CH05 Logistic Regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# CH5 Logistic Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "直接使用Pandas的DataFrame还是很简单的"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import progressbar\n",
    "import logging\n",
    "progressbar.streams.wrap_stdout()\n",
    "logging.basicConfig(level=logging.INFO,\n",
    "                    format='%(asctime)s - [line:%(lineno)d] - %(levelname)s: %(message)s',\n",
    "                    datefmt='%Y-%m-%d %H:%M:%S')\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2018-06-24 08:53:00 - [line:1] - INFO: test\n"
     ]
    }
   ],
   "source": [
    "logging.info(\"test\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.017612</td>\n",
       "      <td>14.053064</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.395634</td>\n",
       "      <td>4.662541</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0          1  2\n",
       "0 -0.017612  14.053064  0\n",
       "1 -1.395634   4.662541  1"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"./Data/testSet.txt\",sep=\"\\t\",header=None)\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def loadDataSet():\n",
    "    dataMat = [];labelMat = []\n",
    "    with open(\"./Data/testSet.txt\") as fr:\n",
    "        for line in fr.readlines():\n",
    "            lineArr = line.strip().split()\n",
    "            dataMat.append([1, float(lineArr[0]), float(lineArr[1])])\n",
    "            labelMat.append(int(lineArr[2]))\n",
    "    return dataMat, labelMat    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MLiA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "西瓜书上这样说：\n",
    "1. 直接对分类的可能进行建模，无需事先假设数据分布，避免了假设分布不准所带来的问题\n",
    "1. 能拿到近似概率的预测结果\n",
    "1. 对率函数任意阶可导凸函数，有很好的数学性质"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sigmoid Function"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "这个函数有个推导，[李航,2012,P78]\n",
    "- 有时候为了方便，将权值向量和输入向量加以扩充\n",
    "$$w=(w^{(1)},w^{(2)},...,w^{(n)},b)^T$$\n",
    "$$x=(x^{(1)},x^{(2)},...,x^{(n)},1)^T$$\n",
    "\n",
    "这时LogisticRegression 模型可以表示为\n",
    "$$P(Y=1\\vert x)=\\frac{exp(w\\cdot x)}{1+exp(w\\cdot x)}$$\n",
    "$$P(Y=0\\vert x)=\\frac{1}{1+exp(w\\cdot x)}$$\n",
    "\n",
    "用这个去理解“为了计算方便，该函数还将X0的值设为1.0”\n",
    "\n",
    "还有，李航和MLiA里面讲这里都是在78页。。。。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def sigmoid(inX):\n",
    "    return 1/(1 + np.exp(-inX))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Best W"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### GB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def gradAscent(dataMatIn, classLabels):\n",
    "    dataMatrix = np.mat(dataMatIn)\n",
    "    labelMatrix = np.mat(classLabels).transpose()\n",
    "    m,n = np.shape(dataMatrix)\n",
    "#     print(m,n)\n",
    "    alpha = 0.001\n",
    "    maxCycles = 500\n",
    "    weights = np.ones((n,1))\n",
    "    for k in list(range(maxCycles)):\n",
    "        h = sigmoid(dataMatrix*weights)\n",
    "#         print(np.shape(h))\n",
    "        error = (labelMatrix - h)\n",
    "        # return array\n",
    "#         weights += alpha*dataMatrix.transpose()*error\n",
    "        # return matrix\n",
    "        weights = weights + alpha*dataMatrix.transpose()*error\n",
    "    return weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dataArr,labelMat = loadDataSet()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "wei = gradAscent(dataArr,labelMat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "matrix([[ 4.12414349],\n",
       "        [ 0.48007329],\n",
       "        [-0.6168482 ]])"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wei"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6.685832121185382"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "-wei[0,0]/wei[2,0] # 最后可视化的截距"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "绘制决策边界"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plotBestFit(wei):\n",
    "    import matplotlib.pyplot as plt\n",
    "    weights = wei if type(wei).__name__ == 'ndarray' else wei.getA()\n",
    "    dataMat, labelMat = loadDataSet()\n",
    "    dataArr = np.array(dataMat)\n",
    "#     print(dataArr)\n",
    "    n = np.shape(dataArr)[0]\n",
    "    xcord1 = []; ycord1 = []\n",
    "    xcord2 = []; ycord2 = []\n",
    "    for i in range(n):\n",
    "        if int(labelMat[i]) == 1:\n",
    "            xcord1.append(dataArr[i,1]); ycord1.append(dataArr[i,2])\n",
    "        else:\n",
    "            xcord2.append(dataArr[i,1]);ycord2.append(dataArr[i,2])\n",
    "    fig = plt.figure()\n",
    "    ax = fig.add_subplot(111)\n",
    "    ax.scatter(xcord1,ycord1,s=30,c=\"red\",marker=\"s\")\n",
    "    ax.scatter(xcord2,ycord2,s=30,c=\"green\")\n",
    "    x = np.arange(-3.0, 3.0, 0.1)\n",
    "#   sigmoid = 0,从图像看 0是分类的分界点\n",
    "    y = (-weights[0]-weights[1]*x)/weights[2]\n",
    "    ax.plot(x,y)\n",
    "    plt.xlabel(\"X1\");plt.ylabel(\"X2\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3XuU3GWd5/H3t7tzI+kKxFzoJB0S\nIJI0co8dlEHxoEzIuOLsODOwoszoksSjLO7q4u14GR13NOroLI5CGFgZRZ3ZVZGjQUEcD+qRkAQT\nLrlITIBOusmFkL6QW1+++0dVd6qrq6rr111Vv0t9Xuf06a5f/br6W13dv289z/N9nsfcHRERkVLV\nhR2AiIjEixKHiIgEosQhIiKBKHGIiEggShwiIhKIEoeIiASixCEiIoEocYiISCBKHCIiEkhD2AFU\nwsyZM33hwoVhhyEiEhubN28+5O6zSjk3kYlj4cKFbNq0KewwRERiw8yeL/VcdVWJiEggShwiIhKI\nEoeIiARS8cRhZveY2QEzezrr2GfMbJ+Zbcl8rCzwvSvMbKeZ7TKzj1Y6VhERGV01WhzfAlbkOf5V\nd78487E+904zqwf+GbgWaAFuMLOWikYqIiKjqnjicPdHgcNj+NZWYJe773b3k8D3gevKGpyIiAQW\n5hjHB8zsyUxX1hl57p8HtGXd3ps5JiIiIQorcXwTOAe4GOgAvpLnHMtzrOA+t2a2ysw2mdmmgwcP\nlidKkRhr62zjlvW30HpXK7esv4W2zrbRv0mkBKFMAHT3/YNfm9ldwE/ynLYXaM66PR9oL/KY64B1\nAMuWLdNG6lLT2jrbuOiOi+g52UPvQC9bXtzCfU/dx9Y1W2me3jz6A4gUEUqLw8yasm7+OfB0ntM2\nAovNbJGZTQSuBx6oRnwicbf2t2uHkgZA70AvPSd7WPvbtSFHJklQ8RaHmX0PuAqYaWZ7gU8DV5nZ\nxaS7np4DVmfOnQv8i7uvdPc+M/sA8HOgHrjH3Z+pdLwiSbBh34ahpDGod6CXx/c9HlJEkiQVTxzu\nfkOew3cXOLcdWJl1ez0wolRXRIpbPm85W17cMix5TKibQOu81hCjkqTQzHGRBLrtituYNnEaE+om\nAOmkMW3iNG674raQI5MkUOIQSaDm6c1sXbOV1ZetpnVuK6svW62BcSmbRC6rLpIUbZ1trP3tWjbs\n28Dyecu57YrbSr74N09v5vaVt1c4QqlFShwiEaWSWokqdVWJRJRKaiWqlDhEIkoltRJVShwiEZK9\nTMjxvuM02PDe5KiU1Go5k9qmMQ6RiMgd02ioa6Df+2mwBvq8LzIltRp7EbU4RCIid0yjb6CPhroG\nls5aGqmSWo29iFocIhFRaExjSsMUNty8IaSoRtLYi6jFIRIRy+ctH5rpPSgqYxrZ4hKnVI4Sh0hE\nxGWZkLjEKZWjxCESEXFZJqQacapqK9rMPXl7Hi1btsw3bdoUdhgiMga5VVuDLZooJtEkMbPN7r6s\nlHPV4hCRSFHVVvQpcYhIpKhqK/qUOEQkUlS1FX1KHCKSV1gD1JWs2tKge3locFxERgh7gHpwH5Jf\nv/BrBnyAOqvjygVXBtqPJN9jatC9sEgNjpvZPWZ2wMyezjr2JTPbYWZPmtmPzOz0At/7nJk9ZWZb\nzEyZQKRKwh6gbp7ezG1X3MYLnS+w49AOtu7fyp2b7+