVYJPsWQty7kOrEVnG7nYlx0Mz4Txn7F4EX7Wd/HwDyK+17BrVm7J8THoY/Ox2IdvAxY29jpjnqzFT1LhHpqXLzQ4kMTFHVzSKynSi7dV+5FwWh1UAk+xbCSEZraXHR3QSZrkI2K1kMX5q59ouwW/SptEOomb7H7AsU8NHXbY8Dw1IPPKfTx8OHB/qdTEyogXeLyJuJmk2/T1W3AQ8BrxGRrwMHAC+OfwahZZWi+qskPYzPdBUE1hJK1GWNwYHZpwdn+NZ5DnYCGMjbH4ruzNq3FfBgs9HqYB8AtjvwAxsU+/cJKO8MX4bPAx8j8n35GPAvwEXADcCfEYmvJ4BfAxX7dwWh1eAUZ7mgdKZrUvVc/faf3lvtl6PQMs/+cBVAHvuFZh4MSz3Qh/2CfoB2B55m2z0UkFLRGX6i12wq/C4i/wH8MF4+DLwnse7XwGOV3i9cXQwSZ7qK/bMKJC0dStZtyf7Pn+YIp58dLY+nHULN7HzYx03FA1kHsw49OMN7sXfYR+y3ORsW+9+nqVJwho//PJvYGV5EOgFR1T0icgowrKqPVHq/ILTs0g8sZfwQYtKsNEuJNjzau2Pag5tu9huyn9bOttvfB/BRo+XBGsEDexwMuQH0iY9sdTMgIjcTta5bKCLrgA8Dy0TkSKKhw7XAO+PNFwE/EZE8sB54UzWfEYSWUQr1W0Q1WRsYXwRf0qzUC1uy9lPzrXN9PDF6aP3iAQ/2DoNOBO+gg0JyD8Of1aCq55dYfP0E264FDpnsZwShZZhk/VaJhtJPMkFR/Ib3/WAmwptW/s9nX5F2CHXgUL777t+nHUTNDOmatEOomfW9W9IOIQCsZwvL5i9NO4ya2R/7Q4cvGGoOoTUTNKXQmoIRqBWqMizd7z8unKl4po3XvL7YW84o9kuDOCBr3y7Eg8Gk/RxKxC4Hs1iHW+xnF8883L6FTrWISAvRTML1qnqmiJwIfApoA1YCb4sL4adEUwotoENVl6f14TMg9O4u+nuM8GpZctQ0fvTMMDtzc9ohBGK6HNSjeKgN8iK0POChT+Bdv9kv7RDqwuuq2+xS4FGgW0QywH8CJ6nqH0Xko8BbmGA4sRqaVWilzbQLvYSYW4zDeq2Q1G4cPIgUDx5UGQc1WoHGIe+g00A1iMj+wBnAx4H3AguAAVX9Y7zJHcBVBKFVX2Yg49QzA5/bo6oXJmq3LgaWA9BiPwMRnt4bh33E/ij8M9qfdgg1E86JQD3ZlWkOoQVcC7wf6Ir/3gK0isjRqno/cA6R+/uUCUKrNNOacSrh9F73z004yhe8tOxXLCcIGa3GYSjc4huCcE4EAuMp5wwvImcCm1V1pYgsA1BVFZHzgM+ISDvwU6pwfy9HEFpOKJMNSxbIA5Dfan8Ucf3wrrRDCMS0ZeelHULNPD28O+0QasaDn9np2X3TDqEu7FvTbbkxmJP3MRRdwRn+eOAsETmd6P7ZLSJfVdULgJcBiMipVGprV4EgtNKhf4KsVg9MeQixVDPpkrMPM3P3nuRbNx4LW2anHUIgZtBB25SFLZ1phxAAfq3b0w6hLrwqY//h49kW+8K9Eqp6FVH9FXFG63JVvUBEFqnq5jij9QGi+q0pE4RWCpTqXwhjhhSnPIRYJNLWJeq0loxuZH+Qwf4eBAL1xUNGCycF/QMODkWTX2OviIcVM8DnVfUXtbxZEFr+GBFpiTqtxcARhQ20z/6wW6gLahyG1P6x8CBS7B+FiJyDPXnhgP2xw4fbm0seqOoKYEX8+xXAFfV67+b6n2x8CkOKPdVsPMEQ48hrC5mzcRmtjH2XTA83xkAgMBZFXZzb/Q5GDezfJRqHILQaiIQwWl7lS8YNMRZeWyTCFpM0MW0Jhz1QP9rE/iXZQ1ZOnQy7eWBDq32h1Wn/lKgKETkE+EZi0RLgGlW9tl6fEe64/khmxVbhs9WQC4NJL3gQKR4I50Tj4EGk9NrXilWhqn8AjoSRVjzrgVvr+RlBaDUmE81KLKaneEHRcOHSItNSonVNcgYFAlUSzojGwIE+AaDVQXJxZ3OeFCcBj6vqE/V80yC0GpCJZiUWU06Mqeon4mL424k8QJ4cXWf/chaGSQL1JO8gG+ThvujlvPbQDqld7Z8TU+A8oO6NdIPQckaJAvl7iI7zyLHWnVtmOqy605e331/PC/u2zEk7hJrZmLffBNhDEXnOwUMgwH75wbRDqJnccHvaIdSFcs7wRdu1AWcR+2rVE4tCq9phtXL01CGOhqKoiXTSuHScaal02Df7zDoowA40Di0O8kH29wDEyXn9aNa+SNln2IforeAMn+Q04AFV3VTvGMwJrWqH1cpRB6HWKCRFZ7KJdDKjtTr+OSK+NGff48XLEEMgEBjFy3nd7aB9jQ+ZNSnOZxqGDcGg0AqMkhSdCXPSHlW9MLkcOJlkryYHQivQOPTX1m81UCdyTkSKB7IODkWHk2HcahCRTuAU4J3T8f5BaDkhOdsw4Qa/If45TDSbIiJvvzddmMreOOxyUC/XIva/TzkHN3cvDDj4Pu0j9u8T1aKqvcCC6Xr/ILScUcINfkP8886RjVrt1w8EGgcPhqXD2jw3lcD006r2Ve9WWtMOwQ1BaDmlYO9AVK+1mmSN1uY1qcVVL96Yn7aHjxnjX3KrK29kgBYHvmzq4MbogRvb7Z/XAP+v3f6swyNy4YG8XgShVYYJegkWGDebr8EZqdHSzU+lGUddOPvsrWmHUDM3fLcz7RACMR6Goj0Ukl+W6007hLpwVNb+uf1YS3PUXorIDcCZwGZVPSxethdRW54eYC3welXdNtXPsCQU0mBcL8ECIrJ8onXTQQXRV8xxwKsSf68hYVhKtq1+gaWEDoWhnkbBg9WGF/+mQGNgX/L6EO5V8mXgc8BXEsuuBH6eGBm6EvjAVD8gCC07TCj6SlEkzMZm33ZOWZgH6kivgyJygE77I4cujoUHw9K8k5v7ocP2b61Pt/g4FpVQ1btEpKdo8WsYbVv3n8AKgtAKlMh4JWcdjmXIfv2AtNsv1FT60g6hLnjIaIF9oZVrRuejBmXAvubFw1kN1TvDF7GPqj4NoKpPi8iiWmIIQsswReJqjH9W0TaQLIbvtV8HIVkHaRQnDOabo5YjEKiWJYP2z4mfzUo7gvowCWf4aSMILduMDCeW8M9KsoGobgsAabWfDcrv6k87hJr5+RkdLPuhfdGbydh/fM/n7WeDvAy7eaDfwUzcxXn7+1ADm0Rk3zibtS+wuZY3C0LLCUX+Wcmeh+NmR+rgwIzHV29ktv2px6/8xnYfdTXBGqEh8PBd8oIHiTLU3F+n7wNvAT4R//xeLW8WhFaDUqLmqqfUsuLXJQTXbUSWDl3ArjHvfcDz6h3ujDPw/d+mHULNiAP3aICMg/3wcCw8NMb2Ume20UFpQ5ePQ1EREbmZqPB9oYisAz5MJLC+KSJvI5qxf24tnxGEVuMyZpZhPCxYvOxnRQ2yk8OGTzF+CDFi1476RpoCGfsOFW5MMvfk7U+u8HAscmL/zujhOAB02j8UtDtojF0Nqnr+BKtOmmD5pGlWodVfJFCK6ZmhOGolWyS8khmvA4GLi7aPtu2cPQOhTS+ZDvtPjFkHWRTw4QyfF/s3FRcixccpgQeXv9/br85oGJpSaBWG1yaigghrJNYm/yhqu/Mk8JF41VgD0w77rsX5QfuPjC0ubBF8kHUw7OZBt+ecFPTvtv91ot/JMG4lRGQe8CXgMCKv2YuAy4BD4k3mAdtV9UgReSNwReLlhwNHqeqD5T6jKYWWI9ZCaZuHxLJil3jYs3MGQ5wexMFMt+BG3jjkHagUFxktJ/Q46FyxrnmeA68DblfVc0SkDehU1TcUVorIvwA7AFT1JuCmePmLgO9VElkQhJZ1CkOgIx5aCZuH8QKrQLv9jNbuJ+1fBQbV/uxPgL2y9oeid+Tsm8cG4d44ZBxo3v1z9q+xlRCRbuDlwIUAqjoIDCbWC/B64MQSLz8fuLmazwlCyw79wNLkgqSlQ2K4s1AQf0/8bzwOhg7FwZWsTXycfh5u8B7c7T3MOvRCX8b+sdjQYv+8roIlwDPAjSJyBLASuFRV98TrXwZsUtXHSrz2DUSteiri40rfBMT1V8snWlf4veyQYVwMr49UzHQ2PN1H2jdd/emRsOz7eypv2ODkHcxOyjsQi8GwtHHY1mJ/KPoVffaHP6FiC54scBRwiareKyLXETWQvjpeXzJrJSIvAXpV9eFqYghCyyAlTElLUchojTMsZcD+dHyZZd/f4eVf35p2CHWhxcHT+1BoIxSoI/bPCMg5mQJaoQXPOmCdqt4b//1tIqGFiGSB1wEvLvG686hy2BCC0LJKh6ouL2FgmmQxUVp0PkWGpR6Q1vDVbRQ6HAyBDoj9ptKhGL5x2OlAaS3yobPKoqobReQpETlEVf9A5J31SLz6ZGCVqq5LvkZEMkQGpi+v9nPsXyGbmHI2FaWaSY/Q3TVdIc0Yum5T2iHUTGvGx+nnwbDUQ42WkwSEC3Y58GUbap4v1CXATfGMw9XAW+PlE2WtXk6UBVtd7Qf4uNIHRpjAtHQNka9WhIMGuh7sHQby9rMoAEvaFqQdQs2sHbQ/jBtqtBqHLrV/fdrpoM6sGmJ7hqNLLL9wgu1XENVAV00QbZIW0AAAIABJREFUWk5ICKweYBVjTUvH1GlJq/1Cch20X1PjwVEdoF/DsWgEJAithqHfgUaxf0Y0DkFo2WLEN6vEumTd1lIisbWEqLH0WBy04FEHhoAuhqvwIbQyDoSW/T3wgwOdxY4m+kKJSAtwP7BeVc8UkXcTucMfBOytqlvi7eYCXyUaLcoCn1LVGyu9fxBahkj4ZpUqhO8pbBOvWwLcHa8ba/XQNXcmwp1WdNC+0PJQRA4+fLQ6M/azvPlQDN8wPC32Hz72dVJDWiWXAo8C3fHfvwJ+CKwo2u5dwCOq+moR2Rv4g4jcFBudTkhT/U86o6OoofTyMrYPY8xL9ak1MxXjtJF97t5ph1Azfbo+7RDqwl4t9jOk23K9aYdQMxqGDhuGOQ7yi/PsPz9VhYjsD5wBfBx4L4Cq/iZeV7y5Al2xY/wcYCtQUVUHoeWHgnP8xEOGMdLVPdEqM+TX2hcp//2mhZz8X/aLsDscDIF6qNEK9g6Nw24HDZm3O/DHq5JrgfcD1UzH/xzwfaJERhfwBtXKKf0gtGxSsh1P0ZDhaUDpBm45+8NuHvAgsgB25n30bAwE6sVc7D98ZJ3o9nLO8CJyJrBZVVeKyLIq3u5VwINEvQ8PAu4QkV+q6s5yLwpCyyATteMpasWzmKhgb1xjaR1wcGMcsl8D0etEoHQ4qOXozdk/Fh6GDg9v3zftEOrCJux7yx0+ZL/7BlR0hj8eOEtETicqu+kWka+q6gUTbP9W4BMapY//JCJriJIe95WLwf4VsnkpzECcqA3PBkabS4+h5aWnT29kM8DKa76Qdgg1M6/FfnNvgM6M/QtyvsW+SPHAk8M70g6hLvxZ615ph1AzmSYYilbVq4CrAOKM1uVlRBZElkknAb8UkX2AQ4hMTssShFaDUeSHNSGJGYjl2vBAqV6HWfs3xkG1n5oPNA6Z8UWv5hAHpgIeZrACtDsohu9tnhqtcYjI3xHVbT0H+K2I/FhV3w58DPiyiPyOyMXjAwXrh3IEodV4FPywllezcdFwYSnR1Q0cM+Y1e7bVGmPqzG6x76qeczDUAzCo9mv+hpzc4AONQZsDobUoZ788YzLEju8r4t8/C3y2xDYbgFMn+95BaBmnSFz1qOqFRct+TFS8tySN+KYL+8/ufsg4OBoe9iG04Gkc+h3MOrR/RjQOQWgZoMLwYE+hJ5OIXJlwjl+VeM0SYP7IK3L2s0FtWftZlHzOx42xz0HPRg8ixYNYHHYgUAD2Ufu31j0OGmNXQkQOAL5CNESYJ5qReF287hLg3UQ+WT9S1fdP9XPsfxuagzHmpEmSQ4xFdVtLGC2Sv3v8K23zzMCstEOomUx2V9ohBGI8iJQgFhuH3WJfMO6fLe0O5Ixh4H2q+oCIdAErReQOYB/gNcDhqjogIotq+ZAgtOzTnxBbY1vtjCfazsHY+2EHbUo7hJr5FnDuE/ZvLHMczDr04AXmRaR4EIweiuFvc9DxAeDoMutU9Wng6fj3XSLyKLAfcDGRjcNAvG5zLTEEoWWcZDF8gcRQY2nh1WL/sO/aWm6ipQ3esrMPHNxU7D+7+7i5e8DLrMPuvH3R29Fkp4SI9AB/DtwLfBJ4mYh8nMgg/HJV/d+pvrf9O25gDLHIWpooii/4aB03stGQfTO9WXPs74Pssn8xBhh2MOvQQzbIg1gs0VvOJAMO6ptaHZwTUN4ZPrHNHOA7wGWqulNEskR1zccRzdr/pogs0Sn2uQpCqzT9iaJya3QAa4vqtBZTwrjUMurgidFLb7p2BxnSXgcTRDzg5Zyw/+gBXjx8KzjDIyKtRCLrJlW9JV68DrglFlb3iUgeWAg8M5UY7F8hp4FEUfnylEOZKv3EWS0YyXItG1nbYX/sfaDP/ldXxMPlGNod9HULGa3GwEtGq89BRivn5FiUQ6Iv3PXAo6r66cSq7xL1M1whIgcDbUBFY9KJsH+3CgBjHeULw4YicjtwMFHbgFEcOMMP9LemHULNZBz0QwMYclGlZZ+sgwJsD2IRcHAkIOPjUFTieOBNwO9E5MF42QeBG4AbRORhYBB4y1SHDSEILU+MOMonRNc9RMd47HEe2JNCePWlvcP+UE9+sDmuZIGZwYsHlQd2OzgW83Me5GJ5VPVuJvZmLdfzcFIEoeWPkWHDWHDdTfHsQ7F/ArU4KCDwkIEA6Ff7diEehg497IMXOh2c2zvtVwQ0DEFoGae4CbWqfiLhEP9q4AdEQ4cfGXmRgynU4qAGwsswSZvYvyL3YT9D6kFoeTknBhxktDIOxGIlJnKGF5FziXwn/ww4VlXvj7dvA75AZM+VBy6NeySWJQgt+4xrQp0o5u+J140thndgWOqB/9oryxu32r/BDzsQ7l5u8B7wcCyyDkTvHB9zdSoxkTP8w8DriERVkosBVPVFsVv8bSJyjGr5i2AQWr5ZFQuwsUOHefs3xh077LfguXhoa9oh1IV2Bw7S27Q37RBqxos1ggf2y9vP8nY5uE9UYiJneFW9A0rOgj0U+Hm8/WYR2U6U3bqv3OcEoeWICZpPn0B0nO8sLNA9O2YyrGlhr4X2C/qj09s+ux20r/FAxkHtZd5BdhSg08Fu7MrY/z5NhiJn+Il4CHiNiHwdOAB4cfwzCK0mobgIviC4VjPaXBoAmT13pmOrO0P99p8