{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "# K-means & 階層的クラスタリング"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "このHands-onでは下記の2種類のデータを用いて，K-meansと階層的クラスタリングを体験する．\n",
    "* 人工的な2次元データ\n",
    "* とある購買データ"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "Hands-onに先立って，必要なライブラリを読み込んでおこう．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "tags": [
     "remove-output"
    ]
   },
   "outputs": [],
   "source": [
    "# 表形式のデータを操作するためのライブラリ\n",
    "import pandas as pd\n",
    "\n",
    "# 機械学習用ライブラリsklearnのKmeansクラス\n",
    "from sklearn.cluster import KMeans\n",
    "\n",
    "# 距離行列の計算のためのライブラリ\n",
    "import scipy.spatial.distance as distance\n",
    "\n",
    "# 階層的クラスタリング用のライブラリ\n",
    "from scipy.cluster.hierarchy import dendrogram, linkage, cut_tree\n",
    "\n",
    "# グラフ描画ライブラリ\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns;\n",
    "sns.set(style='ticks')\n",
    "%matplotlib inline\n",
    "\n",
    "\n",
    "# 警告文を表示させないおまじない\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "\n",
    "---\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "## 例題1: 人工データに対するK-meansクラスタリング"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "K-meansクラスタリングの効果や特徴を知るために，実際にK-meansを使ってみよう．\n",
    "\n",
    "例題1で用いるデータは[University of Eastern Finlandの計算学部が公開しているデータセット](http://cs.joensuu.fi/sipu/datasets/)である．\n",
    "このデータは人工的に作られた5000個の2次元データで，各次元に特に意味はない．\n",
    "\n",
    "まずは以下のコードを実行して，``s1_df``変数にデータを読み込もう．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>x</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>664159</td>\n",
       "      <td>550946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>665845</td>\n",
       "      <td>557965</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>597173</td>\n",
       "      <td>575538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>618600</td>\n",
       "      <td>551446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>635690</td>\n",
       "      <td>608046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>588100</td>\n",
       "      <td>557588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>582015</td>\n",
       "      <td>546191</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>604678</td>\n",
       "      <td>574577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>572029</td>\n",
       "      <td>518313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>604737</td>\n",
       "      <td>574591</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        x       y\n",
       "0  664159  550946\n",
       "1  665845  557965\n",
       "2  597173  575538\n",
       "3  618600  551446\n",
       "4  635690  608046\n",
       "5  588100  557588\n",
       "6  582015  546191\n",
       "7  604678  574577\n",
       "8  572029  518313\n",
       "9  604737  574591"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = \"http://cs.joensuu.fi/sipu/datasets/s1.txt\"\n",
    "s1_df = pd.read_table(url, sep=\"\\s+\", header=None, names=['x', 'y'])\n",
    "\n",
    "# 最初の10件のデータを表示\n",
    "s1_df.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "データの傾向を見るために，XY平面にデータをプロットしてみよう\n",
    "下記のコードを実行すると，データの散布図を表示できる．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAETCAYAAAA4W80CAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAADMnElEQVR4nOz9e5BcWZ7fh33OfebNd9a78CgA3QAK3dPd8+idnuklh0txh5oVtTJNiqYpkaakCFnhCIdkhuwIK0IKUpZCMv+xRNmhoBmUbEo2SZOiN0xyRW5rZ5e7nN3p7pnu2WlMP1BA41Uo1DPfmffmfZ7jP05moqpQAApA4Tn5jelBVVbmzfs453d+5/f7/r4/oZRiggkmmGCClxfGsz6BCSaYYIIJniwmhn6CCSaY4CXHxNBPMMEEE7zkmBj6CSaYYIKXHBNDP8EEE0zwkmNi6CeYYIIJXnJYz/oERlheXi4DPwR+dWVl5cZ93rcM/A2gBmwCf25lZaX1VE5yggkmmOAFxHPh0S8vL38L+D3g/APeJ4B/BPzVlZWVrwJ/APwHT/4MJ5hgggleXDwvHv3/GvjfAv+v0QvLy8t/EfhL6MXo4+HfvwL4KysrvzF8238OVJ/miU4wwQQTvGgQz1Nl7PLy8g3gjwIF4P8O/PGVlZVweXn5/wz4wBXg3wB2gK8DPwP+3ZWVleYzOeEJJphgghcAz0Xo5gD8C8A54IPl5eWfAn8SuIDegfxR4P+2srLyFnAN+C+e0TlOMMEEE7wQeF5CN/thAn9/ZWXl3wNYXl4uos/1beDKysrKR8P3/V3gHzybU5xgggkmeDHwvHr0vwP8qeXl5blhAvavo+P1PwRml5eXvzp837+Cjt9PMMEEE0xwDzyXhn5lZeUT4P8E/DbwGfo8/+rKysoA+FPA31xeXv4M+GPA//6ZnegEE0wwwQuA5yoZO8EEE0wwwdHjufToJ5hgggkmODo802Ts8vKyC3wT2ACyZ3kuE0wwwQQvEExgEfjxyspK9KA3H9rQ30+iYHl5+WvA3wQqwD8H/jcrKyvpIQ77TeAHhz2HCSaYYIIJ9uA7aFWB++JQhn4oUfA3ubdEwf8b+LdXVlY+WF5e/m/Rla5//RCH3gD423/7b7OwsHCYU5lgggkm+LnH5uYmf/7P/3kY2tAH4bAe/V0SBSMsLy+fAryVlZUPhi/9LTRj5q/ve1+Vu+UKFgEWFhY4ceLEIU9lggkmmGCCIQ4V8j6UoV9ZWfm3AbRw5F04xt5VZQM4yGr/JeCvHOb7JphgggkmODocRTJWHPCaPOC1v4b29nfjBJMY/QQTTDDBE8VRGPrbwO4A+yKwvv9NKysrbaC9+7V77BAmmGCCCSY4Qjw2j35lZeUmEC4vL/+h4Ut/Efinj3vcCSaYYIIJjgaP7NEvLy//E+AvDwXG/jxalqCEbgbyfz2i85tgggmeMOrtgCurbTp+TKXgcG6pykw1/6xPa4IjxEMZ+pWVldO7fv4Tu37+BHjn6E5rggkmeBqotwM++HSTomdTK7kMopQPPt3k228sTIz9S4TnVaZ4ggkmOELcy2u/stqm6NnkczbA+N8rq+2JoX+JMDH0E0zwkuN+XnvHjzEErG728MOUQs5iYTpPGE8USV4mTAz9zyEmMdmfL+z32tNMsrHj8/d+8zICRTdImK3mKXoWcSL5/HqD105PHerYl282+OHFDeqdiJmKyy++tcj5U9NP8nImeARMDP3PGSYx2Z8/dPyYWskFoOtHXLrZImcbGEIQp5LtVkDRs3FsF1BIBYiDymP2Ogn9IOLT6w3mqnnmajn6g4Rf+92r/OlfYmLsnzNMZIp/zrDbuxNCkM/ZFD2bK6vtZ31qEzwhVAoOg0hrDK7v+HiOiRAGBc/GsS1OL5bp+jH9QYJtmbz56jQHtakYOQlRklEruXx2rckgTDENA8MwKBdcynmHH148lPzKBE8RE0P/c4J6O+D9i+v8/sUNbm506fp3lE0916Ljx8/w7CZ4kji3VKU/SAjChP4gRimI4pTFmQKFnEXetamVc7x9YZ7FmQI3NrpcW2vz/sV16u1gfJz9TsIgzih6Dq1eOH5P0bOpdx6omjvBU8YkdPMCo94O+PjSNldvtVHA2RMV3n5t/q4QzO5wzWw1hx8mXLrZ4sKpGuWCDt9UCs6zuYgJnjhmqnm+/cYCV1bbDMKEejiglHfYqPsU8zarmwMKOYd2L+Tz6w2kgjdfnSZKMj74dJPzJys0OhG/f3GD2WqOY7NFKkWXct4mjFMyecf97w8SZiruM7zaCQ7CxNC/oKi3A37ro1Vub/uU8zYg+OJGk1Yv4rvvLO0x9rs9sWOzRa6stjAE3N7uYy0a9AcJb7w6iam+LBjF0dd2evhBQiHvcGK2yHTF5cRCidvbPkma8ZOVLXpBTClv8/Xzs1y51aaQczhzvEy5cCem/w9/cJ1qwWWn7bO60eGTL3c4f7LG0nyRn1zeoejZSCnpDxK6Qcx3v3nyGd+BCfZjYuhfUFxZbdPtJVQKDq6jH6MQgl4Q38WB3p2MqxRdzi3VWN/ps9MOOb9U441XpyeJ2GeMwzChDvueDz7dRCnFjfUuvSAhSjLWd3rEseTCmSnmqjl+7+IGpiEo5R1yjslGc8DidJ4Lp6YRuxKxt3f6bDZ8co6JEJAqSRxJ1rZ7TFU8zh6vALDdCpmpuHz3mycnidjnEBND/4JibafHre0uhjDIuSa1co68a9EPsrvi7ZWCw1bTp92Lx1zpWtnl/FKNd986tue9E+rl08dhmFCXbzZ478NVpFQ4tkGSSH7wyTpfPTu9J1w32r1dutGk3Y/wXBupFDc2evT8mN4gIkkVrq0TsrYFwjAo5x3WtvqcWqiMaZgAa1s+lYKDP0goei5516HVDWn2Qk4tllmaL/G9d8888PomY+rZYmLojxhHNajvd5x6O2CzEYASWJYgzRTrO31mqh45x7wr3j5dcfndn65RzjsUPZv+IGa90edP/1Ltru+cUC+fPh5UnVpvB7z34SqmIXBsg5sbPYSAhak8q1s9kkyNn9Fo97bR8PEcCymVTrwrQangsNEY4IcJ0+UcniOIk4wokRgClBJstXy2GgH19oAsU2y1fM4v1QhCiecYYBlMV3IkqeSts7O0evdPvE7G1POBiaE/QhzVoH7Qca6stjm1UCJOJI3OAM81QcHNjS5zUx5rOz3e++AGKIVCsFnvc2q+SJKCHyYUPYeT8yUa+9gRj1MOP/HaHh27Q2sjeK41NqJXVttIqagWXW7v9MnnLEDQHyRYljWmx85U83uolFGSsdMaECUZrmNSdm2SROKaJkGoqZQK8ByDL643yeVMPr0Ss9HwSVIwDJBScX29y3TFI041vX4Qp5ycKx0qiT+RWHg+MDH0R4ijGtQPOk7Hj5mfKuC5Nldvt9mo+0RJSiolr52eJueY/OxqA0PA62em6fgxUTsdJth0/NW1zbtCPGs7PYJBSqMbEkYpOddkuuyR9+4/TCZe2+NhZJx3h0x2G9GOH1MtusSJJIwzPMcEoBekHJ+19iwK55aqfPDpJvmcyZdrHeJEYpkCw4B2P6ZcsCh4Nqtbfdw4AaVotlMyKZmpuNxoBsQp2JbAsgykUgzaKVIqCjkLxzaRSu8Sfna1zve+tXTfa3vQIjbB08HE0B8hjmpQP+g4lYLDdiug1Y1QChZnCtza7FJwbVrdiCBKhkZCsNkIcG2T9XqfKJbM1Txubna5+OUOrxwrU2/PjcMDm42AOMmGHHtBN4jZaQ3IpKKcdw6kbsLEa3tcjIwz6Oc8iNI9TKhKwSFOUta2+piGIE4z0kx73IszhT2LwohKuXKzgZKaLx8qRZbZ5ByTfM7i5HyZct5hvTGg248wTYFp6oRsGGuqZBYrzEwhUOQcC8sU5D0bf5BgmQaNbojVj/h737/MO68v7Bkbu3d3m/U+SZoxP1UYX++Ezvv0MTH0R4gHeWZHdZzdMXfDgJWbLbZbA+aHRrzVC1k+VSPv2vQHehtvGYJeEBGnGUmSEUYJ1ze6/N3/aYXvfWuJRifi1EKJj77YRgiDNEvZqPsAnJgtsrLa2hML3o2J1/Z42M1zb/UiKgVnDxPq3FKVm5tdkiwjCBMa3ZBy3ubdN49hW3vpsaPairXtgFrZxRCKQSwRQo+rjp+wstrCsQ1mqx6ea9ALEsIoJYzvdABVgFIKpcCPUpyBSSoHKKU4d0IXYIEgiNLx2Dh/ssL1jR6fXNmhVnI5tVBmqpLjs+sNAOZq+bsWsXthEgo8WkwM/RHiQZ7ZUR2n0Yl4/fQU63Wfn6xs0+1HZErR6kUUPIcokdza7HPmWIVCzsIPU+anC1xda7PdCOgFMQiBbRpIqfj//fOrnDlWZWm+RLXo0vUjbu/42KbAdUxyrsntHZ8zxyoHeukPWpgmk/bBmKnm73lPmp0Bt7Z6tHsRtmXw6vEyhjCQwwpX2zKGCzS0uiE9P6FWdGh0QhzXZrbmMIgyOkGMKQyOzeTxB6n29iNJP4iJk7vbPGcSDAFSQqng0A9i0uxOzN62DMIw5tLNJgCfX69TK3pMlVyEMPjyVptzSzVePz1Foz3Atsy7FrGDMAkFHj0mhv4I8SDP7GEQRgk/+nSDKJGcWijuKYLq+DGea+EPEqIoJedYJElGqx/hhwmWEGy3fLp+xMKUx3o9IIxTgighSSUYBkpKwiRjbadHL0go5mxmqx4zVQ8/TCjlHRzLxDQFQgiKOZtmJ8S2zLvO9X4L02TSPh5GjJuCZ7MwXSBOMgZxxqmFEo5tkqSSoqfj9Bev1PHDGCkV5aLNtY0OKEgSSd41aHZDDGGw1fQRAnKORc61aPeTe36/VOBYBrPVPF0/QilFqxsRpxLXNrFNA8s0afdiun6EZRjM1vJjLv5G3Wf5VA3HtviVd08f6ponocCjx8TQHzHu55kdBrsrXpcWNLPh1o7P/+f7l/nWVxZ5+8IclYLDp1cbBIMUECSZRAmBAYSR9swMA9a2e9zc7FFwTQZxRjp22hSWoROyWaZodUMuXquz3ghwTMFGM0AoiJSilvNIUsWxmTztfsT5pdpd53y/Be79i+uTSfsY2M24EUKMi+OanRB/kHByoTTUkk/YbPjMVHO0ehFJKinlHcIopdsPaXYUmVRkMkUpzZ4JkwT8exv5EYp5iySVSKWQUpJkiu3mACH0cVzboNkZkLNNmt0Brm1RyNucmCngus5Dhy8nocCjx8TQP2fYXfGaSkW9E+JYJnGc8cHPNvjkSp1XjpW4udlmpzUgCBPiTIL+3xhKQSp1rLUX3t1EIpMQJ3eMf6MdUivm8FOFaxn0BgkqVoBguuLSH8Q4tsW5peqB532vBW4yae/G7lCWEIxpsAeFtXYzbtwh28axTZrdkCTLuLHRJe9qimUmJV9cbxInkpxjUvC0gdYGXv+HAtsySLO7QzX3gswUHT9ktpJno9kny+Qwhg8oiBJJGEm6JNi2QT6n//75jTavHC/RHxQfKnx5VLmuCe5gYuifM1y+1eTi1W2khDBKcRwTQwj8MGGu5jFVcfmDSzvc2OpqLz1RHKAoizzoxV1QMDbyBjoWa1kGWaZYmC5gtAc4lpae7QYxm42A/8UfO/vQXvhk0u7F7lCWacBHX2zpxh+VHKWCw83N7p4w3W7GDYBjG/SCGMMQ5E0Lhl6+HyZkGWRKkUlJtZSn3g7xowQhBLYFSaQHRZwe3sgDhHFKozMAKQjjbI9DIYb/jVwJJSWmYRMlEoEiCNOHDtMdVa5rgjuYyBQ/R7h8s8EfrOzg+ym9IKbjxzQ7A/wg1sY+SFnb7tHohwh0qOYB9vxQsCwo5ByiOCPvWqxu9ci7Ft0gpudHuJZmaHx+vbVHthbuyB//xvs37pK1hb0SuUqpoVRucs+dwcuO3fHnq2sd+oMExzKJkgxDGGw2fD7+Ymv8/nNLVQzD4MR8EcsUNLshmVR871tL5D2bjbrPxSs7fPzFJs3uAFMIbFOQpJI4zRAo7cE/aOW/D8JE0QtS/CghzQA1qsbQ/44M/2g8hnFGybM4OV+kXMg9tHMwCgW6tkmrF+Ha5iSn85iYePTPEPvZKD+8eFvHPHMmUT8DKYkyCOMYzzUZhAnNLiSpnmlx8vhmXgBxCq2+jvmWiw6dfkySKWSWsRVlSAVTJf36b320yi//wtKYe/+gROtRJqjvhxeF2bM7lDWSKbAsgzDW1atKOXy51uF7w/fvvn+ObXF+qca5pSrNzoAv1zp0+iFBmBLGGTnbwPUc0kyyttNjEGbYloHMMuL08c5bKciU3vkJsXfHqKQeR6YJtmVSHgrtRYnkzCNKFj9urmuCvZgY+kfA4xiVOxKyWhXw9GJ5zC++tNpmvpqjWtJt2ZQhUKlu7WaZgt4g1ZNK6Il1FN786BhSQZhkhK0BpgmDMEEYkCSKnGPS7kXkPYertzo02pd55XiVzXqfqUrugYnWJz1pdy84hoBPr9X5wSe3+eq5Wd6+MHdk330Ui8kolJVmknY3RKJj5iNZYFBjb3mEZmfAys3muC+raSg+/HybqZJLox2glKZARokkTiNc2yCKdfI0iI62ybccxuXHvw//NYAkA6UyBmFMFKVMVXL84luLR/r9EzwaJob+IfE4dMHdnw0GCaYhuLnZw3P1Y0hTydqOz2wtj2sZuCWPYBATJtpjk1JPKNPQk+pIIUEY2nNzLBPDEGRSYZmCTEoyJWh2AjaGBiSKM+rtAbWOg+faVIraUD2NROv+hisCxcmFMkkq+fJWG9exmCq5rG70SFJ5JNv+o6KJnluqjllVpYJDvRuSZhLHErR6IUkq+cord2LRP/r0Nv/Db3+JaWqFyfV6wgefbWIKmK3l9e4OSLOUYKAwDOihDb/xFAOzEu18WJYgiiU51+DVE2WmKt7TO4kJ7omJoX9IPA7Hd/dngyjDNASN5oDf+vEqpmVQLthstQY0OgP6QYIQiiSVFPI2cSwBnQiTR23kYRx0lQoGYUaUZMOJa2CYgpxt0R+kGMIgk4pr622SVNHsaW72H/7aCQC2WwGN9oDfeP/GEwmhHNRw5dLNJv0woVxwcR1rGALRFaNKKf7+b14eLgjw6snqHi//sF76UXG7Z6p5qgWXbj5GAGmmF1OFpkxeOFPj7QtzgM7Z/Pf/5BKp1Luqbj+iN0iQUpIbxq+TNCOTCse2SLNEh1iGbrZ8uJzrQ2EUwgHNAkrTDMMQGIZB3rU4f7LKG6/MTWi0zwkmhv4hcS+64OpWj/cvrt/XYOz+rECxutXDGzImaqUchiFwLAMDLUKFEpimgSUEQSYZvnRg4+bHxe4t+e7FJMkklgGlnE0vSlEqQxiCQZTh2gaZhJXVFm+8OkOUZHx2vcHrp6eeWHHUQQ1XKkWXZidkbcunVs7huSYFz8a2DC7daLLdDFg+NUUQxvzOx7f40eebvPP6ArWizYefb4956nGiBd2etMyDQvDW2VmEEHT6ERt1H3+QIJXak/9478NV4kySd0x6QcIgTnBtk0QK/DAjTAaEUaoXC0tgmQZKgVK76I9PCFKBZejdX7Vk0+gMz0EYnJgr0fYT4iQjjJ+EVzLBw2Ji6B8SB9EFt5o+mw2f2ap3XwMnUFz8coc0g826T5xkWKaBAGxTN5NYmM4zVfZo90I6foxtGrT7IWpIdXiSk/deSCU0uhHpMAnsWAZ+ENMHSnmHNNPXNVPxeP30FAvTReDxiqPu5WmPGq6kqSJOMxzLRCnJdisg52rxrUGU0uyGLE7n6QXa08+kotGNcWwTlSkuXW+yttNjYbrAXC1PnEjWtvqcmC8+kszDw2D3sSpFl0rRJQi1Ed/dQERKRSlv0+lpITEDA5lpOq1UEpWJcTw/TRWGociyo8nd3A+WoXdHrm2ihMAfZHiuRc4xUWjZDMc2ubnZ5Y1XZoAXJ1n+smJCr3xIHEQXvLnZ4/RimXzORghBPmePNcJHqLcD2n7ETmvATstnuxXQ7cf0goh8zkEBpbwODZyYK3Hu5BRfPTvL/HSeJFM4loWxP0v3hLD7awyh/0tSbWCEAplJgjAjSSVRnHJ8poBUgkLe2aNSCNrr3S+H/CCM4uFRklErueMm1b/5wTV+9ye3uXq7w43NDl0/wo8SOn6MaRiU87bWaDcNTs4VqbdDekGsC4luNOgFEYbQXPNmN8Q0DeJEDitOTVzHotWNDjzfo6SJHuZYo0Kp+ak8UaobcBtCa8wrqSh5tlaVtE1sy0DXUom7ErlPApYpdLhGKgwgSjPyrkmUSFxbECUZ1aJDqxdxbql6z+e5n4o7wZPDxKN/SBxEF1yYzjNX2+ud7N/WX1lt4zkWrmORhQm2bSIziWNbvPP6HLfrAUkiMYQgijOiOOXcUo1jUYG17T4C2Gllj8WHPixGjB7T0IyQZFhiK3ZR63SkR784P1MgZ1v4QXwkXu9B8fCNeo9//JPblAsOnT4EsQ4LFMIUJWC+luf0sSqvHq+OQyGDMCEIU6IkJY4z8q7NTiekmndIpSRJJddut2l0B9imSblgYxiC80u1Az3Q8ycr/PDixpj98otvLY7DLA/jrR6GcjoqlPIHCYtTeZq9CKl0pqFWdscxcdsQdNKYMAXLVJjm3hyOMXyYRxWuNw2wLFMrW0owDEEt75KzbRxHMVXOUSk65Byb86emxjIYSqmxVEMhZ1Mru5P4/VPExNA/AvbTBd+/uP5AA9fxY1rdaOylzdXyWiM+yej6CaeGujaDKEEqydmTVWzLYHUzIOcYrG35hEfAmz8MdhuFNJV3wkXDjKYQAscEYQjyrt7FnD5Wot2PhvK1j1fReFA8/NrtLmkqh7sIiWkYSKkbcdiWgWkKbmz0yLs2x2a1Rvv1jQ55gV6oEHQHMbZpEpgJjm2RSUWUZIRRSmxI0jRDCTANdRfD5rc+WkVJOH2swmtn9LVdvtUB4PKtzkOzcR5EOT23VKXRDTkxXySM9a4l79mcXijR7IRs9GMqRYf5WoFkvUMmoyHt1sA2JekwhCME5PMWKAgG6XiRfhSYQyXLOJFMlXU47NhskbMnyvSClJmKS5wo2v2IMEk5s1gCdEObzUYwlmqIEy3dED4uuX+CQ2Ni6I8AhynZrhQcVm42mSrnACh4NjOVnG7u0Q45v1Tj3/pVzTkeeYf1dsDtus+JuRKNTkh4gJTsUcLYVwgzkqkdGXoxjM9bpg7SZpliftrjwqkalmlwYrbEuaXqYxdH7Y5hd/2I9R2fq2sdwjhFosi5FlGSIKVCDkv+r6y2cR2TjbpuzlEp2BQ8B8sQtHu+5q6nGXnPQYQwZZt0/ZicaxHGGWmakM9ZfOdrx/n8eovTxyp7dhTdXgJC8crx6vg1gB9e3LjrvfD4om0jr