SiOy4acysh7OeUJNXoqvoWsCLn2MPAa9z9\nQuAPwMeKfP+b3P3iUjOhiAw3lu6ZKAxQl/tCH4XnlBQVTxzu/ihwOOfYQ+7el7n5GDC/0nGI1KLB\n7pk7N9/JxvaNJb9rj8IAdbkv9FF4TkkRhcHx9wAPFrjPgYfMbLOZrSr2IGa2ysw2mdmmgwcPlj1I\nkTj65C8/yZHjRwK/a4/CsiLlvtBH4TklRaiJw8w+AfQB9xU45Qp3vxS4Fni/mb2h0GO5+zp3X+bu\ny2bNmlWBaKXSVPFSXm2dbXznqe/gDC+AKeVdexSWPyn3hT4KzykpqlJVZWYLgZ+4+2uyjt0ErAGu\ndvejJTzGZ4Aed//yaOeqqip+VPFSfresv4VvbPwGAwwMO24Y777w3TROamTDvg0sn7d8XNVKlTRY\nXfX4vsdpndca2TiTIEhVVSj7cZjZCuAjwBsLJQ0zmwrUuXt35utrgM9WMUypomIDobevvD3k6OJp\nw74NI5IGpBPH/Tvv52jv0cjv4Nc8vVmvfwRVoxz3e8DvgPPMbK+ZvRf4OtAIPJwptb0jc+5cM1uf\n+dY5wG/MbCvwOPBTd/9ZpeOVcKjipfzyjRHUUcfiVy0eShqgslQJruItDne/Ic/huwuc2w6szHy9\nG7iogqFJhCyft5wtL24ZljyiXPEy2IUS5a6e2664jfueum9E99/khslK0jIuUaiqEolVxctYS1yr\nrdBg8JULrlRZqoyLlhyRyIjLQOgt62/hzs13jmgdrb5sdSz646NUiJDbcrvxwhv5zpPfiXRLLqmC\nDI4rcUjNaets45O//CQP7noQDK4991o+96bPlXyBar2rlY3tG0cen9vKhps3lDvciohCks5NYA3W\nQL/3U19XT99AnyrrqizyVVUiYWnrbOOCb15A54nOoWP3br2X+3fcz1Pve6qkC1TcxmPyiUK1Um4l\nXV9mMYm+gfRnVdZFl8Y4JBFKnTy49rdr6TrRNeJ494nukquK4jQeE2X5KulyadA+mtTikNjL7fIo\nNi9hw74NI2ZSAwwwUPIFanDQOeyunrjL13LLFbeWXK1Qi0NiL8gqqsvnLcewEcfrqAt0gRrs6tlw\n8wZuX3m7ksYY5LbcGqwBw2ioS7+fVUsuupQ4JPaCTB687YrbSE1KjTjeOKmxJi5QUVoPLLdceM2y\nNfzuvb9jzWVrtJZUxKmrSmIvyGB18/RmnnrfU0WrqqI8uW88seV26f2+4/fc9cRdLJm5ZNy7641V\nvkH65fOXVzUGCU7luBJ75ZyXEKU5DuWOLd/8k0FRep4SjkhtHStSaeVcLjvK24uON7ZiVUxRep4S\nfeqqkkQo17yESiy2WK6ur/HGNloV02iPFeUuPKkuJQ6RLOWe3BekVLjSseUuepir2GOV83lI/Kmr\nSmKt3FVC5Z7cV86ur/HGlt2ld9Gci5hUP6nk0tcod+FJ9anFIbFViXfB5Z7cN57upXxdQ+ONLbtL\nL8h6VdovRbIpcUhsVWrXwGLjJUH7+VtmtbC5ffOwnfhK6V4qlhTLtW5TkHGhJKzPJeWjriqJrWq/\nCw66D0dbZxv377h/xPatp004bdTupah1DSVlfa4oTYCMMyUOia18W6NW8l1w0Iv52t+u5Wjv0WHH\nDOPt57191O6lqHUNlbPkOSxx2YArDqqSOMzsHjM7YGZPZx2bYWYPm9mzmc9nFPjemzLnPGtmN1Uj\nXomHar8LDnoxz3e+42w/tH3Un1XtpFiKuK/PFbVWXJxVq8XxLWBFzrGPAo+4+2LgkcztYcxsBvBp\nYDnQCny6UIKR2lPtd8FBL+bjufgnpWsoSqLWiouzqiQOd38UOJxz+Drg3szX9wJvz/Otfwo87O6H\n3f1l4GFGJiCpYdV8Fxz0Yj6ei3+1kmIt9flHsRUXV1Vbq8rMFgI/cffXZG4fcffTs+5/2d3PyPme\nDwOT3f3vM7c/CRxz9y8X+1laq0oqJeiWq1HYorWQKK/LVQm19nyDStLWsSM3TiDPLjyAma0CVgEs\nWLCgkjFJDQu6tEk1tmgd61IglSpnjiptwFU+YSaO/WbW5O4dZtYEHMhzzl7gqqzb84Ff5Xswd18H\nrIN0i6O8oYpE03gmQUahz7+UpFfONbKisNd6EoSZOB4AbgK+kPn84zzn/Bz4X1kD4tcAH6tOeBKE\nFsALR6FWw59998+Y3DC56GtRaNHDY33HaOtsq/jrV0rS0xpZ0VStctzvAb8DzjOzvWb2XtIJ4y1m\n9izwlsxtzGyZmf0LgLsfBj4HbMx8fDZzTCJE9fHhKdRqeOrAU6O+FoOD94PrVQ3afnB7VV6/Uspj\nVUIbTdWqqrrB3ZvcfYK7z3f3u939JXe/2t0XZz4fzpy7yd3/a9b33uPu52Y+/k814pVg9M8dnnyV\nQtkKvRaDLcTm6c2kJg7fSrfP+6ry+pXSVfboC4+G3p0mI0V9cFxiIAp95bVqtKXSYeRrkdv9U8r3\nVMJo61+1dbax89DOEd/XYA0qoQ2ZlhyRcVN9fHhy53tcMPsCGmz4+8Hc1yK3hZhPNV6/0ea5rP3t\nWvq9f8T31dfVayJkyJQ4ZNwKXQBuvPDGWE8ui8LkuFJiyJ4E+dP/8lMaJzUWnXRYbAvZQt9TCaNN\nctywbwN9A30jvm/JzCUaGA9Z1SYAVpMmAFZf7kS3Gy+8kWvvuza2k62iMFlsrDGMNunwlvW3cOfm\nO0d0ES2ZuYQpDVMiM7+hUJyrL1utktoKCDIBUIlDKiLu//RRiL9SMUQhKZYiLnEmRZJmjktMxX3A\nPArxVyqGuMygjkuc1XTsZD8793ezrb2LpU2NXLIgnDVflTikIuK+Y1wU4i81hrFMvozLDOq4xFkJ\nB7qPs70jnSS2dXSxrb2TPYdeYSDTSbT6jWeHljjUVSUVEfduhijEX0oMUYhTxqd/wNlzqIdn2rvS\niaKji23tXRzqOTF0zrzTp7C0KUXL3BQtTSnOn5ti/hlTMMu3nN/YaIxDiSMSorwybCmiEP9YB7rj\nMpZUa3pO9LGjo4vtHZlWREc3O1/s4nhvenvhCfXG4tmNQwliaVP68/TTCk/yLBclDiUOqRGtd7Wy\nsX3jyONzW9lw84YQIhIAd6ej83g6QWS6mrZ3dPHcS6e2Ej79tAlDyeH8uenP58yaxsSGgLMkUino\n7h55vLERurpKfhgNjovUiLKMxZTpwlOrevsH2HWgZ1iS2NbRxZGjp16Tha86jaVNKf7zpfOHkkTT\n9Mnl6WrK99oVO14GShwiMZa75MiYJu+FcOGJq86jvUOJYTBR7DrQw8n+dFfTpIY6ljSluPY1Zw61\nJpY0pZg2KVmX2mQ9G5Eao5LVynB39r58jGfahyeJfUeODZ0zc9pEWuZO58pXz6QlMxaxaOZUGuqT\nvyCHxjhEKiBW+5MU6y5J4PUh1/Hefp7d38O2js6h8tftHV10n0gvd1JnsGjmVM6fO32osmlpUyOz\nGyeHHHlGmV4/jXGIhEibD0XXSz0nMiWvnUPjEX88+Ar9mckRUyfWs6QpxdsvmTdU2fTqOY1MmVgf\ncuTRosQhUma1tpd3FPUPOM+/9MrQnIjB8tf9XafmRjRNn0xLU4prWs4cShILZpxGXV355kZURWNj\n4eKGClHiECmzKCxXEkgIF55yOnqyjx0vdg8re93R0c2x3vSS7A11xrmzp3HFuafGIpY2pThj6sSQ\nIy+TECrflDhEyiwKy5UEEpOSW3fnQPeJYSWv29u72PPSK0Nd+Y2TG2hpSnF9a/NQglh8/iImdb48\n8gFVbjxmoSUOMzsP+LesQ2cDn3L3r2WdcxXwY2BP5tAP3f2zVQtSZAzKUiKbAOMpEOjtH2D3wVeG\nDVhv6+ji8Csnh85pnjGFlqYUb7t47lCSyLsMR76kASo3HodIVFWZWT2wD1ju7s9nHb8K+LC7vzXI\n46mqSsIWheVKwhRkDa2u473s6OhmW3tnpqupm537uznZl54bMbGhjlfPmTbUzdQydzpLmhpJTS5x\nGY4arxorVRyrqq4G/pidNETirJZXdYUCBQInevjMI/+bvz7vQ6fmRnR00Xb41NyIGVMnsrSpkb95\n/UKWNjXS0jSds2dNZUINzI2Ik6gkjuuB7xW473VmthVoJ936eKZ6YYnIWDy2dxP0NTN1YBETB85m\nop/NhIFFPPL4NB55fDNmsOhVU7lw/ulc/9oFmZZEitmNk8q64qtURuhdVWY2kXRSON/d9+fclwIG\n3L3HzFYC/+Tuiws8zipgFcCCBQsue/55NV5EquHlV05mrfaaLn/dub8T93QrYYDj9Npz9NU/z+UL\n5/N319zMkjMbOW1ild63qquqJLFaHdfMrgPe7+7XlHDuc8Aydz9U7DyNcSRXrGZkJ8zAgPPC4aMj\n5kZ0dB4fOmdOalJ6kHoGrHvy0/QM7OCYv8CE+vrw9gnRIo4lidsYxw0U6KYyszOB/e7uZtYK1AEv\nVTM4iQ7NyB6HgBfP47397Hzx1KZC2zq62NHRxSsn03Mj6uuMs2dOZfmiGZklONIfM6dNGnqM1Vet\ni0aBgJJD2YWaOMzsNOAtwOqsY2sA3P0O4B3A+8ysDzgGXO9hN5EkNJqRPQ5FVsDN3qJ0sBWx+2DP\n0Bal0yY1sLSpkXdcNj8zw3o6iy84m8lHDo98vKxEVOsFAkkWauJw96PAq3KO3ZH19deBr1c7rsRI\nWBM9djOyK2Ecr2m/1bFnxlyemX0222cvYtvss9k2+2wOff6RoXMGtyhdeUHTUPnr/DOmjFyGI1/S\nAM2NqBFR6KqSSknYPguxm5FdCSW+pj0n+tj5YtbGQu/6CjtnncXxCekVXSf097L40AtctWcTSz92\n61CSqMYWpRJ/ShwSG5qRPZIDHY0z0y2IR54tvEXpyeO8c8vPWHpgNy3793DuS21MHEgvG85Pv5b/\nwUUKCL2qqhJUVZWRwDLEWp6R3ds/wK6mc9iW3dU0ZxFHpqSGzhncorRlaN+IzBaldUUm0I3lbyGB\nf1u1LlbluJWgxJGhf+4xC7vsd3CL0sHB6u0dXTy7P2uL0t4TnHfoeVoO7KFl/25aDuzmvF1baSy0\nDEe5x7v0txVcxMcclTiUONL0zz0mQdZZGq/cLUoHK5uGb1E6Kb38xtwULR9aQ8uBPSw6vI8GH8h9\nsLLGVlTEL4KRFPH/x7jN45BKifk+C2GpVNnv4Bal2zuGJ4ncLUovWXA6N15+Vv4tSv/699F4TZUc\napoSR5Lpn3tMylH2+1LPiVNdTXm2KD1tYj1Lm1Jcd8lcljalOH/udM4rZYtSvaYSAUocIjmClP0G\n3aI0nSRK2KI0CV1BSXgOkpfGOCQywh6Qzo4j3xjHY+/5PT3HUsMSxM4Xuzl6cvgWpdkVTUubUswY\nyxalEe8PL0kSnkM5Rfz3UbbB8czqtLPc/Y85xy909yfHF2blKHHETzUHpEfj7mzZt4d/+NW32dbe\nSaq+hfr+s9j78slhW5QOth4GE8W5s6cxqWGUrqZSBb3IRPHdfcQvlFUXxdcoS1kGx83sr4CvAQfM\nbALwN+6+MXP3t4BLxxuoyKCw1qHq6x9g96FXhrUitrV38dIrJ4H0/9AZM6aw9MwUf3HpqSQx7/Q8\nW5SGKWGrBCRSBJJDuRQb4/g4cJm7d2RWpv22mX3c3X8IROg/RpKgGutQZW9Rur0jvfJr7hal581p\n5M1L52TKX4tsURrxd4+hS6VGP0diq1jiaHD3DgB3f9zM3gT8xMzmk17pQKRsyrkOlbuz78ixoRVf\nt3V05t2itKUpNfYtSvUOvzj9HhKtWOLoMrNzBsc3Mi2Pq4D7gfOrEZzUjrGuQ3Wyb4BnDwwuCd6d\nThLtXXQdT8+NyN2idGlTI+fPnR79LUqTPAcnCc+hxhVLHB8hp0vK3bvNbAXwsYpGJTWneXozW9ds\nLboOVb4tSncd6KEvMzdiyoR6ljQ18taL5g6NRVR1i9JySnJ3V7Wem7oTK6ZgVZWZ7QbuBL7i7n2Z\nY3OArwDnuftrqxZlQKqqireBAaft5aOnlgTPDFy3Z21ROrtx0lBV09JMklj4qqnUF5sbUU5RqxiK\n2kUyCr+fKMQQI+VacuQy4B+A35vZrcAFwP8A1gLvHneUIsCxk/3s3N89NMN6e2a2dfYWpefMmspr\nF80YNj8ie4tSQe+gpaoKJg53fxlYk0kavwDagcvdfW+1gpNkOdh9YsQM60JblA4uw7F4zjQmTyjT\n3IhySvIYRDno95NoxeZxnA58EVgOrABWAg+a2a3u/ssqxScx1D/g7DnUw7acfawPdp9ahmP4FqXp\nqqa8W5RGld7hF6ffT6IV66p6AvgG8P7MGMdDZnYx8A0ze97dbyhHAGb2HNAN9AN9uX1sli59+SfS\nieso6YmIT5TjZ8v4jdiitKObnS92cbw3PTdiQr2xeHYjb3z1LFoyS3Boi1IZs2qM5URtvCiCiiWO\nN+R2S7n