YvXgGzcrYt9rYk7dvteGhRkudnBNb7F+e2M9JgnQqzvBl3u4Gorqt+4EngF8TDT+WJQgtJxQVwS9m1BF+A6NDh8vBR0aro8t+bVNum4PHXnzc4D20EWpxkNHy0uuwzYFIac872Amm7Aw/0XsNA+9JvPbXwGOVYghCyxHFDaYTRfDjG0sbp7073BgbBQ/Fy20SLoUNgX3NDkDOwX5sbHWwExUo4ww/0fadgKjqHhE5BRhW1UcqvS5cXRqX/sRMwp7JvjjR63C8aamDsfe8g4mT4uSu4mHWoQc8CF4v58QsB6dEn/3bRDVM5AzfDvwrsDfwIxF5UFVfBSwCfhL3Plwfv7YiQWg1KMns1GR6Lhb5aq2Kf18NHDiyTbv9WWL9O+zXBbU46XW4y0ExfKsDL7DgDN84DDvQi3Oa4OtUwRn+1hLbrwUOmeznBKFlkKJi956i1QVfrZ8Ba+NlG0gUxOtg33SHOO20dtgXKV6m43swLB10UKPlYXKFl3Oi3/6hoLsJhNZMEYSWTTpUdTmUzXatjn8eR3Fj6WH7M6zaF9i/CvxoAZz5gP38fMbBDd7LkJV1PIhFL2yzf2mqChGZB3wJOAxQ4CKgF/h3YA5RwuKNqrpTRBYA3waOAb6squ+u5jOC0HKKqo5MZxWR24BRh08HRdgeapdffV+LiyYXQw5qtDz0nfQgUXIO6swAhhzsh4fhzyq5DrhdVc+JW+x0AncAl6vqnSJyEXAFcDWRjdLVRKLssGo/wMHtKjDRUGJi+TBJrw8HN0bJNM9VoNHpcKB6PTTGDgTqif1Hj8qISDfwcuBCAFUdBAZF5BDgrnizO4CfAFer6h7gbhF5/mQ+x/4VMgBFQ4nJgviEgemSka0H+lMJsp5Ih/3LwJCDuiCAvVo60w6hZnYN2y/o9+Bn5mHmJPjILnr4PlXBEuAZ4EYROQJYCVxK1OvwLOB7wLlE7u9TJggtnxQK4r8YF8WPPc4O7B0Ytn9B9nIh6w2u6oE64eU4zFL7+zFkfxeAis7wWeAo4BJVvVdErgOuJKrT+qyIXAN8H6jpIheElj/6gaXx74WC+ANJmpbO7p7hkOpPrtf+8GergyE38JGF8DBz0v5R8MP2jP3rk5fvUwVn+HXAOlUt9Df8NnClql4NnAogIgcDZ9QSg48rfWCEuBVPcvjwCKCPRFNpD8Xwefujn27wYFjq5OE90CC4yGilHcAMoKobReQpETlEVf8AnAQ8IiKLVHWziGSAvyeagThlgtCyTzKDlaQwfLiY4oyWhxujfa3oIhMEMCfTnnYINbPbgelqoHHoE/vn9hx1cJGtjkuAm+IZh6uBtwJvFpF3xetvAW4sbCwia4FuoE1EXgucWqkNTxBaxilksIoW9wNL46xWwaj0IyNr8/aFlge8NND1cDn2IHq9mH16wEOtWWuTfJ1U9UHg6KLF18X/Sm3fM9nPCELLIbH4uhJYmph1uKywXtpmTfhaKwzusn97v6UH3vCE/f3Y7aAYvtWBXHQwWhVoIII4qB/h/9IpBbElIreTHDYEdMPqCV5lh9kH2b8xnvQz+wIF4MC2+WmHUDNPDm5LO4SAI3Zh37rl8CH7E0SqIfbM+kZi0RLgGlW9VkQuAd5N5EP5I1V9v4icAnwCaCOajXiFqv6i3GcEoeWEItPSJFmShfAAbfZraoa3BoPJRmFHLsxMCASS7Ksebq3NMXYYF8EfCSAiLcB64FYReSXwGuBwVR0QkUXxS7YAr1bVDSJyGJGZ6X7lPsPDtyEQ12Qlhgk7gMVE9VmrSTSUBqDDvsHk0C774yQZDxX9+DBezTqwd/DQJ9BL3eIWsX9OvNDHoZgsJwGPq+oTIvJJ4BOqOgCgqpvjn79JbP97oENE2gvblSIIrZQpkYnqmex7FBXEF2YbFuqyxjaUBhiyP2TVNt/+01beyU1lQda+cH9qaHvaIdSO/VPCDW0OiuGfzfp4EJwk5wE3x78fDLxMRD5OlMy4XFX/t2j7vwJ+U05kQRBajcBI+xyIWuiU2Ka/aHlPmW16YER8QXSMxx7nFvuHPbSmaxw8ZFICgXoy24E1Qq/9XQAqOsMnt2sjartzVbwoC8wHjgOOAb4pIks0nt4rIi8E/onY2LQc9u+4TYCqfiL5dykxVthGRH6WWL8YuJvoizJaEN9eqpTLFg4murkYrgLYlutLO4Saac3YvxR6sBTwMnTowUdrLyfTWCs4wyc5DXhAVTfFf68DbomF1X0ikgcWAs+IyP7ArcCbVfXxSm9s/+oSKGZdosF0YfjwVeVeYBEP5U0tHnbCCR5Eige8nBMeZh1mm08enM/osCHAd4ETgRVxG542YIuIzAN+BFylqr+q5o2b7n+yCVhVlNF6kqRZKZB57uEzHVPdefjBn6cdQs20tTqoC8LHzbHDgdDyYLrqhb3UfrZ6fUvzfJ9EpBM4BXhnYvENwA0i8jCRjcNbVFVF5N3A84GrReTqeNtTC8XypQhCyxnJYcY4o3UyUVGfKzL274t8ZHge12Tt+zd1SVvaIdTMNrU//BloHIYcXJ+axRkeQFV7gQVFywaBC0ps+w/AP0zm/YPQckhiJuOrgR9QXKOVs98uVBw8vV+d9ZHRGsJHXU0gUC/6HZwTCxxk5RqFILTKUzzbL0nPDMZRkSKbiJ7YU2sxxSILkFldMx1e3elzYAjopRi+zcF+tIn975MHvAx/ehBaz3gYNqgCEbkBOBPYrKqHxcs+RmRWmgc2AxfGBqVLiRpMHwV8SFU/Vc1nhKtLGYpn+yUpI8DSoiNRBF/w0doQ/3uSpGmpg5k9HjJaXhhy8H3ycIMPBf2NwywHvTOHHJwTVfJl4HPAVxLLPqmqVwOIyN8B1wB/DWwF/g547WQ+IAgtn/QDS4FVRFmuJSTqtNSBYensTDDSahQGHTjDe8CDWPRCqwPR2+sgK1cNqnqXiPQULduZ+HM2sR1wXPC+WUTOmMxnBKFlnMSQYU9hWaGhNJHA2kDUEHNJKgFOEy0OfGpi3zvzeBhh8CAVPWS0vPhoefg+ZR18n2ohdoR/M7ADeGUt7xWEln0KLXeWJxcmDEyvBO4hcax1z9YZDXA6aMvav5T1OpiUANDuwOxzT85+lteD0PKSlet24Aw/14fmrdoZvhhV/RDwIRG5Cng38OGpxmD/ChkYQ4neiacBY+auS6Z1RmOaDsRBRqvVQR0HQM5BZs7Dscg4aIWUd/BdAh8i5bEWH+UZk3CGn4ivERmUBqEVGNPrcBWjw4YPkSyEd0L/kP2vrpMOF+Qc1HL4uL3bx8txaHewI13Yn008VUTkBar6WPznWUT31Clj/24VAMYMFS4nymitZnz7neUAmrc/ZNU9q2yzdBN8gQ7+ur8/7TBqZjBv/8m3xUE2SBwMHWZFXNRp7bafICXnRvaWR0RuJrpXLhSRdUSZq9NF5BAie4cniGYcIiLPAe4HuoG8iFwGHFpUPD+OILT8kZxxeA+RtcMSEsfaw9DhcM7+lezdQ71ph1AXOhx8n/odPHz4yQfZx75UbB5U9fwSi6+fYNuNwP6T/YwgtJyRmHF4LpEr/BJgPrBrZKOMfZHiQWh5qKkBGHJg7+ChX6MH1IlYtH9GQKuX2oYGIAgtmyQd63uKV8ZiaylRdutuoubSruq0PGiUfN7HTaUtY7+WY9iBWPRSSO6BAQeTddodDEU3CkFoGaSocfTyCTZbBSyNW/HcDlxMXKPlgZaMg+S8/Xs74GPG3rCDg+EhQ+pFLA7YPxRknRyLahCRFqLaq/WqeqaI/BIo9KpbBNynqq9NbH8MUWnOG1T125XePwgtpxSGEGORdTBw58jKvH2R4uCe4mSQxAceLCocnBJuzolWBzvSZIallwKPEhW5o6ovK6wQke8A30v83QL8E/CTat88CC0HlPDOSlIwKx01LO0vO0HCBN3z+ypv1OhsSjuA+tAss5ManXAUGoc5Duqbns7Yz/JWg4jsD5wBfBx4b9G6LuBE4K2JxZcA3wGOqfYzgtDywUhD6QIJ8bU4lYimmUyL/duK/UtxhIdieC/HItAY9Dqo0fLQrxGqcoa/Fng/o0OFSc4Gfl6wbxCR/eJlJxKEVoDR1jxXjlvTYn86fj5n/yLgoUcg+Jix52W2W6Ax2Cn2yzNmOWgjBOWd4UXkTGCzqq4UkWUlNjkf+FLi72uBD6hqTiZRvxKEln0KvllA6SbTRDMOT5jZsKYXDxmt/9oX3vy0/YuZB4NJD2afgcah3cEEkd0OxGIVHA+cJSKnE903u0Xkq6p6gYgsAI4lymAVOBr4eiyyFhIZmw6r6nfLfUgQWsaJi96XJxaNNJmORdcS4EDgpJEtHGQghgft78Nbn1U8VNZkxb69gwex6AEvmcWsh91ogmcPVb0KuAogzmhdrqoXxKvPBX6oqv2J7Z9X+F1EvhyvLyuyIAgtzyQd4g9MrpBseyoB1ZOhQfs3d3FgKQDQ4uDp3UO/Rg94ySz2OajRsn9W18x5wCcqblUFQWj5YJyBacIhfglRG56PFDb20OuwxcHQoZcsiodaMw+ZFC8eVB7wUkjeTKjqCmBF4u9lFba/sNr3DkLLARMZmCYaTV9J1DQzYte2mQtumsi22c8GDTnJovSpfeE+4GDmZKBxeN6QfaH1VFAHdSP8VzqkyFfrOOBVyfXq4KbS0T2cdgg1k9nkIznvo0bLfjbI/q3djyfbsIOD8bTYf4CqBhGZRzSz8DCiotmLgD7g34nuo8PA36rqfYnXBGf4RqKCmSiU6FVYI/1Ew4UbgNOIvjB3ln2FQfIOrmQZF7dGGMzbF70e8CBRvJwTHlrwzGseeXAdcLuqniMibUAn8E3gI6p6Wzwj8Z+JR4WCM3xjMs5MNEmZXoVTomgYMdlM+riRjdavqedHpsLsQxx8ddf4GDqc2zIr7RBqZkfOfqcBD70OPWQWAb6RfzrtEGrmqsG90w5h2hGRbuDlwIUAqjoIDIqIErfjAeYyeh+F4AzvgyqyYBA5vm+osE1y/T0jv7U4GOrZbj+tvWWoN+0Q6kKLgxv81uE9aYdQMx4K+k+YdWDljQzQif1r7GDOR2lDBZYAzwA3isgRwEqivoeXAT8RkU8RTcB8KQRneG+UzYJBSTF2AtUez/ZKGq7x6d9k/+b+i4Nnc8pj9jMpe/KDaYdQMx6yQXn7Ootf9T3FKbN60g6jZgYcTHTpcpJdrNCCJwscBVyiqveKyHXAlURZrPeo6ndE5PXA9cDJBGf4pmc+8IOiZeMK4QGYM28m4plWPLTgOX1Nv4v2Nd0Z+8J9mwOrDQ9eYEd27MtW7GerZzvIaG3J+JAH5VrwAOuAdap6b/z3t4mE1glEmS2AbzHahic4wzcZY7JecT1W8XBillKF8Dn7xcstrfZvKsMOZn96wUNBv4ehw+fQlnYIdaHfwbHY3WL/YbYSqrpRRJ4SkUNU9Q9EHVQeIRpSfAWRr9aJwGPx9sEZ3hsVarV6kn+o6jtKbL+aSHiNFWBP/qmucaZB2wL7FzIvhqWTSaE3Kh6OhQehtV4H0g6hLsyT1rRDqJljsjvSDmGmuAS4KZ5xuBp4K/A94DoRyRLN5H9HmddXJAitxiPZJHrCWq1EL8Ny4zYFcbUEOHhk6azZtUeZMgPP2B9y8+LkvX3YflG/B5Hi4fvULT5uSfupfaHVN2R/H6pBVR8kGhJMcjfw4gqvu7Daz/DxrXZEiSbRE5H0y4Kx9VhriNrulKbTvtDKOXBe9kJXi/0arZ3D9icleMCLj9YesZ8h3ZML8qBehP9JoyT9spLEWa6TKXNsdeP66Qprxpjz/LQjqJ3uJ+37TwHsyvVX3qjBmdvamXYINeOhIfM9A5Uca2xwSvsBaYdQMwtn+X/4EJEDgK8AzwHyRDMSr4utHv4dmAOsBd6oqjvj4cUvEGXA8sClcY/EsgSh5YTEMOJiorQnTDDrUPZaOIORTQ9Dm+0XkucdDFcBLMzOSTuEmtkyvDvtEAKBhmJgqCnkwTDwPlV9QES6gJUicgfRLMPLVfVOEbkIuAK4GrgYQFVfJCKLgNtE5BjV8kWeTfE/6ZWiGq2e4jHjeP09xa+jrX3aY5tucg56XHiwdgAYcjB70kO/Rg91Zl7w4KOVU/vX2Eqo6tPA0/Hvu0TkUWA/4BDgrnizO4ja7VwNHAr8PN5+s4hsJ8pu3UcZgtBqTPrjOq2eCtuNFMsX6rqKxNfi+OcSksfawwyrvP2LwPeXtPDqx+3bCrQGkdIQqINieC/Yf/SA9qyHvageEekB/hy4F3gYOIto9uG5QGEs+CHgNSLy9XjZi+OfQWhZo1B/Nck+iCPiLJnZioUXwGhvi2H7N/e2+fbF4qvut78P4COjFURKoJ4sUPsPH88M+KghreAMX9hmDlH/wsviWqyLgM+KyDXA94FC+4sbgD8D7geeAH5NNPxYliC0nJAQZ1cmBNpiirNZADn7N0YPePCfAhh0ILQyDoZxPWTlvLBd7J8TczL2HfqhojM8ItJKJLJuUtVb4tesAk6N1x8MnBEvHwbek3jtr4nNTMsRhJYRJvDM6ineLjkbMZHNGjuVZ8E+dY5u5hl41v6NMetDZ9HqQKR4mZgQaAz6HNRo7cn7lwcSPe1eDzyqqp9OLF8U12BlgL8nmoGIiHQCoqp7ROQUYFhVH6n0Of7/J/0wzry0eGixSIydBhTm5x6c3I5W+20uBnvtf3VVfTwx5oJIaQg8eFB5cOgHH70OmyRDejzwJuB3IvJgvOyDwAtE5F3x37cAN8a/LwJ+IiJ5YH382orYv1s1ORPNPIx7Hx5IqabSg/Z9jzzgJYsyoPZr/jygDoSWl3Ni37x9oTVcufTIPKp6N0x44lxXYvu1RDMSJ0UQWo1NfyJr1VNqPbA0Ia4K9VmF3oYbiI7x2OOctZ/R6ttjvz2E4qOvW4eDvm5BLDYGXrIo/WJ/P7oknBP1IgitBqao3mp5qfUTzExM9jbsAnZNR3xp0tZuv9g0t9vHMMmQg8nseQdDVsMORMpzW+enHUJd2OmgBc+gg5mT1SIia4nukzmiuqujReRcYDnRLMNjVfX+eNtWIkPTo4g01FdU9f+We/8gtOzTXzTLcAOjzvDJZSPkV/xsZiKbRha+2P7T1uDt9vcBYG6r/Wngzw7bzy56aCr92MAzaYdQF05on5t2CDXz/P18HItJ8EpV3ZL4+2HgdUQtd5KcC7TH7vCdwCMicnM8rFiSILSMMcHswwKlGoUls1vReyy2P+tw+IFNaYdQM3ce3c4rVw5W3rDB8VCE7aFPYMb+Lrhhtn3Ny46t9h+gakFVH4WSNjwKzBaRLDCLyGNrZ7n3CkLLDv1F9VdJCstOoEmOaabD/l3lpb/sTTuEurCwxX5D5o1DO9IOoWY8ZLS80Gl/5JCtgxM9z7tEgZ+KiAJfKDY0LeLbwGuIWvd0Au9R1a3l3rwpbsqNToUsVTHlslZV1WO1LHtd9cE1KGv+8cm0Q6iZea32xSLAppz9hszzsrPTDqFmPMzYO7bVfrYdYK0D0fs/Tq5Pr6zCGR44XlU3xI2i7xCRVap6F6U5lqiWazEwH/iliPxMVVdPFEMQWo3BOI+sakgItLuLVpXKeo2Std9Uevug/X3IZu0bfYKPhsw5BzdGD0O4O/HhLdflwMR3LyfyoJIzfLzNhvjnZhG5lUhMTSS0/g9wu0ZGiJtF5FdEjaWD0LLAJDNbUNTXMPE+t1FsUppc32L/sO8S+/vgxVJgodgfOtyp9r3lPAitbuxbhQAMOcguzlL7YrEaRGQ2kFHVXfHvpwIfLfOSJ4ETReSrREOHxwHXlvsM+3crX0wqsxX7Zn2R8dmrp+Jlx1HKsDRj/wSa5WA6focDsQiwM29/xp6HY5Fx0DvTfm40woPo3SQ+sotVsA9wa1z0ngW+pqq3i8jZwL8CewM/EpEHVfVVwP8jcop/mMjs9EZV/W25D7B/dWliYh+t4ixYcWH8neNeN2R/ppuHXNCQg35oAEMORG8W+w8fHo7DPCfeTa0OhNZWB6ar1RDXVh1RYvmtwK0llu8msniomiC0jJM0NYUxw4+rmSCrlX/oFzMW33Tx4kOfTjuEmjlo9YK0Q6gL9/Y+kXYINdOXs5+VUwd1ZrO7FqcdQl3otq952ZYN8qBehP/JBmeSdVvJIvjCz3vif6N0zatLbIHa8DBLDKAtY/8y4sFHy0ND5l1OsihDLfa/Tx7qzCohIgcAXwGeA+SJZiRel1h/OfBJYG9V3SIiy4DvAWviTW5R1XL1XEAQWo1AP7A0uWCiRtGVEJHbgYvLbLIcgHb7xcu5AftDPZ9vpImLAAAgAElEQVTabzvvX2+/5Uh7xn4BsweRgoPhz6/1/oGTZi9JO4yaaXNwLA7IN4U8GAbep6oPiEgXsFJE7lDVR2IRdgpR8XuSX6rqmZP5kKb4n2xkJuhXOFIUX1hXZWarkLkaXwCfZOvGyYbZcHQ+N+0Iauea/1nEXvYffF2IlIGc/bpFDxnSv+w6xMV+eKg0G3JwbaqEqj5NZDxKPOvwUWA/4BHgM8D7iTJYNRGEVuNTzhG+9KzCCmQOPqYOYaXL2l/fU3mjBud/hzenHUJdWNxqv6/bnJamcsFuWJ5R+7VyAO1iv31Nh329OylEpAf4c+BeETkLWK+qD5VowfMXIvIQ0f34clX9faX3DkKrwSkudi9mCt5bqIMGusM5+6n5obyHuZPwrNpvJaQOsigeiuEXOxAoAPurg+H0tAOoE1KFM7yIzAG+A1xGNJz4ISI/rWIeAJ6rqrtF5HTgu8ALKsUQhJYxSgirHlW9MLG8kOVaw/ix5Ygh+0Ir46Bodkh9XMrmOKjR6nUw69ADc10MusGzYv/cPjjn41hUcoYXkVYikXWTqt4iIi8CngcUsln7Aw+IyLGqOlJ3o6o/FpF/E5GFqrqlXAxBaNljjKlpbFq6nLGC6x7gtAnfIWP/BOpos58Nyjgpgmhz0IKnz4HZpweTzD4n3nIejsVO+4MGFZFISV0PPKqqnwZQ1d8BixLbrAWOjmcdPgfYpKoqIscSzUB5ttLnBKHVGBTqsHom+8LC0GIJwfUSoHQe3kFGK9tm/4lReu1fjMFHMbyHG6OHInIvQsuDYek8JxmtChwPvAn4nYg8GC/7oKr+eILtzwH+RkSGgT7gPK1izD4IrQYgIZaW1+E9CoLrKeBAShTL67P2zT4XPL8v7RBq5/60A6gPfdo0rToC08x2tT/7E+A4nZN2CDWzsQnUgareDeVVsar2JH7/HPC5yX5OE/xXNgclarc2xP/GT8/r2z1DUU0frYs9zBLbk3YAdaHFgWfQoIumTvbpFvv1fhClOqyz3UGdWaMQhJZRKhTFL2Fsv8Ox5OyfQDKrLe0QaubON7Rx4jd3pB1GAMg6qDPzMHPygcFNHNm2T9ph1MwusT8E2gzO8AAicgjwjcSiJcA1wDwiA/Bn4uUjQ4oicjjwBaCbyFH+GFXtn+gzgtAyRCyieuI/JyyKB1Yx2u9w/PBht3038vw2+5YCr7oj5yIb1CL29wEHdWYVRkBMcFjbQheCcZcDc4QOB9emalDVPwBHAohIC7CeqJn0W4HPqOqnktuLSBb4KvCm2GdrAVC2fiIILVt0AGtLrUjWaAHLKGNkKnPsC63df7R/MR5UH8NVs8R+dnHYgdWGh2L4/bU97RDqwjYHw25tDoT7FDgJeFxVnyhhVFrgVOC3qvoQgKqGWYfNRtzSB6LarFJu8miv/eGq1k77FzIPN3cveBG91ulwcnO3PxAN7erjWEyS84CbE3+/W0TeTDR16X2qug04GFAR+QmwN/B1Vf3ncm8ahJYjEnVbhXqt24i+FGPJ2s9AeBit8jBsCNDi4ObookbLgTP8bge1TYCD3CL0OzCFhuqc4ePt2oCzgKviRZ8HPkZ0OD8G/AtwEZFuOgE4BugFfi4iK1X15xPFEISWLzpUdXmiXmsB0ZdhLP3265uy3fZv7mVS06bwYO/goc4s58CDao+DfQBYqPaF+zYvoreCM3yC04AHVHVT/LpNhRUi8h/AD+M/1wF3FtzgReTHwFFAEFpNQk/R3z8mUuNjZxtn7N9UNG//actDBgKgw0ELnt25CScMBWYQD0afAAMOdsNDzd8kOZ/EsKGI7KuqBdPJs4GH499/ArxfRDqBQeAVwGfKvXEQWr7oiYcPlyasHsYf447OGQ+s3uTsW4GRddAKCXz0bPRyLKyz28FsPYBOtf8w2+4gy1stsWg6BXhnYvE/i8iRRMmKtYV1qrpNRD4N/G+87seq+qNy7x+EVmNRTSueStt0AKvibRYDKykqiJes/Zk9w30eLgI+bipN+OQbmCYWYj87CjDsoL7Jg0VFtahqL1GpTXLZm8ps/1Uii4eqCEJr6hQETyV6qn3DalrxVNhmmFEfrQJLKCqI12H7vQ498P0lLfzln+wfizkt9l36e3P2j4MHwfu9/jX8Zcdz0w6jZhY4qNHyUgzfCAShNUUKgqcStfQvrPL9kw7x64hEVnLocLxDfN5+kWNLu/19eNXvh100M/YwdOhhYkLGwX3xlI4DHchFH8ci10RCKzYqvR9Yr6pnisgvga549SLgPlV9rYi8EfhAvHw38DcFP61yBKFlnxGHeBFZHvtoXSkiX2RUZM0Hdo28ImffMyjj4JvrIQPhhZwDZ3gPMyc91DYBDDgQKfYfPSbFpcCjRC11UNWXFVaIyHeA78V/rgFeEddpnUY0m/Elld7cwe0qUExBbDHahmesaamDWYc5B9N6PNzcARdZOQ9tXzx8n/qcWApkHAjGRdoc8kBE9gfOAD4OvLdoXRdwIlE7HlT114nV9wD7V/MZzfE/2TycEA9VHkfUgmcN8CRFNVoyq2v8K42x51n7pqsdGftDbuDD4b5NwqWwEfAg2gH2ydnfj99l7fvjVcm1wPsZHSpMcjbwc1XdWWLd24DbqvmAcHVpcIpqsHoqbL4uMYx4JXAypY5xi/3Dnmmxn4HwUBcEPlzVPWSDPGTl+p0Ylu5osZ/Rsr8HEeWc4UXkTGCzqq4UkWUlXn4+8KUS7/lKIqF1QjUx2L/j+mdMDVaFbdcmhNliolmIJ43bysNNxf49hYG8jyfGOQ7sQrbm9qQdQs14yAZ5aOcEsNPBEOg+ToYOKzjDHw+cJSKnE903u0Xkq6p6gYgsAI4lymqNICKHE4mv06ppKA1BaFmmlJ9WP+NnHEI0jDiKg1mHbQ6aSrc7cFQH6HfQkLnNw+wKB+QcZOUAcg4EowexWAlVvYq4t2Gc0bpcVS+IV58L/FBVR9pGiMiBwC3Am1T1j9V+Tri6GKWUn1ZixuFyRv207iGq0xotiHdQDD/UZ3+4Sh30CAQYciC0PAzjemjpNNvJgFWb2v8+7XYwc7JGzgOKbZyuITI2/bf4mjGsqkdXeqMgtJyREGBXAkuJxNY401LrZLL2n7Zu2a+N1663b5Q5K2N/YoKHYVwPdiE/HniSU9oPSDuM2rGvs9z0nawWVV0BrEj8vazENm8H3j7Z9w5CyykJi4clFNdqZe3fGAd77X91z9/e58L7qNXBPgw52IesgxvjAa1zWa29aYdRM0cyJ+0QasZDzV+jYP9uFQDGzU5czOhQ4RKi43znyMaD/Vinc95g2iHUTMt2+zd3gCEHkys8CF4PQ4de8CBSmqFGC0BE5hEVtx9G1CT6IqAP+Heie+ow8Leqep+ILAVuBI4CPqSqn6rmM4LQss9IUbyqXgjjRNd4w9I2+73phvrt12iJ2C/oBx9DVh5ujDioM+sSHxNENmO/btFLvVwVXAfcrqrniEgb0Al8E/iIqt4Wz0j8Z2AZsBX4O+C1k/mAILSMk6zJShTGH0dUk/Vk/PfY+qwW+xezx9fvlXYINdPVVsoDzx4bh3akHULNBC+wxmAZ9h8CARw0ruAXLbvTDmHaEZFu4OXAhQCqOggMiogSt+MB5hInK1R1M7BZRM6YzOcEoWWMomzVRNxDdGxLH18HF+S9OuwPf24dtO/dBNDhoBjeg9mnh6zceid3pE77Xyfmif3zugqWAM8AN4rIEcBKor6HlwE/EZFPEXm3vrSWD3HytXZHfyI71VO0bsTAtBQJIXZ3vKjQjmeUIfsz3Tww5KB1DYD9/KiP4U8PGa29fJwSDNnXvA5ke0Q5Z3giDXQUcImq3isi1wFXEmWx3qOq3xGR1wPXE3VamRJBaDUgheFAqMoNfoTELMMNjBbEj89uOWjB095mvwaibcj+cfCCB0dyDwX9u+3vAgB7HHhQeTGPreAMv46odd298d/fJhJaJxBltgC+RYk2PJMhXOlt0U/kjTXREGJVBfG6x35NTdde9ocOdbePC1mgMfBguurlhtTmQLjvcZJxL4eqbhSRp0TkEFX9A5EN0iNECYtXEPlqnQg8VsvnePleNwWxN9by+M9xQ4gisjwhsMYPGcJynOBh1uG3FrXw6o270g6jZrqys9IOoWZ2DfelHULNZBwIresHH+f89oPSDqNmuuyP4rK4xX4P0yq5BLgpnnG4Gngr8D3gOhHJEiU43gEgIs8B7icqlM+LyGXAoapadmZTEFr2KNXjcGQdo0OH80l6ZyVx0IJH8/ZvKn/1zB5aHcx28+Cq7uE4eMhovaW1Bw8jVh5mHe7Cf0YLQFUfBIrb6NwNvLjEthuB/Sf7GUFoGaNUj8PidfH6pGnpmOyWdMye3iBngGeete+83J/fnnYIdeE5bXPTDqFmNg/Zt9oQB/31djqobQKY5+BB8HnaFLMOZ4QgtOzSX0JsJcVV0qT0nvgfANpv31agvdV+MbwO+7ip9DnIaOUduKpn7N/bme9AoADsyNj/Pi1wciwagSC0jJLMXhWoUJ8FhRqt9WumM7QZYf4C+/3QetfaL+gH2Lu1u/JGDU5vzv6xcNGCx0kSZYn9Zw922q8wqQoROQT4RmLREuAaVb02Xn858Elgb1XdMpXPCELLABPMMJxITFUk/8ijNceUNgvPsO8Mn/38s2mHUBf68vb7TmbF/qVQHQy72T8KETkHsw6fzDqo6K+CeLbhkQAi0gKsB26N/z4AOIXRLitTwsv32jsjMwwTomvMcCBjhw1PoNyx7bQ/S4wh+0OH7RkPVp8+jFc7HLSl8jD8udu+PgGgz8GdtTlk1jhOAh5X1Sfivz8DvJ9oFuKUcfB1cM+Id1ZMSWf4Mr5Z4zJf0mr/pjK8dmPaIdSMh+EqgLlZ+9nFTQPb0g6hZjy42/c5yMoBLM7ZV4zzcz7GDis4wxdzHnBz/LqzgPWq+lCtM3qD0GpwiryzxlEksApZrQkL4QFk74V1jnLmGf5lTf5xgTriwZFcHOyDfYMKaHUw5AY+skED9k8JoKIz/Aixj9ZZwFUi0gl8CDi1HjEEoWWf4mHFZUxcu7UcgKz9jJaDsiBWHjufl95vfwbo1qHdaYdQM20Z+5dCD0OHN/Su4oNtSytv2OB48NHa7GDm5CQ5DXhAVTeJyIuA5wGFbNb+wAMicmzspTUp7F9dmoOSTaZjYdWT+L24dmvKBfONTssc+1eyl6ywL1AA5mSL52nYw4MzvAfeMWspU5rW1WDMdZDSGnAyjDsJziceNlTV3wGLCitEZC1wdJh16JgyTaY7gFUFp3hVvbAKiwd0u/16lJZ59ueBDznwnwLIOnBVH8rbL+j3wJD95ycAWh1olPnqZOywCuKhwlOAd07H+wehZZyEU/yVseAq1Gk9CXyk5IscFMNnuuxnUdpb7ItF8OHfNMvBsVAHxfAbxf5sYgCy9m+tnfa/TlWjqr3AgjLre2p5f/vfhuaj5DBiCZYAB5daIR32RcrwJvvDbuKk8He3g9mTHkSKhxqtA/I+bkm9DobdMg5aOjUKPr7VTUTRMOLPkqIrHjq8jUhgdQG7UghxRsjMsZ+V8zBbzwsejkXWQVPpZzMOipuAdgciZdj+LlRNbFR6P5Gdw5kiciLwKaJeBSuBt6nqsIi8EfhA/LLdwN+o6kOV3j8ILdtkkzMOY9G1CVgFnMxEx7fTflPp/E770w49ZFEAurOdaYdQMx6K4dWBqYB9uesHL/VyVXIp8CjQLZHXy38CJ6nqH0Xko8BbgOuBNcArVHWbiJxGZBvxkkpvHoSWE5K1WkTDhqvjn+OPcd7+BdnBPcUNA06K+gPp4+W0nusgo+Vh+LMaRGR/4Azg48B7iWq1BlT1j/EmdwBXAder6q8TL72HyPahIkFo2WZt8YIiwQVjzUsjMvafG4e22r8ke5np1uagT+CwgzZCHmq0npO3f20CH4LRw/BnlVxL1GanK/57C9AqIker6v3AOcABJV73NuC2aj7A/hWyuVkLE7rDE/9M2jwsB2DWnJmKb9oY2GnfUuAXS7o5eY39MjoP9U0eJiZk7O8Cn+//A+/sOCTtMGpmjgOl9XSLfeEO5VvwiMiZwGZVXSkiywBUVUXkPOAzItIO/BQYLnrPVxIJrROqiSEILR+Uajo9MQ4yWh172c9AnPSbIReNpWvtA9YIdDg4Dh56Hb6+/Xn0OtgPD6p3vdivg4WKLXiOB84SkdOJ7pvdIvJVVb0AeBmAiJxKYga/iBwOfAk4TVWfrSaGILRsU7B66JnUq/bsnI5YZpTWRfa/ujkdSDuEutAh9kXKTuz7N+XUfhplt5O6oIyDDGlXE8gDVb2KqP6KOKN1uapeICKLVHVznNH6AFH9FiJyIHAL8KZEDVdF/P9POqaEWSlEomsVo1mtcU2lPZDvtZ/R8jLrsF/tF8N7ECkevk8LndRoeSDvRPROkSviYcUM8HlV/UW8/BqiYvl/izP5w6p6dKU3C0LLAUXeWlcCS4nE1jKi+qw1RE7xEW0ODEu3278IeKgLAh8ixQMevk+D9ncBgC4Hp0R7k5ltqOoKYEX8+xXAFSW2eTvw9sm+dxBazlDVTyQsHgpteE7A2bFusV/P76KI3AvhWDQGe8SBQgE6HHyfBlzMnWwMXN18m5Eyxe8biATWuDY8u677wXSHNe3MOffItEOomV2/vCvtEOqChz6BfTkfhb/WaXNiKdDpQKOcPWi/brFaRGQtUSeVHEXDgSJyOfBJYG9V3TIVd/ggtBqMqmYNjiVp5wBVZK/mvO5FU4issRj61cNphxCI6WxpTzuEmunP2a8z81Cj1eFg+BPgqax9pbX/kP1JLpPklaq6JblARA4ATiFZejMFd/ggtGaYKoRUj6peOInXFxuSro6XFQuwEfKPP1lqsSmyBy1KO4SaEbalHUJdGMzbf/LNOLCoyNvXWexwMnToweyz24m9Q418hsjM9HuFBVNxhw9Ca+YZ8bwqRdEMwlKME2JF4qsgspZQYtjQCzJ7Vtoh1MzvPnAkh/3Tb9IOo2bmttg/Fh56HXrg5t2Pcs6cpWmHUTML1X6N1rqM/Uz1JFDgpyKiwBdU9YsichZRk+mHyngFVuUOH4RWg5GcQViKgggrElcj4itRCD/hsc3vse/fJA7qBw799INph1AXcg6GrNTBzEkPhqV/1XUoww72o93+LrCPg+F0KO8Mn+B4Vd0gIouAO0RkFfAh4NQy71u1O3wQWtNPf1GGqqdO75cUV7eJyIp4fRdRUd+EZA87qMYQ0ie/Zl3aIdRMe8Z+ETnAoNoXve0OCvo99Dqcr/ZbawFsaLEv3Ftb7Z8TUNEZvrDNhvjnZhG5FXgF8DygkM3aH3hARI5V1Y2TdYcPQmuaKc5QVRgWrPr9ioYYn2K0HmvC2qyR93jWfm2QzOuqvFGD46EuCMKMvUbBw/dpwIlJ5mwHQ4cesnLVICKzgYyq7op/PxX4qKouSmyzFjg6nnU4aXf4ILSMMtEQYzx0uIzRRtJJlgPkt9pvZJxd9tK0Q6iZRa0b0w6hLni4wXvIBnngrqFNaYdQF17cunfaIdTM7lb7YrFK9gFujTNXWeBrqnp7me0n7Q4fhFYKTMHCYSKOo7SgKkvu2d46fHS6ZB78bdoh1Mz2nP3jEAgExnNOzv5wujjJLlZCVVcDR1TYpifx+6Td4YPQmnn6iYrVyw7vVaAwPFiqj2FlF/hh+yeQ7Gv/iRGeSDuAurCotTvtEGrmmSH7WV4PPlpeeKzFfpuzA4dDSUC9EA0p88A0ICLvKDGzwxRhHxoHD/sR9qFx8LAfYR/s0DSDsIEZ5x2VN2l4wj40Dh72I+xD4+BhP8I+GCEIrUAgEAgEAoFpIgitQCAQCAQCgWkiCK3AdOFh3D3sQ+PgYT/CPjQOHvYj7IMRQjF8IBAIBAKBwDQRMlqBQCAQCAQC00QQWoFAIBAIBALTRBBagUAgEAgEAtNEEFqBuiEizxWRk+PfZ4mI/c7PRhGRFhFZLCIHFv6lHVMg0AiIyHwROTztOALNQ2jBE6gLInIxkfncXsBBwP7AvwMnpRnXZBGRg4HPA/uo6mHxBfksVf2HlEOrGhG5BPgwsAnIx4sVMHVzEZFzgdtVdZeI/D1wFPAPqvpAyqFVTfx9ugJ4LonrraqemFpQk0REOoH3AQeq6sUi8gLgEFX9YcqhVY2IrADOIjoGDwLPiMidqvreVAOrEhEpG6eqfnqmYqkFEdkH+EdgsaqeJiKHAn+hqtenHNq0EjJagXrxLuB4YCeAqj4GLEo1oqnxH8BVwBCAqv4WOC/ViCbPpUQ3wheq6ovif6ZEVszVscg6gah5+n8SiWBLfAt4APh7IsFV+GeJG4EB4C/iv9cBZh48Yuaq6k7gdcCNqvpi4OSUY5oMXRX+WeHLwE+I+vUC/BG4LLVoZoiQ0QrUiwFVHRQRAEQkCya73Haq6n2F/YgZTiuYKfIUsCPtIOpALv55BvB5Vf2eiCxPMZ6pMKyq1sRhMQep6htE5HwAVe2TohPEAFkR2Rd4PfChtIOZLKr6kbRjqBMLVfWbInIVgKoOi0iu0ousE4RWoF7cKSIfBGaJyCnA3wI/SDmmqbBFRA4iFokicg7wdLohTZrVwAoR+RFRJgKwM7yQYL2IfIEo8/BPItKOvSz8D0Tkb4FbGXsstqYX0qQZFJFZjJ4TB5HYFyN8lCiTcreq/q+ILAEeSzmmqhGRz5Zbr6p/N1Ox1MgeEVnA6HfpOHw8FJYlGJYG6oKIZIC3AacCQnRR+5Ia+4LFF+AvAi8FtgFrgDeq6hOpBjYJROTDpZZbeyqOa4P+Evidqj4WZyRepKo/TTm0qhGRNSUWq6oumfFgpkj84PT3wKHAT4lKBC5U1RVpxtVMiMgg8DDwTWAD0TV2BFX9zzTimiwichTwr8BhRPuzN3BOXKLhliC0AnVHRPYC9rd28sRi8Zw4tT0byKjqrrTjambi+qwXqOqNIrI3MEdVS4mXwDQSZyGOI7rB36OqW1IOaVKIyD8T1ZX1AbcDRwCXqepXUw2sSuL//3OBNxCVMnwD+I6qbks1sCkQl5UcQvRd+oOqDqUc0rQThFagLpSa1QOYmdVTQETuUtWXpx3HVBCRa1X1MhH5ASXq41T1rBTCmjJxZu5oosL+g0VkMfAtVT0+5dCqRkRagb8BCt+pFcAXrN1c4tm3PYydOXlLagFNEhF5UFWPFJGzgdcC7wH+W1WPSDm0SSMi+wHnA+8FPqCq/5VySFUjIq8rsXgHUdZ680zHM1OEGq1AvZirqjtF5O1Es3o+LCKmMloxd4jI5URPjHsKC43U1BQuuJ9KNYr6cTbw50Sz9lDVDQa92T4PtAL/Fv/9pnjZ21OLaJKIyA1E1iC/Z6xdiBmhRXQMAE4HblbVrfbq+UeG3s4HTgFuA1amG9GkeRvR7NX/jv9eBtwDHCwiH7UkGidDEFqBemF6Vk+Ci+Kf70osU6Dha2pUdWX88860Y6kTg6qqIlIonJ2ddkBT4JiirMkvROSh1KKZGsep6qFpB1EjPxCRVURDh38bD0P3pxxT1YjIR4AzgUeBrwNXqaq12dAQCfU/U9VNMOKr9XngJcBdjD4suiIIrUC9KMzq+ZXFWT0FVPV5accwVUTkd5Sx1DDopfXNeNbhvNgQ9yIinzNL5ETkIFV9HEYmW1ibzv4/InKoqj6SdiBTRVWvFJF/Anaqak5EeoHXpB3XJLiaaDbxEfG/f4wzckI0ucLKud1TEFkxm4GD4wyjqeH0yRBqtAKBBCLy5lLLVfUrMx3LZBGR55Zbb2nmZIF4xtvITFZVvSPlkCaFiJxEZPi5mmgfngu8VVX/u+wLGwgReTmRVctGIlsHazf3wgzW9xK527/Dmru9l3NbRP4NOJDIyBfgr4gMcK8Afqiqr0wrtukkCK1AXRCR/Ymm7R5PlFW5G7hUVdelGtgkEZF/TfzZQdRC6AFVPSelkALGif2/CrOsVqmqKQ8qEfkTkUj5HaM1WmZu7gAi8g2ieqY3x621ZgH/o6pHphzalBGRhcCzlix0YqPb1wEnxIueBfZV1XdN/Cr7hKHDQL24Efga0RRkgAviZaekFtEUUNVLkn+LyFyM1Q2IyC5GhxDbiAqB96hqd3pRVY+I3K2qJxTtB4xmUhp+P0TkRFX9RYlZVgeJiKkZe8CTqvr9tIOoEdPu9rGx5yeArcDHiK5JC4GMiLxZVW9PM75qiWsuHyeqyXo9kU/hd9KNavoJQitQL/ZW1RsTf39ZRDz0sOoFXpB2EJNBVcfMzBOR1wLHphTOpFHVE+Kf1mYYJnkF8Avg1SXWWZuxt0pEvkY0fJh0t7e0D9bd7T8HfBCYS/S9Ok1V7xGRpcDNRN5gDUvcXP08ohmTzxLN6havQ4XFBKEVqBdbROQCopMeRk8oUxR5UGWI3LC/mV5EtaOq3xWRK9OOY7LET/G/L5jGisgc4IWqem+6kVVGVQvu/B8tNlgVEWsTLmYRiZJTE8usicUPE4mRA0TkJmJ3+1QjmhzZQkeE2AbhHgBVXWUkMbcK+CXwalX9E4CIvCfdkGaOILQC9eIioqeuzxBdhH/NqFWCJZIeVMPAEwbrzJLDVRki008zdRwJPg8clfi7t8SyRuc7jI/328CLU4hlSqjqW9OOoVZU9Q4ReYBRd/tLjbnb5xO/9xWts3Bu/xVRRuu/ReR2IosKEwqxHgShFagLqvokkTO8de4H+lQ1H6e7jxKRTcacvJPDVcPAWmxNZS8gyULf+JiYuGbFQzovBOYWCd9uokkWZvAy0YXo/30b0X3v0LhW7q6UY6qWI0RkJ5E4mRX/Tvx3w3+fVPVW4Nb/3969B+lZ1mcc/14bDgkUJEKCRxSQgwE5NgNBqwKiQ8dTCwoCioeqiApoPQzoKBYFhop1RKQekEHagREHUWq1YWKFIofUcEZAK4eCRQVBiECEwNU/7mez766bw0XWXeUAABMUSURBVGb3ee88+16fmZ3N/bzJzLWTbN5778Pv19TCG67Mv6Wks4Dvdql/6drIrcOYEpLOpfzn+4dmPBs43XanVrUkLQH+CphNqVj8M+Ax24dXDTaAJF1EaVlzVvPoaGBf22+sFmoNSXoD5Q3l9UDvQfKlwAW2r6wSbC1IupRy0WX4UsgRlEbrnbno0tTQOoQx1e271pZqOml64r4JOMT2frXztCkTrZgSkq6zvfvqnq3rJF1rew9JHwRm2T6ta19H1xvoDpM0F/gSsB9lJWUR5evoTE80SQtsX1U7x2QM9wlc3bN1maTbgV26Vlojpoeh2gFi2hhqVrGAFT+tdGKbZwxJWgAcDvygeda1r+PVth+htOy4F9ieUhCwU2z/zvahtufa3tL2YV2aZDWOkrTZ8EDS7KZ3YJc8IOkISTOajyPo3kWXOxjpdxjRV117A4l11+nAlZK+04zfBHyuYp61dRxwPOXcwC1Ny5TOVPFuTJcGunOAdwMvpOf/qo5tR+8yvJ0OYPshSZ1ZHW1Mh4sujwHXS1rE6BIVx9SLFIMiW4cxZSTNo2zzCFjU5d5oAJKGgL9oVoc6Q9KplPNBj1PqZ21GaW+xV9VgEyTpSsqV8CX09Ae03ZkCh00D6VfafqgZPxO4zPZL6iYbLJKOHO+57XP7nSUGTyZaMSUkbTXe8+Y2Ymc0hRmPoryxL6EUCPyC7X+sGmyCmm3c4Qa6GwGb2v5N7VwT0bVzQONpemceTynpAM1Kr+3OdBuYDhddJO1pe8mYZ6+zfUmtTDE4MtGKKSHpJkbqucwCtgZut71TvVQTN/zmLulwSq2jjwNLutRAF0DSPvz5lts63xi7l6TPAlfa/vfaWSZD0k7AvnR0pXc6XHRpamgdafumZvwWysWKTq3yRjfljFZMibFbIZL2AN5bKc5krC9pfcrW25dtPympUz+NSDoP2Ba4npEtNwOdmmgBxwInSPoT8CQd6nU4xm2M1G9C0lYdW+kdkjR7zPZn1947Dga+0/wA9TLgbYyudB/Rmq59s0RH2L5W0vzaOdbCVykFPm8ALpf0AqBTZ7QoleDnuePL1R3vdQhAUybk08BvKZNeUSa9XVoh7b3oYkoz4JPrRpoY23dIOhS4GLiHcjN3bIX1iFZk6zCmhKQP9wyHKG1HNrf9mkqRpoyk9Wwvr51jTUm6EDjG9n21s0xWcx5oO3qqX3eomjeS/gfYy3bXyiGM0tWLLmOONADMBR6muXnYtSMB0U1Z0Yqp0rv6sJxSg6ozt8OGSdqS8tP6c2wf2LzBLADOrptsQrYAfi5pMaOvsneqCrakv6NsHz6Psg26N3AV5Q2/K+6hvLF3lqTzbL8V+Pk4z9Z1r60dICIrWhE9JP0QOAf4hO1dm95613XpOr6kV4z33PZl/c4yGc1qxHzg6uaCwo7AZ2wfUjnaGpN0NrAD5QeP3knvF6qFmqDhbgk94xnATbbnVYw1IZL2Bm6xvbQZb0LZXr+mbrIYBFnRikmRdAmr6B7ftVUUYAvb35Z0PIDt5ZKeWt0fWpd0bUK1CstsL5OEpA1t3yZph9qhJuh/m48Nmo/OaL4HTmCkifFw1dsngK9VC7Z2zqIcZxj26DjPIlqRiVZM1ufHeTY88epeOXJ4VNLmNF9D85NwJ7Z+JC1l/ElvV2/r3du0r7kYuFTSQ8D/Vc40IbY/UzvD2rJ9CnCKpFNsH187zySp93KI7aeb1eqI1mXrMCZF0huA59k+sxkvBuZQ3vA/bvvCmvkmqilLcQawM3Az5Ws52PaNVYMNuGY79BnAj2w/UTvPmpL0n4wz+bXdmXNmkl4+3vOOXUq4CPgJZRUL4GhgX9tvrBYqBkYmWjEpkn4KHGr7nmZ8PbA/sDFwju39a+abiKblzt7AYsq5GlGKrj5ZNdgAa84DbcnowqudqUElac+e4UzgIGC57Y9VijRhzfGAYTMpbZ2WdGyyOBf4EuUihYFFlIKlXWtSHh2UpdOYrA2GJ1mNK5qr7L+XtHGtUGuj2U443fYC4JbaeQbdmBpUTzePO1WDamzbF+Cnkjp1hs7263rHkp4PnFYpzlppJlSH1s4RgykTrZis2b0D2x/oGc7pc5apsFDSQcBFXS/4OQ0cC+zQ5RpUTRX1YUOUtk7PqhRnqtxL2Vpf50n6mO3TJJ3B+Fu4x1SIFQMmE62YrGskvdv213sfSnovZQuuaz5M2fZcLmkZ3T1IPh10vgYVpTG5Kf+OlgN3Au+qmmiCxkxShoDdKZ0TuuDW5vPPqqaIgZYzWjEpzdmHiyk1gq5tHu8JbAi80fZva2WLbpsONaimA0nvA2ZQJlsPA3fa/mndVBHdkRWtmJTm7MM+kvYDdmoe/8D2jyvGmrBmwngC8CLgRuBU213rcTjddLkG1cm2T2h+fYDtS2tnmqim/MHJwDspfw8Cng98U9LiLl0SkbQ98BHghYy+WNGZA/3RXVnRigAk/YiyzXM5pW3HJrbfXjVUdFZvNfWxldW7QtI/UVprfainovqmlNp5j9s+tma+iZB0A/DPlO/xFQWIx7msEDHlMtGKoJSlsL1bz7iTb47TyUq6DjxMOW/zVdvL+p9qzUyTidYvge3HXgppSm7cZnu7OskmTtIS23uu/ndGTL1sHUYUkjSbkWr2M3rHth+slmxw3UG5uXp+Mz6EUuphe+DrwLrc1HiupA9T/v0M/3qFjpwz83g3b20/JakTP6H33Pq8RNLRwHcZfd4v39fRuqxoRQCS7qLUahqvbZBtb9PfRCHpctsvH++ZpFts77SyP1ubpE+v6vUutOaRdDGlzMm3xjw/AnhzF/qYSrqTkVufY+X7OvoiE62IWCdJuhV4zXAleElbUVrwzJN0ne3d6yac3iQ9F7gIeJyRMhXzgVnA39j+dcV4a0TSAttX1c4Rgy1bhxGs6HG4UravXdXr0Yq/B66Q9CvKisTWwNFNx4FzqyZbQ81tt7OALW3vLGkX4PW2P1s52mo1E6m9em4UC/ih7UV1k03ImUDnzsfF9JIVrQhWNP+F0svtLykFGUVp93KN7ZfVyjbIJG0I7Ej5u7htXT4AP56m3c5HKYf3d2+e3Wy7E5XVuy4rn7EuyIpWBGB7XwBJFwDvsX1TM96ZUn8n+kzSRpRK/S+w/W5J20nawfa/1c42ARvZXiyNOiK0vFaYAbS1pO+v7MUunDOL7stEK2K0HYcnWQC2b5a026r+QLTmHMrZoAXN+F7gQqBLE60HJG1LU6ZC0sHAfXUjDZT7gdNrh4jBlolWxGi3SvoG8C+UN8cjGOmXFv21re1DJL0FwPbjGrM01AHvB74G7Cjp15Reh0fUjTRQltq+rHaIGGyZaEWM9g7gfcBw1evLKYeZo/+ekDSLkdWgbempgdQFtu8AXtUc4B8arrAefXNX7QAROQwfMYakDSjNjA3c3qWebtOJpAOATwLzgIXAS4G32/5JzVwTIWlLSr/A59g+UNI8YIHtsytHGziS9uHPex1+a6V/IGKKZKIV0UPSKymlA+5ipInukbYvrxhr4DRbhM8DHgP2pvxdXG37garBJkjSDylnzT5he9emUfN1tl9SOdpAkXQesC1wPSO9Dm37mHqpYlBkohXRQ9IS4DDbtzfj7YHz0yet/6ZDfzpJ/217fm+ZgbF9NaN9TfHbeeO1FIpo21DtABHrmPWHJ1kAtn8BrF8xzyC7WtL82iEm6VFJmzNyzmxvSmPs6K+bgWfVDhGDKStaET0kfZPypnhe8+hwYD3b76iXajBJ+jnlrNxdwKOU7UPb3qVmroloOg6cAexMebOfAxxs+8aqwQZMU5B4N2Axo5tKp45WtC4TrYgeTSXy9wMvo7yxXw58xXanbrtNB5JeMN5z23f3O8vakDREOV+2mDJhFLlcUYWkV4z3PKUfoh8y0YoYI7cO65I0EzgKeBFwE3C27U5WU5d0le0Fq/+dETFd5YxWRI/m1uEvgS8DXwF+IenlVUMNnnMp/SZvAg6k25W9F0o6qIOFVqcFSVc0n5dKeqTnY6mkR2rni8GQFa2IHrl1WJ+km4bLHzTlEBbb3qNyrLUiaSmwMaW/4TJGzpltWjVYRPRNVrQiRsutw/pWbNV2dctwmO1NbA/Z3sD2ps04k6w+k/SucZ6dWiNLDJ604IkY7WeSzmb0rcMlFfMMol17tnUEzGrGnVsNam4djvUwcHfXJ5Edc7CkZbb/FUDSV4CZlTPFgMjWYUSP3DqMqSTpamAPynkzgJcANwCbA0fZXlgr2yBpemZ+H/gm5dzfg7aPq5sqBkUmWhERLZF0AXCS7Vua8Tzgo8BJwEWpEN8uSc/sGW4CfA+4AvgUgO0Ha+SKwZKJVgTlADZN9e7xdKlIZqw7xmu3M/wsrXjaJ+lOyve1xnwGwPY2laLFAMkZrYjitbUDxLR0u6SzgAua8SGUkiEb0nPoP1pzCHCP7fsAJB0JHETpNnBivVgxSLKiFbESkrYAfp9GtLG2mrNBRzNy5u8KSn22ZcBGtv9YMd60J+la4FW2H2zq4V0AfJDSjufFtg+uGjAGQiZaEaxo9nsq8CDl/Mx5wBaUEihvs/2jivEiYi1IusH2rs2vzwTut31iM87WbfRFtg4jii8DJwDPAH4MHGj7akk7AucDmWjFGpP0bdtvXtnZv5z565sZktZrSmnsD7yn57W8/0Vf5B9aRLHe8FV7Sf9g+2oA27ele0qshWObzzn7V9f5wGWSHgAeB/4LQNKLKPXMIlqXiVZE8XTPrx8f81r212NChg9f2767dpZBZvtzkhYBzwYW9py3HKKc1YpoXc5oRQCSngIepalEDjw2/BIw03ba8MQaa3ocrqpcSGeq20fE5GRFKwKwPaN2hpg+bG8CZRsa+A3lcoUoLZ02qRgtIvosK1oRES2RdI3tvVb3LCKmr6HaASIiprGnJB0uaYakIUmHA0/VDhUR/ZOJVkREew4D3gz8tvl4U/MsIgZEtg4jIiIiWpIVrYiIlkjaXtIiSTc3410kfbJ2rojon0y0IiLa83XgeJoG0rZvBA6tmigi+ioTrYiI9mxke/GYZ8urJImIKjLRiohozwOStqUpXirpYOC+upEiop9yGD4ioiWStgG+BuwDPATcCRye1jwRgyMTrYiIlknaGBiyvbR2lojor2wdRkRMMUl7SbpB0h8lXQVslUlWxGDKRCsiYuqdCXwE2Bz4AvDFunEiopZMtCIipt6Q7Utt/8n2hcCc2oEioo71ageIiJiGNpP0tysb276oQqaIqCCH4SMippikc1bxsm2/s29hIqKqTLQiIiIiWpIzWhERLZF0rKRNVXxD0rWSXl07V0T0TyZaERHteaftR4BXA3OBdwCn1o0UEf2UiVZERHvUfP5r4BzbN/Q8i4gBkIlWRER7lkhaSJlo/YekTYCnK2eKiD7KYfiIiJZIGgJ2A+6w/QdJmwPPtX1j5WgR0SdZ0YqIaI+BecAxzXhjYGa9OBHRb1nRiohoiaSzKFuF+9l+saTZwELb8ytHi4g+SWX4iIj27GV7D0nXAdh+SNIGtUNFRP9k6zAioj1PSppB2UJE0hxyGD5ioGSiFRHRni8B3wXmSvoccAVwSt1IEdFPOaMVEdEiSTsC+1PqZy2yfWvlSBHRR5loRUS0RNJ5tt+6umcRMX1l6zAioj079Q6a81p7VsoSERVkohURMcUkHS9pKbCLpEckLW3GvwO+VzleRPRRtg4jIloi6RTbx9fOERH1ZKIVEdGSpgXPYcDWtk+S9Hzg2bYXV44WEX2SiVZEREtSGT4iUhk+IqI9qQwfMeByGD4ioj2pDB8x4DLRiohoz3Bl+C17KsOfXDdSRPRTzmhFRLSopzI8wI9TGT5isOSMVkREuzYChrcPZ1XOEhF9lq3DiIiWSPoUcC7wTGAL4BxJn6ybKiL6KVuHEREtkXQrsLvtZc14FnCt7RfXTRYR/ZIVrYiI9twFzOwZbwj8qk6UiKghZ7QiIqaYpDMoZ7L+BNwi6dJmfADl5mFEDIhsHUZETDFJR67qddvn9itLRNSViVZERERES7J1GBHREknbAacA8+g5q2V7m2qhIqKvchg+IqI95wBnAcuBfYFvAedVTRQRfZWJVkREe2bZXkQ5pnG37ROB/Spniog+ytZhRER7lkkaAn4p6QPAr4G5lTNFRB/lMHxEREskzQduBTYDTgKeAZxm++qqwSKibzLRioiIiGhJtg4jIqaYpC/aPk7SJZRCpaPYfn2FWBFRQSZaERFTb/hm4eerpoiI6rJ1GBHRIklzAGzfXztLRPRfyjtEREwxFSdKegC4DfiFpPslfap2tojor0y0IiKm3nHAS4H5tje3PRvYC3ippA/VjRYR/ZStw4iIKSbpOuAA2w+MeT4HWGh79zrJIqLfsqIVETH11h87yYIV57TWr5AnIirJRCsiYuo9sZavRcQ0k63DiIgpJukp4NHxXgJm2s6qVsSAyEQrIiIioiXZOoyIiIhoSSZaERERES3JRCsiIiKiJZloRURERLQkE62IiIiIlvw/P5gTQj4Uz30AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x720 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sb.clustermap(df_dia,z_score=0)\n",
    "# 각 요소별 연관성을 찾기 위해 clustermap을 작성함"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(12,input_dim=8,activation='relu'))\n",
    "model.add(Dense(8,activation='relu'))\n",
    "model.add(Dense(1,activation='sigmoid'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "seed = 0\n",
    "np.random.seed(seed)\n",
    "tf.set_random_seed(seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "col = ['Pregnancies','Glucose','BloodPressure','SkinThickness','Insulin','BMI','DiabetesPedigreeFunction','Age']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df_dia[col].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = df_dia['Outcome'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "768/768 [==============================] - 1s 661us/step - loss: 0.5203 - acc: 0.7409\n",
      "Epoch 2/200\n",
      "768/768 [==============================] - 0s 245us/step - loss: 0.4713 - acc: 0.7656\n",
      "Epoch 3/200\n",
      "768/768 [==============================] - 0s 232us/step - loss: 0.4761 - acc: 0.7526\n",
      "Epoch 4/200\n",
      "768/768 [==============================] - 0s 221us/step - loss: 0.4959 - acc: 0.7500\n",
      "Epoch 5/200\n",
      "768/768 [==============================] - 0s 225us/step - loss: 0.4665 - acc: 0.7891\n",
      "Epoch 6/200\n",
      "768/768 [==============================] - 0s 234us/step - loss: 0.5063 - acc: 0.7656\n",
      "Epoch 7/200\n",
      "768/768 [==============================] - 0s 240us/step - loss: 0.4905 - acc: 0.7669\n",
      "Epoch 8/200\n",
      "768/768 [==============================] - 0s 237us/step - loss: 0.4633 - acc: 0.7721\n",
      "Epoch 9/200\n",
      "768/768 [==============================] - 0s 254us/step - loss: 0.4687 - acc: 0.7852\n",
      "Epoch 10/200\n",
      "768/768 [==============================] - 0s 339us/step - loss: 0.4972 - acc: 0.7747\n",
      "Epoch 11/200\n",
      "768/768 [==============================] - 0s 258us/step - loss: 0.4826 - acc: 0.7617\n",
      "Epoch 12/200\n",
      "768/768 [==============================] - 0s 236us/step - loss: 0.4691 - acc: 0.7760\n",
      "Epoch 13/200\n",
      "768/768 [==============================] - 0s 239us/step - loss: 0.4730 - acc: 0.7682\n",
      "Epoch 14/200\n",
      "768/768 [==============================] - 0s 253us/step - loss: 0.5048 - acc: 0.7734\n",
      "Epoch 15/200\n",
      "768/768 [==============================] - 0s 324us/step - loss: 0.4635 - acc: 0.7799\n",
      "Epoch 16/200\n",
      "768/768 [==============================] - 0s 524us/step - loss: 0.4551 - acc: 0.7786\n",
      "Epoch 17/200\n",
      "768/768 [==============================] - 0s 258us/step - loss: 0.4683 - acc: 0.7865\n",
      "Epoch 18/200\n",
      "768/768 [==============================] - 0s 209us/step - loss: 0.4652 - acc: 0.7812\n",
      "Epoch 19/200\n",
      "768/768 [==============================] - 0s 222us/step - loss: 0.4802 - acc: 0.7799\n",
      "Epoch 20/200\n",
      "768/768 [==============================] - 0s 242us/step - loss: 0.4914 - acc: 0.7682\n",
      "Epoch 21/200\n",
      "768/768 [==============================] - 0s 253us/step - loss: 0.4734 - acc: 0.7799\n",
      "Epoch 22/200\n",