//jL7ZIpeArZ6axbR1q6vgxS/NFhDAwDYFpGDi2iWUa2LZJkkjCOBkX1CWJHNNy1WPE8LPhrk4Nizlc2+Av/ksXOH9qek/z8mrRpVZ2uXyrw1TFww80lXiUPNfFVBl+8GBBtQmOBhNDfwQ4zFb83FKVjy5t0QtiSnmHOMkQhsFbZ2eZqXi8+9YxgD1xy5WbLQo5i/mpIo7VJOdmJInkIfSoDg2dENax6jTTXPkwVohhjH5EyjGE5tfblsHsdJ4//NXjWOad5hcPUxw1CnlcvtVkbauPUoITcwVeP1Njs+FzZbXFl7c7pJnEDxJsS+u0CAPiWI4XJTk8tzTM8MMME2h2QxxL4Lk2jmORUxAoxSCKkUqhhk2zXdvEG+ZUHNtkpurxyZUGr52x7jBiwpSNeo9q8W7WTb0T8doZ667Xj6KWYKaap1xweef1+T27xfd/to5jGwgh2Kj75FwDgU2UZHiuDgmCNuwCiGI5Nu6Pm5RTwx2BieCXv3mSqYrH+xfX+YOVbXKOyenFCkLA+o5PqxdSbwUgFGpYe+EMFyulFIX8z6fe0bPAxNAfER5k4Gaqeb73rSXe+3CVZjekWnQ5taB1u0dJuP1FOa1+TD4xMY2AKJW6zNzgiRn6atlFSR12sgyBUhk51yRNpabw2YI00WGT2ZrH187NkEko2uZDee6jgqdPruwQRgnXbnf0rkEINptdfviz22QZtHohhiGwTDGU1YWDos37PdRs+LY4ViRpTM3QDVaUUmTS0Pr9mcQaJmMzGdMXglrR4SeXtkkzye/85Bb+IGG2mqdackEJdjohnX40Lg4bRDpc8SRF2w4KY1WLLu1+xLtvHoNl6PoRP/58kxsbXSzDwMvZw57A4dirdyy9iCeJJEwkjgWGYehn+xDjyRxSKr2cyVtnZ8bjFaFTwRe/rCOV5vn7g5i17R6uY3F6oTTshpVSyFnM1krMHkF8fsLmORwmhv4p4vypaaYqHh9/scWXax38QcKrJ6vjv+9PQk6VXNq9kKvrXXK2hW0Y+FFCnB49N1miC3YMw6CQs4bGU+kqR8cgZwuEYRCRce5Ehbdfm8cwDH7htftLDOyvYp2f8siUottLyDkmF69sE8QZjmmgUHT70dhwO5ZBmiliJR+oxnkQMvQAD+J0uBNS2Cb0/QSFDvtEqcSWgCE0EypI+GPfOMH1zS5KQr0zABSVkjOMLXd46+zsODz3i28tjmP1T0Jp8SBa59Qw5BeECZ5rYZkGr5yoUCk4JJmiVspxbLbAb/14lesbOrfhOSaJVEg0UyZLwbAVwtD36H7GXnAnEe86JnnX4sRckUYnGo/XoueQpLrVYRSl9IcNaqbKORzbZGW1xXe+dpzlU7XxPXpcYbuRY6SUotkJ+emVLf7H37/G2ZNVlpemJkZ/Fw5l6JeXl/914D8CHOC/XFlZ+a/3/f0bwN8Y/v0W8BdWVlbaR3uqLw+STPGVV6bHhmGUvBt5b6O4NAI2GwGOYzJXzXFru49jmhSrNn6YEITZmCFzFGnaTIJhKIQQyEzHWtNMYRiCIEyolmzmp/L84a+dGHO/7xeLvnyzwT/859fYbAZU8g61isvHKzs4liBKJD0/wh+k40KfJFNk8o4i4iB+/K1LJiHLJK5jMIh00wypJKYxlOsdinWZUmsKSaVIMs0eiWKJP0joBQnvfGUBpdSB4bmpivfERNsOyv8IIcZ9fkff+cu/sAQw9rA918IQBkJoWYIwkZimwEBLFkulxsl9Q4Dn6nqOKJJk3Nk57laptC0Dz7GZqri8cXZ2z27j2GyBSzdbhHFKb5BQK5mA0hIISlNzP/pii/mpAmdPVI6kiO7jS9t8drXOjc2uZoUZUHBtrt7uMFXO3bP47ecRDzT0y8vLx4H/DHgbiIAfLi8v/7OVlZXPd73tvwL+8srKyj9dXl7+vwD/B/TCMME+3KuU/uNL22zW+/z0yhb+IGWulufkXImNnT5+mBKniuOzRT0ZU0mUZJyYc2l2IwZRdiSFVKYpqBZdMgmeZ+I5JoMoJedagA6hfOsrC4fStam3A/7h715jbbuHBHY6AzZaPlGYoVC4tkU3iLTuvpRjGuBRLVojaN19SRrqEIUhQA75hnJoyEZFa1Jqw39js8vrp6dJ0gxnrkh/kFIu6IXt68tz43zKCE9KtG0UlhhEKfVWoLs2zZbuu5DszhUdnyuw0wmIopTUUBhCkKQZathiyjKUTqyjF/dy0WU7CsY3zrYEQimyoWgaAspFm6+8Ms3bF+bG55bP2ZQLLhdO1Wi0B8RxSM4xxjo3q1s9yp5DqaDbU37ypW4W/vZr8w8tRbH73vzgp2vc3tIN7NNMC0EJJRAGtHsxSwulCYVziMN49N8FfntlZaUJsLy8/A+APwP8J7veYwLl4c95oLn/IMvLy1Wguu/lEw93ui8+Doq5JmnGJ1d2uHCqxtXbHaRU7LT0hKuVPU4fs5mtFrhwegqAIEz4+IstZqoe1253aHQHhLHeNgeDlOwhLaVl6iRbIWcxU/VIhjF51zGxLZNjs0UGUYJpCo7Plcafu18s+uMvtthuBWRKJ3k7vqaNxoluf+gPUtLsTgjqSZYHGEIQD79Ae7XobkvohUBKiW2bOKbmh/f8eOyhRklGIWePi5qeVvOL3fmapfnSnnDHYWmb9XYASvH7P9u408lK6mc9U3Fp92O0kTcAgcwktqVX3ELeJYpS4jTDEJDLmbz56iy/8Pr8Hq2g3bsNyzS4cKZG3tWtBPOuxY2NLkqBlzNp9ELKBYecbfDhZ5t88mWdr56b5cxi6aEpqh9f2qbZDZFI8q5NL5BIqbRjkjNpdAYopdhphwA/92Gcwxj6Y8DGrt83gHf2veffB35zeXn5rwE+8K0DjvOXgL/y8Kf4cuGgmOuN9d5QCjjGD1OEAFMKen7M18/PsrbTp9ULUUrdiQGfncG1tRH+rR+vIlWEwB4O9uxAw2kYurI129U4YiT25eU048R1LMI4olJ0mJsqcOFUjXLBZbPR5+ZmbxwXflAs+su1Dq5j0u7HtKJ0nMTzE4VCYlsmpmkyiLKjq+Y5AAJt6Ef7hDTVMsDC2PeuUSNV9CJQyjucWijpVn2eTm4+CR39e+EoRNRmqnn+7B9fZhCnXLvdRQGea5JJ3WgmThSDKBk2EpEEYaqrbB2Lct4mMAVeZuM6Jt995yR/5peX7zr+frbZL//CEs3OYEw6SNKMxZk8O62Q+apHmika3ZAsg2MzWmH08mqLUwulA6919G/HjxHDAj2l4IcXb+uGLWZGmKR615ZJkjgl75m0eiGuYzJbzVHvDPjof9piYbrAidniz6XRP4yhP6iqejw1l5eXPeC/BX55ZWXlR8vLy/8+8N8D//K+z/w14G/te+0E8IPDnuzLgINirpvNvmZEpBmzVY8wTlFK4bkWx+dKWJZBoz3YEwOGO/HY2VoOyxDc2ulTK+WYnzIJwoTOMLEZp5pHPUqoKam9+EyC55gUPZtzp2pUCi5bzQFpBsW8zamFEqW8QxAmGIZxV1z4foYvjFLCOMWyBNkgQ2DQDxMdDzfANARx+uQLwBRap33377ZtaCmEeNRIXZBmEsMwyFJJP4i5dLPB+ZNTvPvm4jMxCkclojZTzfNv/MtfGSctL17Z4cZGj04Q47q6r2wUS1IpsU2DUt4mTiXz0wWtv5RKWv2IUuHgBiIHha125y3+YGWbVOqdXL0T4ocJtmVQGSqN9ge6HqLVjcYaSaNrXd3q0ehqB2dtq8fV2x1s2+CbF+bIMq3smndM6t0IJRVZJskyuLXps1H3Wdvuc/5EmSu326SJHNJpjZ/L2P1hDP1t4Du7fl8E1nf9/gYwWFlZ+dHw978B/Kf7DzJMzrZ3v7a8vLz/bS89DvKCpsu5ce/P6UqOtZ0+WaYIIq2H4g8SFmeLe+Lwu4/jOTblYy7TVY+ZqkcQpvz40ibFvEMmJWaSEceSLFPYtsGpuTyZFERRwvx0nrdfW6AfJPhhyuJMnm99ZZ63X5t/rASjl7MwDIOZSo4kyQiilExKBBIhTDKlKOR04cyTxv78RZrKve6LUsQpOJZuql30HG5u9ijk7GdG2ztKEbWZap7zJyu89+GqHg/DRR4lcWyHYt5hrubh2Carmz3OLZQIE8kgzjCF4OyJykPngEYLwHTF5dd+9yqObWENk/pSKhamCsTDsFjetdhq+ohhb+SRRIIfxOQck5ubPVodTUlOMskfXKkzU80RRAntXkxhqKeUyV31HqZBkqT86PNtKiWX6UqOTk/dV7juZcZhDP33gf94eXl5Fh2W+VeBf2fX378ETi4vLy+vrKysAH8S+PGRn+lLhP1e0NpOn+2mTxRrFcDZqsd2M0BJQRSnukAoyWh2QlZuNvno0hbf+9YS509p4zvaJWzU+0RxpumXUuniGWXimBaRlZHPmWO+erlgM3u8hB9mTJVzHJ8tjsMxoyTZbh2Xj77YfiiDNzvsmhWEKbNVj1Y/Jkkz4iTTnlyQ0A2iJxq2uRdGErujr45STTmUUssyS5mxvuOz3Qz45msL95UvflK4X7X1o3DHG52I04tlhBBMlT122gN22gF+EHPyzBRLC2XKBZcwTnEci+mqQ7sfsd0K6PgRm/U+9Xbw0Nff6ES8fnqK2zs+Nze6GAgw4MtbbSpll7fPzxHGCdvtAZapWyre3OySZZJXjpVpdgSeY1JXitxQwK03SPByNsdmirS6dfxIYloGZiaxbIFhCKRURJkaSmprKeckSViv99lqBsxN5X+uQjgPNPQrKyu3l5eX/0Pgn6Hpk//NMETzT9BMm4+Wl5f/TeDvLy8vC2Ab+Lee5Em/bDgxW8S1Ddq9mP4gpVJwObVwp6AkTqXuROWYTJVz9IKY9z5cBfRE0nLGglo5x+fXmzTbA4p5mySFomfy6rkq6zs+Qag9eM+x6AUJ2+0QKRX1dkAx797luT9OV6Xd1+SHKU69T5gktHoxrm1S8iw2GgOUTDUD5inJL1vGSANmXzNGMezIZECrFzNb85BSNwZ5Fl7gvaqtgUd6JrrRfIg3VOmslnKcPVGl2Q155/WFMZPozGKJ9z5c5Xa9Tz+ImSrlcB2TqUrukXoLdPyY+akCC9NF5qfy/OizTVr9CMNQzNc8bm52idOMC0s1bmz1QEHONhkoySdXG5Q8h7MnK+SGFdsgKORslIITc0Wu3GqRZIqi59BMJfmcTZpKgjhBCZ1IDuOMvh8TJpJWP8SydJh0d//jlx2H4tGvrKz8HeDv7HvtT+z6+Z8C//RoT+3nByMBq6WF0h7v7dxSlY++2N4zQUFrwG/U+7z34SpvvjoznvCtXsSJuQJRpLnpmdTsmUGYsjid5/JqC1MItloBhhC4tlZs3GgE/Gtvn7xrwD9OQnD/NX2QpFimx5lFgyST3Nzsks9Z2JYgyxRRIlFIkietcyV0XsIf7A0ZmYZASS1hbJmCOJFUSy6u1v+l1Y1w7KdbX7h/51dvB/z6D64N4/c5FmcKY6rrg57Jfq0lYChCF/P7FzXX4txSdVzU9+s/uEbHsfZ8z4PqJu71vaMQVD9IOHeyRhAl9PwYyzTJ2QZpEGFbJmcWyqSZ4vp6h2CQ0u1F1JsD1rb7HJv2MEwDxzKpFG1ubXZZiTPyORul1LjTVt+PMC0DBWTZsEoa2B6ybwwDcq5iEKdcvdWhWtzme98+fejreVExqYx9DnA/rZyDJmicSJJM4kq1xwj3ghiUYPnUFEmqQyTRUN1xaaFEECYM4nRId9OTOJ+zaHbDAyfw4yQED7qmqUoOz7W5stoi51gcmynQ6kc4pokfxrS6IZlQuLY2/kfZnMjgTl9TpbSkgmsLpBQohgqdQnv4lmkQxRkn5jWV1LENml0tPPesMNpdjZ5JkkqurLY4t1SjXHAe+Ez2ay21+xG3tnrMT3nMVHJjjfiRx74wU+S1M+5w56PxKMng3SEof5Bobr4QvPOVBQBub/eptwa0OtopuLHRYW27RxBlyExhmoI0y7i108cUuuVimklc22S25pGl+rhxqsiGRW8HaYSM9m5CCGzDYBBpOvLVW2349kNd0guJiaF/TnCvopu7xdB0sw/TMO4S2UpTTV08M1vm0s0WoKsZW72QqUqON87OculGk6X53HgCR3FGtege2GzjcROC+zndH3y6iW0ZnD1Zpd4Z0PFjzh6v4Lk2CkiSjHpnwCDScrpRHLPRGIz1Wh4VArBsQbXo0A9SXj1eYXWzCwL8MKXkORimoO8nJJmkkHcoeTYlT3uLvSDGMMRjl+w/Dka7q1opN17EATbq/pDFsveZHBTH36211B/EzE95eI7NsdniXbu1o0oG717wpVIoZXDhlF4wL91sYQiYq+X48naHjc98en407HWriQNCCAwBgygbttfUz9NPU0DXaQSDDNM8mB64H64lyHsOmVTstAcIQzyx/sbPEyYdpp5zjMTQMqm7Ilmm4MR8kYKnmQm7YVli2EBDVynalm4gbVsGtino+jE7rYCdVjBWE4zilFrZPXACH2VXpdGEd20TqeBbX1ngna/M8/Xleb6+PMdrp6eYn8nz6vEyrV7IzfU2G40ByRFo8OttvMKxTfI5C2EIcq5FyXM4OV9EGAKBweyUx/kTVV49UeGbr89hmdqTz6Tie996trHcjh/juZbW2h8+N9sStHrhXc/kXh2dpioe/9q/uMw7ry/g2jYzlTznlmp7Kp1HC/5RP/t33zrGH3/nJGGS8tm1Bh9+ukEcpwyijEzBsekiQZQQJopkmLARQ6G0QajrQrJMa/Jkw1qQ/iAjDLOhl8+hCgXjTOk+BqEOj85WPEwDPr3a4P/xjz/nvfevv5SdryYe/QuAUdx0t4d24VSNy7c6ewqYSnkHYejK2VLeYWnBwHUNlNQa4EulHEmS8pPLO6SZZH6qwOx8EcMwDpzAh5FffhgcFHMeHVsI6Ptam98QgjgddrB6DAiGRWJornyrG3P+ZIV/7899nWZnwK/97lXKeQdDwHo9oBtEvPPmIm+9Oj1Ocp9fqj0Xnt7Iwx4t4iMZ4ErBuStBer/cyrtvHRu/N0qye3rsR/3s6+2Ay7c6nFoo0epGrG33GMQZ0yXNp+8FCdWSS5oOSFJFkkEmJQg5ZkeNRsNYnvpRhO4yRaMTAopSwabRHfCPfnANzzUpei4rqzq5+7Lx7CeG/gXBQaGd/WJa331HC1vtfq06LEwZTeizJ6co5h2anZD56cIDt6xPSsdl/7Hfv7iOVNAPEpIMcu6wavYRMRZ7U9rYe46FRPHWudkxVfSPfUM3Gal3Il45VuYX31rk/KmnI3HwsNgd6x4t4lOV3IEG6TC5lcM0yznKZ7978VmYLiKEoD+I2WoG5BwLxzLIuzalQsYgTPDDoWLpPmO+vznOw0INheu0IqfBtdsdbMsgkxaubXJ7J+HMscpLx7OfGPoXEPvjr/ulgnf//Bvv38Bz9z7m+akCjm3xK++eflqnfBf2X8PaTo80VXSDCAMtUfw4hj7nGHiupQu1MoVtC6olj5lqfo9h+9XvvPJCTOiH8bAPE18/ao/9Qdi/+CzOFLi8GpGmkn4W4zkWQglsw6QZ3Tvh+7g03DGrVujq7SjJMISg00/GfSJ+/PkmC1OF52Ind1SYGPoXDA/LbT/KCsujQr0d8P0frdILYtJUYVmCdjfCdU2k1Nm2QfxwbeZG89cAcq4BQhAmGY5tUqnaZJI9nZqOquXf08RhPezDeOsPc7yjwP5xWCm6nJwvIYTgxmaXdBDh2Aalgs3GXZKIRwfT0PF8pSALtSOhDIWUujo6ywaYhuDkfOmR6gaeV0wM/QuGh+W2H3bSPw4etlLz4y+22Gz4mIZBfxDT6kW0ewMMoTn2/UFM+hB8egMwTIGXM0mzjCyDpfkCfpiSDTVsKgWbV47vpUceVcu/5w1P21s/DA4ah4NYM6DyOZOPvtjBELqC9Ulit+LGuOf9robpWabbIM5Ue6AEH196OXj2E0P/guFhue1PetI/SvXsl2sdDCGod3RlbhhnOJZFlKRDSeAU01BYQy6dVFpdctQFybUNLFPgDz0yCahMYQkDOaTlfePCwvj7un5Eqxs+dzubJ4mn6a0fBvvH4UhcL0q04J1tCdo9LYX9tLE7HKQb7cDatk8hZ9Hqh3tkmV9UTAz9C4ZHCcU8yUn/KNWzAl2VaVuCrq/VDEedrPM5i1ePVVjb6SMVVIoOJc9irR4gM0kYp5jG3cbANCAeKhiW8vZdjaiPzxXpD3Q46EntbCa4P/Yn36M4ZW2rj+tYlAsOjW7EIEiwjPu3NnySsC2BbRlIpbTA29LUCxXeuxcmPPoXDEfJbz4KjPjdu7Gbj30QXj1ZpePHKKWIE93xKEoySnmtFmmaBtPlHHnXotkJubnVR2USIdSwKEaRZtl48OZsQcGz8WwD2zY4PnenEbVtGZycL7G8NDXm8bd6Ea5tvjTx1xcRHT+m1Y1wHYs0k8SxwhoW8dnWszNLcaqwTYFlGqRSkwLuN5ZfFEw8+hcMz1v89VF2GG9fmOPilzv0+nqxylIo5W1t4CsefpgQppLZqq4CjZIML2chpSKfc+gHCZkEywLHERjCoOA5FPMWxZymkh6kG/S8hTN+nrFb2mOnHVDIW8zisdmQDMInLXh0fySZomgJ5qoecSpfivDexNC/gHieDNa9kr3HZvK8f3H9wATtTDXPn/zOK7z34Sr5vDVWSdxqBRyfLbJe95lyLUzToJh3mM9ZGKZBqxuSz2ld85HnbwrBwkyB189ME8UZUkmWFspjz/1ZL4QTHIzd0h6jcJzrWLzxyjSXbrSQMiJ6Co1pTENrG8WJvFOQJRWtTohtCAZxRjeIH0mi+XnCxNBP8Fg4aIdxbCb/wB6gUxWP80s1rt5qY5mCnGMyW8vjOTZzNY/Zah4hBNdvd0DoydjpRVSKOSpFl9kpRTBI6A50c4owSukPYhamCy9F8uxlx0ja470PV0kSieFYzFWHDXhck3xmIwfxkaqZ7hOmBu4kYg3jjhZalulG8vVOyB85M4Vrmy881XJi6Cd4bOzfYbx/cf2+CdrdTJ23X5sf7wLOn6xw+VYHdtANV4RBPmeCMBAovJxJL4gwDYOFKQ/XNigXbDxXd6p67fTUuGnKBM8/RtIeH3+xxSdfNnBti+mKw0ef67xNKe/S6kaPKYRxB6PjGLtaBEsFSSrHBj/vGroxjWVwbKZIsxu/kDUX+zEx9BMcOR5EAb0XU6fRifj2Gwt8fGmbT67sUCu5vHl2ljBOubnZ4yuvTOMHKUGUYJkm37gwPfHeX3DMVPN8790z49aVf7CyTbXo4A9iEqmwLYiP0Ku3TYFjaxnMNNPSxlmmjX/Zs8h5Dr0gYmYoC94N7jC1XuSai4mhn+DI8aAE7f0Wgplqnu99+zRvX5gbF2HNVvP84i4xrglePox2hR0/puDZbLUCPr9Wf2yapWmAIQSWqRvA25Zuo5lzbKSCeNjechAm9MOMfhRgGoJMKgZRyvGhwX/Ray4mhn6CI8eDqnEPq8UyMew/f6gUHOIkpd4JyTk21aIgjlOCKMMwdJQ9PaQEkiF05zDTMDANgVRQ8Bymynnmah75nM1WM+DWVhchBEIoDEOHCTebAaW8zYVT1TGF+UWuuZjw6Cc4cuzWnj+Is/681QJM8Pzg3FIVwzCwTMi5FnIoP1op6CYw2dDIGxzcaMQ0dF9g2wLXNsnnbAp5CwzNzzcEVEsuf+irx7lwegpDQJzpGL3rmNim0GwuA2YqHnnPfSlqLiYe/QRPBPfzyJ+3WoAXBQ+rKfQiYjQ2Vje6zFZd/EGEcLTh7QaJFq4T4LoGtYKHH0a637DSzKxCTstRx3GGa5tDoUpBrehyfqmGaQh2WgMtw4ACQ+A5FlMllyDKSBJJuWCT92ymy/lnqvB6lJgY+gmeCSahmYfDo2gKvaiYqeb5+vIcn16rUy7kWN/pc2Ozi2Ua5ByBZQhM08SywLJMQDBTyVEt5UilYhAnTBVc1nZ8TNNgpuqxOF2g4NmEUcr8tMGvvHua9y+uszCVp9EeoIBqMUcmtWSGbRrMVNwHneoLg4mhn2CCFwCPoikEL+4u4NxSlR98cpupkku56FAruyilMITBIEoI44yNeqxDMWUX2zKRSlHK2ziWwZvnZpmqeNzY7JIkGQrdOrM/iDkxX+L9i+v8/sUNCjmTStFho6Hba5qGwB+klIsOv/jW4rO+DUeGSYx+ggleANxPU6jeDnj/4jq/8f4N3r+4Pu55eq/esS9CT9SZap6vnptFKd3r2HMtZioFXFs3pHFtk4JnUS27FFybxdkC0xWPnh9z9kSFheki55ZqzFQ9TFOwvuMjlaRctElTSZRkzFZzKKCYdzg27eEHCTutENc2+BPfPvXcdht7FEw8+gkmOGI8jhd9r8/ei6kkBPcM6TzqLuCoruVx8faFOZJUglAMwpStVkCzF1MruziWSS+IWZwuYBgGYZTx2ukyV1ab/OCTdVZutqiWXPxBzGbDJ4wlrm1wYr7IwnSBfM7m2GyRK6stDAGZFHzllWmavZBi3uHDz7eplnIvjbGfePQTTHCEeBwv+n6fvRdTCaXGxlwIQT5nU/TssXF+WGXRo7qWo8AoMbs0XyJMJHPVPK6t9Y0yBbVSjoLn4LkWt3d6vPfhKv1Bgj9I2GgEXPyyzs3N/jhJ2wtifucna/xkZYtOP6JSdDm3VKPTj9io97my1oZh4tY0BO99uPpC7H4Og4lHP8EER4jH8aIP+mzXj/j1H1xjYaaIbRlEcUoYZ2Om0kdfbB9ozEdspsdptnIUO4LHxf7K2Xp3gFCCE/NFmt2QNJMEg4RWL2K67GHnta59nMhxP1jPNUFBnEpKnsPN2z1ytsW5Jd1xLAgzinmH6UoOIQTr9YBjM3mSVL7Qsge7MTH0E0xwhLhf1e+DwiC7P7u23eOTy9tsNAKEgOVTNUzDwDAE3/vW0jikcD9j/rhtJB+2m9mTxIilNV1x+bXfvUqaSapFh+sbXZrdCENAmKSkmcQ0QRiKONHhGseyMAyQEqYrOW7vBCBgfadPnGbYtoFhgGOZw+7hGdvNAUsLpZdCix4moZsJJjhSjAzvbuyOpd8vDDL67Np2jw8/3WCnMyBOUjq9mJ9e3mEQxXeFFO5XfPagwrVHvZZnKQVw/tQ0f/qXXiXvWvQHGafmSyxOe1imSZYpKgWXnGMjEJjmcBcyrKxybC1YtjiTp5Cz2WmHJInkmxfmcB2LIEpBKZQCP0yold0XWvZgN34uPfoXlXI2wfOPe3nRtikeqOh5u97nx59tst0KsC0TfxATJxm2ZZKkkp9eaXBmsYQhzHHT6nvJRF9ZbfPRF9uPNb6fRmP5R8H5U9N7kqR/6x9/Ss6xuLXdRwGFnEWjHWIIQc41SZMMULgFnax+940FamUtkw26b+03zs/yk8s79AYJjmWyOKOTvOeWqi+Fvfi58Oh308/e++AG3//R6pElmO5FbZvg5xMjwxvFKR99scXn1xrYlkE3SO5Lj/ytj1ZZ3exxfLaAVNBoB/QHKUmqm6eHUUqUZPSCBNcx+OTKzniszVTzvPvWMX7l3dOcW6py+VbnSMb34+4InhYKeZtSweWVY2Vs0yCMMwqexfJSjW++Ps9U2QVhIoCFKY/VrT4/u1pnuuKOd0S1sscf+dpxTs2XybkWF07V+PYbusH8i0pR3Y2X3qPfX1F48UodP4yZrnjkc+KRE0z1djDW0a4VXU4fK1HvDPi9f3wbIfRu0cvZzNY8TsyWXkgv4HnCi+ZVJZniK69Mjz3hzYaPaxvkczbrOz5+mGAZBkuLJa6stun2EioFB9exKOQs2r1orJGeZcMfpGK7pWPHrm2Mk7S778fjJFDvdY8P+tzz9DzKeYeNuk9/kFLI2Zw9UaVSdIjijPmpIudPTmEaig8/30ZKRbXoMlXJcflWh29XvPGOKIwz3nh1es+13Ku3wseXtinnnefi+g+Dl97Q7x/4qZQUPT0wKkWdaLpfgumgAQ16ld+o95kquQgBn1ypEyUpXT/GNgwkkLYGBGFCzrFY/ahLteCiEC/EwHiesHuxNgR8eq3ODz65zVfPze7Ro39ejM9Bxvb0YplLN5uYhkE5b2ObBv1BTKsbslHvcWu7q9UZlSCMEtQ9um0MooyfXakzW/OoljVLZOVmk9/75DZnjle4sd5jtprj+FyRcuHu8X2ve3QviYXzJys0OtGB4/95kGOotwPafkSaKU7OFQFBvT1gs+WzfLLKVsPnp5e3WdvuMzflcXqhwrHZAuWCSxAmXFlt8+59JLAPSkgnacYnV3Z45/WFZ379h8VLb+j3P6hCziZOUvxdDYj3J5hGk2Ftp8dmI+DUQon5qcL4gY7irWkGQkjavYjbO338QYJjG3T7MTnXpuBZdPoxt3f6BGFKNx/z1tnZ+06i53WgHCUe1iB//MUWG/U+/SCh1QuZreaZKrmsbvRIUrlni/08GJ+DjMNcLc/lmy1KeZskg0LO5CuLMwyihE+v1jGEltJt9SKkEvftqhREKZvNgHo3IueYVAou6/U+zW7I0kIJP0y4dLPFhVM1ysO4dKXgUG8HfP9Hq/SCmDRVWJbg5maX776zdE9q53sfrvLmqzN77qltGc+cdjnCldU287UC02VvvFOKkoy8a9LoxigpGUQpaSrZbgZMlXJ0g5gLp2qU8s4DGURCwAc/26DZDUHA4kyBTi8iSjJ+cmmbME7xXIti3ubjL7b43rtnntKVPxxeGkP/oIrCNJOs7/jUOwE77QELUwWUUnuaWb/3/nV+dq3BdivgxGwR29ST78pqi9s7PiCIk4T1HZ/TixU2Gz5BlFAp5pBS0gviMY3LSFOCdkKrGxKECa+dniLJGBe13GsSPc9ewVHgQeJc+5/jdMXlky8bTJVcwjjV3Oa1NoWciWNZzE97XFltAzw3xucgyuO1221ubHYp5mxKBYf5qSkqRZf1nT7FgoNtGjQ6oa4EVWrc7m6/wRfo1+M4o1xwWdv2CcopRc8hSnQsv9WLidOMTi/krXOzCCF449VpPv5ii82GTynvUszrhtibDZ+Pv9hCIe5anFrdCCnVgWGLmUqOINKhkmOzhUMZTdjtRPXxg5hC3n6s0OZoURVCjHcwH32xyWYjwMspbm71iCKJQhGniv4gYbaqF4WlBWO8AO4+J2GAkvpmr2512aoHmtWjJLd3+oRRyrmTVTp+BAjCOMQyBZ982XhuW1keytAvLy//68B/BDjAf7mysvJf7/v7MvA3gBqwCfy5lZWV1hGf6z1xP+NxbqnKb320yu1tn3Leppx36fkJt3d6fP/HCcdnCrx+psZPVna04R6kCBSfXWvQH6S4liDJFDnHxHNN2v0If5CyXu+RpgqFIIxSBmGKQCAlwxkqsAzdk7LRCQnClGopNz7ne02i/YbpqCfG08BBxnq0c9ms95muegdeN9ztlb/34So5WwtZrW33CaIEUwikNJmfcri11SNKJKW889xwvvezVa7dbvP+Z5sUcjb5nEWUSD78dAPegHY/YnGqwPG5Il+utfniegOpwLEFAkGU3OlnCoCATCqUgDTT3mowiFmcKWKZgs3mgBOzBXpBTKsXcWOjy/e+tcRMNc+Xax2KnkMmJbd3AsJYFxT97FqdP/Tm8bsWp3Y/wrEMLt1o4ocpQiiCQcwXN9ocmy5wfL5AkmZcutni1EKJmYp33/symqdSSrabPkIIegMd2mx8Gh7Kydk/toRgz3l3+hH1Vsh6vU+7r+mTxbxNpgyCQUKzG3JspsB6o0/HjyjlLT66tEXONljd6hOECe1+RKXoMIgysjQjSiXlgoFlGiilGTw77ZCF6fyYEdXuxyxMFZ7bAqsHGvrl5eXjwH8GvA1EwA+Xl5f/2crKyufDvwvgHwH/u5WVld9YXl7+q8B/APwfn9xp78X9ElDvvnWMasGlm49JMgBJ0bOYrXkUcjanFsvjJE0p77LVHNAPUoIwJYoT/IH2rjp9cBxBkihMA4IowzQMlEzp+DFxKjGMYdNhQ5BmGWmmMAyBYQhWt7q8duYOJazdj6gW72+YjmJiPE3sTlDnbAPbNuj4EY12yNfPz+pk4Vqbz280WZzO4+VslBLkXZO8p3c5G/U+aaYpcoszBaRUZFnGZnNANDRMCEHHjzl7Qpf9+0HMidniPQuHnmTs/l7HHiX4Vrd6fPTFFpYh8HIWfpRS9GxQFp9erVMr55iq5CgXXL6xPM/ZE1X+v799Reu7RNleI88dL187EAPyroVUBlvNAbWSTTHvUit75HMOtbKHZQq+/6NbNDoRYZwhpWSnE+JYBp5j4ocJ262Y6YqrG7Nzh0oZJSmDCBzbwhCKm5s9Wr2QWtEhU5KbGz1OLZQwBNzY6PLum/dXexzN09XNHjnHxnVMojil1Y1YWigdSolzvyPQ6oYIA+ZrBeIk4/PrDeyhlHGaSuI0IwjBtiymKyZZJrmx0WW7GXBivkS3HyEEfL7VJ+9ZdPoR/SAhjDLyOYsgTLEsvbNfnClCvY9lCjr9iPkpb/hAFK1eyEw1x+9f3AB47hyxw3j03wV+e2VlpQmwvLz8D4A/A/wnw79/A/BXVlZ+Y/j7fw5U9x9keXm5esDrJx76jA/Agyr4FIK3zuot7KUbTRzLxLEN+oOEfM5GSkWjEzJVzpFkGUGYDieY7h056lwWxXrWyQyMTJGSIYE4jSm4FhJBmmYkmUJxpxFxtWAziCWDKKE83NYbhmCqoj18rbXh0+qFY8O0m0ExmhhpJmm0Az5qbzFVdlnd6PLKiepdXvPTjPePDN3KaoMv17pEcYprmdTTDCV0iKHrx/zjH1zHtg0c28Qx4WetAbZjsjidx7EMOv0IP8yYq+WZn/Lo+BGfXa/T9SP8QcbitFYh7A9S0kQhDMWn1xqcWigxe3rqnpzvYzP5Jxa7P4xGfKMb6mddzpFJ3QQjyxQKRRBm/Mq35/jw822u3e5QLbokacYgycjnHIoebDb8oYNyN9IMeoMUywQlJX4YszQnuL7eIckyojjFNnVV6E47QGaSm/WAgqeTwUmmyDLFVCnHDy9uUMjb1FsBhbzDidkirxyrsLrVBxTNboRtGcN2fDaLM0W2mwHrdZ9XjlXIe/YD7+donvphStHTpsexTfqD5J67r90L6UG7wYXpgh5ztslHl7bwg4RSwaaYd/DDBNPUgmdzBRfTFDi2SdePOXO8wlwtz8pqi61Gn46fsNXM9DMSWi4hjPU89XI2/YHO6WVSkaYSIfROPZ+zMYRACMHGTgBC8enVBqvbXX75F5aeG2N/GEN/DNjY9fsG8M6u388Cm8vLy/8d8HXgZ8C/e8Bx/hLwVx7tNO+PB2l6CBQXv9whzWCz7nNsNg8oCsP3V4su282AOJHYlkk4bDufZffuTLz7L1JBmGgPXne00e3MEIpMShami9i24KMvtuj5CWkmmSq7hGFKz490txuhPf/pqjc2FrsnRhjFXLmlKWBJmnJ7Wz+6QZxiGQadIOLr5+eYq+WfeLx/dzhps+EzU8lxda2LkopOP8YfRESxJJW6ylCgJ0+mBLZlsNkMybkW1ZLLTmtAlGSUCzaWaZBJyfUNrSEeRClpBnGSDnMkkCQZpmni2oIsley0Q14T4p5dq56kXsuDjr3775v1ADHc3eVck4XpAijYbkecWihxe6fP5dU2t3Z6VAsWBS+HZQo26v59z0Ep8ByLKNE7zFQqen7E2nYP0zKxDUHes/nZlR1My6TeDqh39GJjWQLXMRlECfVOyJuvTDNd9RBCcG6pSsePefNVl416MHYgFqfypFJR8GxOHyvTHyScWizj2uYD79donhZylpYncEziJKOQsw+suN2/kK7cbNIbLgq7GUVhnI3164/PFnAdCz/QRjrvWvSChFLeIc0k9VZAEKYkiaQXJIRhQsePSRJJpp3zYWJktJVSyM4Ayza4cqtJuxcjM4ltC8LIRhjDvAmQKcnSXAnDgNvb/rio7XnAYQz9Qa0Zd9s5C/ijwB9ZWVn5aHl5+T8F/gvg39z3mb8G/K19r50AfnCIc7gv7lfBN6Jf9YKEct7GNAXX1jvMVD2+fn4OgKlKjkrJoRdEeI6FEBDF2dho34sBIYYJMwCp1K6hoV+3hL51n9+oMz+VJ04klWIO2zSI0ww/SvjplToFz0JKiJOMenvAdDmHbRm7StAVl262iBKJbQoiKegFCY5t0vNjTEPvTm7v9JmfKjzRROTucNLl1SaNTsgfrGwP771JszsgTtTd90wBqSIYJFimQZJkbDZ9rS0+FOuaLnu0eyEdPyZNJfmc7ghUK+dIEkm7HyGVwrUhSRUgGYQxN9d1yOEgzvdHX2w/sdj9/p1kpx+xvtNnpx0CsLbTp1rUidYwyfAcEyUVt7b6CAGvn5kesrckmYRXjpdpdgeYhkmUpOy0ItJ7+xpjBFE6Zu00uyECyJTCyBSYBlGccW29y3QlR5bJ8Y5TYNAJU1JPcWLOJc0kKzdbOLbBzY0uhZzFdNXjwukpQNMKgyhlpxXo+aEk1nDsHaZadjRPq6VRbiUlk4rTU/k983XkRPzsyx3twLkOizN5HNtEKcX6cNFf37mzC/74EsOkrIEQgvnpPN1bEa1uSLWUo+DZbDb7YBgU8w5S6d1Sq6cT4Pe7z3EGSkganWG1bc7EtizCOEEiSVM4vaiZeQVPz72yUly91YZvP/j5PQ0cxtDfBr6z6/dFYH3X75vAlZWVlY+Gv/9d4B/sP8jKykobaO9+TedwHx/360H6/sX1PfSrYt5mECe4tkkp7xCECUII/ud/5FWur3f58eebZJlEGAILiPYHSXdhN9d599vE8P/CRGGkCqkgTvQWWEqF69gEodYnSbIM2zTIORZezsaOJcHQy/hffvccl2916PRDOv0YUKSmQZJlmKb2jlv9iFopR8mztfc3vKVPKhF5ZbWNlJK1rT4dP6bXD+n6CVHyYIuUZtDPsjEzSQ39hUwqgn7CIEqZq+UR6K1z1B2Qcyxc20RKRZpJpIROmmCZYNsmQQSfXW9y+WbjQO3wx1VwvB92H7vTj7iy2gIBs9UcUZJxY6NLqzMglTqv0/EjpIRS3uLUfJEbGx0uXqnT7keU8zYnF8rkPYtBlA2pluGhziMd5p6SVO6hDRtI3Ezq8Qz0wxjbNMk5JpalFwCpIE5TnVeSOtbsWiaztTxTlRyfXW8AWh/mZ1cbGAK+em6WZjek1YvuqmW4H3bP0yiRY3LBTMXbw8+XUrK62WW9HuDaxlDeoIeJIkoVUZpx+ZZDreSQcyymKjk+/GyDYs7i1nYAKKzh/esPErpBxJW1FigoFx0qBZdOP6brR0TxPeJiQwgBjqkZPUGUYlsGs9U8SZYRximOZRLHCdWiOzbyw0/elyL7tGs+DmPovw/8x8vLy7OAD/yrwL+z6+8/BGaXl5e/urKy8gnwrwAfH/mZPgD3quA7iH61ttXhp5fr/OCn68xUXH7xrUXOn5pmquKRZJpyee12m35896Pa7eGbArLdbvyuH0dRn9ECECX6h41GQM7VmXqldNVjoCT9QUo+TKmVXWSsO9hfX+/SbAd8drXJINJ5A9vMyBQUhEkmFX6QgNJbC8+78zgfx5iNBuHlW02urnXYavokqaLoOXg5k0reJpOC67fbhAfcowdBDu9NJiWWIUgyqR1+qROQgyglSyUIsCyBVIpePx7fU2N4X5VSWKZAAO99uMpUxbtrDDxJvZbdx17f6YPQ53V8rkiaSdI0Y6sV4A0Na15YWKaBaRj83icbJJnEMQVRktHsDLi52UUq3WTjXnH5+2H/k5AwNuYApApTSBzHQESCNFO4tjGkbApa3RDPsYYSAjYL00UAGu0BCzNFvvLKtBb9QnBspvhIxul+vYJHVairmz2CMKWc10nfKMkQwHorwLYNUDqs2unHLE7nMQ2fbj/GDxJytkEQpqzV+wziTM8x7VWQSkXakXT9mDDOUPLOPD0IpqHzbKZhkHMthCFA6d17EKbDbTu4rsnqVg8hBNWSS5xI+oOY14Y7of14Fv1/H2joV1ZWbi8vL/+HwD9D0yv/m5WVlR8tLy//E+AvD8M1fwr4m8vLywVgDfhfPZGzHeJhVsP9Hl3Xj7ix2efkQmlcvHT5VoepiuZjK6UwTAPLMsnZkjDRx7HN4aBQ4FoQy70e/WHNXSbBH+hZbJniTpm7hP4gJU4zSsNE0o8+36LTi9CpOzAEIPQC44cZTiaZny5QzltstgIsW9Dph9iW+cjG7PLNBu99uEqrO+D6epf+ICJOQShodkJMU19swbMfycjvRppBmt05Rpbp5yOVQkkwDOj6CVJCojLM4U4AAYYQpKliQMrJuRJSqgNDVffb7T0udh97px3uqUi9dKNJwbPJMkU/TMc1GWGU0A9iglhiijsMmieBoV3ag0zBINLfDRCkepG9vNqg6NlMVzxsy6SUt7l0o0l/EIOCX3jtcF7742B3TirNJNWSQ70zoBfExHGGlGAaBq5tEiUZlYJFnGR8erVBux8SpRlzVU+PzSQbOgPsGadZJhGJHNcp3AtC6O/KD710xzYAxSCWBFGi1YyFQAnBhaUaYZzR6oVYpoFlCRamC7z92vyBx34WOv+H4tGvrKz8HeDv7HvtT+z6+UP2JmifGA5aDX/ro9V7ygvs9+iu3+5iCDi9WNlTvPTrP7hGvRMRxQmOZTCIMwxhYCCR6MnoOQZ+JAnT+5zgQ2C3kYNRbF/R7evGCUmSjXnUjmmQSh3LNYVgFAUvFRwc22JpvkSl6HJltc3Xl+ceyZj96NPb/D//xy/o+wlBlJBm6i4jlA2vPe4nj3rZ94VpgG2YKMA0BHGS0R8kmjFh6QUxSRXp8MRsNCuiWnTvqR1+Py/ycTE67srNJhtNn9s7fYQBtzZ7xIk2ogXXwrYt6i0fP0zHXmT2hAz8CHKYCD+o+Eru+z2KU4JBShAmfOdrx7ld9/EcE8cyUainUsy3O1mbZYpekJCmEn+QkElF3rGYKrnUO6EunksypBL6X6mQmaLZC9luDYbhrLsxyp89aHHVJAId9rKtoUMmoJQ36fWH55OzWJzO8+bZWZTSjsaZ45UHOp/PQuf/hauMvUu7JpNcvdUhzSSz1fyesu7RBN/t0cVpxutnpsc6N51+xK2tHmmmKHgm12636AUxKlMoE1zHJEkzEBDE2hN4Uh6Y4I7HKgQkmSSTEpmpYaFNSpzqeG8hb5OzLZbmy3dVJ7771rF7fsf+3ZBpKD6/3uJnV3e4vtElS0EY99/SPklkEkxTYA47A51aKLHVGNAN9M5GJ9v0A9DJcDGWnngW2uEjx8OxDVqdAUrpiRyGKRJFztZcdTlI6AfpeGf2pMbQfiggZ2snIUn3vj6CQHPlpSVJM8XNjR6vHK+gK8FTzi3VsC3jiRcDjZwy2xL4YYw/3AnlHItBnGLbAqkEhikwDUE/1GFLPS4EOcfCsgRBcDTNQpIUMpVQLToIFKYhsEwTz1Xk8zZnj1c5e7I61s35+vLcfefeCE8yb3QvvHCGfv9q+OWtNv1BgiEMinl7T1n3SHdit0f3/sV1ol0B0I26Tzjkeo9id3EiMUxNB3RMncDp9mOEAsuC+Ig8+v3QcWpNzXRsg0EsUUoRZxlCKBzbwrYUlmUyP5Uj7zq8feHO9jAIk3sOloPUNm9stHn/sy2qeZvbOwFZNqRTPSMjDxCnCqkyco6BwKSQc5ipKrpBTM62dBhLDlkSShdWnZgrUO+E/PFvnXrq5ztyPJqdkBNzZW5udTFNbYwswyBMJaYwUDIbJ3ielpEfYbgsYpnaIJqGIIzvPGTb0lWfrmURJpJWL6LrJ4RxSs6x2Kj7LEznCR+QuHwU7Hc8zp+s8MOLG2NtqTjJ8AcppbxNEKU4tkExZ5OO6MyZbh5uWyZxljGI1dENX6FlNTIJzV7E8skq89PaoRoVaXmuNW72cthQ6bPQ+X/hDP3+1XCj7mObBl7OQgjNC1bK4cu1Dt874PP7b/JGo0+9rdkd0+Ucji1o90Nsw8TLaZ5vFOtkkGnquPGThCGgUnZxLItBECEznQwYxBlmkpH3bKbLLl7O5uR8kSBM7ioQev/i+oFqg1fX2oRRyk0/YnW7S5pmlHI2W60BoDCETnw+a1gGTJVzBFFGoxuSdy0WpjwkWjrCy9njGOpUJcdMJX+ogp0ngY4fYwi4utZBGIokkczV8rR6EVkmUSol55q0uymOLcZFd08LptD6So6lyJQAqQ29uWfXps8pTDK9Myo6GAbMTxVIs4ybG11+9uUOZ45VxsV8R4GDwrCXb3Uo5B3+hTPTiCE9udOP+OTKDp9dq7O21SMIdVixUnDoD3TFuxyunskROmFS6X6yeddkkGT8wZUdzvRj3r4wS7OlpRbCKMXLWczW8mMZjwfdnyeZN7oXXjhDv99Q6yIlybFycde71IHkf9A3eeQ11DsR282A6UqOOFXYlsFUOc9MJaQf6BJoIQTHZgts1gOCKCYIn6y7qxR0uhGpl6FSXXTkOhbGUNukmHd49XiV775zEmB8HTMVl9fP1Lh8q3NXNt+2DKSUbNR9Cp5NydPe0dq2z6vHywyiDMcy9ux0nhX0dSqyDI5PF5idyhMlGSfmiqRScXOjS5Rk5BwTIeCNV2aH+uwPLth5EhAoPr/exLJ0ptw0BVvNgGrR1ZO46DJdzuFaejdyc6N7KG78452TrmkwDKgWc+NqbCkVQaTL+0fzw4Chh5+CgqmKx2zNGxq5mHpHUzwtS4d/jjJW//GlbTZ2fFIpx+HHoqerc/eHNvpBjAIcx8QwdMW6aRpEiRyqEKh7xuUfB3EqSTNdCauk5t5/8GnGsdkiXz8/x+fXG4RxRrXojJuSHOb+PMm80UF44Qz9/tXw+FyBTj/CMoQOcyQZ3SDRVLADUG8HXL7V4fSxCq+dsfhnH9+i0RngOpZmlwjBdDmPY8cIoWNyJoKcY9DqPfmYxih8EwyNr2saVEoutaKLMAR/5GvH+d67Z8be0Og6BlHKh59vc2qhdFc2//NrDQqeTWHoCUeJ1NWBmeTmZh/HEliWPeRgP2OPXui8yNfOz+rdhYILp6eIk4ybmz1OLZaptwPSVCEMQa3sPpP2dqOQw8+uNWh0BkyVcvQHMTnbpBfo/q151yKMEla3EqbKLqWCZlNtNgZP9NwU2kAVchbz0wXmah551+JXv/MKH3+xxa//3nXCOCFJJYM4Jc0UQoHrWnzt/DSlfI5q0eHHn2+RZYpiXhthKXUo4yhi9fV2wCdXdpgquRQ9mzjR4mjLS1UKeZv+QCf7PdfixkaHKM34ypkpeoOUdlezW5rdwTixKp+gjzJKao8KKcMoJcsUm40A0zTodSN+75N1Xj1epVZ2x57989AbYYQXztDD3tVw1Iat20voBbog5PhcgbcvzB342ZG2+Ug4q1J0sS1Bt5/QH6QUczaLMwUWRYG1nT45y6DVTwii9KkmKJNUkaQZjiXoB3pSvn56arxXOcgbklLR6kZj/jPoiaLQImpzUx43Nnr0ggjDEFSLDu1eQilnaj7+IWCgw1dPyiu1DIP5qTzfuDBPEOrCttEu7tRCiWYnxA8T+n7M2RPlcbHN05xEu0MOBdfGmTbYaQ/wXBPXtigWbLr9BC9n4ToWx2bzeI7FdmvA4rR2TLJMoZS6q97CNrVa6uMizRSZVLR7IWmWce6kvkffe/cM3SBhu+kjpRZGa3RCUik5Plvkf/ads8OCpoxa2ePkvN7VRnGK7ZhHxg65streU8nqOtoU3Vjvjbs8jZy5JJHMVHLM1QpUIs0MavdC2n3dLN01D1D5PGIopRVFk0xXyCqlaHZCLEtQ9BwEQit53mgSp5IffLI+zoU9jKf/pPBCGvrdmKnm+eVfWNoj5WsIcWC8rN4OxtrmRc8kTiTNdkA3SJBK8crxCnEsCRPJ/FSO2zuKL293UEA/OPoM7P3kFUaQSuvlDCLFlVstFIpeEPHjz7c4MVugVvbG3pBjGbT7eyfhIEo5e6LC5Vtt/EGCP4jxh97S7FBAbKcdkiiFk5nYpiLONNPHHNYOKDUMqcg7O44nAROwTJifzu9JcO3exdmWyfml2jP1kHYzvwqeTZKaFBZsbMvkwukpgjDhxrreNSapDpn5YcpszWNpvsTidJ7fu7hOkkgUEnsYgsgPNfZ7gwgkpOr+XO/9GHmddzTr02G1sUfXT8bx9ROzRVzboN2LsSyTV45XqZYcZocOVLMz4L0PV9ms+7iONey5Kji1WD4ydkjHjzm1UObLW22CYW1BP9A6UL/0jWN3ESg+vdogTjLyOZtzJ2p88NkGpiHIOSbhsCr7MPPpcTAqesykotmLaPZi8q7JqWMVCjmbVCrqnQFRJHnj1WmEgJXVNhdO1Y5sJ/SoeOENPdwx5o1uyGzVGycn96+iV1bbOgQyTFD1gohb2z7+IMa2TEAwN5Xn6+emeP+zLZrdaNi05OjP2TKhUnBpdO/vHRm6MoM0zWj1FJ9ebWohsCjl1nZfC6rFmmuuhv0w9ydov/3GArWSw//w21/ihwm2VlyjF8TUCkW+cX6OXpBo7ynN2GkN8Ifc9WCQMBzfmpNtMC5EOdL7YcD8lEep4FAt5HBtc0+C6mnHNO+H3cyvxZkCV1ZbQxXGeLxAFfIOnmuRz4kxlVcpRasX8crxKh1f79J22gFCCTYa/aGom8IwDNJMPvQ9tkzdQa03SHBtA9s2MU3BTmuAaQr+/m9e5uvLc1rpdNiNavc4ObdUHYc2Ty2UsAzB1dsdBnHCNy/MYZmH17V5ECoFHdNenCnw8eVtXYjoWMwV7XEBIzB24Np+RJpmzFY9TEOgpMI0daxeDUN8T9rQ74ZtGgRhRieTfLnW4vzJKfqbPUzDwDS1YNsomby+47N8qvZMeiOM8FIYeji42mxUCDVqoLy20+f0sRIrq22CKOXyrRZBFCMMwfG5opYgzTJ+66O1oT6NHBraox8+MoPuAQU+u7/NNgW2qROpQghsS5BJSTeIKXoOzU5E129yYq4ISnsThgEff7GJEoK8Y/HqySoAmRR8Y3mOH/5sg2wY7jENwepOH9MyWNv2GUQJ7X6ot6GJRArN6x95iLulHR4XI++zlLco5BxeOVHBtnTY7XmSdz0Iu5lflaLLuaUaNzY62lgNF6grq+17cqU7fsw3lufYbAQUPZtGL8QfOHSCGNMyyNkm0tKxftgntXEfaAVUA9cyQGgjaJmaOhnFilIeoiTj8q3OuI3lQdpQo3m0MF3k7El9bZuNAYszpSNjh4zCcfV2yOn5EkJoYbsRZ//jL7ZIMkXRs1ma183QV1ZbdPoxSii8nEmSpGTKwDYU0lBPLbRqCnQnMIZ1L1J7+Y1uyOmFIplnj9U5RzLMT5on/yC8NIb+ICXBUSHUa2c0A2Wz4ePaBhdO1fjRZ5v4g1QbUNOkN0gwBAiluL3jc2qhSJLIe7J3HhcKrXNjDLfapinGMdtRqMQ0BZalG3j4QTquxss5tu75aerF4tZWjyRVzFVzzFY9NhoDZms6dm1bJh98ujnkJEvOL9WotwfYlollaE3tlZstpFT4YUycKpQcqhs+gsTDvWCbWnscIM0yFqeLFPM63OHYJiXP4dWT1UMLZD1L7Gd+2ZbB4kzxrhjsvbjSoxj4hdNTdP2IL2+12dzuA2Ao/dzlQ8THRiJ6I6Pu2CZxKpkq55gq6eYjYZJR9JzxwtPoHFxYt38eVYoub52dfWAh3sNiFI77e795GUMICp7B0kKNStFFKd3h7SuvTI/Pd2FaS0tESabVJhPFtaxLnKS0+8lTM/I6R6UFyywTkqGMRyFnkaQ27X7MO68vjOWlH1bh80nhpTH0B/HrhRBjQbN8zub0YpkbG13efHWGaskdVy1ahqLva8++I/UWcaM+IJXZoRJjj+Lzu45ADSd1GOuYuG7MobXCp8oO/TAdl2ALoYbyBwaeY7DTDhmEKa5jUvAcojjFME2a3QjT0FIQG/VgLDFbb2lN8alyDsc2aXVD+oFubKxVFRVJqqv/kmFWyzDuGJFRwwzD1LuRh5lXhoC8a2GaBpVyjr/wL57nnTeOP+Qde35wWB60bRl8fq2BAs6eqOxZCD74dJNeEHNjo4tpCKZrHksLJTYaAXEqcWyLKE5JpQ6V3a+aVgGW0PmcME7I0AV3pxdKBFFKPGyUcWy2ANy/3P5pVm3OVPN8Y3lON/Pe931ieJ674bkWH3+xSbno0vEjynmTzWaMZRmoWOI4YJkW/iB9YiEciXYGs0zPD9MA1zbwh41JTFNQK2tZ5Jub3YdW+HxSeGkM/X4vq9XTnX1GgxtgrpYnHHajQenQgVS63Z9pGsSJlh61HYMwSccr9/1gCrBtgyiWhx5cOVtQLeXo+TEgsAytYZNkkoJtszRfwHEsqmlKox0ihKDoGRgGeDmTINSFLXEicIca3fNTBUzTYKvhc2y2MN4yju5HIa9jt70gppR36BqCrh/RHQpGmaahy/MRGIYWFVMKco5JlEiEoQe5iUAJdeiVLe8a2JZJpqBWdF94Iz/CfubXldU2H32xPe72NapnePu1+bE3v/uz335jgV//wTWkVFQKLqcXKziWyfG5kg4LCPjwZxv0BjEyG1Z73ueeS6nHYqWovd6vnp2hUsxxebWFRPHOa/Nj9db7Ge6nXbV5r+979WT1rgVnq+mz3dK70TjNaHYipILZskt/kJEqhUDt6RPxJKBGTCkJOUfv5ka7M9DhuzDOeOOVmWdOqxzhpTH0B3lZ01VvPLhBD/ATsyXefeuYpm+ttciG8e9kmLn3HItSwdESs83BUEb37lFjm1p90rZMMgk5h7v0REYwh13lTVNo79az8QOtymiZ4FgWtm2CgJlKjj/1R89xbqnKR19sYxo6c49S3NruEoQZ4VD/emEqT861dWPyYSA9lbqoatS5Z/d1f/XsDO99uMr19Q63d/qU8w6lvE2SSuJUUsxZxNkwfMQdtcxS3hp7WabQRSuDofztfsEsy9A8eNMUzFbznF+qYZsm5ZL93MfeHwUHVXe+9+EqpxfL91UnnKnmWZgp8toZveMc6dlnUrLTHjBd8Xj1ZJVuL6TR1brqEoHMtFZ9mNzp3mUaOixWLuUwTYNfXJ7DNA3mpwssTOdp+xG1sodS6oGG+2lXbd7r++Du0NfNzR55z2TlVhslJWGipbu7QYrjmLiGQZKkj0yzfNDO3AQwdPHYSONfoRu/XF/vkHNMTh8rH2mI66jw0hh6uNvL+uDTzbsYKKNBNFPNc3yuRLsT0htoHd5aKUcwiDEMQamWJwgzJBIlodOPsW0tNeu5Fq5rYQowDEPr48QZhmnQH0QgFVLpmLshIJGaqujautlD108AhWebWuNaKk4ulMhZBn6YjgfKiJlw4VSN9R2f+akCPT/RzRrKOaaHDKMvb7W5vtHFtUzOnahorW4/5s1Xp++iKU5VPP7Gr12kVs5RLbrkXItstUWjHZBKHWKRUpEisYdNH2pll24/xjJBokvos05IlmplT881MU1Dx6tNwWunpykWbI7PFA9UFH2ZcBAJQErNsZ6furObHIVL9vdAjZOUhekilaLLwkyBj77YRGBQLbqcOVZmEKdUCy4bjT7X1jqsbvVQCAqeIk0VQZRhWYK5iseJ+RK9QaK1kKYL/Mq7p4E7O45WL0KgK8BHu4+DnsvTZjjd6/v2LwDlgs3alm5wnndtokQOaxIyip5FmkmiNHukUOrQp7nrc4a400nOtnWy+/zJKZrdgEYnHGvPx0kOyxJ7HMvnCS+Vod+Nw3gmb74yzRc3mpw57g5DIZIvb7cxFCxM6XLsL2+3kZnCy1lIpROoJ+YKhEP9G4SgWnBodCMqRZvNhtSefaLIuTr+7grBwnSeVi8mTiU526QfpoSxJOdCpeSSJBmWELqxwhDnlqp8/0er9IKYNFWU8g7HZop8950lQHs8lmnw9eU5js0WuLnZY2E6TznvDFkXUDyApug6NifnSxgj4Z4l+CxOafdjXNfEtQ2COCXv2iwtlJirFjAtRTnvsrLaYrMeMF3NoSRMlV2KeYcsk8Sp4tzJCudPTr20hn0/DpKcrRbdA+sZhGCP95+k2biD0/xUgUZnwEw1z1tnZ8YGIwgToiRjplrg9GKVn1za5PKtjm4wbwoqRQMpBZZt6tBfzqbdjzi/VBt/98iQ3tl9WPekID9P2L8ArNxsIoRgYaqAHyZ4jkk8pDHatoUhMvoDcB1jj2jbfowYX56rw6AKxv/thil0Tx97V1jTFIJWN6DV02HXatEl71lkUiF4wjGjx8BLa+jhwZ7J26/N0+pF9IKYfqA9o7PHy6w3ApTSrIVlo8ZWS2uXoMDLWbpl2yChVHBwbZOtRkC5IPEchzdeLXBqscTNjS5r232CMKFWcmj3Y9JMUi46OJZJmkmSTGKaWo2vP0iQUvHtNxb2nKMwGEqxSlCG/p27F7LZap5ffOvY4Vq6VbRswMiYzFQ8LpyZptMLmZ8uIoC56TylvINS3NPze9rt0J5HHJS8rJVdukF8127SNsUe73/k8TfaAxzbIkkkb746vccr9FyLz681eH3IQPFyLl8/P8utrT5r2z1ytoUwIEpSXc1atDAMMRaz241n0fDiKFHIOygkjW5IlmXEqcS2TQwBUZRg2brpupSSrWZAFMs9pIGRxz4qKhtp9Y8is7uF3gwYO126yU2GlbMp5R2iRGIaguKw8tk0BZ5jUSk491HZerZ4qQ39gzBTzfPdd5buMlYfX9pmdaM3NoYXTk9pGddd5fhSyiF9U1Itu3xteRYhxB7vqN4O+PUfXKPjxwRhh5kFjzjRrdGOz5Xo+hGDMKU/0OyZUwtFSgWX33j/BpWCQzeIma8VOLN4x4gEYTKemI+6xf7Ftxb5td+9Cmjtkv4gIckkf+Ffeu3Avqv3u38vgoF4kjgomWgYBt/71tJdPPWPvti+i0kyV8tjWya/8u7puyS0Qe8EFHcYKIWho7F8qkY+Zw2L5WIEUPJsbNvkW6/P7UkOjxbgZ9Hw4ihRztt4jk277zMYaMliyxSUCg4LU4VhN6+ErWb/rkYrhtCGXAiBHNKHDcPAsSVRrMYNSXZDSYlj6Z18rZzj7PEq682AVpDgOgaWZVCr5DgxW8RzLZrd8Jly5e+Hn2tDDwcbq7cvzJGkkqJn3xXff1CD493Hmqnm+dXvvMIHn26y3QxwbQshMnpBzGzVo+BZtLsx89N58q5Bux+zutEb62N8cmWHr5+fBe4Y+qOYmOdPTfOnf0krX263QmYqLt/95smHMvLPC571ruJhkpcj73+3LIJlwtJ8Cbg3A+Xsicp41zCqxI0SOD5bpFZ2xyG7E7OlPYyf/f1In0XDiyOFEJSLLqlUBG5Kkul4/YnZIgvTeT78fAvHMpmvFfD9hDAZqnNaw5oQofNrUilsYWDaBplSFD3ddDxOJKahP5MpzZG3LMV0xeXEfJkT82UKeYcbZgdTCHphMq7E7wXxPXdSzwN+7g39QXjQ5H0YT3Z0rNWNDtfWu1SKLstLVXpBSqcfceZ4ieOzJXpBTKUo9uhj1EouN9Z7fPV8bny8o5qY509Nv5CGfTdGMeeRwNTKzSYfXdrie99aeqrXdtjxMMq5bDZ8ip6DbUI3SGj70ViH5n4MlK4f0epG9AYxvSDh7InyXSG73ZWtsDc88zDUyWe9gB4EpeDNV6f5vU/WKRVsPNejVnKRShAniqmSSyah3h7guBYnSi5BlCFQJJkaN2mfqeRIMm3YGUocR5mWBsm7Jn6U4pqCqZKLP0hJMs2Wi2LNePpDXz3O6mYXgaCcd2h2NZX7e996flllE0N/DxxlWGKmmufP/vHlPYnVwrTDuZNVaiUX17Fo9WKK3l59jFMLZX56pX5P5tDPA+5ncEbN3G9u9vAck6lyjl4Q896Hq0xVvOdu0s1U89TKOXp+Mk6cvnW2gmUaDwzHnT9Z4b0PV5FSsTBV4PUzOYQQdxng+4VnDrv7OIgy+iyStvufvRBgWyavHq+SpBmuYxHFGbZlsN7oEaeKM4tlzp6ocmOjSxAmLM07HJ8rUW8H9IOUQl437pmq5MgyyT//6Ro7rRCUwrEMEqXwXItq0cEwDEqGwanFMo1OyFTF4+zJqq48TjJq5RxK8cxF9g6DiaF/SrhXPmAUty3krLv0MRzb5Ktnp3Ft86l1onmesFuCOskybNNkdbs75uN3/JhmJ8RzzLHMbWnoYT2vCUal4K1zM+MFXb+mHhiOa3Qi3nx1Zk/YZXe+ZoR7hWeEYE/nsV947d6VmldW20gpWd3s4Ye6WXe15DzVe3rQYtPqhghDJ7tvDBvQKKWYrZW4vJpQK7vjcTA3lefa7TZJqlg+VWNpoTQW+AO9Q9pqDPjKmRma0yGfX28SpxkoKBcsFmdKDKKEk3Mlvr48x+pWlxOzJTp+jGub457ULwomhv4p4iBvbTQxR7FX2KuP8bxS344S9XbAx5e2uXqrPZYLePu1eX7n41v8+NMtojTDEFoGIohSPvx0i7denaaY11TCqfKd0FacSKpFl84BgnHPAw4yxFtNn2YnHCfhD/IOD5tIPSg8s9XyUVJXbB7GQ1/b6bPd9Mk5NkVPOyC3tnpEw6LCpxHWGTGEklSyutka5zNqJVd71HFG24+oFhxcx+LsySr+ICGKddW4ZRjMVvOAOtBJ2q2zszBdZGGmwOdXG2y1B/iRJIwSaqUcZ4cVuqNCyxcVE0P/GDiKAT+amEXP5uzJ6nOlj/E0MPLab2/7lPM2IPjiRpPb9R5/cGmHVEo812EQxjR7kRbuMgRBlHJjq4slBP0gptkLCQYplmXy1qtTz22Ccb8h3mr6fH6jyVfOTN/XCFcKDltNn3Yv3uNlu7Z5V4/g/eEZUwhaQUSrFyPQDJM4yai3An71O6/cNcb8IB73XwZd6RwlKX4QP7WwzqgX75e32riORdHTYZovbrQ4d7LK26/N7wlnLk4XiJKMdi+mP9D35/xSldlq/kADvVtnJ0klV1ZbHJ8vYVqCejskk/DK8fKRSjM/S0wM/SPiqAb87rjp86aP8TRwZbVNt5dQGXpmoClwX1yvD1u1CQI7I80klmlgIIhTNeaar252WN3qk2VazdNzHX76ZZ2vnpt5lpd1T+yPkzc7IV85Mz3m1B/Eba+3A27X+/zeJ7cp512OzeTpD2JubHQ4MV/Aday7xuDIuNXbAT/4ZJ2pkoshJF+utRmEKcW8zXbT4Lc+Wr1LmmKkixTF6TgenUlFIW8/NS5+peDw6bU6rmONFxwhdBPwXhDzyvHqnu+PkgzDMA7U2L8XRovuRr2PY+sexIszJX7hwjz1dsRGPWBxuvhShEsnhv4RcZQD/ueJj75/F7S20yPJMkquQxAmNLshO+2A1c2eLhJD910dxBLLBCEMwjjhdz5eRSlFoxtpFUHTIJOCJJXkXJOPLu3cVzztWbJKdj/vv/Xrn3Lxyx36wQalvM2F01MUPZvPrjXo+NoDb/sRm3WfJM349Fqdn15WWAaUig6uY/DGK7ptZpJKPrtW5733b5D3bE7Nl5ib8sbNdjYbuum2OZTasE2DH322RaMd8me/e358TidmS+Qci1Y3YqcVEMYZlinwg4S1nf6YDjrCk+Din1uq8oNPbjNVclFKCw4O4oxi3iJN9xLePVfXE3z7jQU+vrS9RzH0fhgtuv/dr39Ox48QQrA4U6BUcDk2WzpyaeZniYmhf0S86MUnzwK7d0GGgE+v1bmy1sYxDFxXSy8rBe1uqMXShEGa3ilR17VEEtsU+FHKILzTx9cUCmnoxiwAq5u9Q53H02CV3GtRuXyzwWfXGjiWSaVgM4glP/jpGpWCZoV8fm1Hs64GKWkq95TppxJa3ZifrmxzbKaIP0j57HqDdj/Ec2zmah43t7p8fqPBt16fpxOkmnZoaamEdi/i9GIZxzK4ve3vuf5zS1Uan4ZUSw69INJ6SFIxXfW4sdHFtY09fYmfBBd/pprnq+dmx4WLhZzNqcUy12939zZK2Pf9SSp5/ZXpYTP5Lp98WX9gGNRxTBZyBUpDMcBLN1ucWigxM+xy9TJgYugfEc+y+OR55Djvx0HnuDvBNoq9npgtcm2tzfXNkJxtYhhi3NM2k1oa2dhXVZ5mimRXD99MgiEUBhI/1DHXgnfvof00pQAOWlS+/6NVLMvgd3+yRphk2IbAMLQ8QqOdcWu7x2ajT6MTESbyLrEtY/iCBIJQ8vs/W+f4dIlBlGIJk0wpwljfh16Q8PHKDr/09eNcXm2RpJIwTqmWtEZRkmhvfX9PU9sy+P1PNsgyxcn5EudP18bhshsbXcoF94lTfg8qXCyXbJTkQMrxQeNrquSyutEjSeWBC/mV1TanFkqsbfWJE4ljm0RJxo2NLu++ufhCzLXDYGLoHxHnlqp3Uf9GUryHwf4BNF1xaXSiBw6o54XjfD/c6xwHkZZX/r2frtPoDDAtg1rJHXZGEnT8eNgH9E5jEwNdui53KQMcJBuVZGBkOtEYygzHNMZFSPvP7Scr28OuRrrStFJ079qNHdUE37+oJKnkxnqHjp+QJhLXMugFCR2/g2PpEEU0rNBUSuoevfv0uaTS90Wg71O3F1Mr6ubaQoCDwe16n6lyjrxr0u7H/OiLLSwD+pFuGG5bBltNH9MQvHq8iudarG71eO+DG3xyZYdayaVadCkXnHHzbbjT0yFKskOHSB4VB/H+R/ProFqAj77YplZyWd1sjWP7Sulk6r2ac3f8mPmpAp5rs1H3h4lcm7ynn9fzPtcOi4mhfwwo3W5Gq9YJNW5IcBB2G45R3NUbxkEvXtlmuz3gG+dnefVE7b4D6kUQphqdY5pJVm62hl28DHbaPo1uyHZjgG1p/dd6e0AUZ6SpxDAEjmVgWiCHDvuDGm7shkQn7HKugWWZ/P3fXOGVE7WxoQY9cR3LRAg1ZluM+pSOdmP3WqhGfVYfxvjvD/Ft1H3iYScv1zVp92JtzBNFFEv8UOupm4a+bFMIDmpps3uoRUlKoxNhGlqIK8sgTeXwnhrkHK3AGBvGuLdCkkqiRHcWOzZT4Opai0Y3ot2NmCq5CKElOWzLJJ+zWN/xKRf0vSgX3HGI5EmrYN4rf3U/iQk/TCkOd3RxkiEE3NzostMOAfY8t9FnKkV33MQ9CBNc23wh5tphMTH0j4grq20Wpgvj7D8cXMAC2nCMqmJ7fszqVo84ySgXXI7PFocSp/DBp1tsNgbMVL17Fqi8CLmBtZ0etza7XFnroKTCNAVxIml2B7iOjWUKBkmGUtogxYk2ZTJTRElKlu7VBt/v0d4PxZzF8bkiYZJxe8fn7dcWxobItgyKns2Z42Uu3WzhOQLHNrmx0WFxpjgOPxw0wUcVt2++OnNo767eDtis91m52aRWynFstoAfpsRJSsGzkVLS6kXEcUYUpwhDYA5bBo5yD4fpW6bFuQRpZuEPEtJMopTi9k6GacCx2SLHZ4us131OL5a5tdFlo+kzVfYo5W1a3ZjbdZ+3l2dp97X3K4Rgtuax0wpYmi/TH8Tj3gb7VTifFwM4YtFYppYrEAI2mz5dX9NFS3mbemdA49NwTz7iXrIQox3Cbjxvc+2wmBj6R8TDGNyPv9his+FjGgYdX2vSB5G2Zm7Hwh/EJGlGKiVRokWvdheo7MazFqbaszMZdrXa3Vyk2Rnw+5/cZqPhI5SBQo7bNVqGQZpJ4jhDCQOZyXE3K9D/xgd06DosHMugVs5RzDuEYUIvTPnJyjYC/v/t/XmQXVme34d9zrnb25fcE0gklgLqVVVXV3VP9fTCmeZwpptqikNxxFlEkSPTok3Rthwa07YcoQgyuNm0GQ6FSZmbKY1CJKUY0+RIIhXN8bSnZ+2Z6a2qpwvdXcADqrAj98y33vvufvzHee8hM5EAEkBmIgHcT0R1I9+Sed599/7uOb/z+32/hLEO/G+cmeDkTGls5tIfhKDYEbDvrPfwBjFeoJfxJ6aLbHV80lTtO7iNVgUT1RyuH9EfhFy6EQwN4BWlodztwnSRK3faehYPGKYgidRjuSSFUcrddZdSzsI0BV6g3yxIMW1Jq+vz7Q9WsW3Jiaki0tAlrF0vwB2EVIo2c5NFwkhRzFmEkZYXqJWc4SawloJ0ht4Ge6lwHocAOEr1vHdplfc/3CRnSzr9AMvQGjeVgs3NlR6n58o7JCceJAvxrK+1gyQL9E/I45wEH97pUMrbbHQG41nlwI8IQ2383fNCbNOgmNOuOdsbVHZz1J6e29me0jAkfO+jTaSAN85OEkQJX/3WLX5wbYPNtj+0ZkxRCqQhCfwU04REd5ljmSlpuu+szL4oFUwqRRt3ENLqhczW80gBt1Z7pKl2+3L9iMs3W5yc0sE5itMd39lG29P1+1LsqMLoDwLmJ0o7/t7Dgtv2VcEo/9vq+ViG4MJilW4/QgrtVeyYJigtie0OQtRjmlekSu9htN1ox+MKGIQpcRpRzCsMafJ7F5foDpuRdHpI0HZD/DBhEMb8yFsnuXxz1KENpYLF/FRpx43wOAfAqVqBL33uLO+8PsuXv3Zt/BkmKnmKeYsgjNnq+FimseM9e92sn+W1dtBkgf4JeZyTQBeN6EqInG1QytssrffpBiHutZAgTMcmw9puMB43qOzmcWRxD5rtwevyjd7wwhasbHq8dmaCtZbHjZUelmFQyOmLP0kUaZoOj4G2V9SdmQc/viDUiqCDob1eqWBz9baWk5ZDRyGUft17zTUmyjlavYBBEPPffPkDzs6XWdsakKYpYQSOda8Ko+9G1M/sXME9LLhtX/GN8r+dvtbgKRdy+H6CYUDfTyjkJGASx6neVN5mgPG06FWSwokUnorp9iPS4eNxoquVDBPWOx7tfkjONjkxWeDaUo92P+TCQoVXT1V3nF/PQwCcqmlPXiEEcaLGTVe2ZbDV9Xc4cD3sdzyra+2gyQL9E/I4SoAIuHSzRZoqkrxFGCeEQ4MJpfSFHcSKW6tdum5A1w2ZnypqS8AH/G1gnEK5equ94/HDYnvwcn2dywXoD3S+ZaPt61mqJcgZJlGsgGS4+ahveKaxt9n6QRCEKQgQQtF3dTB3DEG9kiNRgiBKmZ8q8sGNTfqDCIRgtp7Htg0+utOh1R0wWc1RLebo9H39PcVaZfL8qRpSyvvK+k5MFe6TINhr2d91A7730SblgsXibJnpWp7+IOLVU1X+5W9c5cNbLe1drEZpl3tOSE9zuOTQmL7n78yJpdv/P1GgJPMzeW6v9FjedDl3osrnPj6PbRlcud3ZoQb6vATAatEmjGLurPYBsC352LrxL0oz474CfaPR+LPAXwFs4O80m81/8IDX/STw95vN5tmDG+Lx5VEnwSjVsTBTYhDErG15LG+6xHGCbRsU8xZhmBJGMXGSAoKCowOD6+/UKd/r9x512df24DXK5YKgmNOn0cCPkAIGfkwKOKa4NzNVYEg9u3oS8+b9EKc6X91zI5JUp4ZCBYOwz3S9QL3ssLrl4Q1i0kThGAa2ZdLq+pQLOm0WhCmiBLVyHsuUvHZmYlyFMeoFGAW3E1OFB5p8jGa9XTfg7lqfH1zbIkoS5ibz3FntEcRaH/3OSodBkFCvFlBiQBwldIY9AlI8/cz+URvZSunUT84WTNcK3Frt8eqpOqW8Ta18Tyxu917EcQ+AG22Pu+s9vnVpFZUoCjmTwnA1epx14w+LRwb6RqNxEvhbwDtAAPx+o9H4zWaz+cGu180C/zmjTEXGnnnapY0+3/9oAyEESawwBExW88OKhpg41Rt99VKO2XrxPs2Tq7fafKe5hm0anD1ZQQhxZFUP25fs81OFHTn6lc0+UaqbW6JEjWULDCEwpcLJmUggn5N0+uGBpSZ2s9ke6I3MkSko+qbiehF5O+SDm1sUc9rnc6vns7zpglBMVvIUcyY522QQJuQsiTtIWN1yubHcZW6yOD4G+zH5+NxbJ3j1VJV//TvXWNnyAEUhZ/DhnS4Fx2Cqlmd5c8D3Ptpgpp7T5Y5BqqVyhxzWMdqNLkk16HohQZiwsukSxT28IEIpKDgmhYc0oB03Ntoe/9PXPuK7V9Zp9wLCKMEwJGfny/ypH3vluTfceRL28+19EfiNZrO5BdBoNH4Z+Fngb+563S8CfwP423v9kkajUQNqux5eeIyxHhv200yzV2POa2cmKOZNLt3Y0rlqpXCDGC/SM0zbMsg7BkmiWN50eWWhih8m4993Tz5AS/ZevtnitdP1cZfiYVc97BZg+9i5SVC6QmSr4/OZN2a5u+7R7vn0PIN8nBDFCfWSxcqWr83QhcA0BMnjlJU8BnGqm4lABzAESKH9PYMowbYMPnZmgt+9uIxlSHK2pOuGrG4NOH+qwmQ1z4npItfvdnGDgBvLXU7PlZmdKN63cnpU5dVmJ6BWyjFRybHeHnDtTkerIXoR7V5IiiKKFMsbHkLq/Yt4p2XskaAURLE2cCnlLS33GyXcWulxer6M60d0h8qVx3kmPLouf+97S3xwbZMgSsg7FjnbJIhiljc8Ln60mQX6B3ACWN728zLw6e0vaDQavwB8B/jGQ37PXwL+2mOO79ixvSY+jhWmKbi50t1hRDAKyrsbc+amirx7eY2pao6tXkC3H+iAN1yiR2mCYZRI0pS+F/Gr37jJKyeq4xN4NHssDtu887bY0chyFFUPD1qy/+rXb1AvO0zXiyytu7h+BErxnStruEGKKQWDIGUrCA59yTduslJgCIiTlHY/pOuGukPyToeTU0U2OgPCOMW2DW3u0g35xIUCpiGZny6yaJRwbPOBJZWPqj7puOFYsG2ikuN7wSZhFBGGurVVDauO4hSc4c3vsNJaD0MBYZhwcrpIzrG4u95jpl5ASrh0fQvHNjkxWeS9y2t86bNnjnh0+2N0zaVpyrU7bba6ujkKpSjmtSFJGCZcvLrOz/7Eq892sM+A/QT6va7L8aKy0Wi8CfwM8AUePkP/u8A/2fXYAvC1fYzhmbF79n53vcfKpku54FAqSMIoZWXT5b1Lq7zz+uyO1MpUzeHuhjduzPnO5VU2uwMmyzksCYYpSSJdZ24agpxlst7yiJOUvGNiGno5reUD4rFq4MikRDtR3Wtk2V31sN+Vx0G0+leLNmstj1Y3wPUjBHD5VotW18eQAiEElYJDnKZakwW9eXrYQS0ZBnyBIhXQ8UI+utOiUnIoOBYiSTGGOjM9L+Lq7TbnF6p89s25R9aLP6r6pFq0sQwt81vIWViGwBtoTU52pWXSVOk90QP+/AJdRpk8pJQ1ZwsmqznCOKVWMagUbNr9gLX2AFNK3j5fppg3ef/q+rH0SNhoe3z5azpF1urpzmUUSCkYRClKBRRyuglsr96Ul4H9BPq7wOe3/TwPLG37+eeGj72L3qw90Wg0vtZsNre/h2az2Qba2x9rNBqPP+IjZK9Nz299sMqpmdK4XCtOU7puxFe+cZMrt9ucma+MUyt3NzxOThVY3Rpw5fYWt1b6SAndXkAY6zTCzEQBP4iHG5uKds+nVLBJEl0E99HdNkLoDtKRamC15HBhsc6lG5tstAYodb/eyH42bA9yU3ey6vDb371DpWAjBVy51ebuWh/TEKSpDmRuEGIIA5XqVn27IPHD+KmapPZir1mxHOZzVAodN0ZKiWMZWKZu4qqWHM6drHF6vqIrcnh0vfiDqk9A5+/vrPdp933iNGW6msc0xbiCZrswGXBolUiGhGLewvOjofrnXgjmJ0ukShGEMa6f0BsE2KbEsiWrLR/D0O5Uu/eBnrXo1+gcvrHaZWmtT7sbECWpLiFVCpEoEinw/Hgs3fwysp9A/1XgrzcajWnARc/e/+LoyWaz+dcYpmQajcYZ4Ld2B/nnlZH59K2VHhsdDz9I2Or6RFHCa0OTjKX1/jC3qmeGN1d6usU/TGj3dMWFH0asbQ2Ik5RkW/14FCe4AxcpwLEkYZAOlRh1hUq7F5Ikig+iDWqlHN+9us4ngNmJIn6oOzc/9cbsnvnj/eh0HKSWx2Yn4I0zE7R7IR/d1ZaAuWHdeiIVUZwQhYCVYNmCdHgxHkaqXho6oG//3Vo4TeIH+q7Sc0MMKZmfzGObFq1+wFsXpnccg/3Ui+9OZW2/eS7OlnEsyftX1vnwTocg1GJlKJ0XF7sEy6Q4+OORpnoyYhqSdJgrGv1dwxAkiUIO6zi7/YCl9b5eXSSKVEKqFBvtAUopfuJTp3ZYNB6kJtCTcvVWm54bcHOpQ7sXEm87nqO+szhJMQzB7ESeL3761KGM47jzyEDfbDbvNhqNvwz8JnrG/ovNZvNbjUbjV4C/2mw23z3sQT4r7qz3WNn0EECrFyCFFt3q9ENurXSxDD1NjBNFMW+NOym7/YCVLQ8pBXfX+wz8eIcC4G5SpTsYR2x29cUkgLwjCWNFnEKSpNxZ62NbJpvtAW+cmRjrgu8O0vuRaHhc3ZyHKW5+dKfNhVM1XjtTwvVjwEWlio3OAITQwcTUnZgogWWC5ycPFYJ7UkYrhjQZNiABpqlVwlKlA6ptG1RLNuudkBOTearl3FiGd3QMnqRefPfNM+9YlIs2c5MlTs+X+e0/uEuSaA/c3WkE29SBN0oZq1YaT9k8pYBKwSJNFGvtAAXIRN8MGfZwmIbAG0T03JAwShFCYA7PbTWUuRBCNxs51r2O0oPSBHpSRgUPX//eXTru3stCU+ob2vmFGv/BH3vtpdyIhX3W0TebzV8CfmnXY398j9fdAM4cxMCOA64XYUhBd+j8bpkGlaKDIsAQktWWthqz85IwTGjebFHMW3Rdn2LB5PrdHn03fOK8q+5eVBhKYQCbbR+Vwr/3xVd59xIPDdL7aVN/nFb23bO3tZbHb3/3Dm+cmWB2oohtGnxwfZOPnZuimDP1noNjUC06hHHK+iAiHQYw2xAoaWiZ3WHgVcP89EHMahUwU8vTc0M6ru4ETRM1NiUxhNZtef3sJGubHlGS7jCZ2J2e2W+A2qvSannDpZS3iZKU8wt1vnN5lXY/1WJvUmBIiFOFLWGyWiCIEy3ChUIZ+gA9TSGOAtr9cHxMRxtuSQKmoSg42gDcHWgvWseS5ByTONHHIYoS8kUTKcR9q5m9JgqPqwn0pIzOx+X13gODPEAhbzJdK/ALf/qTx25v4Sh5fopjD4mNtsd7l9f46HZ7rK39zuuzTNUKFAs2rh+x1RmQKoU7iHSVhmngOJK8beLYkrUtj66r9cBH8rI5x8SyJIYph7ovT0YYKwo2mJZk4Mesbrn83//bd5FCqxK+fmZyLK+6PUDtJ+3wOK3su2dvrW6gN+16IXOTJc6erHDxww1uLHc4PVdhab1PGKfMTua4u661Y0ZVJWGiMEmRhkRsy10MJ90P3ZEUwGQtR9/18XfJKIxmv1Jo0a65qQJB1CWIdGotVQrb0rPVuckCBcekWrK4ttylXnFQSj1xO/+DKq38MKFS1A1meh+gygfXthBS6PPI1nLAtimxLIllCi20hsBEgCGIkhjLgPAJI75QkLcMLSKX6NTNSOsGFH6k5YxPTZfoeSFtN8KUgpytS1J7npZC2MuwfPdEod0PqJUOX/HxvUurfHSnRXPYFf4g4jjl1GzlpQ7y8JIH+o22x6+/e4u7ay6VggUILt3