7FuD1ZnZzmeN4k7sfKnDftcDizMdy4JuZz1JF7k5H5/FTYxGZZJG9Ren0KRNoaUrxzuVn\nDY1HnDNrGhMbSpwbIbVjrBfnIPNnxtpdpjk6oyo2xlFwLMPd76pMOHldB/yrp8u/HjOz082saXBy\nopRfb/8Auw70DF+Go6OLI0dPTc4b3KL0Ly6dP1TV1DR9crTnRkh0VOPirNZBxUShwN1Jd4M5cKe7\nr8u5fx7QlnV7b+bYsMRhZquAVQALFiyoXLQJ03msd9ieEYNzI4a2KG2oY8mZjaw4/8z0DnRNKZY0\npZg2KQp/OhGnLg9JqCj891/h7u1mNht42Mx2uPujWffnews7ogg7k3DWQXoeR2VCjS93p+3wsaHW\nw2CyGL5F6USWNqW48tUL011NTSkWzZxKQ6nLcFRCnC++6vKQhAo9cbh7e+bzATP7EdAKZCeOvUD2\nutrzSZcGSwGDW5Ru6+gcWq8p3xall551BjdeftbQftbDtiiNCl18RSIn1MRhZlOBusxSJlOBa4DP\n5pz2APABM/s+6UHxTo1vnPJSz4mhNZoGk8Sugz15tyhtaZpOy9xUaVuUikRRNeaHaA7KqMJuccwB\nfpQZUG0AvuvuPzOzNQDufgewnnQp7i7S5bh/G1KsocrdonSwuyl7i9IzU5M5f26Kt7TMGZphfdZo\nW5RKvLvD4mqsF+dqvB56zUcVauJw993ARXmO35H1tQPvr2ZcYTt6so8dL3YPG7TOt0Xp68+ZObQD\n3dKmFGeMZYtSUXdYGHRxjrWwWxw1r/NYL0+88PKw0tc9h14ZtkVpS1OKv1rWPFTVtHhOGbcolcpR\nl4cklBJHyDY9d5j33pteybd5xhRamlK87aK5Qyu/Rm6L0mqL88VX76oloZQ4QvbaRTP499WvK7xF\naa2L68VX4yaSYEocIUtNnkDrohlhhyHlpnETSTAtIiQiIoEocYiISCBKHCIiEogSh4iIBKLEIVIJ\nhcqFq1VGnEqB2ciPVKo6P18STVVVIpUQdsmtqrqkgtTiEClX60Dv8qVGKHGIdHWB+8iPoK0Gvcsv\nTRQTbBRjijAlDhGprigm2CjGFGFKHCIiEogSh0gShV3VJYmmxCEStkr0r5dr3EYkDyUOkXIZ67t8\n9a9LzChxSH6qMgmuEu/yo/j7H+/fRhS70aIYU4SFljjMrNnM/sPMtpvZM2Z2a55zrjKzTjPbkvn4\nVBix1iS9C46mav3+iyWH8f5tRLEbLYoxRViYM8f7gA+5+xNm1ghsNrOH3X1bznm/dve3hhCfSO3S\nGwcpIrQWh7t3uPsTma+7ge3AvLDiERGR0kRijMPMFgKXABvy3P06M9tqZg+a2flVDUykGtSPLjET\neuIws2nAD4APuntuh+ITwFnufhFwO3B/kcdZZWabzGzTwYMHKxewSLnl61+vVSrKiIVQE4eZTSCd\nNO5z9x/m3u/uXe7ek/l6PTDBzGbmeyx3X+fuy9x92axZsyoad01QlUm4ovz7r2RsGluJhdAGx83M\ngLuB7e7+jwXOORPY7+5uZq2kE91LVQyzdqmaJFxh//4bG/NfrBsbw49NQhdmVdUVwLuAp8xsS+bY\nx4EFAO5+B/AO4H1m1gccA653r+V2vEiZFCqrHUwMSg5SRGiJw91/A9go53wd+Hp1IhKpIeoSknEI\nfXBcRETiRYlDkkeVOfEV5aIAGaLEIcmjbpjqK1ey1tIfsaDEISLjp2RdU5Q4RGqRuoRkHMIsxxWR\nsKjrR8ZBLQ6RYjTQLjKCEockTzm7YdR3LzKCuqokedQNU33FliiRxFGLQyRuoth9pjLamqLEIRI3\n6j6TkClxiIhIIEocIsVovoPICEocIsWo7768ojg+I4EpcUg8JO2Ck7TnUyqNzySCEofEQ9IuOON5\nPuo+k5BpHodI3KibTEKmFocMV6tdKCJSslATh5mtMLOdZrbLzD6a5/5JZvZvmfs3mNnC6kdZY5LW\nJZQkSuoSEaElDjOrB/4ZuBZoAW4ws5ac094LvOzu5wJfBb5Y3Sil7HTxG7skJHWNzyRCmC2OVmCX\nu+9295PA94Hrcs65Drg38/X/A642M6tijFJuY734Je2Ck7TnUyqVNydCmIPj84C2rNt7geWFznH3\nPjPrBF4FHKpKhBIdSbuwJO35SE0Js8WRr+XgYzgnfaLZKjPbZGabDh48OO7gREQkvzATx16gOev2\nfKC90Dlm1gBMBw7nezB3X+fuy9x92axZsyoQbo2o1S4UESlZmIljI7DYzBaZ2UTgeuCBnHMeAG7K\nfP0O4JfunrfFIWWiPujoUlKXiAhtjCMzZvEB4OdAPXCPuz9jZp8FNrn7A8DdwLfNbBfplsb1YcUr\n45RKFR8A18VvdEreEhGhzhx39/XA+pxjn8r6+jjwl9WOSyqgWNJQI1IkVjRzXCQuNAdGIkKJQyQu\nkjABUBJBiUNERAJR4hARkUCUOKQ6VEoqkhjaj0OqQ6WkIomhFodIXKjVJhGhFodIXKjVJhGhFocI\naI6ESABKHCKgORIiAShxgN5tSuXob0sSSIkD9G6z1lXyIq6/LUkgJQ4RXcRFAlHiEBGRQJQ4RIrR\nHAmREZQ4RIrR3AmREZQ4QDNypbDxVkXpb0sSSDPHQe8qa11jY/4B8kLHofQBdf1tSQKpxSHxUMn5\nEF1d6e1rcz900RfJK5QWh5l9CfhPwEngj8DfuvuRPOc9B3QD/UCfuy+rZpwSIZoPIRIZYbU4HgZe\n4+4XAn8APlbk3De5+8VKGiIi0RBK4nD3h9y9L3PzMWB+GHGIiEhwURjjeA/wYIH7HHjIzDab2api\nD2Jmq8xsk5ltOnjwYNmDlBqlqiiRESo2xmFmvwDOzHPXJ9z9x5lzPgH0AfcVeJgr3L3dzGYDD5vZ\nDnd/NN+J7r4OWAewbNkyH/cTEAENkIvkUbHE4e5vLna/md0EvBW42t3zXujdvT3z+YCZ/QhoBfIm\nDkm4YiWzIlJVoXRVmdkK4CMHNWcPAAAFEUlEQVTA29z9aIFzpppZ4+DXwDXA09WLUiJFJbMikRHW\nGMfXgUbS3U9bzOwOADOba2brM+fMAX5jZluBx4GfuvvPwglXREQGhTKPw93PLXC8HViZ+Xo3cFE1\n45IESaUKd22plSIyLlGoqhIpP00YFKkYJQ4REQlEiUNERAJR4hARkUCUOEREJBAlDkkmLRUiUjHa\nyEmSSSW3IhWjFoeIiASixCEiIoEocYiISCBKHCIiEogSh4iIBGIFtsKINTM7CDwfdhwBzAQOhR3E\nGCju6oljzKC4q2m8MZ/l7rNKOTGRiSNuzGyTuy8LO46gFHf1xDFmUNzVVM2Y1VUlIiKBKHGIiEgg\nShzRsC7sAMZIcVdPHGMGxV1NVYtZYxwiIhKIWhwiIhKIEkdEmNnnzOxJM9tiZg+Z2dywYyqFmX3J\nzHZkYv+RmZ0edkyjMbO/NLNnzGzAzCJfOWNmK8xsp5ntMrOPhh1PKczsHjM7YGZPhx1Lqcys2cz+\nw8y2Z/4+bg07plKY2WQze9zMtmbi/ruK/0x1VUWDmaXcvSvz9X8DWtx9TchhjcrMrgF+6e59ZvZF\nAHf/SMhhFWVmS4EB4E7gw+6+KeSQCjKzeuAPwFuAvcBG4AZ33xZqYKMwszcAPcC/uvtrwo6nFGbW\nBDS5+xNm1ghsBt4eg9+1AVPdvcfMJgC/AW5198cq9TPV4oiIwaSRMRWIRUZ394fcvS9z8zFgfpjx\nlMLdt7v7zrDjKFErsMvdd7v7SeD7wHUhxzQqd38UOBx2HEG4e4e7P5H5uhvYDswLN6rReVpP5uaE\nzEdFrx9KHBFiZp83szbgncCnwo5nDN4DPBh2EAkzD2jLur2XGFzM4s7MFgKXABvCjaQ0ZlZvZluA\nA8DD7l7RuJU4qsjMfmFmT+f5uA7A3T/h7s3AfcAHwo32lNHizpzzCaCPdOyhKyXmmLA8x2LRGo0r\nM5sG/AD4YE5PQGS5e7+7X0y6xd9qZhXtHtQOgFXk7m8u8dTvAj8FPl3BcEo2WtxmdhPwVuBqj8ig\nWYDfddTtBZqzbs8H2kOKJfEyYwQ/AO5z9x+GHU9Q7n7EzH4FrAAqVpigFkdEmNnirJtvA3aEFUsQ\nZrYC+AjwNnc/GnY8CbQRWGxmi8xsInA98EDIMSVSZpD5bmC7u/9j2PGUysxmDVYzmtkU4M1U+Pqh\nqqqIMLMfAOeRrvZ5Hljj7vvCjWp0ZrYLmAS8lDn0WNSrwczsz4HbgVnAEWCLu/9puFEVZmYrga8B\n9cA97v75kEMalZl9D7iK9Iqt+4FPu/vdoQY1CjP7E+DXwFOk/w8BPu7u68OLanRmdiFwL+m/jzrg\n3939sxX9mUocIiIShLqqREQkECUOEREJRIlDREQCUeIQEZFAlDhERCQQJQ6RCsistLrHzGZkbp+R\nuX2Wmf3MzI6Y2U/CjlNkLJQ4RCrA3duAbwJfyBz6ArDO3Z8HvgS8K6zYRMZLiUOkcr4KXG5mHwT+\nBPgKgLs/AnSHGZjIeGitKpEKcfdeM/ufwM+AazLLoovEnlocIpV1LdABxGIzI5FSKHGIVIiZXUx6\n577Lgf+e2WFOJPaUOEQqILPS6jdJ7+nwAukB8S+HG5VIeShxiFTGzcAL7v5w5vY3gCVm9kYz+zXw\nf4GrzWyvmUV2ZV6RfLQ6roiIBKIWh4iIBKLEISIigShxiIhIIEocIiISiBKHiIgEosQhIiKBKHGI\niEggShwiIhLI/wc2b34RsambOgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28702e2e978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotBestFit(wei)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "#### SGB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def stocGradAscent0(dataMatrix, classLabels):\n",
    "    m,n = np.shape(dataMatrix)\n",
    "    alpha = 0.01\n",
    "    weights = np.ones(n)\n",
    "    for i in range(m):\n",
    "        h = sigmoid(sum(dataMatrix[i]*weights))\n",
    "#         print(np.shape(h))\n",
    "        error = classLabels[i] - h\n",
    "        weights = weights + alpha*error*dataMatrix[i]\n",
    "    return weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1.01702007,  0.85914348, -0.36579921])"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataArr,labelMat = loadDataSet()\n",
    "wei = stocGradAscent0(np.array(dataArr),labelMat)\n",
    "wei"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCYglDmOMMQH5/9s9mn9jubgvAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28702ec4828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotBestFit(wei)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def stocGradAscent(dataMatrix, classLabels,numIter = 150):\n",
    "    m,n = np.shape(dataMatrix)\n",
    "    alpha = 0.01\n",
    "    weights = np.ones(n)\n",
    "    for x in range(numIter):\n",
    "        for i in range(m):\n",
    "            h = sigmoid(sum(dataMatrix[i]*weights))\n",
    "    #         print(np.shape(h))\n",
    "            error = classLabels[i] - h\n",
    "            weights = weights + alpha*error*dataMatrix[i]\n",
    "            if x == 0 and i == 0:\n",
    "                result = weights\n",
    "            else:\n",