      "768/768 [==============================] - 0s 264us/step - loss: 0.4726 - acc: 0.7708\n",
      "Epoch 23/200\n",
      "768/768 [==============================] - 0s 246us/step - loss: 0.4974 - acc: 0.7617\n",
      "Epoch 24/200\n",
      "768/768 [==============================] - 0s 251us/step - loss: 0.5050 - acc: 0.7786\n",
      "Epoch 25/200\n",
      "768/768 [==============================] - 0s 322us/step - loss: 0.4843 - acc: 0.7826\n",
      "Epoch 26/200\n",
      "768/768 [==============================] - 0s 266us/step - loss: 0.4965 - acc: 0.7656\n",
      "Epoch 27/200\n",
      "768/768 [==============================] - 0s 280us/step - loss: 0.4771 - acc: 0.7799\n",
      "Epoch 28/200\n",
      "768/768 [==============================] - 0s 221us/step - loss: 0.4628 - acc: 0.7799\n",
      "Epoch 29/200\n",
      "768/768 [==============================] - 0s 257us/step - loss: 0.4670 - acc: 0.7708\n",
      "Epoch 30/200\n",
      "768/768 [==============================] - 0s 231us/step - loss: 0.4664 - acc: 0.7852\n",
      "Epoch 31/200\n",
      "768/768 [==============================] - 0s 228us/step - loss: 0.4718 - acc: 0.7799\n",
      "Epoch 32/200\n",
      "768/768 [==============================] - 0s 242us/step - loss: 0.4979 - acc: 0.7708\n",
      "Epoch 33/200\n",
      "768/768 [==============================] - 0s 270us/step - loss: 0.4768 - acc: 0.7812\n",
      "Epoch 34/200\n",
      "768/768 [==============================] - 0s 262us/step - loss: 0.4641 - acc: 0.7799\n",
      "Epoch 35/200\n",
      "768/768 [==============================] - 0s 196us/step - loss: 0.4543 - acc: 0.7826\n",
      "Epoch 36/200\n",
      "768/768 [==============================] - 0s 196us/step - loss: 0.4567 - acc: 0.7826\n",
      "Epoch 37/200\n",
      "768/768 [==============================] - 0s 219us/step - loss: 0.4503 - acc: 0.7891\n",
      "Epoch 38/200\n",
      "768/768 [==============================] - 0s 214us/step - loss: 0.4704 - acc: 0.7721\n",
      "Epoch 39/200\n",
      "768/768 [==============================] - 0s 235us/step - loss: 0.4944 - acc: 0.7578\n",
      "Epoch 40/200\n",
      "768/768 [==============================] - 0s 245us/step - loss: 0.4920 - acc: 0.7721\n",
      "Epoch 41/200\n",
      "768/768 [==============================] - 0s 244us/step - loss: 0.5137 - acc: 0.7760\n",
      "Epoch 42/200\n",
      "768/768 [==============================] - 0s 242us/step - loss: 0.4571 - acc: 0.7865\n",
      "Epoch 43/200\n",
      "768/768 [==============================] - 0s 259us/step - loss: 0.4624 - acc: 0.7734\n",
      "Epoch 44/200\n",
      "768/768 [==============================] - 0s 292us/step - loss: 0.4746 - acc: 0.7773\n",
      "Epoch 45/200\n",
      "768/768 [==============================] - 0s 230us/step - loss: 0.4702 - acc: 0.7826 0s - loss: 0.4586 - acc: 0.790\n",
      "Epoch 46/200\n",
      "768/768 [==============================] - 0s 232us/step - loss: 0.4762 - acc: 0.7617\n",
      "Epoch 47/200\n",
      "768/768 [==============================] - 0s 225us/step - loss: 0.4701 - acc: 0.7786\n",
      "Epoch 48/200\n",
      "768/768 [==============================] - 0s 236us/step - loss: 0.4557 - acc: 0.7852 0s - loss: 0.4372 - acc: 0.8\n",
      "Epoch 49/200\n",
      "768/768 [==============================] - 0s 226us/step - loss: 0.4497 - acc: 0.7917\n",
      "Epoch 50/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4773 - acc: 0.7669\n",
      "Epoch 51/200\n",
      "768/768 [==============================] - 0s 203us/step - loss: 0.4801 - acc: 0.7747\n",
      "Epoch 52/200\n",
      "768/768 [==============================] - 0s 206us/step - loss: 0.4631 - acc: 0.7956\n",
      "Epoch 53/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4543 - acc: 0.7839\n",
      "Epoch 54/200\n",
      "768/768 [==============================] - 0s 262us/step - loss: 0.4554 - acc: 0.8008\n",
      "Epoch 55/200\n",
      "768/768 [==============================] - 0s 221us/step - loss: 0.4725 - acc: 0.7878\n",
      "Epoch 56/200\n",
      "768/768 [==============================] - 0s 194us/step - loss: 0.4641 - acc: 0.7799\n",
      "Epoch 57/200\n",
      "768/768 [==============================] - 0s 197us/step - loss: 0.4782 - acc: 0.7839\n",
      "Epoch 58/200\n",
      "768/768 [==============================] - 0s 191us/step - loss: 0.4556 - acc: 0.7799\n",
      "Epoch 59/200\n",
      "768/768 [==============================] - 0s 205us/step - loss: 0.4602 - acc: 0.7852\n",
      "Epoch 60/200\n",
      "768/768 [==============================] - 0s 203us/step - loss: 0.4626 - acc: 0.7930\n",
      "Epoch 61/200\n",
      "768/768 [==============================] - 0s 201us/step - loss: 0.4543 - acc: 0.7799\n",
      "Epoch 62/200\n",
      "768/768 [==============================] - 0s 260us/step - loss: 0.4603 - acc: 0.7891\n",
      "Epoch 63/200\n",
      "768/768 [==============================] - 0s 223us/step - loss: 0.4863 - acc: 0.7904\n",
      "Epoch 64/200\n",
      "768/768 [==============================] - 0s 205us/step - loss: 0.4643 - acc: 0.7826\n",
      "Epoch 65/200\n",
      "768/768 [==============================] - 0s 258us/step - loss: 0.4708 - acc: 0.7721\n",
      "Epoch 66/200\n",
      "768/768 [==============================] - 0s 230us/step - loss: 0.4640 - acc: 0.7786\n",
      "Epoch 67/200\n",
      "768/768 [==============================] - 0s 236us/step - loss: 0.4940 - acc: 0.7643\n",
      "Epoch 68/200\n",
      "768/768 [==============================] - 0s 222us/step - loss: 0.4711 - acc: 0.7786\n",
      "Epoch 69/200\n",
      "768/768 [==============================] - 0s 234us/step - loss: 0.4433 - acc: 0.7930\n",
      "Epoch 70/200\n",
      "768/768 [==============================] - 0s 238us/step - loss: 0.4869 - acc: 0.7708\n",
      "Epoch 71/200\n",
      "768/768 [==============================] - 0s 245us/step - loss: 0.4597 - acc: 0.7826\n",
      "Epoch 72/200\n",
      "768/768 [==============================] - 0s 235us/step - loss: 0.4570 - acc: 0.7930\n",
      "Epoch 73/200\n",
      "768/768 [==============================] - 0s 232us/step - loss: 0.4644 - acc: 0.7826\n",
      "Epoch 74/200\n",
      "768/768 [==============================] - 0s 247us/step - loss: 0.4760 - acc: 0.7617\n",
      "Epoch 75/200\n",
      "768/768 [==============================] - 0s 243us/step - loss: 0.4538 - acc: 0.7786\n",
      "Epoch 76/200\n",
      "768/768 [==============================] - 0s 229us/step - loss: 0.4732 - acc: 0.7786\n",
      "Epoch 77/200\n",
      "768/768 [==============================] - 0s 240us/step - loss: 0.4528 - acc: 0.7956\n",
      "Epoch 78/200\n",
      "768/768 [==============================] - 0s 243us/step - loss: 0.4522 - acc: 0.7956\n",
      "Epoch 79/200\n",
      "768/768 [==============================] - 0s 238us/step - loss: 0.4614 - acc: 0.7812\n",
      "Epoch 80/200\n",
      "768/768 [==============================] - 0s 230us/step - loss: 0.4637 - acc: 0.7878\n",
      "Epoch 81/200\n",
      "768/768 [==============================] - 0s 226us/step - loss: 0.4540 - acc: 0.7773\n",
      "Epoch 82/200\n",
      "768/768 [==============================] - 0s 253us/step - loss: 0.4570 - acc: 0.7839\n",
      "Epoch 83/200\n",
      "768/768 [==============================] - 0s 251us/step - loss: 0.4477 - acc: 0.7982\n",
      "Epoch 84/200\n",
      "768/768 [==============================] - 0s 199us/step - loss: 0.4616 - acc: 0.7773\n",
      "Epoch 85/200\n",
      "768/768 [==============================] - 0s 193us/step - loss: 0.4464 - acc: 0.7995\n",
      "Epoch 86/200\n",
      "768/768 [==============================] - 0s 189us/step - loss: 0.4737 - acc: 0.7852\n",
      "Epoch 87/200\n",
      "768/768 [==============================] - 0s 197us/step - loss: 0.4533 - acc: 0.7839\n",
      "Epoch 88/200\n",
      "768/768 [==============================] - 0s 196us/step - loss: 0.4469 - acc: 0.7891\n",
      "Epoch 89/200\n",
      "768/768 [==============================] - 0s 206us/step - loss: 0.4790 - acc: 0.7747\n",
      "Epoch 90/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4620 - acc: 0.7799\n",
      "Epoch 91/200\n",
      "768/768 [==============================] - 0s 266us/step - loss: 0.4540 - acc: 0.7878\n",
      "Epoch 92/200\n",
      "768/768 [==============================] - 0s 182us/step - loss: 0.4679 - acc: 0.7708\n",
      "Epoch 93/200\n",
      "768/768 [==============================] - 0s 201us/step - loss: 0.4539 - acc: 0.7930\n",
      "Epoch 94/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4664 - acc: 0.7852\n",
      "Epoch 95/200\n",
      "768/768 [==============================] - 0s 198us/step - loss: 0.4517 - acc: 0.7852\n",
      "Epoch 96/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4597 - acc: 0.7865\n",
      "Epoch 97/200\n",
      "768/768 [==============================] - 0s 199us/step - loss: 0.4751 - acc: 0.7865\n",
      "Epoch 98/200\n",
      "768/768 [==============================] - 0s 196us/step - loss: 0.4584 - acc: 0.7891\n",
      "Epoch 99/200\n",
      "768/768 [==============================] - 0s 186us/step - loss: 0.4560 - acc: 0.7995\n",
      "Epoch 100/200\n",
      "768/768 [==============================] - 0s 190us/step - loss: 0.4564 - acc: 0.7982\n",
      "Epoch 101/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4641 - acc: 0.7812\n",
      "Epoch 102/200\n",
      "768/768 [==============================] - 0s 208us/step - loss: 0.4585 - acc: 0.7799\n",
      "Epoch 103/200\n",
      "768/768 [==============================] - 0s 201us/step - loss: 0.4487 - acc: 0.8008\n",
      "Epoch 104/200\n",
      "768/768 [==============================] - 0s 234us/step - loss: 0.4551 - acc: 0.7839\n",
      "Epoch 105/200\n",
      "768/768 [==============================] - 0s 198us/step - loss: 0.4583 - acc: 0.7956\n",
      "Epoch 106/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4686 - acc: 0.7799\n",
      "Epoch 107/200\n",
      "768/768 [==============================] - 0s 226us/step - loss: 0.4528 - acc: 0.7799\n",
      "Epoch 108/200\n",
      "768/768 [==============================] - 0s 200us/step - loss: 0.4551 - acc: 0.7839\n",
      "Epoch 109/200\n",
      "768/768 [==============================] - 0s 190us/step - loss: 0.4577 - acc: 0.7852\n",
      "Epoch 110/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4514 - acc: 0.7839\n",
      "Epoch 111/200\n",
      "768/768 [==============================] - 0s 192us/step - loss: 0.4641 - acc: 0.7943\n",
      "Epoch 112/200\n",
      "768/768 [==============================] - 0s 240us/step - loss: 0.4762 - acc: 0.7826\n",
      "Epoch 113/200\n",
      "768/768 [==============================] - 0s 189us/step - loss: 0.4489 - acc: 0.7917\n",
      "Epoch 114/200\n",
      "768/768 [==============================] - 0s 245us/step - loss: 0.4526 - acc: 0.7891\n",
      "Epoch 115/200\n",
      "768/768 [==============================] - 0s 225us/step - loss: 0.4518 - acc: 0.7917\n",
      "Epoch 116/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4430 - acc: 0.7943\n",
      "Epoch 117/200\n",
      "768/768 [==============================] - 0s 191us/step - loss: 0.4489 - acc: 0.7865\n",
      "Epoch 118/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4933 - acc: 0.7760\n",
      "Epoch 119/200\n",
      "768/768 [==============================] - 0s 187us/step - loss: 0.4735 - acc: 0.7865\n",
      "Epoch 120/200\n",
      "768/768 [==============================] - 0s 189us/step - loss: 0.4444 - acc: 0.7917\n",
      "Epoch 121/200\n",
      "768/768 [==============================] - 0s 238us/step - loss: 0.4512 - acc: 0.8060\n",
      "Epoch 122/200\n",
      "768/768 [==============================] - 0s 219us/step - loss: 0.4551 - acc: 0.7904\n",
      "Epoch 123/200\n",
      "768/768 [==============================] - 0s 198us/step - loss: 0.4489 - acc: 0.7839\n",
      "Epoch 124/200\n",
      "768/768 [==============================] - 0s 186us/step - loss: 0.4566 - acc: 0.7826\n",
      "Epoch 125/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4615 - acc: 0.7812\n",
      "Epoch 126/200\n",
      "768/768 [==============================] - 0s 182us/step - loss: 0.4419 - acc: 0.7982\n",
      "Epoch 127/200\n",
      "768/768 [==============================] - 0s 190us/step - loss: 0.4408 - acc: 0.7878\n",
      "Epoch 128/200\n",
      "768/768 [==============================] - 0s 190us/step - loss: 0.4389 - acc: 0.7865\n",
      "Epoch 129/200\n",
      "768/768 [==============================] - 0s 190us/step - loss: 0.4454 - acc: 0.7865\n",
      "Epoch 130/200\n",
      "768/768 [==============================] - 0s 210us/step - loss: 0.4422 - acc: 0.7982\n",
      "Epoch 131/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4638 - acc: 0.7865\n",
      "Epoch 132/200\n",
      "768/768 [==============================] - 0s 217us/step - loss: 0.4440 - acc: 0.7839\n",
      "Epoch 133/200\n",
      "768/768 [==============================] - 0s 193us/step - loss: 0.4570 - acc: 0.7852\n",
      "Epoch 134/200\n",
      "768/768 [==============================] - 0s 215us/step - loss: 0.4364 - acc: 0.8008\n",
      "Epoch 135/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4409 - acc: 0.8125\n",
      "Epoch 136/200\n",
      "768/768 [==============================] - 0s 245us/step - loss: 0.4505 - acc: 0.7786\n",
      "Epoch 137/200\n",
      "768/768 [==============================] - 0s 214us/step - loss: 0.4475 - acc: 0.7878\n",
      "Epoch 138/200\n",
      "768/768 [==============================] - 0s 222us/step - loss: 0.4430 - acc: 0.7930\n",
      "Epoch 139/200\n",
      "768/768 [==============================] - 0s 219us/step - loss: 0.4592 - acc: 0.7891\n",
      "Epoch 140/200\n",
      "768/768 [==============================] - 0s 212us/step - loss: 0.4552 - acc: 0.7969\n",
      "Epoch 141/200\n",
      "768/768 [==============================] - 0s 222us/step - loss: 0.4521 - acc: 0.8008\n",
      "Epoch 142/200\n",
      "768/768 [==============================] - 0s 198us/step - loss: 0.4375 - acc: 0.8047\n",
      "Epoch 143/200\n",
      "768/768 [==============================] - 0s 187us/step - loss: 0.4400 - acc: 0.7956\n",
      "Epoch 144/200\n",
      "768/768 [==============================] - 0s 193us/step - loss: 0.4693 - acc: 0.7891\n",