YotUL+OKnF6kULK7dDVlrDUhSrZeedywsU5KzTQo5g8s3Wwz8iCRJx5UTSQp9L6aQN3Tp5FPOUP04YWm9hx+mWIYkTrSA143lLp4f80ONmWEFzr0AtZ+0w+OkJnbP3kYSCO6wtb5SdPj40OUnVfDJxjQ9L+TdS2skSYplgEoFhinww5Q01ZvL22fvcapn2w9DAaW8SZo6xEkwzskaQqdr5HBjdbqeJ4xSZicKQyMPiRvETFdzTFRzTFa0BEHesfnRt08yVc0/VTv/aD8nUSnX73Yp5S1qJVsHbQFvndeb5TnbYm6qyFY3IE5T8sLk4+cmubbcRSLo+RFFx8K2DKaqOdp9H0ifqlokBRACQwoKjknOMekNAtJUO0nNThZwvZAwSUmUwpQ6Z69SMA2D8yervHn+/mqbvSYKUgrqlf1rAj0JV25u8uXfu8Zmx39oStSQYJrGsdLgeVa81IH+6q023V5EtWjjDDdXhdBqku9dWqXtBrT7IfOTBVa2BiQqRUqdj19ruSSpbvsPo2SsyjjWUAf6g4PpfklTcH0tzOXYkjCKWd6MWJgq0HMDrt5u80ONmT0D+aMC1n5TE7tnb8WcRX8QUsrfu4At0+CTjZkdhsqrW4Oh9AFIEZGoe+mZNNWG39rAZZiLFmAoSLi/MsUx9bG4vdrHsSSWKVFhOhbpSmJdljpdzbO2pZt7zpyoc/ZEhb4XsbTRp++FvHV+mpl6YTx7P4j88Wg/p+CYnDtZYW1zwO21PtWSzcmZIqYh6fR91lp67+bTb8wSxinX73aG5YAKMdQDOj1fZqKSx/NjOm5IMWcRJxFKpU+Ur9d7QDGmIZGGpFTQWviJUuRsi3Mnqtxe67Oy6YJSTNZydN0IIQUnJgv86CdO7vl395oofOkzi1y53dnT6u8g2Gh7/IuvXmGrG4w31h+EKQUz9TzvvDZzIH/7eealDvTbG1pG2Jak7yV8eKfDx85NMl3rE4QpW72QJBUIJJWiQ9f1UVIRDTevhFSPZZDxJEipZXVty8Q2oTeIMC2DmYn8obvV75691SsOS5t9FmZKD015CNCa4JbENGyCKCFN1dhsxDQljmmQDpP0KfoGECdaHiKMFUKNSiYFlqXt9vwwxTbvuUo5liSfMzk9p7tblzZcXjlZ5cR0aZzaapyuc2u199Sz970Y7ec4tokDnD1p03UDCo7JFz61yHuXVvm9i0vEiU6ThHFKHCf0vID1Vqo3R9Hib36Y0OoFGEJwfqHKnXWXyWqOla0B3X74WCtEKSBJFAJFrZYnb2uxuSBKqJZsZieKFPM2p2bLBEHMRndAMW8xWc1TyFm8dX5KS2tv24Tdzl4ThYlq/tAEz37ld6/xneb6I8tRTQm1isOf+aOvvvRpG3jJA/32hpbRjD6MUkxTEEWKvGMyVS0QxdqtJopTklQxU8/jhzGb7QFIiOPHM4p4UuIEVJoiREohZxLGKTnLxPUOQfN3F7tnb1PVPD/9Y6+w2QkeekG/cqrGdz9c1zlmBaYpkDGoSKe0qnkblA74eVuy2fFRiHH+GPQSXEihq2UsAyFh4CcYhslk3iBVetNtdqLATL1Akip+4p2FHV2toFMIC9OlQ7kpjvZzgjDBtnQjnVKK4lCBNEoUk9U89bJDpx9yZ61LqxtiWQYgmKrl2ez4BHHM0nqfhdmyVvs0BfWSTbngUCvnad5qYRuCtZa/u+fqPiyp93ZMQ/cMJKkiZxucO1nlo6UOjimZm9IaPsWcxeJcGWkIqiWHetnh9FwFy5SPPSM/LMGzb33/Ll/55s1HBnkBvLpY42d+/DyffvPkgY/jeeSlDvQXFmvcWuvqHL3SSlj9QcjcZJF6JccgiDkxXeTysJpmZaOPYcih85DJjUFEzjLpRYcfaEfolIei3QuwLYlji3EwOWz2cwGPSjDvrPdxvRAhoZS36LohpiERCCoFiSl1A8vJ6RKgiKKEtdaAvGPih3rWnyrIOXrVlHcMwjDFNARxrGv0HdtgqpajP9ANUEGYDI2dTd55ffZINdMXprWXQLsXDg2mTabrZaa39TTUyzmt/VPJ0e4FuIMY2zDI2YJaOUffC4liQaIUa1seM/UChZw1liYo5S0WZgp84werGCbIoeGIYegbpR/c6zQ2JExUcjiOvhG+fmZCWwpOlei42tyk62ovX6X0CkvfIE+NO7w7w2qzE1P6M7x7ae2ZNEWBPq/+0f9wka736O66UsHi//Q/++FsJr+NlzrQT9UKelm9rerm9TMTvPP6LHDPAb6xWOPGUo+cYzFZcZieKLC+5VErOWx1BghxrznjsFFAEOv687wtWN3yqVe8YyE4tV1vZG3L1QqdqeKNM3V+cL2FaYjhKinPetsbzhxz2rd1qct0XR/HetnhzrqLH+i67jjRDTxTtRyGIYkTxWTVYRAkuIH2X71wqkaq4PR8Bccyjlwz/cJijZsrXYJIa8oHUYznSy4s1sbeo6NJA+gmJhS4QczJaT2rToZNXtP1PJWiw+JsBT+MUIjxZ5mZKPNTf7jM73337tgPNU112kuiZ/CObeh0WpRQreT40bdOcGK6RKsXjFczG+2ZsWZT39NGLXOTxXHF2WF0Tz8pG22Pf/jLf8BG537Htd0IYLKae+bXwnHjpQ70MLQe++wZ+Oz9z21XaqxXbKI4xgtSXC9CSKgWLVa3PBxLEiXpkSkPCqU3ZRECx5LEcXrkF99ejGaut1Z65GxLWyKGMZZp8OPvLLDV8ZmbKlEt2nS9kK3OgOtLXXpeSMcNKDkW/SCmUrQ5d6LCIIjpuiGNgsW1ux0cR7swzdTz5GyDE9MlNto+p2bLGFIglbiv8ugoj4eQgBIoUlBS/8y9jexK0Rl71cZxSrVkk6RoYTulCCOdIizlbfKOSZKmrG4OWNnyWJgujWfSG22PK7daVMs2m+0Bt1Zd/DCiWDCwLAs/iCkXtTS2Nksv4/nRjsqXqVqBL3568ZHyBQfZPf2k/LN/832+fWljX6+VEj71erb5upuXPtA/iO1doP1ByAfXt5iq5Zmp5+gPIt67tEYxbzFT12V6XhiReskj86ZPi21oMTRp6CW+G0QsbXrkHeuZqwuOSjBdP6Y01C8flX02TtexLZM/9jk9vis3N/nmD5aJo6Fc8yDCDxJm6jn8MMUPA6ZqeU5Ol1mcK/NDr89RKdjcWe+xvuWx2Q2ol/KcX6ix2fHZ6gW8fWH6mYluXb3VZrZe5Oz8vT0Bz4/uk1IoF2wW5yS2pW8EaZLyg+tbtPoBUgqqOWuYS5fcWO4CiqlqgSBKxjfzq7fanJ4rc2e1T95JODsvuXy7RRApbAtqZRsp5NjFanXL5fLNFhLFd5prO3pFHnWs9jJKLxfsIzMC/7VvXOM3v7P86BcOOTNf5id/5Nwhjuj5JAv0e7B7ufqt76/gBhGzRpFBoDsX/TDBDSJmankGQYxAYJmCID7cHE6YQNmW2JaJUil9N6TgWNxY6XD1bouz8+VDb/N+kJDVaOZazJmEQ5PzMNJ589211NeXe1iGRKDouAmlvIUfJMNNbUUcp6xuuZwZioztXq1sH8P8ZOmZ5I238zA5ib3SSF/89CKgbxCmaeB6ERudAbZpYNsGyxuuFmJTI8mNHrWyPf7MOdtESMHyVl8rVAqBaUum63niRNH3AtxBqFVEb24RRXoltLtX5FEOaXsZpZ+eK+/oJD4sNtoe//C//96+Xy+Af/sPncvSNnuQBfo92L1cHYQRpZzF8kYfpXRAz9kGrp+y3vaHrexiaAd4+Mn6VKmxzVwYp+RzJo5pYMQJX/nmrR3+ngfNw3K2o5lrrWxze7VHEGmT8zMThfs2Qj+63Wa6lifnWCQp5G1JbxDS7UfMTdgMhpKWU9X8nkH8uFnZ7Ufpcq/xPkgQzR3ojmwEnJguEMXp8JimVAoWH1zfolxwODVTZmXTG1s7hlEy1mRKUsVnPjZHqxsiJff1ijwq/bJ95RBG94zSbyx3+dzH5w/q0D2Qf/Zvvr9DpOxRlPJmVjP/AOSjX/Ly0XHDHTrk5YJNnChdDmcKLNPAcQyklNiWQSlvUS3pjcLDRl+weqYcxAm1soNjGUSJ4uR0iTRVY7Pww2D7TXBkY1jKW+Og8dk355iuFZiZKFJwTOYmC0xV8/fNyO9ZYEPOlkSJImeZ1MsOP/qJBX7krZP8xDuLfO6tE8cqoD+IC4s1ncLzI5RSY4+A/ZpQw70SVscy6PQCDEOyOFuhmLdxbAMhBK4XghDjlc9EJUcQJli2wexkQQfxQcRkNc/Hzk1QKjhEiZbEHmFbkjhWO5Qo96LjhsxOFLmwWB+WWer0zdxk8dC/k299/+5jpWwAPv+J5+NceRZkM/o92D07e/3MBF97f4kwjonjhPWWxyCIsU2JUimbHS1clhySprgYdogqwDYkxZxFraSFxeqVHKYhma7lMA1BrWQ+8gJ+Gh6leLnfmfb5hSqXbmwhhKBWdri90iNOU87MVR5opHKcOagqn9HxG3XamoYYlj8mJKmiWLBQCj7+yiTLG95QTyhPkiiSFN46M8H8VBHLlGOlyQf1ijxKlmB0HVRLzrjpbGSUfphstD3++VevPNZ7pqo2P/fF4+1v8SzJAv0e3N8FmuPNV+pcudHmzppHuWDx9vkpljdcbq70iFOtApgYKWaa4B90nB2ZXSs9mz85U+ZTr89y8co6xcI9eeTWsLa+P4j4+sWlQ8lbP47i5cN45/VZWr2AnheiUpiZKBBEMVP1PI5lHGop5GFxkOmkhekyOduk1Q3oDyKKOYszE4VxbjyIEl47MwFA161x8cMNygWLxun6fT0DD+oVedRq43FE7w6Sf/N717i+1H2s9/xvfvqt5+58OUqyQL8He83O/uTnz/Pe1CqXbujcqG1JNtqB1g+RglLBoj+IkEKSL+nWfduUCAFbvf1FfoFOzRhSX1hxooaNQwqBoFwwqdfyLMwUWV7v0xsE3Frr6WqgWp4oSTANwWunazuqNA7yAjioi3+/5X0vKxcWa2x+32dxrrzjOI+C8/bvwDQkJ2eK1IrOnquJB/WK7EcH6Sh7EUDP5n/7O3cfq1Q5b5N1wD6CLNA/gL1mZwrBG2cnWdn06A9ijKFoUsfVy9m8YxFEEUGYYpnas9LzY4x9Kljapm6w+eE35rhyq8Xd9T6lgjYCtwxJpWjT6fv81nfukrcNTENL3bZ6PkmaYgoDx1SsbHqcmC7uyJ0f5HE5qIv/uG2oHicedZx3P/eFTz24guZhvSL7GcdRfke/9Z07bI6MvffJf/rznzqk0bw4ZIH+MagW7R1L5ss3TD683aaQt5FSVzoEocX15S4oxYmponY6Uor+4P7WbdsUWnY3Z5EqRb2S58R0CcuU427PYs4a51ZbPZ8PbmyCEuTLOQDCOEEpxdKGy2ffmNNu98MyuMZiDf9JRcwfQhagj4aHHecX8Tv41vfv8j/+1tXHEgf8uZ84m83m90FWdfMY7K6sqFccTEtrfE9Xc6SpzlfnbUm94tAfJPhBymTFIWfrChMp9EEXQrtHCXSX60w9P9RZV2y2B3zpM4vMTmpLQD+I8IOYG0tdBn6ClALXj0hSpUvegpgoTBFCIoRWUMzbBjeWegeqA56RcVhstD3+y3/1PboPcYvazWffnOXP/eRbhziqF4dsRv8Y7KXg+O9/8VWuL/f46HabyWqON85N8AfNNS06Zgr8ULHZDbEtk1RFlAsOaaobguI0JedY5HMWZ0/WqBYdGqfrtHoBr56eZKKaH+dWvSAiVSnFnIll6PK6/iCilDfxo5RywSIY1p7blkQpaPWDxyrve1oe1EiVkfEo/uVXL7Pa2n+37dvnJ/m5n7hwiCN6scgC/WOy15L51dOT4/zn1y8u0eoEXI23CMOUOEkRQmEZBmbOJklS6uUchbyFKSWOJQniFEMI5qeK91WwVAo25xZqrGz0sU0DyzBY3fKwhdbB7/S1g9HibIULi3WWN1z6gxjT0BfDUQpPPWvxq4znk422xzd+sLbv1xccyZ/746+/tEbfT0KWujlgOm7ImRNlbNtEKZiuFZis5kFoQaiJco43zk7yxR9eZGG2RNvVfqHnT9XG2t8XFmvjwBlEidYwd0PCOKFatqmV9Y0gDGOSNOXj5yY4u6C1wxun67x+ps78VGmswnkUPKyRKiPjYVy91SZ6RJmNAGwTJso2P/eFV7Mg/5hkM/oDZrRh+86rM/z/vnWTVCkUilLOxh7a37X7AeWCzWunJ6hXHGpFh1Sxo3786xeXdsgw1Ms5bd1XsJms5lne0AqIJ6dL/Ht/9FWAIy2D282jGqn2Q5b6eTnpuCG2aQAPzs9bJkxVC7z5ygR/5J1TRze4F4Qs0B8wozrzeiXHW69McWOlQ6sbMlMvsDBbwh1EtLo+t1Z7LEyXHlgWtztwzk8VuXIrIAhTPvvmPK+dvt/v9FkGxadtpMpSPy8nV25u8vsX79L3HtxrIgXM1Iv8uz/2yr7q/zPuJwv0B8z2Ddupep67Gy6N00Vm6oVxG/pn3pxnqvpwn9fdgbNacjg1W2ar4x/YrP0gZ9BP20h1HHTPjyMv8irnys1N/off/oi+F2mlzl1IIJczmKzkOHeyzpc+d/bIx/iikAX6Q2D3hq03iMdt7KfnK/vS894rcEop+ROfPxgZ1oOeQT9tI9VBpH6Omo22x3uXVvnwTgeB9sc9SD38F3mVc+XmJv/4f/werV6IOwju03yVAixTMlnNgYKpqrPn7zlOHOebchboD5mF6TJBlOxIaex2+9mLw24/f9QM+klO2qdp4jkoDZ2jYqPt8a9/+0M+vNvFQGDbko4bcHe9x8mpEgrx1Bf7i7rKGc3ku16EICWIU9JkLOcEaEtJ09BWlHnH5A+9dbCyyHud38ATB+rjflPOAv0h8zQpjcPsfnzYDPogTtoH3Sge9PizEtDaL7vHfXe9x0d3O4D28N3sRnh+TPPmFudO1vjsm/NPrTf0PK5y9sPvX1zGNiQqVbT6IWlyv4uDkNpisVq0+dkfP3+gVTZ7nd+//u4tVApzk8UnOueP+005C/SHzLMQhtoPQsDFqxvEaTq2iDMNSbVoP9VJu9H2eO/yGu9fXadedjg9VxkHvFdPVblyu/PAG8hxPE6jz7Q7MHzrg1WiOBn6vTL0DNb/9f2ID2+3ubBYfyq9oedtlbNf7m70GQQxliUBtSPISwGGARPlHCemS/zCn/7kgZ8De53f3V4EQnHuZG382Oi1+/n7x/2mnAX6I+C46ZJstD1aXR/XDynlbcIo5uKHG5ycKfKFTy3y7qW1JzppRwFxed1louwghNwR8H7/4jJnTlSJ4pRbKy1cXzd2vXdplS997uyxO04jRoEhTlKaN1tsdDzW2x6DQUSpmEMChpQgQBqSNFU4tsnyhjvudH4Sjvsq50nYaHtsdXzcQUycpBQci14SkSod5GtlmzSBWjnHjxyS6cxeQTlKEgSCTj9gecPF9WMKjkEhbz3gt+zkuN+Us0D/EnL1Vpu5yeK4Ht/1Y8oFC0MIrt5qc+1OG8uSnJmvjg0n9nPSjgNimlLK68YpgOUNl7nJApdutNjq+rT7IdP1PLWSQxAmvP/h5rEum+u4IYaE969u6ECw3qfnhgSRIk59pNAuZKlS4+7lKEm4tdql1fOpFm022t4TmZAc11XOk7DR9vj1d2+RJCl9X/sQxElKOW/ihQmGEBhSUs6bVEvOoTT8bbQ9Vjb6NG9uUS/nmJ8qUi05WIaBF4RcvdXCsU1KeZOeF9L1wn19d8f9ppwF+peQ0YymkBMIAUvrLrdXu3zn8hqnT5SpFhw2OgPcQcQbZyfHPqFzk0W+8vXrIARKcd+G1ej3jvxL41Sx1Rmw2fG5fFNiSNjq+vS8iI2OT71oIYa52i9/7dqBVRQdNNWize989zbX7/bY7A7ww3vyikGk/x0ngU4HOAaGIbi90iVJFevtAZ2+z9/953/A+YUq85PFBx6/vTiuq5wn4b3La9xdc5msFbBtk432gEEvQkiTM3NllBLYtsHJqSKNMxMHVl022lsRAlpdn8lant4goj8IuXIr4NRsmUrZot3zsQoGtiUJI21Uf2a+sq/0zXG/KWeB/iVktMyMk5TLN1sopVje6GNIcAcxpZyNZRhIKbj40TqOZXJ6TjsefXB9k3RoZbd9sxHg2p0215Y7DPyYbj8gShRCQJKk5B2DailHq+9TL+cwJNzddLEtg9dP1+i44QM3v466bG132WQpb/L9a1ukqdoR5LeTpArLFHTciJ7XxjQM6hWHyWqOrhehUsWlG1usbLikCuoVhzBK+d2Ldzk7X6FUcPa1af0889HtNpWCRXGYBjt3ska3P2Bly0dIyYWFKgszZYQQT2XyrZuwlrmz5uKHER87N8G5k3UuXt3A9UM+dm6KhekSl25s0er5rG55vHVhmjBJCLsJfpgwVc3vuxR6xHG+KWeB/iXkwmKNX3/3Fpevt+gPQlpdHy+ImakVQCn6g4jpWh7TkHpWf26SQs7i8tBdCxTLG95Yl/+9y2u0uj5rLY9ePySKEtpuNP57UoBpSoIoQQgJKQzCGCkEcxMF+oOE03N7G6UcZdnalZubfPXbt/ng+iZKwem5MhOVHO9eXsN1w4eax6QKXF/XCEZRQmIpep7g2t0Y0zTI2YbeF/Fi4jRlvT3g9GyJ9daA2ys96mUHL4hxbJMf+8Q8QQxKKbY6Ps2bW7x7eZUvfWbxudZ4GRnCF3MmJ6ZLtLo+tmkxPyX4iU+d3vcq52FcubnJP//qFeJI0er7hFHCb7x3hxsrPfwgYaqW48M7bVSqmJ0oMlF1uHq7w53VHnnbYL3tc2ejz1rLw/MjTs6UxvaND+O435izQP+SolLoeyF9PyJOwDIFUZLSdSPiRHFyusRW18c0JHnHpOsGfHS3DUDONnFsCUyQd0zeu7SK60dsdnyEgI4XjGuiDamdtuIkpdsPqJVzIAW2JSk4JrZp0BtElAsWN5e7rLe1u9DoQjmqsrVRbbfnxViGZBAlfHBjC9sUbLQDon2YYRgS7Q+A0B6+YYBSioJjooBBmFB0EqbrBaI45dpyD0Mo1lse3iCiVLAZ+D7//Neu8oc+Ps8gSsnbBhOVHD0v5CvfvMVENX+sAsij2B4A/SDk+tKAnG1RzBuUCjY52+D1MxPaAesA/tZ/96uXWd5wKRd1kUEYK5IkZWXDZapWYGndJYpTqiWHJFV0egHFoYvblVstfbM1DNqJz3d7Ad+7tsmP/9DJh+bpj3sNPewz0DcajT8L/BXABv5Os9n8B7ue/yngb6B7Hq4Df77ZbLYOeKwZB8RoM7Ze0XLJgyDGHUTDGbd2rep5IVIKzi9UWWt53FzpkcQpXhCz3vYwpOTsfA8/jPQMCfDDiChWxLE+ERBgSqmn9ArCJCVnG1imwUAppKF18+eGUhFSwHQttyMldFRla1/99m06vYCVLY/+IEApEAg2H5Cq2QvXjxkEw1l9rOevUtzL40sUfT/C6PrkHBMS6PZDBBKFwLIMpCFo9QPe/3Cdty/MjN3FygWbra5/bOqy98P2ACgFRAkEUUzONnAHip4X8srJ6lNtuo5uJHfW+6xsuiytu+Qcg/4gotUJMC1JwbHoehFvnCvQvLnF2pZHEqfEqaLrhqRpwkbHI05SJis53EHEVi9gqpajaBt8cH2La0td3r4wvWfn83GvoYd9BPpGo3ES+FvAO0AA/H6j0fjNZrP5wfD5CvCPgB9uNpt3G43G3wT+OvC/O7RRZzwVHTdECugNQtq9QAdjFFLqKpkgTLix1KVacvjetQ3WtjyqRZskVfhRjERQylv83sUlojihUrTxgphWNyGMU8ahcRjcZTKa3QuCWHcJz02WWdrw8MKIYt5ADpcAJ6ZLOy6Uwy5bG+Xjv/mDFdI01W5e46zTPox+t5GkkKAYHkZAp3TCOB1ruQjAjxIKeZNWN8AdxBTyBoYUCKH/yzsmW90Q2zLGvyeMUmolLVf9vLA9AN5a6TFdK1DKW3TdkHolhyklC7PlJw6G228k3iDS51eUjDvRDUMQxyn9JERKuL3So90LGIQRm11BPmdRyJu0vUiXxJoGOcfEj/Q5bRmSfhAhhhOV3/nOHa7cat2XQjvuNfSwvxn9F4HfaDabWwCNRuOXgZ8F/ubweQv4j5vN5t3hzxeBn9/9SxqNRg2o7Xp44fGHnPEkbF9CX7vTpusGFGwLsypx/ZhuP6aYNynlLba6Pv1BQBgl2I5ks+Oz3vKwTRPTFBiGJIpTNts+hgH1So6VTZcoSUl2JbKVgpHSuCkUPVeX1QGcmCpgmgbXl7q8uljj/Kn6uJxzdKF86vWZAy9buzcL7HH5xharmx59LySOFY9hV/pAlLr/Z6H7qsjnTeJYsdUOQOjNapVAlCYEYQRCSydEsU/PCykXbMIoJQhjpmdLx6Yuez9sD4CuH1PKm9iWg2FI3nltFqXUUwXD7TcSL0goF2yKOZN2PyCMtR9zGKUoBbYp6XkhPc9HCu3nmaT6uSCMiWOFyMFWxyeMEyYqObpuRM6ReGFMuWBjmfqGvDuFdtxr6GF/gf4EsLzt52Xg06Mfms3mJvCvABqNRh74z4C/t8fv+UvAX3vCcb6UHNQGz/aZjyHh7nCZO1nJkShFnCikIbAtveTNOaZOPQjwgwTTkGz1Q2YnbU7NlOn7EatbA6ShUEqwtOHS9UKUetQcWOAOYsI4pVKwUEhePzM5DuDLGy4f3mlTzFnUKw5Tw4vpIMvWrtzc5CvfvEWaKm4ud7i91ieMUtLHm7w/lN2dniiQBggkcrgZKaXAD2MMmRAlCT1XEQQx9UoeDMFnPzary1O7PrWSw/RsCSnlkVpDPi3bA2AxZxJGKaAoDgPikwbDUff1b7x7m5xlMD9VxA9i7q718IKYINSVM8mOu7ZidauPH4JpKPwgIokVUZIQRym2bYyLDxKVstXzSROFY1nUKg5BlDAIY2gpoigdN/nB8a+hh/0FerHHY/dNfBqNRhUd8N9vNpv/dI/3/F3gn+x6bAH42j7G8NLxtBs8O2fwLaJE4Q0iljZdBn6EEHBztUsYp6D0rHN5w9MbiqagXHCQUpuNe35EFKcsbbh65iMhjRP6foxtShI/JtlDr2SEKfVzUayQQw2TpQ2XlqubUcoFi6u328xOFKkULTw/ZGmzz0//2CvAwZWtbbQ9vvLNWxhSYJuSG8tdoocbGz0RhoRUH1ZSfb8kikHKlCiROLagVLAAbQ4/CGIEEMSpNqkpOHzhhxeZqOaPdSXHo9geAOcmCztKcz0/eqJgOGq6urvmkqYpN1dcPri+SRynOI4EIUkV9924g1gRxPp4JymoSCFliiElqaFI05QgiFFCoVJ9rVSLjk6fKcVG12eqWsCQEmnLHU1+x72GHvYX6O8Cn9/28zywtP0FjUZjHvgK8BvA/36vX9JsNttAe9f79j/Sl4zdbfeuH2FKyXuX1/asUNjRGIKi7QbM1otIAdeWuggB/UGIISVBnOD7MYnSJ/4w1gP6IhiECj/0Kea1HaLrxxhSl/t1+j6DMEENLyZfJI+cDcfbpgUCkAq8IMEwBR030PsEhkEcJ3x4xyOKE+Ymilz8aPOpygl3r4i6XkiaKmolh2t3O4cS5IFdM0l9bJUCEqjULHK2Sa8f4gYxoHRlU84kSWG6WuCd12fY7GiD+OMULB6X7QHQDxNePzMBQpCkUNrmprYfRt/ld5prLK33GQQRy5seYZQihymwgZ8i5MNXZ2r4P6mCMFIkcYJCaz8FUTx0gdOdsa8sVLl0fYuNtk73xEmfct7i7MkaOcvcsdl6nGvoYX+B/qvAX280GtOAC/wM8BdHTzYaDQP4MvAvms3m/+VQRvkSMmq7b95qk7cNSnmLIIx5/+r6fTv/u2f/Fz9cp+dF2Ibk0o0Wfhjj+wlRmjI74dDqBSgESikeZNWpgP7gnrXbKHj1k51veJKUR4qe9UaRYkCCbRuUixbrHZ9a2cEYVqr87vt3eeuVyScK9nutiN6/uk4cJ1y93WJpw338gT8lhqEDimlI2q5PMW+TJnrKrxDMT+YAmJ0oHquNvKfhIALg9u8yCBM22gN6gxAhwbYknh/fSxvuc5PFkBDH6Q5ZZIFeiflBhOdH9NwAaRiAwjQlKtXXZbvn89k355+rjfFHBvphJc1fBn4TXV75i81m81uNRuNXgL8KnAI+CRiNRuNnh297t9ls/oXDGvTLQLVo8/2PNsnbxrjETghJvezcV7Z19VYbpRS3Vnq4fsTKpksxZ/KdK+tDg/I81+52COKUQRCTJLoK5BF+zIeCnt0qEgVxqIWj5id0fbOU4FgmUZTQH4TkHMm//PWr/K9++vFrx0crou0CautbLiubHkiIH9b9dEhECXT6EZYhkVKSJkqrh+ZtSjmLME7JOcdvI+9Zs33TtdMPdPVXEJOMgvMTfJXxHjcEha6QAnBMwSBIkVLh2AZSSHKOiSF1KvPGcpcoSvn6xSUmqw7Xl7qHZkBzEOyrjr7ZbP4S8Eu7Hvvjw3++i3b9yjhALizW+