    "                result = np.vstack([result,weights])\n",
    "    return weights,result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### 迭代过程"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3YAAAHVCAYAAAC9umnkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDC7E5AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28702f44710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "wei,result = stocGradAscent(np.array(dataArr),labelMat,numIter = 200)\n",
    "\n",
    "x0 = result.transpose()[0]\n",
    "x1 = result.transpose()[1]\n",
    "x2 = result.transpose()[2]\n",
    "y = np.arange(0, len(x0), 1)\n",
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(15,8))\n",
    "\n",
    "plt.subplot(311)\n",
    "plt.plot(y,x0)\n",
    "plt.subplot(312)\n",
    "plt.plot(y,x1)\n",
    "plt.subplot(313)\n",
    "plt.plot(y,x2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "抖动是来自样本之间的差异，数据集并非线性可分"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### 改进"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def stocGradAscent1(dataMatrix, classLabels,numIter = 150):\n",
    "    m,n = np.shape(dataMatrix)\n",
    "    weights = np.ones(n)\n",
    "    for i in progressbar.progressbar(range(numIter)):\n",
    "#         print(\"Iterations %d\" % i)\n",
    "        dataIndex = list(range(m))\n",
    "        for j in range(m):\n",
    "            alpha = 4/(1 + j + i) + 0.01\n",
    "            randIndex = int(np.random.uniform(0,len(dataIndex)))            \n",
    "            h = sigmoid(sum(dataMatrix[randIndex]*weights))\n",
    "            error = classLabels[randIndex] - h\n",
    "            weights = weights + alpha*error*dataMatrix[randIndex]\n",
    "            del(dataIndex[randIndex])\n",
    "            if i == 0 and j == 0:\n",
    "                result = weights\n",
    "            else:\n",
    "                result = np.vstack([result,weights])\n",
    "    return weights,result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100% (40 of 40) |########################| Elapsed Time: 0:00:00 Time:  0:00:00\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3IAAAHVCAYAAAC0UcmeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzs3Xd8XFeB9//PudMljbosyZJt2Y57\nmksSJ4EUUilLILCUhQDLsoEFloe6Czxhlx/LsoF9WMrC0utuQkJZQiCN9B47dhLHduJeZfWu6TP3\nnt8fM1bcq6SR7O/79dJLM3fuzD1z5mp0v/eUa6y1iIiIiIiIyOThFLsAIiIiIiIicnwU5ERERERE\nRCYZBTkREREREZFJRkFORERERERkklGQExERERERmWQU5ERERERERCYZBTkREREREZFJRkFORERE\nRERkklGQExERERERmWT8xS7Avmpra21LS0uxiyEiIiIiIlIUq1ev7rHW1h1tvQkV5FpaWli1alWx\niyEiIiIiIlIUxpidx7KeulaKiIiIiIhMMgpyIiIiIiIik4yCnIiIiIiInBastcUuwqiZUGPkRERE\nRGRistZijBm5nXE9Ao6D45jDPsfz7H6PZ3IenrWEA77j3uaJ6B5OUxbyEwkevD1rLZ4F3xHKfzhP\nbulhQ8cwl82ro7kqQsdgCgCDYW9xjWGk7JGAj4pIgFgqh+NAPO1SHvFTEswfim/vieMY8Gy+XBbI\n5w2LtYXlWDa0D7NnIEkq65JxPRY0lGOxlAT9zK4rAyx7BlIMJbP4HYPPMfh9hqUzqikL+UllXQaS\nWaJhP8mMS21ZiO09cfoTGSoiASoiAQCSGZcv/GEdc+ujzKwtHakr17N4Nl+mZNYlls6RdT1czxIN\n+1k0tYI9/UkGk1mGU1naB1O8em4d9dEQ8UyuUO+QyLjcvbadpsoIfp8h51ouPqOWksLnZIyhP5Eh\nlsoR8Bn8PocdPXGCfof5hfe8uy9J20AS11rSWY+X24fY3ZdgVl0pDRVhkllvJLTtzW7dw2kGkhkW\nT6uiqSrCwsZy3rK0+bg//4nCjEYqNcb8FHgD0GWtPbOw7IvA3wLdhdU+b629+0ivs2zZMqvJTkRE\nRORArmdxPUvQ75B1PQI+h0QmRyTgw7MwmMwSCfiwWCIB33Ed/KdzLnv6k2RcrxA0oLYsSCrrccaU\nsv3WzeQ80jmXoN8h5N8/HHieJZl16Y1lcK3FkD9g7YmlmVIeYu6UKN97dCs7euJMrYwQ9Dvs7I3T\nVFnCeTOriAR8DKVyuJ6HMQbHGBwDL7YOMpTKsm7PINOrSygN+rl0Xh3ntVTz/K4BNncNs/fdOo5h\nY8cwsVSOhVPLCQd8zKgpoaokSDjgEE+7bOuJEfA5xFI5fr1qN/MaokyvLmXN7gGm15SQyrp0DaXp\nHM6Hk4DPIeAz7OhJUBbyU18R5onN3XiFQ8igz6GpKkJjRTgfHByDz3HY2DnE7r4kAEumVxINB1i1\no49E1mXxtEpsoc5ca8nkPACmV5diDPgdQ08szaqd/UD+vt9xOLu5grKQnw0dw5SF/GzsHKa5KkIk\n4COVc0llPerLQyxoKKd9MMUTW3oAuH5JE2t2D7C1O86/vvlM7l3XwZNbekbeQzSUD1RlYT+ZnEdv\nPMOUaIj68jA1ZUEaysNkch67+hKkcx5r9wwe8/51OCVBHy01pQwms+wZSB7Xc415JZwUUzjgEHAc\n0oW/nX3tDaZjrbYsiDGGqpIAU6JhWvsThAM+QgEfvn2+BowxdA6l8DxLWdjPzt5Efp+86cqxL+Rx\nMsasttYuO+p6oxTkLgFiwC8PCHIxa+3/O9bXUZATEZFTXX88QyLrUhEJUBY6to4xsXSO/32ulXdf\nMIN0zmNrdwxbOEM/q66MkN8hns5x77oO0jmPzqEU/Yks5WE/DRVhDFAXDeNay6zaUnb0xnlsUzfh\ngA+/43D+zGrWtw0ynMrhepas6+E4hoBjCPgc/D6HF1sHmFFTSnVpgI7BNM1VEV49p5aq0iDTqkp4\ndFM3mZzH09t6WLm9j7ObK5lZWwrAIxu7yLiWsN+hIhIg4HOwWDwv/x7qy8Msa6kmlXEZSmUBRt5f\na3+SPf1JHtzQBeQPuOOZHFOiYTqGUpQGfWTdfOvQXlUlAaZWRljfNsT1i5tIux5DySyprEsy65LM\nuPvV787eBLnDHHE2V0WoKQ0Sz+TDXjL7ynOvWlhPwGcKrRTwUtvQEQ/IKyIBBpPZY/rMx0s44JDK\n7n8AHvQ5+BzD2c0VI8E559qRUNVUGQHgjedOJeR32NWbYH3bEIPJLBWRAKGAQ8615DyPTZ2xkedE\ngj6qS4Oksy4lQT9+Xz6s7m0R29kbx5h8EMy4Hn3xDLPryrhodg05z9I2kOS+9R14Fs5trsRxYEtX\nnNqyIC01pUSCPrZ2x3ixdZDGijClIT8N5WGMgdU7+0kc8Lk7Bpa1VBNL5QgHHIZSORZNLacs5Gdz\nZ4zSkA/XwvaeGOlsviWxtiwE5EPBmxdPpW0gRcjvMK26pHAi4ZUWNQq7lMUymMyysSNGc1WE3nia\n/kSWdGFfyrqWgWSWd543jVDAOaBVLx/m9y4LBxwunFVLOODQMZTiqS29NFaEybgeA4ksPsdQGvLR\nVFmCxZJzLa39CdbtGSLgcwj6HXKuR/tQilTGpbEyzNTC5+l5Fp+T/9sEqI+GOXtaRaGuXjmxsLdM\nAZ+zX8vq7r4EO3rjVJUEObOpYuQ1X2gdYDiVY1pVBKfwxoyBkN9HLJ0DLHVlYda0DlARCYzsD8ZA\nY0UE18vvSz7HYC0MJLLE0lm2dse5bF4dU6LhE9r3U1mX4VSOumjohJ4/lsY1yBU22AL8SUFOREQm\nktb+ROFg1PL8rn4CPodQoTUlkckRz7ic01xBbVmI0pCfoVSWP6/vpDzszx+4OPkDl7KQn0jAx/ae\nOK39STxrR7oZeTbfYlAa8mOtpTwSYG59lMFklvf+dCVBn8Mlc+tI51we35xvIQgHHD599TyCfoc9\nA0myOcu6PYP0JTKkcy6lhS5XW7piIyHjqoX1rNzed1AYONTZ+WjIz3A6d8S6qSwJMJB45bWiofzB\ndcDn4FlL1rXkXO+goHQi5tVHqSgJ0B/PYAFTKPdwKkd7oVvaoZSF/FSV5sPfhbNq6ItnGExmqYuG\ncD1LWciP4xiaKiP0xNKFLl05BhNZVu7oA/IBamZtKeGAQyTgy3dn2+dMvd8xnNVUQVNlhIDPoWs4\nTX8iQzyd48XWQfw+Q2nQT2t/gm09cV57ZgOrdvZjYCSIGGOoLQtyVlMFmZzH9JoSykJ+SoI+4mmX\njZ3DrNk9QNDv8LW3nk1dWYhUzsNXaCV4uX2IcMBHWdg/Uv97u7OVBv0snFpOXzzDQCJDKODjqS09\nrN7Zj7Vw5cJ6zplWgcGMPCcS8IGBzZ3DdA+nSWZd/D4H1/NoqSnF5xgSGZelM6rIFj7jldvz9XXl\ngikn1Z1xXznXw7X2oNbLE5Ut7IcB3/FP8+B5lm09MRIZl9l1ZZQe44kUkfE2UYLc+4AhYBXwKWtt\n/5FeQ0FORGR87O5L4PcZwoUzosmsi2ctW7viTKuOcHZz5UFjU6zNj0nY2hNjYWM5jjHs6I2TdT2s\nherSYH5MCAafkz/ArY2GCPkddvcleHRTD4l0jtpoiCsWTOGLd65nRk0pMwsHlamcy+3P7qZ9MMXU\nijBlYT/nt9SwpnWAurIQu/oSDBQO4IM+Q9Dv4HMczmmuYE59lF8/u5uOoRSprEuuEDzSWZe2I4SE\nfQV8huWzakaC1mhb2FiO35c/6L94di23rNhJ/z4hKhxwiIYDLJ5WSSLjsr0nTknQR0NFmJrSIHe8\n0Day7uvPauRNi5tYt2eQTZ3D1JeHaa6KUF8epro0SE1ZkPkN5SQzLqmsS288wxObuykJ+umOpemN\nZXjL0iYWTa1g9c5+dvTEOau5grn10cOWP5lx6Utk6ItlKI/4GUrm2N4b5771HbiupS4a4oYLZ+AU\nzqKH/A6eBc/akbPvhxNL59jYMTzSspHfj/IBOhLwndAYJsiHmJm1pfhP4KBfRKRYJkKQqwd6yDcs\n/wvQaK19/yGedyNwI8D06dOX7tx5TNe/ExGRYzCQyPDdh7fw8Svn0jmUDzTP7Rrg079Zc8TnVUQC\nDKWyI608ZSF/oQtMceztxjO/IR80sq5lV1/ioDJNiYY4s6mC3X0JklmXxdOr8Lx8CxlYZtWWsWRG\nFcZAOuvhmPyA/c2dMV5qH2JjxzAbOoaYXl3Ct96xGMeYQoubZSCRJet6BP0Oi6dXEQ44+PZ2N3IM\nqaxLutBFbXPXMIPJLKmsR188zTWLGphSvn/3n2TGZUPHEABnN1ceNax4Xr77VUnQd8wTRYiIyORT\n9CB3rI/tSy1yIiIH64vnu3i1DSRJZF0qIwH8Tr47Xjydw7OWSDA/4YPn5cdD7G0J+eYDmw/