      "Epoch 145/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4675 - acc: 0.7865\n",
      "Epoch 146/200\n",
      "768/768 [==============================] - 0s 223us/step - loss: 0.4545 - acc: 0.7865\n",
      "Epoch 147/200\n",
      "768/768 [==============================] - 0s 225us/step - loss: 0.4446 - acc: 0.7930\n",
      "Epoch 148/200\n",
      "768/768 [==============================] - 0s 199us/step - loss: 0.4516 - acc: 0.7956\n",
      "Epoch 149/200\n",
      "768/768 [==============================] - 0s 191us/step - loss: 0.4435 - acc: 0.7917\n",
      "Epoch 150/200\n",
      "768/768 [==============================] - 0s 218us/step - loss: 0.4497 - acc: 0.7878\n",
      "Epoch 151/200\n",
      "768/768 [==============================] - 0s 216us/step - loss: 0.4482 - acc: 0.7878\n",
      "Epoch 152/200\n",
      "768/768 [==============================] - 0s 218us/step - loss: 0.4397 - acc: 0.7930\n",
      "Epoch 153/200\n",
      "768/768 [==============================] - 0s 223us/step - loss: 0.4556 - acc: 0.7734\n",
      "Epoch 154/200\n",
      "768/768 [==============================] - 0s 201us/step - loss: 0.4534 - acc: 0.7839\n",
      "Epoch 155/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4368 - acc: 0.7969\n",
      "Epoch 156/200\n",
      "768/768 [==============================] - 0s 202us/step - loss: 0.4433 - acc: 0.7773\n",
      "Epoch 157/200\n",
      "768/768 [==============================] - 0s 192us/step - loss: 0.4346 - acc: 0.8008\n",
      "Epoch 158/200\n",
      "768/768 [==============================] - 0s 215us/step - loss: 0.4350 - acc: 0.7904\n",
      "Epoch 159/200\n",
      "768/768 [==============================] - 0s 212us/step - loss: 0.4385 - acc: 0.7943\n",
      "Epoch 160/200\n",
      "768/768 [==============================] - 0s 193us/step - loss: 0.4389 - acc: 0.7839\n",
      "Epoch 161/200\n",
      "768/768 [==============================] - 0s 193us/step - loss: 0.4526 - acc: 0.7852\n",
      "Epoch 162/200\n",
      "768/768 [==============================] - 0s 213us/step - loss: 0.4365 - acc: 0.7904\n",
      "Epoch 163/200\n",
      "768/768 [==============================] - 0s 219us/step - loss: 0.4378 - acc: 0.7865\n",
      "Epoch 164/200\n",
      "768/768 [==============================] - 0s 193us/step - loss: 0.4517 - acc: 0.7839\n",
      "Epoch 165/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "768/768 [==============================] - 0s 227us/step - loss: 0.4406 - acc: 0.7799\n",
      "Epoch 166/200\n",
      "768/768 [==============================] - 0s 201us/step - loss: 0.4338 - acc: 0.8047\n",
      "Epoch 167/200\n",
      "768/768 [==============================] - 0s 196us/step - loss: 0.4488 - acc: 0.7839\n",
      "Epoch 168/200\n",
      "768/768 [==============================] - 0s 196us/step - loss: 0.4548 - acc: 0.7982\n",
      "Epoch 169/200\n",
      "768/768 [==============================] - 0s 173us/step - loss: 0.4391 - acc: 0.7943\n",
      "Epoch 170/200\n",
      "768/768 [==============================] - 0s 197us/step - loss: 0.4461 - acc: 0.7943\n",
      "Epoch 171/200\n",
      "768/768 [==============================] - 0s 189us/step - loss: 0.4514 - acc: 0.7812\n",
      "Epoch 172/200\n",
      "768/768 [==============================] - 0s 219us/step - loss: 0.4779 - acc: 0.7826\n",
      "Epoch 173/200\n",
      "768/768 [==============================] - 0s 182us/step - loss: 0.4417 - acc: 0.7956\n",
      "Epoch 174/200\n",
      "768/768 [==============================] - 0s 195us/step - loss: 0.4494 - acc: 0.7760\n",
      "Epoch 175/200\n",
      "768/768 [==============================] - 0s 196us/step - loss: 0.4363 - acc: 0.7826\n",
      "Epoch 176/200\n",
      "768/768 [==============================] - 0s 199us/step - loss: 0.4403 - acc: 0.8008\n",
      "Epoch 177/200\n",
      "768/768 [==============================] - 0s 204us/step - loss: 0.4398 - acc: 0.7969\n",
      "Epoch 178/200\n",
      "768/768 [==============================] - 0s 211us/step - loss: 0.4550 - acc: 0.7760\n",
      "Epoch 179/200\n",
      "768/768 [==============================] - 0s 180us/step - loss: 0.4264 - acc: 0.7956\n",
      "Epoch 180/200\n",
      "768/768 [==============================] - 0s 212us/step - loss: 0.4453 - acc: 0.7773\n",
      "Epoch 181/200\n",
      "768/768 [==============================] - 0s 180us/step - loss: 0.4359 - acc: 0.7878\n",
      "Epoch 182/200\n",
      "768/768 [==============================] - 0s 191us/step - loss: 0.4238 - acc: 0.7943\n",
      "Epoch 183/200\n",
      "768/768 [==============================] - ETA: 0s - loss: 0.4512 - acc: 0.783 - 0s 180us/step - loss: 0.4419 - acc: 0.7917\n",
      "Epoch 184/200\n",
      "768/768 [==============================] - 0s 191us/step - loss: 0.4387 - acc: 0.7865\n",
      "Epoch 185/200\n",
      "768/768 [==============================] - 0s 225us/step - loss: 0.4338 - acc: 0.7995\n",
      "Epoch 186/200\n",
      "768/768 [==============================] - 0s 188us/step - loss: 0.4524 - acc: 0.7799\n",
      "Epoch 187/200\n",
      "768/768 [==============================] - 0s 192us/step - loss: 0.4430 - acc: 0.7826\n",
      "Epoch 188/200\n",
      "768/768 [==============================] - 0s 182us/step - loss: 0.4331 - acc: 0.7956\n",
      "Epoch 189/200\n",
      "768/768 [==============================] - 0s 201us/step - loss: 0.4289 - acc: 0.7930\n",
      "Epoch 190/200\n",
      "768/768 [==============================] - 0s 191us/step - loss: 0.4371 - acc: 0.7839\n",
      "Epoch 191/200\n",
      "768/768 [==============================] - 0s 191us/step - loss: 0.4319 - acc: 0.8060\n",
      "Epoch 192/200\n",
      "768/768 [==============================] - 0s 238us/step - loss: 0.4363 - acc: 0.8008\n",
      "Epoch 193/200\n",
      "768/768 [==============================] - 0s 214us/step - loss: 0.4303 - acc: 0.8021\n",
      "Epoch 194/200\n",
      "768/768 [==============================] - 0s 187us/step - loss: 0.4349 - acc: 0.7943\n",
      "Epoch 195/200\n",
      "768/768 [==============================] - 0s 188us/step - loss: 0.4331 - acc: 0.7969\n",
      "Epoch 196/200\n",
      "768/768 [==============================] - 0s 192us/step - loss: 0.4348 - acc: 0.7917\n",
      "Epoch 197/200\n",
      "768/768 [==============================] - 0s 197us/step - loss: 0.4454 - acc: 0.7760\n",
      "Epoch 198/200\n",
      "768/768 [==============================] - 0s 199us/step - loss: 0.4408 - acc: 0.7969\n",
      "Epoch 199/200\n",
      "768/768 [==============================] - 0s 212us/step - loss: 0.4232 - acc: 0.7995\n",
      "Epoch 200/200\n",
      "768/768 [==============================] - 0s 244us/step - loss: 0.4372 - acc: 0.7917\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x20288ae6c50>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
    "model.fit(x,y,epochs=200,batch_size=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "768/768 [==============================] - 0s 150us/step\n",
      "\n",
      " Accuracy: 0.8021\n"
     ]
    }
   ],
   "source": [
    "print(\"\\n Accuracy: %.4f\"%(model.evaluate(X,Y)[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}