Nr7S0yUHZRS48qL86dq980k7qz3WNn0KDgmUuja7Ot3OxTyJnOTJQxDUCk59L2QnqfTNyJV2BYMgoOoM9k/SaoDPUDOkQgp8YKIvhdSKljEcUK7rxuuZmp52v3goXsTD9qwHpWQfni7PRapGgQJfpRQLdkIoscsnjwYPD9mKUqQUhErSSFnkrMlSihcP+Hsieqx28h7Voy+29+7uEwxZzAIYm6v9Qii5NDSbqBvHkIwvJHohquRBHWaKm4seyAEP/zaDBudAb/2rZvYphyen4ofXNuk3fcf6Af9LMg6Y48pU7UCb5+f5NZqj/4gppgzWZyrY5kSZ1t9NYDr6RriOFEsbXiUCiYbHV0dMAhC0lR3sxZyBlGcEscpd9b6KAUG98ofjwop9CwsSSBnCVq9gFil9L0AN4gxDcnCtHZpqpdzD9R0f9iGdbVo8/1rGzi2iWPr45WkioJjEYYJliEJk/S+Usij+OxCgmNbFBy94ojiBENIijmTxmL9WCt5HhVXbm7yr3/nGu1eQKs3IIp1H0bBMfGDaIcb1UGjAD8aCTZokjTFHUQM/BjHMhj4Mb/27dukqSKOUyZrORZmK/o9QtDtRc9Xw1TGs+Od12eJEkUpbz20bKtYsHH9iLUtD8sQYFqUChZKCSzTJI4TBLDeCYjihHLBYqKcI05T4gS6bnhks1sxnC7lLIGUktW2TzFnMVXNsdnxhzvDKettF6UEb5ydJIx0udxuHtaROFl1uHq7jW1KTEMSD7t6dcWQFhXb7PrjDtajQkjIWbqUL+cYFIcrjQun68+9ls1BoW0ar7HRHVDO23Q9g54XkCpFKWcSRCmmoVOPh/3tRYmWUlYqJVVaaTXnSDZaAwxTIIUgVYrldZeT0x6T1QK2ZdDzwmOVw88C/TFmv2VbC9MlHEtyZ62P60cIJOWCjRSS2Xqej5Y6OJbBbD2PEIJ232cQJaAUhbyNHFbkJOmT5TsfB6V06sYPFYZUY5MO0zCYqRfx/FCrPQ4S3nplkkrR4YPrm7piYxd7dSSGUcK7l1ZIlUANZYK7/VBrxVsCz49JU5AS1GF/2D0Q6JRAPW8ihaCYt6gUbP7Mv9U4NrO/Z83VW23aboBlSrpeSH8QY1sG7iBks6sDvlL6pqmOIPOoZ/gAKaWcoO9FJAocJAo9wUhEyu3VPpPVAmGUYBnGsdpnyQL9MWc/VQsXFmvcXNEllPVSDsuUeIF2b/r+tU06/YCCYzE7WcDzdCBFgWObBH5MZxAi0d2DQZQy9AsZ14AfdDhM0mFt+XBDtNMNGAwibMugVna05ovUnY2DIGK15RGnikrR2VFLvrsjsdMP+OD6Jt4goVa2EQJ6bjgM6oIohmpRq0S2exHS0JtLR7lLkaSg0hQhBFPVPMWCxcfOPd9llAfNnfU+ra5PfxAPN0B12i1OdDlwYShTMKpYOqr7tQD8WJEmCkPq/S1DCpRtYllanM8PIrpexMmZ4rFqbss2UV8ApmoF6pUcs/UifqTrgmdqOXqebkqplRyKeZPVLY87633CMMU0tKJklKaYw8guJDiWxDK0umQ5b+jNwkOYDmy/NkeBNkwSum4ISmAago2Oz2rLY36ySNGxxmJnG20P0De4/kCXwimluLHcIVVQLlq4fkSlpBte4lRhGALDEFTLOS6cqlPMmZycKlHIyQO9CKSAeslktI1iCN0wJtGrCNuUwxVNQrVsc3Km+FTa6y8SG22Pr3z9Ou9+sEJ/EKHQm59xkjDwI0xD4NhafiNJFPERrEC3s718c9QMZxgG9bJDvWxjG5IwSvnYucljtREL2Yz+hUEp+OzH5+m6IcsbLh/d6aAUQ40OiTeIkUNlrSRJiIYa8WGsp9eWKck7WslSCsFEOYdjGbT7Pj7GodRiCu4tvy1Dq1y6Q0Ux09Qlh4tzZUBgmVKbOLsBX/7aNeamtO7Lq6eqbHYCWr2Abj+kkDdZ2fDY6vpM1fKYpkSEUMxbSCFQSi+9c45JIW+Scypcvd0+sGVL3jGolPIEkUuSpFiW3jw3pMA0DeI4RkqDxpk6P/z63HPX7XpYjDpem9fbpEqRpIo0SXGMFBAkw/y8UtpuEBkThSlxejSrsvHKVuibuRQC05RM1XJ4QcziRIX/5b/zyrHdY8kC/QvCKI1RLTlUSw6uHxOnCbWyw1rLIyXFEjolglJYwzRNwdGP5R2LvG2yGrkkSpEqhRdojfREheAfXKDfng7KWyZCKIQUDMIYpRTFvNZF6bgBmx2T/iCiXs7x3uVV3EGAZZq8flZX2Vy53Rk7XF293UIpODFdZLMzYGXL1TcQgd6EzttIqTt9p+o5LYSVtyjmTYIwJozvjU8MB2magvAxNmwnKzkmqzk8P6JSBNfT+eVSzsI0JT1P8Opijf/wJz+WBfhtjGwGwyRhspLDDxNaXZ8gTrAtk5xtknNMKnmbvh/R39RfluBoUm9yKGthSIFjGRTzNpahK90sQ/LTP3Z8gzxkgf6FYbewkmlAqhRT1bxWoTQNuu496dtizmIQ+kgkcaLI50yCICFVijhOGAQR7kDr01vGXnJHT45gNCuChJTpah4/SEjSlLxjDjVFdM35zZUep+cr+GHMR3fbJKni1YUaQogdVTagHaHurPYxDYPzC1Wat9sEccJMLc8gSHH9iLydEscJSWJRzhu4A93AlBoGQiiioaSwFALTkAip9fn3g0SbiziWwcfPTZIC61sepiHpDSIGYUy9YvNTf/h4euM+S0Y2g2GUECcpUzV93g7CmIXpEkmaagG60EdKgSkFiZTIVD3WjfhJEUoHeykFJ2dKvHKyhmlIOm7Ix849Wff2UZIF+heE3RU6i7NlDEPQ7oXkLBPLMCjmde22H0S03QjHNEAKqjlTywsLsEyDMEwYBAlC6IvooJpTpIBaWcvuxnEKQmEIgTvQrlblos2FUzVKOZsLi3U+vNPi8o0tOv2AvG0ihEChWG0N6PQDqiVnh+53zjYRUnBrtQtKcGGhSorg9FyFWytd7q71yTsm56rar/XWWo9qyaGQM7m12idvQ+qlw8+rc8OGgJwlsUxJfxA/NMNTLJhMVLUg2VrL4wfXN3nrwhRRrGj3A6QUWQnlAxhVrdfLOZY2XCxTYBmCAIEfJZw9USFJU3puRLsfYlkGZdvAD2OkTAjD9FBm9nojGFIB9aJNveKwMF0mSbXvwPOyx5IF+heIUYXOqKOw60X4fsLAFPT7ESeni9imgRdIhBScmCzQcn0kgo4bMj9ZRKkUx5J4YTJsFklI0wdfRFLARMUmTdXQBBsKOZNiTqdcEgUFR+L5KY4tKeVtqtM2fpgyN5HjxnKPME4JwkhvABcczi/UqJZ0MK6WHCxDb6g6lkGtZBNEKd/+YIV6JY9pwOJsGYAPrm+N3x9GKT0v4GNnJvjS587y9YtLY0MK0BU6m12fdi/AsQxOzRTpeRHeIMIyIYn1ZptSYNu6YaaQk7j+/UdCANN1vfHrejG3VrssTJf56R97hc1OQMcNeXWxnuXjH8L5hSqXhp7EJ6YKrG0N8MOUxbkyf/iTC1SKDt4gxjJ9gijFMqHnxXiBPucOOsgbw2oe05AUHIOpmsPbF2ZY3fIwTcm5k7XnSlE0C/QvGNu7RatFm3bPxzINPv3GLOWCzbc/WCOIYmYmCtRKDqW+zZ21HnGcEgw1vOM0JW8ZREmKKQVhKpBSYZuSnGUSxilJqtM8xZxFMWdTKztsdAeoBHKOThOVCs4w/50QhAGpgk4/pO+FVIo2CMlULc/puSquH7K86RIM8/SeHyGloJA3OTlVHsslX7ndYuDH2IbBTB26XkTbDTCEGFZg6PeutQa0+wG2ZbDR9nbU3HfdgOUNl7xjDvPwEsc22OxoobGBH2FYgjhNsUxJqnReNowiZms5Wq5POHShEkDekRRy2v90YabMwnSZz7114lmdAs8l77w+S6sX0PN0J/fcVIGcbfDa2QkqRf29TdXyKHSHqlKKvhejUjAMiSAdq7CO8ulPihCgtDcJecekUrSxbRPHNpmp5wmjlD/2uTMH8bGPjCzQv2CMukW3uj7f/MEyrpeASOn2As6erBDGMVGsLxbbNHAsA6Vgpp6nXrFZ3nRpd0OEACklji1JwuHCeti+bxhgWVrC+PRcmVolR8G2+MSr0wB864NVclHK3GQBL4hYbw1Q6JJC05AEkcKQkqu3t7hwqk6c6PG4XsyNQY+bKz3ePj/FZ96Y4ZsfrI2dlgwpMaXENiXdQcjttT7zU0Xytsmd1T4ff2WSj+50uL7cpZSzuHCqShCmfOP7K1imZBDExEnK5Zst8rbBVDXPLb9Hxw2wA0nPi8YbbAjdlGMIgTAMpms5bq/HnJ6vcN6qk7MM3ruySpLoCpFq0aFeyfH6mYlj1RH5vDBVK/DFTy/u0C2arDpcud3B8yPyjkmtbLO02edTr8/wg+tbFIIEIQVJnOIrcEzt8ws62IePqYkzmsUrpYO8aUriJKU/CCk4o1ApnolG0tOSBfoXjJGY1zd/sETXjck7BiiD5S2X7iBCCkWpoCt0BkGfJNY19aW8jRC6m2+j7aMUmCrFD7UejJRgCqFvAEJ7cRbzFpWSw5/80XM78s4KuHKrxcBPSJJhXfsgIkkVjmUObyIghXb0cQcJYaS12aslBz+KmajmWGsH42A/clrKOQZJmnJurkJtmN65vdpjECZYpkEhZ/HqqZpuBgsTCmVJKW+x0Rnw/t02t5Z7IKBWzumyvFRvPrc6ASkpQSxQQm++GULgBQmWkdLpSwwEvUHE+ckihZzFKydrbLQGSCl5dbHO/FRxTy2ijP2xO/V4bamHZUo2OgPWNj0UcP5klXLRQcoWXhBRcExyZYeVTQ+ldLNfkioKORulFNFuc4BdjHSXgB0OaULoCptUpeQdhxRFEGpV1b26tI87WaB/wRiJebmDhLxjYBo6MEohx76lcxNF+oOIvhfRHURcWKjS83RjVRQXWW+5dLyEWIGIwTK0FV6hYGvxMD8i71icP1UjSdV9HpoL02Ucy+DOap+lDZcoTvCDlErRYHGuggCWN10MKdhsB0zWcvT9aNgVKzFNm3YvZHGuTJIK/sy/1RjP9O6s9ZifLFIva+12xzYIopiCrW0QWz2fetkZpotiFufqRHHCB9c3sYfa4iDYaHlEccp0vcC5iTof3W4xCLW9nyklSXJvX8I0dOlnvepgmWKoYphSKdhs9QI++8YsryzUj6WF3PPGbqG6tZbHh3favHFmgtmJ4vgYz9WLoLT9pWXq/pCtno9t6CY/gcC/5/K+J6Pqr5EW/aik1hBQyFvkbYM40e5TpJCqlLnJIu+8Pnvox+GgyQL9C4aWN76L9gfUm4hxojdCE5VSzjsgBNO1PCemijRv6pLFvGNiW5KuGxAPzZRBi54laAPrOEkwpXbgMU3J8qZLGMZsdUP+i//PH/Cjb53gnddnubBYY7PrszBbotX32eymCKFLPXO2SRQnTFbzqFSx1vIwhKCSt3FMgzBOOTlVxPXjcUXNdhmInhewsunR6g7oDyJcX9dTn1+o8Nk359hoebR6AfVyjsU5bTb+/pV1UDBdL4DQIlVbnYBW1x8Ku6VUKzksP6KitNG0O4hIEoVtCvJ5i/MLNT7+yhRhlND3ItZaPlPVHD/69jxJKo6thdzzxm6hulY3oFLQN/68Y7G84dLq+Syt9ykX9H6KckxKBVsbzyvF3GSRnhdywiqxsunSceM9/5YhYXayyCCICKKUvG0QhAmGIcnZJo4lqVVyqFRxYrrEm+emnpvN191kgf4FY6pW4O0L03R6ARtdrQw5WcmxMXS3Pz1f5vxCjaV1fcGcnCkyUcnR80KCMGHgDzVEVDpWBxxVNXTdkDjuMl0vIIDVTY++F1EtWqgELt3YotUL+OKnF8elnq8u1seduVGc4A5CoiSlWnQ4M18hDGOuL/d0Z6sbUinZ3F7tU8iZrLU8pqr5HZ9vYbpMHKdcutEiTXVzlWMZ9Dx9Mf+Jz58bzwjzjonnR7T6AaW8jW1JJio57qz3iZIEKcHzIwo5i4XpEteWuoRRQmNxglbP1zo6jkm15PDDb8xRKdq0egE/+4XMAvOw2C1U5/qRTr21B/S9EMc29Ux/2J8wWc1pNzYhubBYp1KwOLdQH+ocBfzWe3foun1A7y+NNmklehN3up5nvQVn5gsIIREiZXVrwPxkgSRVLEyX2OoF/Ok/+upzGeBHZIH+BeSd12ZodX1uLGnz7yCKsUxBpZDj5HSJcsFmcU4yUc2Nu0rfu7TK+x9uYozql0Md3KVg3MqaptqZKY4Vq66nKx5MgTAkpaJFueDQ80Ku3mrzubdOjC+MjbbHe5dW+d61Tbq9gMlqjo+fm+LsiQrfaa5TLdnkHYOP7nRYbw+YKDs4ts0Prm+ODcJHXFis8e7lVeYmC5QL9rCpJuH0XHn8d3crfo50/cMoHQf1fj9CygRDGkzX8tRKDu4g5Paai5CKatFmbqJAzja5sKhXBp4fHStFwheR3UJ1xZxFfxDScQNcX5KkuveiVDDpeAESwauLdeoVbWa/26DmrfNT/P1/+QdcX3YhvdeoZ1qC+YkCAkG15DBVzeP6EYMgZXaiAGj1U6UEb1+Yfq6DPGSB/oVkVMHw3qXVHfZmZ+fLY12Y3WmGL33uLO+8PsuXv3aNdjegNxgghfbk1ClKNZ4FSSmIYkWaJNSredJUUa/ksC1J30vuqzqZqhX40ufO8qXPnd3x+NcvLjE3WWSymufbH6xQLTtEkZ5yTdeKmAb8/sVlri31dtQsz00W8AY6V1vMWZyer1Au2OPGqd2Knxttnc5Z2XRRSssyT0/kqCU28xMF1toDbq10cWyTn/kj5ygVcmPXroXZEpWijedHWf79CNjd4V2vONxY6dDqBkzXcphS6PRiopibzNMfaEeyrhfypc/cLyQ2Uc3zh95aYLrW4qO7HdwgwhKCC4t1zi1U6boRkxXto2yZks1OxPxkEcc2ODVbRkr5XDREPYos0L+gjIPrY77nT3z+HLfWenS9QFfbCEFMqitJlMI2Dd44N8mVWy022gOkEEzW8xRz2rjZNAUCxdcvLt1n7wf37OHurPe4+OEG9ZLDVE3PnD/+im586g8i5qeKXLm1RZyosa7NyD1qYbq8o/kJeOhse/uN73sfbdB1Q6qlHCenHFZaHqW8zcJ0mYlqjiCGTy7W+NxbJ8ZjzfLvR8fuDu+pap43zkzwgdoiiBSlgkEhrytqgkhxfqHGa2cm8PyIzc79hupXb7WZmywyVctTr+bJ28awukYxWS3wuTe1KJ613sf1QqZqOVSqzXyma4XnNie/myzQZ+xgqlbgMx+bZ7M1YHnLRRqCgmEgpZbXrZVtTCk5PVsaGjMrJioOfhDTH4SUCxZtNxjnUrcHaIBvfH8FpRQrmx5CaSliyzRo94OhQqVJMac33YQQ1MvOfbo2u2d9+6l2maoV7nPsunh1Q9+4zk5SLem8sOffs4DbjxdAxsGz+7j/6tdv8Nk358f+v7dWe5iGHE8IgB1SGNsZ5fybw94JxzZ1s9UgppS32OwEL0VzWxboX1IeZKoN93L8l29s0u5HuINwqMaoZzl313Ut+mc/Pg9KcXOlz612l0rRJghTZib2tvcDKOUtbq30KDgmC7Mlbix3aff0snx5s89UrcDHX5nkg+tbSCk4MV0cj3l7Fc5+nLd2s7uiI05TSnmb5Q13HOgfFDAynh3Vok0QJVxYrLO84QJafO7sfGX8vQ2CeM8V3SjnP9rUBQijlGLOfKm+6yzQv4Q8zFR7NJv64qcXqZcdPrzTwQ8iBkFMpehgmhLLMKiULf7IDy0A7Khy+cb3l7mx3B22jt8fPOtlZ3zRCSE4PVdmacPFMCSTlTwfOzdJkuoLdLKWH/8O2HkxP8lse3dFRzFnEUbxuERz99/IOB6MVnClvEXjdJ1a2eaDG1ucnCmhlHroim70XlNKgjBGCDnur3iZvuss0L+EPMxUexQ8t+f4dwuCwb0UB7Djd9XLOfqDkKV1dxykt19QgyAeBthk7BX7yskai3O6yWq0jB7djEbt7wfRjLS7ouPEdJGLH25QLliPDBgZz47dK7jpWoGf/rH6AwsL9nrve5fXeP/qOvWyw/lTtaEa6cvzXWeB/iVkL1Pthy1jH/X60XOdfoDrR9xc6WJKydxkAdsydlxQ3/j+iq6kWO4SRAlKKabr5fsuuidNzzwsJXW/Zr/k5EyRWtHJNlyPOU+zXzJVK/Clz57hnddmxueGYxkv1XedBfqXkN0zW3h4yuJRrx8EWijt6q0Wjm1ycrpEqxvw3asbvH1+ckdt8yh4+2GM60UPrW543It7Pymp3TeP4+btmXF4vMyb61mgfwl53KqVR73+G99fYXmjP3SvUkgp+ZG3T2AaWuBr+8V1mBfbflNSL+vFnvHyIp/1ADKOntHM1rEMWkPjjd0dhft9/ei5KEoJY60g+drpOpWidn86SsnejhtqjfltHPUYMjKOI9mM/iXlcWe2D3v9VK3AJxsz923YHnVVw+OmpDIyXhayGX3GgXBhsUZ/EOH50dghqj+IuLBYe6nGkJFxHMkCfcaB8LjpoBd1DBkZx5EsdZNxYByHjc7jMIaMjONGNqPPyMjIeMHJAn1GRkbGC04W6DMyMjJecLJAn5GRkfGC86w3Yw2AlZWVZzyMjIyMjOeHbTHT2M/rn3Wgnwf4+Z//+Wc8jIyMjIznknngo0e96FkH+m8DnweWgeQBr1kAvjZ83Z0jGtdx42U/Bi/754fsGLzsnx92HoNldJD/9n7e+EwDfbPZDIDffdhrGo3G6J93ms3mjcMe03HkZT8GL/vnh+wYvOyfH/Y8Bo+cyY/INmMzMjIyXnCyQJ+RkZHxgpMF+oyMjIwXnOch0LeBvzH8/5eVNi/3MWjzcn9+yI5Bm5f788NTHAOhlDrowWRkZGRkHCOehxl9RkZGRsZTkAX6jIyMjBecZ90wtYNGo/Fngb8C2MDfaTab/2DX858A/iugCvwO8L9uNpvxUY/zsNjH5/8pdI5OANeBP99sNltHPtBD5FHHYNvrfhL4+81m8+xRju8o2Md50AD+MVAHVoB//0U6D/bx+X8I/flt4DbwHzSbzfZRj/OwaTQaFeD3gT+xu3fgcWPhsZnRNxqNk8DfAn4UeBv4i41G441dL/vvgP+k2Wy+ig52/9HRjvLweNTnH37p/wj4yWaz+TZwEfjrz2Coh8Y+zwEajcYs8J+jz4EXin2cBwL4n4C/PTwP/gD4z57FWA+DfZ4D/wXwV4efvwn8p0c7ysOn0Wh8Bt1M+uoDXvJYsfDYBHrgi8BvNJvNrWaz6QK/DPzs6MlGo3EayDebzW8MH/onwM8d+SgPj4d+fsAC/uNms3l3+PNFYPGIx3jYPOoYjPhF9MrmReRRx+CHALfZbP7q8Of/K7Dnquc5ZT/ngAFUhv8uAIMjHN9R8R8B/1tgafcTTxILj1Pq5gRav2HEMvDpRzy/cATjOioe+vmbzeYm8K8AGo1GHj2L+3tHOL6j4FHnAI1G4xeA7wDf4MXkUcfgPLDSaDT+KfBJ4HvAf3J0wzt0HnkOAP8H4NcajcbfBVzgM0cztKOj2Wz+Bdghe7Cdx46Fx2lGv9cyPH2M55939vX5Go1GFfgV4P1ms/lPD31UR8tDj0Gj0XgT+Bng/3xkIzp6HnUemMAfAf5es9l8C7gG/D+OYFxHxaPOgTzwXwNfaDab88A/BP7ZEY3tuPDYsfA4Bfq7wNy2n+fZuWx51PPPO4/8fI1GYx6tXvc+8BeObmhHxqOOwc8NH3sXfbM70Wg0vnZ0wzsSHnUMVoCrzWbz3eHP/2/un/E+zzzq878JDJrN5reGP/9j9I3vZeKxY+FxCvRfBb7QaDSmG41GAT1zG+UhaTabNwG/0Wj8yPChPwf8f49+mIfGQz9/o9EwgC8D/6LZbP6lZrP5Ina6Peoc+GvNZvPVZrP5CeCPA0vNZvPzz2aoh8ZDjwG6CmO60Wi8Pfz53wHeO+IxHiaP+vwfAqca93IaP8U+pXpfFJ4kFh6bQD/cZPzLwG8C3wV+qdlsfqvRaPxKo9H41PBlPw/8nUajcQkoAv/PZzLYQ2Afn/9PonOyP9toNL47/O8Xn92ID559ngMvNI86Bs1mcwD8KeC/ajQaPwB+Avg/PrMBHzD7+Pwt4D8E/kWj0bgI/C+AP/+sxnuUPE0szCQQMjIyMl5wjs2MPiMjIyPjcMgCfUZGRsYLThboMzIyMl5wskCfkZGR8YJznDpjMzIyMl4aHiZatut1Ty1il83oMzIyMo6YfYiWjV53ICJ22Yw+IyMj4+gZiZb9t6MHGo3GnwP+EnoC/t7w+Y9xv4hd7XH/WFZHn5GRkfGMaDQaN9ASDkXg/wX80Waz6Tcajf8bWrDtKvA/B9bZJmLXbDa3HufvZKmbjIyMjGfPjwMXgG80Go3voqUdXuOAROyy1E1GRkbGs8dA61j9AkCj0Sih4/M73C9i98uP+8uzGX1GRkbGs+e3gD/VaDRmhhuw/widrz8QEbss0GdkZGQ8Y5rN5vto17TfAH6Ajs1/+6BE7LLN2IyMjIwXnGxGn5GRkfGCkwX6jIyMjBecLNBnZGRkvOBkgT4jIyPjBScL9BkZGRkvOFmgz8jIyHjByQJ9RkZGxgtOFugzMjIyXnD+/+zObQrEIdK5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(s1_df.x, s1_df.y, alpha=0.3)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "散布図を見ると，このデータセットの特徴が分かる．\n",
    "目視レベルでは，データは**15個**のクラスタに分かれているように見える．\n",
    "\n",
    "では，K-meansクラスタリングによって，目論見どおりにデータを15個のクラスタに分けられるか試してみよう．\n",
    "以下のコードを実行する．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KMeans(init='random', n_clusters=15)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# クラスタリングを実行\n",
    "model = KMeans(n_clusters=15, init='random')\n",
    "model.fit(s1_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "たった2行でクラスタリングが完了した．\n",
    "クラスタリング結果は`model`が保持している．\n",
    "各データがどのクラスタに分類されたかを抽出するには，以下のコードを実行する．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 8,  8,  8, ..., 10, 10, 10], dtype=int32)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 結果を格納\n",
    "labels = model.labels_\n",
    "\n",
    "# 表示\n",
    "labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "K-meansクラスタリングの結果を踏まえて，散布図を色分けしてみよう．\n",
    "今回の分析ではクラスタの数が15と多いので，下記のコードを実行して色分けのための準備を行う．"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "下記コードを実行して，散布図を15のグループごとに色分けを行ってみよう．\n",
    "K-meansクラスタリングは，目論見どおりにデータを分割できていただろうか？（[★Quiz 3](#C3-Q3)，[★Quiz 4](#C3-Q4)）"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x13fdd92e0>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(\n",
    "    s1_df.x, s1_df.y,\n",
    "    c=labels, # クラスタに応じて色分け\n",
    "    alpha=0.4, # 透明度\n",
    "    s=40, # マーカーのサイズ\n",
    "    cmap=\"tab20_r\" # カラーパレット（20色対応）\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "\n",
    "## 例題2: とある購買データに対する階層的クラスタリング"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "次はあるeコマースサイトにおける購買データを用いて，階層的クラスタリングを用いる例題に取り組んでみよう．\n",
    "まずは何も考えずに，以下のコードを実行してHands-onで用いるデータをダウンロードしよう．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "tags": [