7ui01\nJbx1aTNTysMEfPnJCYI+hw0dQ6zbM8TK7b2YQpC5bN4U/rgm3yL0rXecS08sg+dZplVHqIgEAega\nTuF38i0frrXEUjnWtQ1SHw3TWBHmnGn5MVP3rGvn+V0DWGt547lN1JeH8DzIeh7hgA9rLY0VEdoG\nkvgcQ3VpEJ9j8qFpn7CzdyzXzt4Eq3f2Ywz85dLmk259yboe/kJXORERkfF2rEFuzDoHG2MarbXt\nhbtvBtaN1bZERMZbKusSKEwIsC9rLX96sZ2WmlLOas4P9u6JpYmnc1SXBgn4HNoHUziGkZYcgBd2\nDdDan2DR1AoWTS3njd99gt19SV5/ViP3v9R50uOTAOY3RPnQpbOxWM5urixMVX2wa89sOOTyefVl\nTK8p5Y3nTD2pclx3bhPXndt01PWmVZcc8XGfY/A5PuY1RJnXcPgugcfrRMbeiIiIjLdRCXLGmF8B\nlwG1xphW4J+By4wx55LvWrkD+OBobEtEJjdrLYmMO+qDzO96sZ2W2hI2tA9TGw2RzXmcM62SqpIA\nsXSObT1xXmoboqokSDyTY3ZdGUumV+LZfAvMjt44T23pJeN6PLOtl0c2dlNVEmDpjCpm1payvSfO\n1u74yPWCNnfF8DuGC2ZVM726lNqyIE2VEf7zoS0js9bNritlZm0pD7zcdVzvZd8pm+9amz8f9uHL\nZlMa8rNnIMmiqeVUlQRJZFwCPkN5OEBTVWQkHEbDAaJhPw++3EU8k+PNi5swMDIhw4n66GvmnPBz\nRUREZHSNWtfK0aCulSITz4HBa33bIPet6+CaMxuoLAmOXHT02R39xNM5GirCzKwtZXPXMLt6kwT9\nr0xlXB4O8NvVrbywe4C3Lm0mmXVHJjS4e207M2tLyeS8kfUN+Wmo66IhLOArzA7nOAafyQcTxxjW\n7Rk85gktjseUaIhQwKF9IDUyy15lSZBMzmVeQ5ThVI4NHcOHff606gjRUAALzK0vo7EiQsBnmF5d\nst/Yq5xn8RnDRbNruWddOy+2DnLX2nbOnVbJDctnsKylihk1paP+/kRERGTiGfcxcqNBQU7k5OVc\nj+8/upUblrfw7I4+tvfEiaVzhAIOf3i+jdlTSplbHyXge+U6P8OpHKmsy9buGI4xXLOoAdez3LJi\nJzt6E0B+xr13nD+Nf/rD+lEra2NFmEjARzrnsWcgSW1ZkItm1+J3DBbIeZZt3bHCta7MyHTvbmEs\n2N6p313P0h1L85p5U6gpC1Ia8tMTS1Ma8lMa9FNamAJ8QWM51lo2d8YKF+1N4/c5+enWnfzv6dUl\nvGb+FIBjar1KZPKTbQwksvzXI1uoKQ3x0decoe55IiIickIU5EQmsT0DSQYTWcpCfqbX7D9OKJ1z\nGUxmSWZcYukcw6kcPic/RbcxcPuzu/nl08c++6sxjFzryPXy47mOpCzk5/yZ1SNh69K5dRgDsVRu\n5MLEQ6n8zHrnTqsaKVf3cJpdfQlm1pbiepaSoI+aslcuwpnKupqJT0RERE57RZ/sROR0kcl5eNYe\nFEL64hk8a6mIBMi6Hp6F3liaFdv7mF5dkr/eU0mQL/5xPYPJLJGgjzW7B2iuivDMtr6R1/E7+e6E\ney867HrHd/LlnOYKbriwhaDfobokyPzGKNGwH2s55GQdOddjS3eMnGupLw/TPZzGs/mL7oYCDlUl\nwRMKXNOqS444eYVCnIiIiMixU5ATOQltA0kuuvkhAD7/uvlsaB9mxfY+SoI+NnfFgHyL1/E0fHcN\npWmsCPPR15xB11B+tsNXxoUZAj6H6tIAkaCfspCPaDhAzrO0DySpiARwHMOChnKaqiIndBFdv89h\nfkP5yP26aOgIa4uIiIhIMSjIyYS1qzfBNx7YxM1vOYuQ/+DWGmvtYccweZ4lnfN4qX2QTM5yXkvV\nyLWlemJpPvCLVUytDLOwsZzashAlIT/JTI5wwIdjDLv6EiQyOXzGEA76aKkpZU9/kt54htqyIDNr\nS8l5lg/+9+qRbX7l7g2EA/kQ5BiYVx9lTn0ZM2tLKQv5cYzBGGisiAD5roRbumPcvbadnYVxaM/+\n3yupLQvq+lUiIiIickQaIycT1l/96Bme2toLwPffvZShZJaBZIZExuWuF9vZ0h0j4MtfwDiWzlES\n9FEXDdFQHualtiGG07mR1yoN+ji7uRLHgSe39I56WT/wqplcv6SZBY3REwphPbE0AZ9DRSQw6mUT\nERERkclDY+TklPKRW5/bb2xY0O+weFol58+sIZ1z6R5OU1sWYnPXML2xDJfOq2Ph1HKCPod0zuP5\nXf0MJXNkXMuS6ZU8t2sAgBuWz+BvXjWTjOuxrTtOfXmIspCfWDpHU1WEKdEwmzuHybqWhoow0bCf\nHT1xEhm3cDFiQybncXZzxUm1otWWqfuiiIiIiBw7BTkZF65nsdbicwzWguMY0jmX7T1xZteVkcl5\ntA+mKI/4KQv5iafdke6Ge58fDjg887krKAv5R7pJjqa59dFDLp9zwPID74uIiIiIjDcFORl1PbE0\nn/nNGoZSOVr7E8TT+Wny93XOtEp29cbpT2SP+nqzakt51/IZLJtRRWVJcKyKLSIiIiIyaSjIyWFZ\na9nSFSPrWjZ2DrGpM8a27hjNVSUEfA5+x3DtmQ3Mb4jy6KbuwjT1sLlrmIc3dh/yNT946SxuW7mb\nnb1xZtWWcnFlhFTWo6okQHNVCfesaycS9LF4WhUA3bE0X37TmRo7JiIiIiKyD012IiNyrseTW3uZ\nVhXh6W29fOWul4ln3IPWC/kdrIWM6x32tcIBhwtm1vDU1h6e+dwVDKVyNFdFCIxBl0gRERERkVOF\nJjs5TeVcj58/tYMlM6pY2FhO+2CKgUSGZ3f0ESyEqKsXNeAYQ1nYT/dwGgOURwLcvbadm+5Yd8jX\nfduyZq5Z1MCyGdVUlORbx17YPcC96zpYsb2XxdOqWDKjcmSq/UjQR1VJkFTWJRoOUKPJPERERERE\nRo1a5Ca59W2DdAym+Pf7NpLMusTTOXpimZN+3c9cM4/L5tWxaGoF6Zx7yOu4iYiIiIjI6FKL3Gni\nr360gsHkKxOG/MU5U9nTn8Dvc9jeE6d7OM2iqeXMa4hy0exaQn6HB17upGMwxRULpgDQE8vgGMNv\nVu2mN57h7y6bzYcunY3PyU+nrxAnIiIiIjKxKMhNUrF0jnvWtu8X4u786MWc3Vx51Of+xTlTD7n8\ns6+dP2rlExERERGRsaMgN0nd8sxO/u2eDSP3z2upOux10ERERERE5NSiIDdJJfaZTXLrV1430g1S\nREREREROfZoL/hSgECciIiIicnpRkJukske4hpuIiIiIiJzaFOQmqb1BrjSoGSVFRERERE43CnKT\nVNbNX//vkc9cXuSSiIiIiIjIeFOQm6SyrkdtWZC6aKjYRRERERERkXGmWSsniZzr8eW7XubnT+3g\n/JnVZF2PgE85XERERETkdDQqQc4Y81PgDUCXtfbMwrJq4HagBdgBvM1a2z8a2zudrNrRx48e38Z9\n6ztHlq3c3gfAtOpIsYolIiIiIiJFNFotcj8HvgP8cp9lnwUetNbebIz5bOH+P47S9k55Wdfj/923\nkR88tm1k2ZsXN/G1t55Na3+SP65pY2FjeRFLKCIiIiIixTIqQc5a+5gxpuWAxdcBlxVu/wJ4BAW5\nY/a+n63kyS29lIX83HbjchZNLceY/PXiZtaW8rEr5hS5hCIiIiIiUixjOUau3lrbDmCtbTfGTDnU\nSsaYG4EbAaZPnz6GxZk8fvz4Np7c0ktTZYQHP3Up4YAuMSAiIiIiIq8o+mwZ1tofWmuXWWuX1dXV\nFbs4Red5lp88sR2Aez/+aoU4ERERERE5yFgGuU5jTCNA4XfXGG7rlHHLyl20D6b45tvPJRoOFLs4\nIiIiIiIyAY1lkLsTeG/h9nuBP4zhtk4JXcMpvnDHOoI+h2vPbCh2cUREREREZIIalSBnjPkV8DQw\nzxjTaoz5G+Bm4CpjzGbgqsJ9OYyc6/HRW58H4EvXLVKXShEREREROazRmrXynYd56IrReP1TnbWW\nT/9mDSu397GgsZx3nK9JX0RERERE5PDGctZKOQbWWq795uNs7BwmHHC44yMXFbtIIiIiIiIywRV9\n1srT3e6+JBs7h5nfEOWhT11GyK8ulSIiIiIicmQKckX2QusAAF9/2zlMrYwUuTQiIiIiIjIZKMgV\nUdb1+K+HtxDyO8ytjxa7OCIiIiIiMkkoyBXRL5/eyYaOYeY3lhPw6aMQEREREZFjo/RQRP/18BYA\nfvn+84tcEhERERERmUwU5IoklXXpjWd4+7JpVEQCxS6OiIiIiIhMIgpyRXLbyl0AXDyntsglERER\nERGRyUZBrkjuWdcBwPKZ1UUuiYiIiIiITDYKckWQcz1WbO/jwlk1TCkPF7s4IiIiIiIyySjIFcED\nL3cCcOHsmiKXREREREREJiMFuSL47sNbAbh0bl2RSyIiIiIiIpORgtw4s9aytTvG685q4JxplcUu\njoiIiIiITEIKcuPsJ09sJ5FxmVZVUuyiiIiIiIjIJKUgN8529yUAuPGSWUUuiYiIiIiITFYKcuNs\nOJ2jqTJCTVmo2EUREREREZFJSkFunD27o4+ykL/YxRARERERkUlMiWKcDcSzeJFil0JERERERCYz\ntciNM9darj2zodjFEBERERGRSUxBbpylcx7hgKpdREREREROnBLFOMq6Hq5nCft9xS6KiIiIiIhM\nYgpy4yid8wAIBxTkRERERETkxI35ZCfGmB3AMOACOWvtsrHe5kSVyroAhNS1UkRERERETsJ4zVp5\nubW2Z5y2NWGNtMipa6WIiIiIiJwENQ2NI7XIiYiIiIjIaBiPRGGBPxtjVhtjbjzwQWPMjcaYVcaY\nVd3d3eNQnOJJZ/MtciG1yImIiIiIyEkYjyB3sbV2CfBa4CPGmEv2fdBa+0Nr7TJr7bK6urpxKE7x\npHL5FjldfkBERERERE7GmCcKa21b4XcX8Hvg/LHe5kQ1lMwCapETEREREZGTM6aTnRhjSgHHWjtc\nuH018KWx3OZE8/jmblbt6CeezvHjJ7YDEA2P1xwzIiIiIiJyKhrrRFEP/N4Ys3dbt1pr7x3jbU4I\nrmf50ePbuPmeDSPLqkuDfOLKOSyaWl7EkomIiIiIyGQ3pkHOWrsNOGcstzFR/WbV7pEQ9+dPXMLc\n+miRSyQiIiIiIqcK9fEbAy+2DvDZ/13LnCll3PGRiykNqZpFRERERGT0aPrEUeZ6lk/c/gIAN7/l\nLIU4EREREREZdQpyo8hayw8e28rW7jgfu2IOS2dUF7tIIiIiIiJyClKQG0Xff3QbX7t3I1Mrwvyf\nK+YUuzgiIiIiInKKUpAbJW0DSb56b35yk1v/djk+xxS5RCIiIiIicqpSkBslX77rJQB+9tfn0VJb\nWuTSiIiIiIjIqUxBbhS09ie4e20HV8yfwuXzphS7OCIiIiIicopTkDtJmzqHedVXHwbgYxoXJyIi\nIiIi40BB7iT9/KkdANz0+gWcM62yuIUREREREZHTgoLcScjkPG5dsYtL59bxgVfPKnZxRERERETk\nNKEgdxLuWdcOwNWL6otcEhEREREROZ0oyJ2gzqEUf3qxnZKgj3ecN73YxRERERERkdOIv9gFmIyS\nGZdXf/VhMq7HG85u1DXjRERERERkXKlF7gTcdMc6Mq7HlQvq+bfrzyp2cURERERE5DSjFrnjlHM9\n7nhhDwA/es9SjFFrnIiIiIiIjC+1yB2nl9qHcD3L195ytkKciIiIiIgUhYLccfrd6lYAFk/XNeNE\nRERERKQ4FOSO029Wt1Ie9jOnPlrsooiIiIiIyGlKQe44rNrRRyLjcvn8KcUuioiIiIiInMYU5I7D\ngxu6APjQpbOLXBIRERERETmdKcgdh7aBJNOqIyxoLC92UURERERE5DSmIHcc2gaSTK2IFLsYIiIi\nIiJymhvzIGeMudYYs9EYs8UY89mx3t5YahtI0VSpICciIiIiIsU1pkHOGOMDvgu8FlgIvNMYs3As\ntzlWfv3sbvYMJGmsDBe7KCIiIiIicpob6xa584Et1tpt1toMcBtw3Rhvc0w8vqUHgOuXNBe5JCIi\nIiIicrob6yDXBOze535rYdkIY8yNxphVxphV3d3dY1ycE9cfz7B0RhWz68qKXRQRERERETnNjXWQ\nM4dYZve7Y+0PrbXLrLXL6urqxrg4J64vnqGqJFjsYoiIiIiIiIx5kGsFpu1zvxloG+Ntjon+RIaq\nkkCxiyEiIiIiIjLmQe5ZYI4xZqYxJgi8A7hzjLc56m5dsYv2wRTVZWqRExERERGR4vOP5Ytba3PG\nmI8C9wE+4KfW2vVjuc2x8FL7IADvvmBGkUsiIiIiIiIyxkEOwFp7N3D3WG9nLCUyLk2VEaZVlxS7\nKCIiIiIiImN/QfBTQSrrUhL0FbsYIiIiIiIigILcMUlkXCIKciIiIiIiMkEoyB2DZMYlElCQExER\nERGRiUFB7hgks2qRExERERGRiUNB7hioRU5ERERERCYSBbljoDFyIiIiIiIykSjIHUXW9dgzkFSL\nnIiIiIiITBgKckexpz8JQNCvqhIRERERkYlB6eQo0jkPgPNaqotcEhERERERkTwFuaNI51wAQmqR\nExERERGRCULp5ChS2XyLXFiOwtPWAAAgAElEQVRj5EREREREZIJQkDsKtciJiIiIiMhEo3RyFH94\noQ2AkF8tciIiIiIiMjH4i12Aie68lir64xlm1ZUWuygiIiIiIiKAgtxRvf286bz9vOnFLoaIiIiI\niMgIda0UERERERGZZBTkREREREREJhkFORERERERkUlGQU5ERERERGSSMdbaYpdhhDGmG9hZ7HIc\nQi3QU+xCnKZU98Wl+i8e1X3xqO6LR3VfPKr74lHdF89ErfsZ1tq6o600oYLcRGWMWWWtXVbscpyO\nVPfFpfovHtV98ajui0d1Xzyq++JR3RfPZK97da0UERERERGZZBTkREREREREJhkFuWPzw2IX4DSm\nui8u1X/xqO6LR3VfPKr74lHdF4/qvngmdd1rjJyIiIiIiMgkoxY5ERERERGRSUZBTkREREREZJJR\nkDsKY8y1xpiNxpgtxpjPFrs8pyJjzA5jzFpjzAvGmFWFZdXGmPuNMZsLv6sKy40x5tuFz+NFY8yS\n4pZ+cjHG/NQY02WMWbfPsuOua2PMewvrbzbGvLcY72WyOUzdf9EYs6ew779gjHndPo99rlD3G40x\n1+yzXN9Jx8kYM80Y87Ax5mVjzHpjzP8pLNe+P8aOUPfa98eYMSZsjFlpjFlTqPv/r7B8pjFmRWEf\nvt0YEywsDxXubyk83rLPax3yM5FDO0Ld/9wYs32f/f7cwnJ954wyY4zPGPO8MeZPhfun5n5vrdXP\nYX4AH7AVmAUEgTXAwmKX61T7AXYAtQcs+xrw2cLtzwJfLdx+HXAPYIDlwIpil38y/QCXAEuAdSda\n10A1sK3wu6pwu6rY722i/xym7r8IfPoQ6y4sfN+EgJmF7yGfvpNOuO4bgSWF21FgU6GOte8Xr+61\n74993RugrHA7AKwo7M+/Bt5RWP594O8Ktz8MfL9w+x3A7Uf6TIr9/ibyzxHq/ufAWw+xvr5zRv8z\n+CRwK/Cnwv1Tcr9Xi9yRnQ9ssdZus9ZmgNuA64pcptPFdcAvCrd/Abxpn+W/tHnPAJXGmMZiFHAy\nstY+BvQdsPh46/oa4H5rbZ+1th+4H7h27Es/uR2m7g/nOuA2a23aWrsd2EL++0jfSSfAWtturX2u\ncHsYeBloQvv+mDtC3R+O9v1RUth/Y4W7gcKPBV4D/Law/MD9fu/fw2+BK4wxhsN/JnIYR6j7w9F3\nzigyxjQDrwd+XLhvOEX3ewW5I2sCdu9zv5Uj/wOSE2OBPxtjVhtjbiwsq7fWtkP+QACYUliuz2T0\nHW9d6zMYXR8tdKX56d6ufajux0yh28xi8mfIte+PowPqHrTvj7lC97IXgC7yIWArMGCtzRVW2bce\nR+q48PggUIPq/oQcWPfW2r37/b8W9vtvGGNChWXa70fXN4F/ALzC/RpO0f1eQe7IzCGW6XoNo+9i\na+0S4LXAR4wxlxxhXX0m4+dwda3PYPR8D5gNnAu0A18vLFfdjwFjTBnwO+Dj1tqhI616iGWq/5Nw\niLrXvj8OrLWutfZcoJl8a8KCQ61W+K26H0UH1r0x5kzgc8B84Dzy3SX/sbC66n6UGGPeAHRZa1fv\nu/gQq54S+72C3JG1AtP2ud8MtBWpLKcsa21b4XcX8Hvy/2w693aZLPzuKqyuz2T0HW9d6zMYJdba\nzsI/ew/4Ea9021DdjzJjTIB8kLjFWvu/hcXa98fBoepe+/74stYOAI+QH39VaYzxFx7atx5H6rjw\neAX57uCq+5OwT91fW+hqbK21aeBnaL8fCxcDbzTG7CDfBfs15FvoTsn9XkHuyJ4F5hRmugmSHwR5\nZ5HLdEoxxpQaY6J7bwNXA+vI1/Pe2ZneC/yhcPtO4D2FGZ6WA4N7u0bJCTveur4PuNoYU1XoDnV1\nYZkcpwPGd76Z/L4P+bp/R2E2rZnAHGAl+k46IYXxDj8BXrbW/sc+D2nfH2OHq3vt+2PPGFNnjKks\n3I4AV5Ifo/gw8NbCagfu93v/Ht4KPGSttRz+M5HDOEzdb9jnxJEhP0Zr3/1e3zmjwFr7OWtts7W2\nhfz3xEPW2ndxiu73/qOvcvqy1uaMMR8l/0fjA35qrV1f5GKdauqB3+e/0/ADt1pr7zXGPAv82hjz\nN8Au4C8L699NfnanLUAC+OvxL/LkZYz5FXAZUGuMaQX+GbiZ46hra22fMeZfyB9YAXzJWnusk3ic\ntg5T95eZ/PTTlvzsrR8EsNauN8b8GngJyAEfsda6hdfRd9Lxuxi4AVhbGLMC8Hm074+Hw9X9O7Xv\nj7lG4BfGGB/5E/e/ttb+yRjzEnCbMebLwPPkgzaF3/9tjNlCvkXiHXDkz0QO63B1/5Axpo58t70X\ngA8V1td3ztj7R07B/d7kQ6eIiIiIiIhMFupaKSIiIiIiMskoyImIiIiIiEwyCnIiIiIiIiKTjIKc\niIiIiIjIJKMgJyIiIiIiMskoyImIiIiIiEwyCnIiIiIiIiKTjIKciIiIiIjIJOMvdgH2VVtba1ta\nWopdDBERERERkaJYvXp1j7W27mjrTagg19LSwqpVq4pdDBERERERkaIwxuw8lvXUtVJERERERGSS\nUZATERERERGZZBTkREREREREJpkJNUZORCamLV0x7lvfQSbnkfM8Lps3hfNaqotdLBEREZHTloKc\niBzRyu19vO0HT++37Nkd/fz6gxcWqUQiIiIiMqZBzhgzDfgl0AB4wA+ttd8ay22KTEbdw2le+63H\nGErluH5xE639SSpLAtRFQ6SyHn7H4HMMfsdQEvKzsDFKc1UJ6ZzLYDJLx2Ca4VSW15/dSHNVyaiW\n7U8vtgHw4/cs4zXzp/Ch/1nNrr7EqG5DRERERI7PWLfI5YBPWWufM8ZEgdXGmPuttS+N8XZFJpU9\nA0l6YhkAbnt2936PVZUEMMaQcz1yniWRcQ/7Ov92zwbqoiGaKiPc8oELKA2d/J94OutRXx7iyoX1\nAJRHArT2J3nb95/mL85p5N3LZwBgjDnpbYmIiIjIsRnTIGetbQfaC7eHjTEvA02AgpycMlJZl5Df\nOakgk3U9AH7x/vOZUV1CPJNjdl0ZGdejPBzYb92eWJo1uwfIeZZIwEfA51AW8rOtJ8buvgRPbe3l\nqa29tA0kaaqKEPb7cJwTL1vG9Qj5fSP3F0+v5L71Hazc0cfKHX184Q/rAbh8Xh1fuf4sGisiJ7wt\nERERETk24zZGzhjTAiwGVhyw/EbgRoDp06ePV3FERkVPLM2yLz/AjZfM4vOvW3DCr5PN5YNc2O/Q\nUls6sjwc8B20bm1ZiCsW1B+0/KzmCgDm1kd5amsva1oHue67T+JzDPd/4lIaKsInVLZ0ziXof2WC\n23ddMIN3XTCD361u5derdtNUFeHRjd08vLGb36xq5WNXzDmh7YiIiIjIsRuXIGeMKQN+B3zcWju0\n72PW2h8CPwRYtmyZHY/yiJwoay3r24YoCfqoKQux7MsPAPCzJ7cfFOQGEhk8C9WlwaO+bqbQIhfw\nn/wVQUKF8Pfp36wZWfbAy50jXSCPVSKT443feZItXbFDPv6Wpc28ZWnzyP0LvvIAu/oSpLIua3YP\n4Pc5LJpafsgw+vTWXjZ2DGGBSMBHLJ3jxdZBPnL5GcxriB5XOUVEREROR2Me5IwxAfIh7hZr7f+O\n9fZERpvrWe5/qYOHN3Szelf/IYNNWWEsWsdgiu8+vIWBZJY/rslPErJ0RhW1ZUE8C7NqS3nD2VNH\nWs/2PieVzQe5oO/kg5x/n26U33j7OXzi9jUkMrnjfp0XWwcPG+IOpakywm9Xt/L75/fgevlzMu+6\nYDr/+uaz9luvJ5bmXT9+Bu8Qp23mTClTkBMRERE5BmM9a6UBfgK8bK39j7HclsiJcD2L74DxY9Za\n/vRiOw+83El/Istjm7pHHju7uYLrzp3K0hlVZF1LTyzN9x7ZiutZPn7b89y3vpNkNj8ZSTTs54wp\nZRhgR0+CtoEk97/UyT3rOvivdy3hrrXt/OixbeQ8S02h1S4wCkEuU+imeencOq47p4lP3L6GoWSO\np7f2ct/6Dj7w6pkjM1t6nuX53QP0xzPkPI+sa1k+q4aqkgA7e+PHtd1rFjUwkMgyrbqES+fW8ZvV\nrdyyYheJjIu1lsFkls6hNAG/g2fh++9eytTKMHXREK5nedVXH+bZnf3s6k1gDPh9Bp8xPLihC59j\nSGZcHtzQRX00hM8xVJUGiYb9WAsbO4bpGk7xxTcuYn5D+UnXoYiIiMhEN9YtchcDNwBrjTEvFJZ9\n3lp79xhvV+SIrLV884HNfOvBzZw/s5qzmip4sXWALV0x4ml3pKvjtOoILTUl/O0ls7h+cTOR4MHd\nBL/3yFaGUjnueKGNpsoIX3/bOcyoKTlo0g9rLZ+4/QXueKGNN/znE0C+hW5bT5zeeH7GyoDv5Gd+\nTOfyQTLod3AcQzTs5zsPb+E7D28B4OdP7eDGS2bheZbVu/p5ftfAQa8R8Bmybr7J7Hd/dxGVJYGD\n1jnQBy+dzQcvnT1yP5HJsaFjiPtf6qS6NIjrWfYMJAFoKA9z9cL6/SZhmVtfxmOburnk3x8+4nam\nREMAdMfS2ANa9R7f1KMgJyIiIqeFsZ618glAc5LLhLO5K8a3HtwMwPO7+lm5vY+Q3+HqRQ3UlAap\ni4a44cIZB80YeSQ/e995vHpOLf7DtKoZY/jLZdNYu2eQxooIH79yDstaqrnxl6v480udwOi0yF02\nbwpvOLuRz752PgDffPu5vNw+hGfhP+7fBMBPn9hOwOdQWRLgc6+dz4Wza/A7Dm/67pNkXI+FjeUs\nn11DWdDP0hlVJ1SOj75mDu9/1cz9Zs1MZHLs7ksyJRo6aCbN2268kEc2do10y8x5lpzr4fc5XDy7\nluF0ljlToiMTr6RzLtaCU5gtdO5N9/Cvd79M0O/wngtn6HIIIiIickoz9sBT2kW0bNkyu2rVqmIX\nQ04DP358G1++62Vuu3E5y2fVkMq6OMbsNzvjsWobSJLOeczcZ7bJ47FnIMmX/rieaDjAzdefddgg\nOBqGUlkiAR9+xxwy6Dy7o4+V2/v4yOVnjFkZxkrLZ+8auX3J3Dq2dA5z2fwpfKUwRi/reod93xNN\n1vX4pz+s44blLSycqhZGEZEDpbIumztjPLihk7efN22/XjCDySzWWnpiGcojfqZET2zW5sPpjaWJ\npXNMry6ZFP9TTlQ8nSOWzhHyO/kTp46hInLoE9zWWta0DvJy+xClIT9vPGfqMW3DWntK1+GJMsas\nttYuO+p6CnIy2aSyLs/u6GPN7gE8m2/y7YmluXpRAxWRABs7hpleU0LY78PnGBY0Rvf7kkjnXObd\ndC8Az/7fK6krdNWTye2pLT1s7orxz3euH1nmGKgvDzOQyJLMunzyqrmT4vIImzqHufobjwHwp79/\nFWc2VRzlGaeeTM7j1hU72dod56mtPUytjPDhy85gyYzK/a5ruFcy42JM/vvh+49u47md/WztjjGQ\nzJ+8uO3G5adlPcrE0zWUoqYsdND4bDm6VNblF0/t4Lld/dz/UufIpFk3LJ/Bhy+fzWAyy20rd/M/\nz+wkV3hwbn0Zf/7EpQBs645xz7oOPnzZ7KOGhz0DSTZ2DPFi6yCrd/aTzLjE0jl29yWIZ/JDGH71\nt8u5cHbN2L3hIhlMZrlvfQf/8NsXD3psVm0p4YAPSz6E7dU2kGQo9crEag988hKqSoKUhvx0D6fJ\nuvkT3nvr3VrLx257gT+uaSPkd/j2OxdzzaKG/baVyrr0xjP0xzMMpbIEfA5LpleN/O3ctnIX97/U\niTEwnMqxoWOYJdMr+f4NSwn5fVhryXl2ZBI417NYRqf301g71iA3bteREzkZ/fEM335oM9sKB3V7\nx2/t6xdP7zzkc3/+1+dRXx7GWrBYfvDoNgA+/7r5CnGnkIvOqKV8nzOFt9+4nDteaGMgkaGyJMCv\nVu7mmw9sIuh3eGxTNzdffzb1FaGRL3vXswynclQdw+UixtreC8QDbOuJn1YBJJV1+eKd6/njmraR\ngyWArd1xHt/cw7+86Uxu2OdSGg9v6OJHj2/jqa29+73OjJoSXjWnFmvhzjVtPLOt97SqR5mY2gaS\nXHTzQ0TDftZ+8Rr2DCTxPMu06pLDPifremzpijG/IXpatlzs7I2zpz/JQDLLtx/czIaOYfyOYVlL\nNUtnVPG9R7by38/s5L+feeUYYEZNCTcsn8H3HtnKps4Yb//B03TH0mzrzk/i9fvn97C7L8HFZ9Ty\n0