     "remove-output"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2025-03-18 15:22:06--  https://gihyo.jp/assets/files/book/2014/978-4-7741-6674-2/download/DM_sampledata.zip\n",
      "gihyo.jp (gihyo.jp) をDNSに問いあわせています... 172.67.22.15, 104.22.59.251, 104.22.58.251\n",
      "gihyo.jp (gihyo.jp)|172.67.22.15|:443 に接続しています... 接続しました。\n",
      "HTTP による接続要求を送信しました、応答を待っています... 200 OK\n",
      "長さ: 41124 (40K) [application/zip]\n",
      "`DM_sampledata.zip' に保存中\n",
      "\n",
      "DM_sampledata.zip   100%[===================>]  40.16K  --.-KB/s 時間 0s         \n",
      "\n",
      "2025-03-18 15:22:06 (111 MB/s) - `DM_sampledata.zip' へ保存完了 [41124/41124]\n",
      "\n",
      "Archive:  DM_sampledata.zip\n",
      "  inflating: DM_sampledata/ch10_hbayes.stan  \n",
      "  inflating: DM_sampledata/ch3_2_2.txt  \n",
      "  inflating: DM_sampledata/ch3_4_1.txt  \n",
      "  inflating: DM_sampledata/ch4_3_2.txt  \n",
      "  inflating: DM_sampledata/ch5_3.txt  \n",
      "  inflating: DM_sampledata/ch6_4_1.txt  \n",
      "  inflating: DM_sampledata/ch6_4_2.txt  \n",
      "  inflating: DM_sampledata/ch7_3_2.txt  \n",
      "  inflating: DM_sampledata/ch8_test.txt  \n",
      "  inflating: DM_sampledata/ch8_test_label.txt  \n",
      "  inflating: DM_sampledata/ch8_train.txt  \n",
      "  inflating: DM_sampledata/ch9_2.txt  \n",
      "  inflating: DM_sampledata/ch9_basket.txt  \n"
     ]
    }
   ],
   "source": [
    "# zipファイルをダウンロードし解凍\n",
    "!wget https://gihyo.jp/assets/files/book/2014/978-4-7741-6674-2/download/DM_sampledata.zip\n",
    "!unzip DM_sampledata.zip"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "次にダウンロードしたデータをデータフレーム化して，`e_commerce_df`変数に代入する．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# データをpandasのデータフレームとして読み込む\n",
    "e_commerce_df = pd.read_table(\"DM_sampledata/ch5_3.txt\", header=0, sep=\" \")\n",
    "\n",
    "# ファイルを読み込んだので，元ファイル/ディレクトリを削除しておく\n",
    "!rm -rf DM_sampledata.zip DM_sampledata"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "変数``e_commerce_df``に読み込まれたデータフレームには，\n",
    "* 書籍\n",
    "* 衣服\n",
    "* 化粧品\n",
    "* 食料品\n",
    "* 飲料\n",
    "\n",
    "を1年間で何回購入したかについて，100名の顧客のデータが記録されている（このデータは[書籍「手を動かしながら学ぶビジネスに活かすデータマイニング」のサポートページにて公開されているデータ](https://gihyo.jp/book/2014/978-4-7741-6674-2/support)である）．\n",
    "\n",
    "以下のコードを実行し，中身を確認してみよう．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>books</th>\n",
       "      <th>cloths</th>\n",
       "      <th>cosmetics</th>\n",
       "      <th>foods</th>\n",
       "      <th>liquors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>43</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>25</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>19</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>31</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>46</td>\n",
       "      <td>9</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   books  cloths  cosmetics  foods  liquors\n",
       "0     43       0          3      4       10\n",
       "1     25       5          5      3       11\n",
       "2     19       0          2      3        8\n",
       "3     31       3          3      4        5\n",
       "4     46       9          2      5        7"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 最初の10件を表示\n",
    "e_commerce_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "### データ間の距離の計算"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "Pythonで階層的クラスタリングを実行するには，`scipy`ライブラリの``linkage``関数を用いるのが簡単だ．\n",
    "``linkage``関数は入力（第1引数）として **ベクトル間の距離（を格納した行列データ）** が必要となる．\n",
    "ここでベクトルとは，n個の特徴量からなるデータを意味する．\n",
    "今回の``e_commerce_df``のデータフレームでは，1行1行が1名の顧客情報を表すベクトルデータとなる．\n",
    "では，距離の情報を用意しよう．\n",
    "\n",
    "正攻法でやるならば，100行ある``e_commerce_df``の顧客データのすべての組み合わせを比較して，各顧客ベクトルの距離を計算する必要がある．\n",
    "`scipy`ライブラリはそれを簡単に実行してくれる便利な関数``distance.pdist``を提供している．\n",
    "\n",
    "以下のコードを実行してベクトル間の距離を計算してみよう（[★Quiz 1](#C3-Q1)，[★Quiz 2](#C3-Q2)）．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>90</th>\n",
       "      <th>91</th>\n",
       "      <th>92</th>\n",
       "      <th>93</th>\n",
       "      <th>94</th>\n",
       "      <th>95</th>\n",
       "      <th>96</th>\n",
       "      <th>97</th>\n",
       "      <th>98</th>\n",
       "      <th>99</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>18.841444</td>\n",
       "      <td>24.124676</td>\n",
       "      <td>13.341664</td>\n",
       "      <td>10.049876</td>\n",
       "      <td>18.708287</td>\n",
       "      <td>14.628739</td>\n",
       "      <td>41.388404</td>\n",
       "      <td>13.228757</td>\n",
       "      <td>9.486833</td>\n",
       "      <td>...</td>\n",
       "      <td>8.426150</td>\n",
       "      <td>19.442222</td>\n",
       "      <td>14.933185</td>\n",
       "      <td>35.270384</td>\n",
       "      <td>5.744563</td>\n",
       "      <td>12.124356</td>\n",
       "      <td>23.452079</td>\n",
       "      <td>18.920888</td>\n",
       "      <td>44.068129</td>\n",
       "      <td>17.378147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>18.841444</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>8.888194</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>22.045408</td>\n",
       "      <td>11.789826</td>\n",
       "      <td>22.248595</td>\n",
       "      <td>25.416530</td>\n",
       "      <td>25.179357</td>\n",
       "      <td>23.302360</td>\n",
       "      <td>...</td>\n",
       "      <td>15.427249</td>\n",
       "      <td>15.968719</td>\n",
       "      <td>17.663522</td>\n",
       "      <td>49.386233</td>\n",
       "      <td>18.920888</td>\n",
       "      <td>16.431677</td>\n",
       "      <td>17.748239</td>\n",
       "      <td>9.110434</td>\n",
       "      <td>26.589472</td>\n",
       "      <td>23.216374</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>24.124676</td>\n",
       "      <td>8.888194</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>12.806248</td>\n",
       "      <td>28.548205</td>\n",
       "      <td>13.416408</td>\n",
       "      <td>29.866369</td>\n",
       "      <td>25.238859</td>\n",
       "      <td>30.347982</td>\n",
       "      <td>27.856777</td>\n",
       "      <td>...</td>\n",
       "      <td>21.283797</td>\n",
       "      <td>22.583180</td>\n",
       "      <td>23.388031</td>\n",
       "      <td>56.797887</td>\n",
       "      <td>24.515301</td>\n",
       "      <td>20.024984</td>\n",
       "      <td>24.083189</td>\n",
       "      <td>12.247449</td>\n",
       "      <td>26.191602</td>\n",
       "      <td>30.033315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>13.341664</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>12.806248</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>16.340135</td>\n",
       "      <td>7.071068</td>\n",
       "      <td>19.390719</td>\n",
       "      <td>29.376862</td>\n",
       "      <td>18.193405</td>\n",
       "      <td>15.556349</td>\n",
       "      <td>...</td>\n",
       "      <td>9.327379</td>\n",
       "      <td>13.564660</td>\n",
       "      <td>11.958261</td>\n",
       "      <td>44.810713</td>\n",
       "      <td>12.124356</td>\n",
       "      <td>7.937254</td>\n",
       "      <td>17.146428</td>\n",
       "      <td>7.348469</td>\n",
       "      <td>32.341923</td>\n",
       "      <td>18.654758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10.049876</td>\n",
       "      <td>22.045408</td>\n",
       "      <td>28.548205</td>\n",
       "      <td>16.340135</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>20.856654</td>\n",
       "      <td>8.185353</td>\n",
       "      <td>39.597980</td>\n",
       "      <td>12.409674</td>\n",
       "      <td>8.426150</td>\n",
       "      <td>...</td>\n",
       "      <td>7.874008</td>\n",
       "      <td>14.525839</td>\n",
       "      <td>12.884099</td>\n",
       "      <td>28.583212</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>13.190906</td>\n",
       "      <td>18.193405</td>\n",
       "      <td>19.261360</td>\n",
       "      <td>43.162484</td>\n",
       "      <td>8.888194</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>12.124356</td>\n",
       "      <td>16.431677</td>\n",
       "      <td>20.024984</td>\n",
       "      <td>7.937254</td>\n",
       "      <td>13.190906</td>\n",
       "      <td>9.539392</td>\n",
       "      <td>18.357560</td>\n",
       "      <td>34.146742</td>\n",
       "      <td>11.661904</td>\n",
       "      <td>9.746794</td>\n",
       "      <td>...</td>\n",
       "      <td>9.055385</td>\n",
       "      <td>14.106736</td>\n",
       "      <td>8.485281</td>\n",
       "      <td>40.484565</td>\n",
       "      <td>8.831761</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>18.466185</td>\n",
       "      <td>13.228757</td>\n",
       "      <td>37.881394</td>\n",
       "      <td>15.459625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>23.452079</td>\n",
       "      <td>17.748239</td>\n",
       "      <td>24.083189</td>\n",
       "      <td>17.146428</td>\n",
       "      <td>18.193405</td>\n",
       "      <td>18.708287</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>23.811762</td>\n",
       "      <td>24.269322</td>\n",
       "      <td>22.315914</td>\n",
       "      <td>...</td>\n",
       "      <td>16.583124</td>\n",
       "      <td>5.830952</td>\n",
       "      <td>16.155494</td>\n",
       "      <td>40.570926</td>\n",
       "      <td>19.672316</td>\n",
       "      <td>18.466185</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>13.928388</td>\n",
       "      <td>27.166155</td>\n",
       "      <td>12.727922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>18.920888</td>\n",
       "      <td>9.110434</td>\n",
       "      <td>12.247449</td>\n",
       "      <td>7.348469</td>\n",
       "      <td>19.261360</td>\n",
       "      <td>10.198039</td>\n",
       "      <td>20.928450</td>\n",
       "      <td>23.216374</td>\n",
       "      <td>23.643181</td>\n",
       "      <td>20.297783</td>\n",
       "      <td>...</td>\n",
       "      <td>12.529964</td>\n",
       "      <td>11.832160</td>\n",
       "      <td>15.394804</td>\n",
       "      <td>47.138095</td>\n",
       "      <td>16.822604</td>\n",
       "      <td>13.228757</td>\n",
       "      <td>13.928388</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>26.758176</td>\n",
       "      <td>19.697716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>44.068129</td>\n",
       "      <td>26.589472</td>\n",
       "      <td>26.191602</td>\n",
       "      <td>32.341923</td>\n",
       "      <td>43.162484</td>\n",
       "      <td>30.854497</td>\n",
       "      <td>41.158231</td>\n",
       "      <td>8.185353</td>\n",
       "      <td>47.021272</td>\n",
       "      <td>45.563143</td>\n",
       "      <td>...</td>\n",
       "      <td>37.960506</td>\n",
       "      <td>29.933259</td>\n",
       "      <td>37.054015</td>\n",
       "      <td>67.216070</td>\n",
       "      <td>42.178193</td>\n",
       "      <td>37.881394</td>\n",
       "      <td>27.166155</td>\n",
       "      <td>26.758176</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>39.597980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>17.378147</td>\n",
       "      <td>23.216374</td>\n",
       "      <td>30.033315</td>\n",
       "      <td>18.654758</td>\n",
       "      <td>8.888194</td>\n",
       "      <td>21.587033</td>\n",
       "      <td>8.366600</td>\n",
       "      <td>36.069378</td>\n",
       "      <td>15.716234</td>\n",
       "      <td>13.490738</td>\n",
       "      <td>...</td>\n",
       "      <td>13.076697</td>\n",
       "      <td>10.954451</td>\n",
       "      <td>13.527749</td>\n",
       "      <td>29.427878</td>\n",
       "      <td>12.845233</td>\n",
       "      <td>15.459625</td>\n",
       "      <td>12.727922</td>\n",
       "      <td>19.697716</td>\n",
       "      <td>39.597980</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>100 rows × 100 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           0          1          2          3          4          5   \\\n",
       "0    0.000000  18.841444  24.124676  13.341664  10.049876  18.708287   \n",
       "1   18.841444   0.000000   8.888194   9.000000  22.045408  11.789826   \n",
       "2   24.124676   8.888194   0.000000  12.806248  28.548205  13.416408   \n",
       "3   13.341664   9.000000  12.806248   0.000000  16.340135   7.071068   \n",
       "4   10.049876  22.045408  28.548205  16.340135   0.000000  20.856654   \n",
       "..        ...        ...        ...        ...        ...        ...   \n",
       "95  12.124356  16.431677  20.024984   7.937254  13.190906   9.539392   \n",
       "96  23.452079  17.748239  24.083189  17.146428  18.193405  18.708287   \n",
       "97  18.920888   9.110434  12.247449   7.348469  19.261360  10.198039   \n",
       "98  44.068129  26.589472  26.191602  32.341923  43.162484  30.854497   \n",
       "99  17.378147  23.216374  30.033315  18.654758   8.888194  21.587033   \n",
       "\n",
       "           6          7          8          9   ...         90         91  \\\n",
       "0   14.628739  41.388404  13.228757   9.486833  ...   8.426150  19.442222   \n",
       "1   22.248595  25.416530  25.179357  23.302360  ...  15.427249  15.968719   \n",
       "2   29.866369  25.238859  30.347982  27.856777  ...  21.283797  22.583180   \n",
       "3   19.390719  29.376862  18.193405  15.556349  ...   9.327379  13.564660   \n",
       "4    8.185353  39.597980  12.409674   8.426150  ...   7.874008  14.525839   \n",
       "..        ...        ...        ...        ...  ...        ...        ...   \n",
       "95  18.357560  34.146742  11.661904   9.746794  ...   9.055385  14.106736   \n",
       "96  16.000000  23.811762  24.269322  22.315914  ...  16.583124   5.830952   \n",
       "97  20.928450  23.216374  23.643181  20.297783  ...  12.529964  11.832160   \n",
       "98  41.158231   8.185353  47.021272  45.563143  ...  37.960506  29.933259   \n",
       "99   8.366600  36.069378  15.716234  13.490738  ...  13.076697  10.954451   \n",
       "\n",
       "           92         93         94         95         96         97  \\\n",
       "0   14.933185  35.270384   5.744563  12.124356  23.452079  18.920888   \n",
       "1   17.663522  49.386233  18.920888  16.431677  17.748239   9.110434   \n",
       "2   23.388031  56.797887  24.515301  20.024984  24.083189  12.247449   \n",
       "3   11.958261  44.810713  12.124356   7.937254  17.146428   7.348469   \n",
       "4   12.884099  28.583212   6.000000  13.190906  18.193405  19.261360   \n",
       "..        ...        ...        ...        ...        ...        ...   \n",
       "95   8.485281  40.484565   8.831761   0.000000  18.466185  13.228757   \n",
       "96  16.155494  40.570926  19.672316  18.466185   0.000000  13.928388   \n",
       "97  15.394804  47.138095  16.822604  13.228757  13.928388   0.000000   \n",
       "98  37.054015  67.216070  42.178193  37.881394  27.166155  26.758176   \n",
       "99  13.527749  29.427878  12.845233  15.459625  12.727922  19.697716   \n",
       "\n",
       "           98         99  \n",
       "0   44.068129  17.378147  \n",
       "1   26.589472  23.216374  \n",
       "2   26.191602  30.033315  \n",
       "3   32.341923  18.654758  \n",
       "4   43.162484   8.888194  \n",
       "..        ...        ...  \n",
       "95  37.881394  15.459625  \n",
       "96  27.166155  12.727922  \n",
       "97  26.758176  19.697716  \n",
       "98   0.000000  39.597980  \n",
       "99  39.597980   0.000000  \n",
       "\n",
       "[100 rows x 100 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# pdist関数を使い，全ベクトル間の距離を計算．距離の定義にはユークリッド距離を用いる\n",
    "dist_matrix = distance.pdist(e_commerce_df, metric='euclidean')\n",
    "\n",
    "# 距離行列を表示\n",
    "pd.DataFrame(distance.squareform(dist_matrix))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "100行あるベクトルデータのすべての組み合わせについて距離を計算した結果（距離行列）が``dist_matrix``として得られた．\n",
    "この距離行列を使って階層的クラスタリングを実行してみよう．"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "### 階層的クラスタリングの実行\n",
    "\n",
    "講義でも述べたように，階層的クラスタリングではクラスタの作り方の基準として\n",
    "* 最長距離法（complete linkage method）\n",
    "* 最短距離法（single linkage method）\n",
    "* 重心法（centroid method）\n",
    "* 群平均法（group average method）\n",
    "* ウォード法（Ward's method）\n",
    "\n",
    "などがある．\n",
    "今回はウォード法を使ってみよう．\n",
    "以下のコードを実行して，ウォード法を用いた階層的クラスタリングを適用する．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "e_commerce_result = linkage(dist_matrix, method=\"ward\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "階層的クラスタリングの結果が変数``e_commerce_result``に格納された．\n",
    "\n",
    "階層的クラスタリングにおいて，各要素（ベクトル）が徐々に併合されていく結果を示したものを**デンドログラム**と呼ぶ．\n",
    "以下のコードを実行し，デンドログラムを表示させてみよう．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1008 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 図の大きさの設定\n",
    "plt.figure(figsize=(10, 14))\n",
    "\n",
    "# 図の表示\n",
    "fig = dendrogram(e_commerce_result,\n",
    "                 leaf_font_size=8, orientation='left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "デンドログラムが表示されると，階層的クラスタリングを実行したという気分になる（[★Quiz 6](#C3-Q6)）．\n",
    "\n",
    "今回は顧客の一人一人について私たちは特に知識がないため，この図を眺めてもよく分からない．\n",
    "そこで，適当な深さでデンドログラムの枝を切って，各要素をクラスタに分けてみよう．\n",
    "その上で，各クラスタに入っている要素がどのような性質を持っているかを分析してみよう．\n",
    "\n",
    "デンドログラムの枝を切り，指定したクラスタ数に分割するには``cutree``関数を用いる．\n",
    "``k``に数値を指定すると，デンドログラムをその数のクラスタに分割することができる．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 1, 1, 0, 1, 2, 3, 0, 0])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 4つのクラスタに分割\n",
    "e_commerce_cluster = cut_tree(e_commerce_result, n_clusters=4)\n",
    "\n",
    "# 最初の10件のみ表示\n",
    "e_commerce_cluster[:10, 0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "変数``e_commerce_cluster``に100名の顧客のクラスタ割り当て番号が格納された．\n",
    "\n",
    "さて，クラスタリングの結果は変数``e_commerce_cluster``に得られたが，各クラスタがどのような購買傾向を持っているかを分析するには，``e_commerce_df``データと照らし合わせる必要がある．\n",
    "そこで，以下のコードを実行して，``e_commerce_cluster``と``e_commerce_df``を結合してみよう．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>books</th>\n",
       "      <th>cloths</th>\n",
       "      <th>cosmetics</th>\n",
       "      <th>foods</th>\n",
       "      <th>liquors</th>\n",
       "      <th>cluster_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>43</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>25</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>11</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>19</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>31</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>46</td>\n",
       "      <td>9</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   books  cloths  cosmetics  foods  liquors  cluster_id\n",
       "0     43       0          3      4       10           0\n",
       "1     25       5          5      3       11           1\n",
       "2     19       0          2      3        8           1\n",
       "3     31       3          3      4        5           1\n",
       "4     46       9          2      5        7           0"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 結合\n",
    "new_e_commerce_df = e_commerce_df.assign(\n",
    "    cluster_id = e_commerce_cluster[:, 0]\n",
    ")\n",
    "\n",
    "# 先頭の数件のみ表示\n",
    "new_e_commerce_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "各顧客データがどのクラスタに分類するかのデータが得られた．\n",
    "このデータを用いて，クラスタ毎に書籍の平均購買数，衣類の平均購買数などを調べてみよう．\n",
    "\n",
    "今回の分析のように，クラスタ毎に何らかの情報をまとめて計算することを**集約演算**と呼ぶ．\n",
    "やや複雑に見えるが，以下のコードを実行してみてよう．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>books</th>\n",
       "      <th>cloths</th>\n",
       "      <th>cosmetics</th>\n",
       "      <th>foods</th>\n",
       "      <th>liquors</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cluster_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>48.363636</td>\n",
       "      <td>8.454545</td>\n",
       "      <td>4.636364</td>\n",
       "      <td>5.636364</td>\n",
       "      <td>3.863636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>26.562500</td>\n",
       "      <td>6.187500</td>\n",
       "      <td>4.500000</td>\n",
       "      <td>4.750000</td>\n",
       "      <td>5.531250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>35.923077</td>\n",
       "      <td>17.730769</td>\n",
       "      <td>4.423077</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>6.923077</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9.200000</td>\n",
       "      <td>14.400000</td>\n",
       "      <td>5.350000</td>\n",
       "      <td>4.250000</td>\n",
       "      <td>8.400000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                books     cloths  cosmetics     foods   liquors\n",
       "cluster_id                                                     \n",
       "0           48.363636   8.454545   4.636364  5.636364  3.863636\n",
       "1           26.562500   6.187500   4.500000  4.750000  5.531250\n",
       "2           35.923077  17.730769   4.423077  5.000000  6.923077\n",
       "3            9.200000  14.400000   5.350000  4.250000  8.400000"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_e_commerce_df.groupby(\n",
    "    'cluster_id' # cluster_idでデータをまとめる\n",
    ").agg(\n",
    "    'mean' # まとまり毎に平均値を算出\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "クラスタ毎に書籍，衣類，化粧品，食料品，飲料の平均購買数が求まった．\n",
    "ユーザの購買傾向がクラスタ毎に少しずつ異なることが確認できる（[★Quiz 5](#C3-Q5)）．"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "\n",
    "---\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "## クイズ\n",
    "※ 以下のクイズの回答にGoogle Colaboratoryを使いたい方は[コチラ](https://colab.research.google.com/github/hontolab-courses/ml-lecturenote/blob/main/content/quiz/kmeans-and-hierarchical-clustering.ipynb)をクリック．"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": [],
    "user_expressions": []
   },
   "source": [
    "(C3-Q1)=\n",
    "### Q1: ユークリッド距離とコサイン類似度\n",
    "\n",
    "以下の図のように，2次元のユークリッド空間上に，13個のデータ点が図のように分布している状況を考える（各点の座標については，下記コードのリスト``data``に収められている）．\n",
    "\n",
    "N次元空間上の任意の点$v_1$，$v_2$が与えられたとき，$v_1$と$v_2$のユークリッド距離を求める関数``calc_euclidean_dist(v1, v2)``，コサイン類似度を求める関数``calc_cosine_sim(v1, v2)``を実装しなさい．\n",
    "さらに，下図の13つの点のすべての組み合わせについて，そのユークリッド距離とコサイン類似度を求めなさい．\n",
    "\n",
    "なお，実装する関数の第1引数および第2引数は，実数のリストを想定しなさい．"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "sns.set()\n",
    "sns.set_style('ticks')\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = [(2, 2), (3, 4), (5, 2), (10, 5), (9, 3),\n",
    "        (11, 4), (12, 3), (12, 6), (4, 10), (4, 8),\n",
    "        (6, 8), (7, 10), (7, 6)]\n",
    "\n",
    "xs, ys = zip(*data) # zipの逆操作\n",
    "\n",
    "# 可視化\n",
    "ax = sns.scatterplot(x=xs, y=ys)\n",
    "ax.set_xlim(0, 13)\n",
    "ax.set_ylim(0, 13)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "user_expressions": []
   },
   "source": [
    "(C3-Q2)=\n",
    "### Q2: 距離/類似度の解釈\n",
    "\n",
    "Q1で扱ったデータ集合において，点$v=(11, 4)$との最も類似する点（最近傍点）を求めたい．\n",
    "ユークリッド空間の各次元の意味，最近傍点の定義が下記のような設定である場合，点の類似性を評価する関数としてユークリッド距離，コサイン類似度のどちらを用いるのが適切か考察しながら，最近傍点を求めなさい．\n",
    "\n",
    "1. 各点は都市の位置情報を表しており，第1次元は「緯度」，第2次元は「経度」を意味する．最近傍点として，**物理的距離が近い都市**を見つけたいケース．\n",
    "2. 各点はユーザの購買傾向を表しており，第1次元は「ジャンルAに属する商品を購入した回数」，第2次元は「ジャンルBに属する商品を購入した回数」を意味する．最近傍点として，**購買傾向が似ているユーザ**を見つけたいケース．"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "(C3-Q3)=\n",
    "### Q3: ポケモンデータ再び\n",
    "[第1章pandas入門](/content/introduction-to-pandas.ipynb)のクイズ課題で用いたデータセット[pokemonData](https://github.com/lgreski/pokemonData)を再度ダウンロードし，その内容をpandasデータフレームに変換して変数`pokemon_df`に代入しなさい．\n",
    "その上で，データフレーム`pokemon_df`から属性`HP`，`Attack`，`Defense`，`Sp. Atk`，`Sp. Def`，`Speed`の値だけを抽出し，データフレームの最初の5件だけを表示しなさい．\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "(C3-Q4)=\n",
    "### Q4: K-means\n",
    "Q3で準備したデータフレーム`pokenmon_df`とK-meansクラスタリング手法を用いて，第1世代ポケモンを7グループに分けたい．\n",
    "属性`HP`，`Attack`，`Defense`，`Sp. Atk`，`Sp. Def`，`Speed`のみに注目して，第1世代ポケモンデータにK-meansクラスタリングを適用しなさい．"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": []
   },
   "source": [
    "(C3-Q5)=\n",
    "### Q5: クラスタの特徴\n",
    "Q4で求めた各クラスタに含まれるポケモンデータの`HP`，`Attack`，`Defense`，`Sp. Atk`，`Sp. Def`，`Speed`の平均値を求めなさい．"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(C3-Q6)=\n",
    "### Q6: 階層的クラスタリング結果の比較\n",
    "\n",
    "Q1で扱った13個のデータ点について，階層的クラスタリング法を用いて段階的にクラスタに分割したい．\n",
    "\n",
    "``scipy.cluster.hierarchy``ライブラリの``linkage``関数および``dendrogram``関数を用いて，Q1で扱ったデータに階層的クラスタリングを適用しなさい．\n",
    "その際，階層的クラスタリングのクラスタ生成手法として下記3つの方法を用いて，クラスタの生成過程を結果を比較しなさい：\n",
    "\n",
    "* セントロイド法（centroid）\n",
    "* 最短距離法（single linkage）\n",
    "* 最長距離法（complete linkage）\n",
    "\n",
    "なお，クラスタ間の距離の尺度にはユークリッド距離を用いること．"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}