/edh7WW1Tv7+d1ze+iPZ0jlXB7Z+MpEZ9WlQeY3RGkIh1k0tYJEJsc96zroGEoe9/ux1nLryl08\ntaWXT149l0jARyrr0hfPX7JoKJmlsiSw3/+ifdtcgj6HadWRUdkX9l5S6ZGNXazdM8iTW3rx+wwD\niezIOhfNruHKBfU4BtoHU2zpipHftMGY/Ml0Y/J1Hg74OKOujK/fv4kr/+OxQ27T7xj8PjMyWzfk\nA9aX/vgS33loC+2DKdI5l6zrkc55B42Fn1oR5vL5U4iGA3z/0a2Uh/00VZUQ9BkGk1ke3tjNvJvu\npSSYr1fP5su37+t87a1nc9m8Ov7xty+SdS3/84ELTroui0VBTiaFJ7f28LMndwAws7aU91/cwoWz\na2ipKcW1lie39BBPu3QMpjijvoxdvQmqS4P8/a+e530/e/ag15szpYwPvGrWOL8LGWsLG8v51zef\nyfkt1cypj3LBrFfOlKazHv/7/B5uvmcDAJf8+8Ps/T944D+KefVRLptXx+dO4iLvJ2PvzKMAifTx\nXzpiMkhkcnzrwc28+4IZ+x3EfvTW53jg5S5Kgz7+6oLp3PT6BYT8PtbuGeRN332SL9yxjmcLY1of\n3dRN13AagMaKMG9Z0ozFcl5LNZfNmwLkW+DvXNPGl+96mR89vo3v/NUSzmupLsp7Fuku7K/DqRwX\n/duDtA2mMAY+fsVctvXEyOQ8sq7HUCrHcCpHJueytRA+br9xOXXRELXR0HGN355MkhmXl9oHSWU9\nXm4f4pYVu9jes/8MylctrOeHNywdCTLxdI6X24e4fkkz5eEANWVBzm+pxnEMc+ujvOenK1mxvY85\nU8qY3xBlQ8fwyKV1ntraQ388w+Vff2QkvMyqK6Uk6OP8lmretXw6ZzdX0lKzfxfKruEU96zrIJbK\nkcjkux+2D6RwrcXz8icG87fBtZZkxmXl9j4e3dQ18nkC3LW2/YTq6fvvXsK1ZzYe8/rdw2nW7B6g\npixI0O/Q2p9kfdsQt67YRU8sv09GAj6WzKhkdl0ZjjG8qnBy9PyZx/d9mXM9omE/Gdfjqa29PLKx\nm0VTy7nu3KnE0i451yuMgbc4Bj5+1Vze+J9PsK0nTirrcvWiekJ+HyG/Qyjgo7EiTMDn0FgR5l0/\nXkHbYIo/rmkbaf27fkkzX3zjIiB//d47nt/Dtp44QZ9DOJB/nYzr4RiDzzH8x/2b9mtpvGZR/XG9\nv4lGXStlUvj1s7v5h9+9yJOffQ1NlZGjP6HgpjvWcvfaDqZXl/ChS2cBhqDfcPEZtYfsniWnLtez\ntA8m2dwZ469//iwlQR/vvagFv2NwjMFay7cf2jKy/tSKMN9791KmlId4emsv1sK50/P/5A6nbSBJ\nyO9QU3ZyLb0rtvXy9h8+A8AX3rCQv3nVzJN6vYnokY1dvO9nz3JWUwXvXj6dPQMpntrSw6qd/ZQE\nfay66UpKgvufa/zG/Zv45dM7iBQuMm+M4fL5dXzm6vlUHGFm1T+92Mbzuwb4yRPbuWH5DL503aLD\nns3ee2Z+WnUJ9eWjO65Gjo21dqTbvDGcVMtD1vUYTuWIBHyHnHV4vD29tZd3/ij/t72wsZxI0Mfq\nnf0jj8+ZUkbQ71Aa9FMeCRAKOPTHMzy1tZfmqgit/fkWoE9eNZf2wRT15SFm15VhgWUzqqiLhgj4\nHHpiaRJpl+k1+ZMk1lqGkjlynseGjmGe29nPjt4Ea/cMcPXCBj59zbxxq4OhVJZv3r+ZdM7FV/j+\ndYzBs5Zfr9pNYp+W+Bk1JVy9sJ5L5tZRXRqkrixEbdnBk2UdjrWWbz+4hbKwf+R7tHs4TWnIx08e\n387X799ETWmQ3niGc5oruPktZ7Og8ejjkpMZlwX/dO9xv/fasiCvO6uR6tIg//PMLmbVlXL94ibC\nAR8VJQFc11JREqBrKE3Oe+WE3r7dET9++wtMrYjQUBHGMRANB4ilcmRcD8/mA9Lebo+etVibn+Dt\nUBY2lvOXy5q5ckH9EVuFT1Qy4/LIxi6uPbPhiH/HGzqG2NGT4FVzakeuy3sof17fQSjg49K5dezs\njXPnC21cd27TyH5+LH742FZ+u7qVKxfUc9m8KccdVMeLxsjJKeGxTd08s62XnGf54WPbeP4LV02I\nrm8yuXUPpwn6nIMO/jd3DvPElh62dsf4n2d2HfS8M5vK+Ydr5vONBzbx7gtm8JalzUD+H+Y96zr4\n8C3PAfnJVv7pDQs5Y8rhQ9++rLWkst7Igebjm7u54ScrAfjUVXP5+0kwru9ocoUzonsPwO5c08bH\nfvX8futEAj7m1JfxnXcuOa5/zMfqzH++j1g6xwcvncXiaVW0DSSpiAS4YsEUXtg9wH89vJWVO/qA\n/DiQhz592aiX4XSWyOTY1ZegqTJCdJ8Wpf54hj+/1MGa1kFWbOulYzC1X7daY/Kz09ZHQ9z/yUsp\nPcKB3oGu+84TrGkdBPInZ4wxOE7+9QyFWW9NvrvX51+3YKQl92hybr6FP57O4VlGume3D6a4dcUu\nfI7hd3930UGTFT20oZP3/3wVt9+4nAtm1ZB1PT7zmzXsGUjyj9fOZ9khWouttdx0xzp29yf3u67p\noTgm/3e0t/5uev0C/ry+k209MXpimf3W9TtmZBzZP71hIcmsy+bOYT5z7XyaKiM8vbWXRzZ1EXAc\n/D6D38lPCPbGc5poqDi+kxyJTL7VLJ52+fr9m1ize4Dq0uBIaM+5HvGMy6zaUj5y+RlMqy6hNORj\nYWP5mHUnvevFdj5ya/47e35DlLs+9upjHrdoreWSf3+Y3X1JKksCfOSyM2iqihAJ+vAVWn72tgD5\nCvvblPLwfieiD3Ud22PxpT++xIaOIda3DTGYzO73WEN5mDOb8nXmGDDk93eDYX5DlIVTy8kWWsTO\nnVbJ1IrIMYdiGV8aIyenhK/eu4H1bUMECwNTJ8IZVZn8Djc2ck59lDn1Uf7wwp6RIPepq+ayeHoV\nt6zYyT3rOnjPT/MBy2cMrzurkX+/byP//cyOkXGbM2pKeGxTN49s7NovyCUzLnevbWd7T5yBZIZo\nOIDrWeLpHLes2EVlSYDnbroKxzH7d63MukxUO3riPLShi2Q2P56hLhqiLOTPHyQb6BlO8+CGLpIZ\nl1U7+1k6o4rf/d1F/G51K78sjGf52V+fxxl1ZUTDfspC/jGdtfUHNyzlXT9eMTJO9kB+x/BXF0zn\n1hW76EtkDrmOHN2qHX388umdnDutkoVTy+kcSvH1P29iV1/ioHX3HbtiDEytiHDNogZm1JRiyR/k\nYy1bu+PctbadLV0xzplWSddwitrSEC/uGWRmTekhW2Q9Lz/+59VzammpKSWZdf9/9u47PI7q3v/4\n+2xT792WZMm9N2zTjTHNtBAIJPBLckNCAoGQe3NTIdz0kEAgjXSSkEJCKAkBQu/d4IJtbOMmd0u2\nepe2n98fu1okW3KT1rKkz+t59Gh3dvbM2TOzs/Od0whHOksTjgYQlsjj59+r5mfPbzlkINfpD/HF\nB1ezdFt9j35EBwhFpnu5flGkCb/DGDbsbeWHT20AIDM5ckPS7XTwsyvnHHSbxhhujY6+W1HTyvId\njdz39i5uOn8y2+ra2VrTRnaKh5xUD1VNnXT6w9zzxnYAvv9EZHsnjMniYyeNITvFg8MYTizPpiw3\nhZ88t5nfvLyV7z7+Xmx7DmOYWJgWa4buMMQGFQH44VMbuXbhWCYVpHHZ3OKD5h3gkVWVfOGB1T2W\nzSrO4OEbTu0RyITD9pgGFRfOLGLaqEUAlGQnH1FQZYzh8RtPZ+O+FqaOSu9xY+JwHe1gN9+8eCoQ\nCX5vengtZ0/JZ3tdB63eAF85b9KI7Ec5kimQk+OaN3oR6w+FSYq2dRaJt0tmj2ZqUTrFWcmxmwfb\n69t5ev0+zpiYR7svyPIdjZzw/efo8IfITvFw4YwizptWyMnjchj39Sdp69a37SfPbuJPb+ygNbrM\nEa1hcDkNbkfkmG7qCHD70xupb/fzwobq2Hv700cuHLb88KkNXDJ79IAN9HH70xt5au1eUhNdrKts\nOaz3zIhue+XOxh7TRCyenM/CCXnHbPS+U8fnct+nT6Sq2UtpdjK5qR6+85/36PAHOXV8LpfOGc2Y\nnBRSPM5ea2RHilDY4g+GD7hx1n2Y8JU7G0lLdDEuL5VvPrqO7XXtbK5uwxcM0Rrtu/LYmqoe7//w\nvGJml2RR2+ojZG2Pzqn56YlcMntUnxfEK3c28sTavVzyqzdwO80BA15NKUrnK+dNpDA9CY/Lwfj8\nVOrafQTDlnOnFvDxk8sO+pkv+PlrrN7dxAPLd5HodtLpDzF9dAZluSks3VpPuy9IizfA4+/uZdn2\nBgrSE/jArFF88+KpuJ2O2HfaEqkRm3/r8zy/oZrnu32Xu+d1dNbhdxHobnx+GuPz07hqQSkQOaZ7\n838XTmFPYycupyEz2X1AM+Uu1y8ax7baNs6eUsAFM4pY+KOXeHhVZez1X1w1h4tnjSIcjoz+N/H/\nnsJaYjdDDhXIPfzOHr744BoAfnDpDMbkJFOandxrM77BqBkqO8ppgwAykt09+mEfay6ngzuvmDVo\n25fjg5pWynHttNtfZFRmEtctHEtBeqJGnZNB1XUhe8/r22N3sG+9dDofPXFMj/W6gpWrTynjf8+Z\nyKzvPAvAZXNG8/ULp5C7Xx+6LdWtnPPT90f4mlOayfyybB5ZVcmC8myWTC9k5c5Gqlu8XDxzFOfP\nOLxO7mv3NHPxL18HYO7a1TEAACAASURBVPsPL+j3ndqKmjbO/skrQKT5aDAU5gOzRvHBOaNxOUx0\ntLEwRGtSrIWsZDf56YnUtHr5x9u7WbmrkZoWLw999uSjuot9LPzkuc3c9cKWASmzoaCmxcvf3t7F\nql2N+AJh3q1swhsIk5vqYXJheqyf2quba/nqkklYC3c8swmIHAevbq6lNDuZMTnJjMtLxeUwzC/P\nJjXBFWsWOaUw/aD9GA8lFLb8Y9kudjd0sLfZSyhsGZuXwvqqFlwOw9Kt9Ti7jbb3wdmjaPUGeWFj\nDb//r3mcM/XgAxrsqu9g4R0vHVZeUhNcrPzG2QftZ93uC7Ii2vfNAD9+bjPZyW5+9dG5fQZVx4M2\nX5D6Nh+BkO11/rWXNtUQClle3FTDfW/v4tHPncqskkwgcuPo0TWVbK/riAXpT67bR0VNG/dcPY/F\nk4f2oBIix5L6yMmwMP/W5zl7SgE/vGzGYGdFJMZaS2VTJ+lJ7l5Hj+te69Tl5vMnc90Z4/pMc+O+\nFh5asYc5pZlcNDMykepZP365xwhnACeWZ/PAdScD0NwR4DuPr+eZdfsI20hNnzGG1AQXf/7UfC75\n5RvRwAo2fm8Jie7+NU2+4e8reXLtviMeMW2o+c3LW7n96Y2kJUYuuAOhMIGQJSPJzfyyLJI9Lj57\nxjgmFaYNck4Hxjk/eYUtNW0kuh3MLM4kye0kM9lNQ7ufDn8o1gxxze6mXt+fmezmPzeeFpfBEg7X\nP1fu4YUN1Ty1bl9sWUl2EkluJ/dec+JhDVyzu6GD5s4Abb4gH/vD27H+Y58+rZwrF5SSluiipsVH\nSXZSrGnkSLWnsYPTbn+J7BQPj914Kqt2NXHbUxupbHp/OP6ueyDXnFrO/100dZByKjI0qY+cDAu+\nQEjNKeW4Y4yhOKvvi9a/XXMiyQlOtte2x0YL++Cc0QdNc3JhOt/Y72LnI/NLeHR1FWU5Kdx5xSxu\n+fda3tpWz7baNjZXt/G5+94hFLZMyE9l0aQ8rIXtde28sLGGJT97DYD8tARqWn10+kP9DuRavUEm\nF6YN6yAOYEpRGqkJLk4sz6EkO9JUzxcIs7aymR11HWyuaeW1LbWcN62QzdWtAHzxnEmMzkwiwe3A\nHwxz29Mb2VbbjsthKMpI5FcfnXtcTUK7o66d17bUUt3iY0tNGyeWZ/P3T5940D6KP3xqAy9sqOF/\nzprAvLIsXttSxwdmjer3cTUQLj+hmMtPKMYXDLF2TzNluSkH1HwfSkl2MiXRxxU/uICN+1ro8IeY\nW5oVW0cjmUYUZyWzZFohT6/fx2m3R2oy3U7DTedP5jOnj9Vk5yLHiGrk5LjV4Q8y9ZvP8NkzxnHT\n+ZMHOzsig+6nz23m5y9s6bHs+kXj+Pzi8bHmWtZaHlldyY66Dspyk/EFIh3il968mKKMo+uX0+WS\nX71BeqKLe68ZupOnDoQP/3Ypy3Y0kJbooiQrmff29t5XcEpROtZaNu5r5fWvnXnQ4P9ohcKWqqbO\n2LDzgVC4z8DqsTVVlOekMKkwjQ/+6o1Yvkuzk/n5lbOZ0y1gETkUfzDMk2v34g+GSXA7OHlczgFN\nMUXk6KhGToa8Z9dHOomP7UdnZJHh5OJZRVTUtuFxOrjihGKmF2cc0LTTGMOlc94fgOCR6MAFnf7+\nj37Z5g1QfATzOA5Xf71mAY0dfvJSE3A5HbxRUcfeZi8Qubj1BUPkpiZw8axRPP5uFTfet6rH3FiH\nq77Nx+Pv7iXB5eAj80uobOqkOCuZt7bV89iaKtwOwxNr9/YYVj7F4+TNm85ie307e6PN3HJSE7jj\nmY0s39HYI/0PzS3mu5dMI9njHBF9AWVgeVyOQ7Y0EJH4UiAnx62u+VEWTzm8eX1Ehrvx+Wn86v/N\nPaL3dNXOeAPhQ6x5aK3eYKzf2EiW6Hb2qN3sa+RAIDbnWdsRjj66encTV/z2zdjojHc+u5m6Nh8v\nfOkM/vDadl7aVENqgotgKMzkwjQWlGfz/HvVVDV7mfXdZw+Z/oUzivjyeROPaE42ERE5vugMLset\nrjvYyZo7TuSoJbojfZ46j3A+urq2SL86XzDEjfet4uRxObT5gqTqwv+IpESbvC7f3tCjrxVE5oGq\nqG0jGA3Watt8PLKqku117bwbncS6a1TIujYfEJk8u8UbYN6YrNigN12uPqWMxT+OjCq6oCybr184\nhaYOP26nA7fTweisJFbubOTE8mz19RIRGQb0iyzHrU5/5A524kGGeBaRg0uK1sj59gvkrLXUtfnx\nBUP4g2Fy0xJIT3TT3Bngc39/h9cr6nqsv6WmjVDYHrdTBhyvuubQq2+PNH+01nL705uoqGnl+Q01\nvb5n4cQ8Lp41io/MK+G0CZHavhU7Grj8t0tpaPezbHsDZ085cCj3sXmpvPbVM9nX4mXemKxem0uO\nVtNYEZFhQ4GcHHdavQEeWrGHu16swO00gzJJqMhw0dW0cv8auZ89v6XHwClOh2FuaWasH1Wyxxmr\nFb/ujLG8urmOFI+ThRP7bkYoB0ryOCnNTuae17dz39u78AZCBMMWT3R0yI/MK2HxlHwcxmCAGcUZ\nvdaWdTWBvPbelQBMLEjtdXslfUy2LCIiw0/cAzljzBLg54AT+IO19rZ4b1OGrn8s28W3Hl2PPxTG\n43TwxXMnDnaWRIa0vvrIra9qYVRGIl84eyLeYIgn1+6lMxDm7Cn5nDutkA/PK+GVzbVkJ3uYUZzB\nzecPRu6Hh+9eMo2l2+oJBC1OR2QI+6tPKcPCYU9JMDoridGZSeSnJ3DhjCKuOa08vpkWEZHjXlyn\nHzDGOIHNwDnAHmA5cJW19r3e1tf0AyPDP1fuobolMsLb7JJM8tISeGx1Fct2NLBsewOTC9P4ynmT\nOH1CHh7NISfSL7vqO1h4x0uMykhkzpgszp9eiDcQ5ssPrWHx5HzuuXr+YGdRREREujleph9YAFRY\na7dFM3U/cAnQayAnw1cgFOb+5btZtr2B/6yp6nO9GaMzePC6k0nSACciA2J0VhIfPbGUR1ZV8sS7\ne3ni3b0ApCW6uGpB6SDnTkRERI5WvAO50cDubs/3AD1mkjXGXAtcC1BaqouK4er3r23jR09vIjXB\nxSnjcvjGRVN5cWMNP3luM6dPyGXhhDw+PL9EI+KJDDCnw3DrpTP45sVT2VbbjssR6XdalJEYm0Rc\nREREhp54/4r3NkpFj7ac1tq7gbsh0rQyzvmRQfLWtgaS3E7Wfvvc2EhqU4rS+dyZ4wc5ZyIjQ4LL\nyZSi9MHOhoiIiAyQeHdA2gOUdHteDPTdrk6GrVW7Grl07uheh8MWEREREZEjE+9AbjkwwRhTbozx\nAFcCj8V5m3KcafEGaPUGGaMhsUVEREREBkRcm1Zaa4PGmBuBZ4hMP3CPtXZ9PLcpx5c3t9bxyuZa\nAEZpIloRERERkQER957u1tongSfjvR05/uxt7uRjf3ibsAWXwzClKG2wsyQiIiIiMixoyDKJm33N\nXsIWfn7lbM6bVhibmFhERERERPpHsy1L3LT5gkCkSaWCOBERERGRgaNATuLmmfX7gMjEwyIiIiIi\nMnAUyEncbK5uA6AsJ2WQcyIiIiIiMrwokJMerB24Odk7/SHOnJSnZpUiIiIiIgNMgZzErN7dxPhb\nnuLPb2wfkPQ6/EGSPWpWKSIiIiIy0BTISczm6lZCYRub9+1oWWt5s6KOpo4ASR7VxomIiIiIDDRV\nl0hMS2cAAKfDHNX7vYEQTR0BVu9u5LN/eweA/LSEAcufiIiIiIhEKJATfMEQ9y7dyfef2HBE76uo\naeMPr20jPz2RLdWtPLVuX4/X//GZkzhhTNZAZlVERERERFAgN+K1+4Kc+9NXqWzqjC17fkMNbb4g\nqQk9D499zV427muhOCuZiprWWK1bF2PgS+dMJDslgcKMBE4el3NMPoOIiIiIyEijQG4Eqm7x8vn7\nVtHcGWBTdSsAM4szeOSGU/nWY+u5962d1Lb6YoHcjrp27n5tG/e9veuAtBZNyiMlwcWSaYVcPGvU\nMf0cIiIiIiIjlQK5EcBaiy8YJhi2PLNuH795ZSvbats4Z2oBEwvTKEhL4KbzJ+NwGE4el8O9b+3E\nHwzH3v/j5zbznzVVTC5M46SxOcwqySAUhvLcFGYVZ+ByaswcEREREZFjSYHcMFfV1Ml3/rOeZ9ZX\n91j+lfMm8bkzxx+wvicalPmCodiyrTVtnD4hl3uvOTG+mRURERERkcOiQG4YW7q1nqt+/1bseWl2\nMosn53PNaeWUZCf3+p4EdySQ66qRe3dPE+/tbeHqU8rinl8RERERETk8CuSGsW11bQDc/qEZXDa3\nGPdhNIFMcEXmffNFA7kn3t0LwPnTC+OUSxEREREROVIK5IahLdWtbK1t48HluwG4dM7hBXEAHldk\nvbte2MJPntvMyp2NAMwvy45PZkVERERE5IgpkBtmdtV3cM5PX409n1uaGQvODkdZTjKTCtLYVN1K\nU0cgttxxlJOEi4iIiIjIwItbIGeMuQO4GPADW4FPWmub4rU9idjbHJkP7mtLJrNkeiGF6YlH9P7M\nZA/P/O9CADr8QQIhS0aSe8DzKSIiIiIiRy+eNXLPATdba4PGmNuBm4GvxXF7I1YobLny7qUUpCfS\n6g0CcPqEXMpzU/qVbrJHFbYiIiIiIsejuF2pW2uf7fb0LeDyeG1rpKtv97F8R2PseUF6AmNyeh+V\nUkREREREhr5jVeXyKeCB3l4wxlwLXAtQWlp6jLIzvNS3+QH4xVVzWDQpjyS3U5N0i4iIiIgMY/0K\n5IwxzwO9jUt/i7X20eg6twBB4O+9pWGtvRu4G2DevHm2P/kZqdp8keaUGUlu0hLVn01EREREZLjr\nVyBnrT37YK8bYz4BXAScZa1VkBYHm/a18tLGGgASjmB0ShERERERGbriOWrlEiKDm5xhre2I13ZG\nuo/cvTQ2TcCRTDMgIiIiIiJDVzyv/H8JpAHPGWNWG2N+G8dtjVjd53pTICciIiIiMjLEc9TK8fFK\nW3qnppUiIiIiIiODrvyHEY/TOdhZEBERERGRY0CB3DCippUiIiIiIiODrvyHOKfDAHDDonEUpCcM\ncm5ERERERORYOFYTgkscPLq6klDY8rGTSvnqksmDnR0RERERETlGVCM3hD38TiUAl84pHuSciIiI\niIjIsaRAbghr7PCzaFIeJ4zJGuysiIiIiIjIMaRAbogKhMK8u6eZnBT1ixMRERERGWkUyA1RDyzf\nDcDkwrRBzomIiIiIiBxrCuSGqHZfEIAPzysZ5JyIiIiIiMixpkBuiArbyH/NHSciIiIiMvIoChii\nwjYSyRkzyBkREREREZFjToHcEOdQJCciIiIiMuIokBuiwmHVyImIiIiIjFQK5IaoaBc51ciJiIiI\niIxACuSGqK4+cg7FcSIiIiIiI44CuSGqa9RKoxo5EREREZERR4HcUGWt+seJiIiIiIxQcQ/kjDFf\nNsZYY0xuvLc1koQtKI4TERERERmZ4hrIGWNKgHOAXfHczkhksRroRERERERkhIp3jdxPga/y/iCL\nMkDCViNWioiIiIiMVHEL5IwxHwAqrbVrDrHetcaYFcaYFbW1tfHKzrATtmpbKSIiIiIyUrn682Zj\nzPNAYS8v3QJ8HTj3UGlYa+8G7gaYN2+eau4Ol9XUAyIiIiIiI1W/Ajlr7dm9LTfGzADKgTXR4fGL\ngXeMMQustfv6s02JCFuLUZWciIiIiMiI1K9Ari/W2rVAftdzY8wOYJ61ti4e2xuJrGrkRERERERG\nLM0jdwgPLt/NZ/66gnZfcEDSW727CV8w1O90NNiJiIiIiMjIdUwCOWtt2VCtjXtzax3PvVfN9rr2\nfqe1u6GDD/7qDb792Pp+p6XBTkRERERERi7VyB3ChTNHDVhaLd4AAKt3Nw9IeqqRExEREREZmRTI\nHUJXP7Sw7f+Aml2BVzjc/7TC1qI4TkRERERkZFIgdwix4GsAJkZ4P63+J2bVR05EREREZMRSIHco\nA1ojN3Bpha3VqJUiIiIiIiOUArlD6Kr1GoDYK2YgavciaSiSExEREREZiRTIHUJXrZe1llc317K+\n6ugHKglFo8GBqJED1ciJiIiIiIxUcZkQfDjp3kfuv+5ZBsDbXz+LgvTEI04rFK2KCw3EYCdhNNiJ\niIiIiMgIpRq5Q+gKlqpbvLFlJ/7gBfY0dhxxWuFw5P+exs5+58tiNdiJiIiIiMgIpUDuEEy0H9rn\n/7Gqx/K6Nn+P580dAdZVHrzZZWgAO9qFNWqliIiIiMiIpUDuEPrqh+YNhHo8/+KDq7noF6/T1OHv\n/Q0MTJPKLgPTz05ERERERIYiBXKH4OgjkvMFw7HHG/a28MLGGgC+9/iGPtMa0ODLgkN7T0RERERk\nRFIocAiHUyP37cfWxx7/6509NHX4WbO7CW8gxC9e2MK7e5qAgauRW7a9gYdXVQ5IWiIiIiIiMvRo\n1MpD6j2Sa+4IALBqVyNvb2/o8drvXt3Gb17eGnv+65e38l8nj2FbXXu/cxMIhbny7qUAXD63pN/p\niYiIiIjI0KMauUPoq0bu9qc3AvDo6ioAnvqf0/G4IsXZPYgD6AyE+N2r23juvep+5+fMO18mbOH6\nReP4n7Mn9Ds9EREREREZelQjdwh9jQzZEB3U5M2tdRRnJTGlKJ0N313Cj57ZyO9e2QbA9z84nRPG\nZDGlKJ3Kpk5Ove3FfuVld0MHexo7mVuayVfOndSvtEREREREZOhSIHcIvQVyaYkuWr1BPnHPMjZX\nt3HdwrEAOB2Gm8+fQl5qAk6H4WMnjYm9Z3RmEtctHMvvXt1GWmLPYl9X2cyexk5OGZ9DOGzJTPYc\nsE1rLYvufBmAWy6c2ucgLCIiIiIiMvzFNZAzxnweuBEIAk9Ya78az+3FQ/c4bnZJJoluB4XpiTyy\nuopXNtcCMKc0s8d7Pn362F7TuvmCKdS3+3mzoi62zBcMccVvl9LZbfCU33x0LufPKOrx3pc31xIK\nW86eks/c/bYnIiIiIiIjS9z6yBljzgQuAWZaa6cBd8ZrW/HUPZBbUJ7N/deezAll2bFlozOTOHdq\n4WGn5zSmx8Tgy7Y39AjiAK7/+zu0eAOx5+2+IL9/NdJc866r5mA0EbiIiIiIyIgWz8FOrgdus9b6\nAKy1NXHcVtx0b1rpijZnzEmJNH28duFY3rhp8RE1c3Q4DKH3p6Bjx34jWXY109xR144vGOKaPy9n\n2ree4c2t9ZwztYBkj1rDioiIiIiMdPGMCiYCpxtjbgW8wJettcvjuL246B7ILZqUD8B50wr51/Wn\nMG1U+hGn53T0nBh8b7MXgBe/dAat3iC+YJjfvbqNipo2PC4HL2ysIdnj5OwpBdy4eHw/P42IiIiI\niAwH/QrkjDHPA721K7wlmnYWcBIwH3jQGDPWWttjVmxjzLXAtQClpaX9yU5cOPZrWgmRQU1OGJN1\nVOk5jYlNDL6+qplfv7yV0ZlJjM1LBaCiphWALz64hl9/dC4Af/3UAuZ1a84pIiIiIiIjW78COWvt\n2X29Zoy5Hng4GrgtM8aEgVygdr807gbuBpg3b549IKFBNtD90RwOQzgayFXUtAHwhW7zwY2LBnQA\nN/z9HQDG56ciIiIiIiLSJZ595B4BFgMYYyYCHqDuoO84Dg30uCLdBzt5dXOkOBZPzu+2PcNXznt/\njrjffHRur9MRiIiIiIjIyBXPPnL3APcYY9YBfuAT+zerHAr6mhD8aDkdhg5/iPuX7eKlTZHxXzKS\n3D3WKc5KAuAzp5cfMA2BiIiIiIhI3AI5a60f+Fi80j9WBnre7UmFaQDc9PBaAC6bMxqXs2fF6MUz\nRzG3NIuS7OSB3biIiIiIiAwL8WxaOSwMdI3cZXOLWfvtc2PPe2s26XAYBXEiIiIiItInTUp2CKMz\nk8hLS2DxpPxDr3yY0hLdpCa4aPMFSXQrlhYRERERkSOjQO4QHA7D8lv6HJzzqM0ry+LlTbUkuZ0D\nnraIiIiIiAxvqg4aJKXRppM5qQmDnBMRERERERlqVCM3SL5x0VQ+Mr+EiQVpg50VEREREREZYhTI\nDRK308G0URmDnQ0RERERERmC1LRSRERERERkiFEgJyIiIiIiMsQokBMRERERERlijLV2sPMQY4yp\nBXYOdj56kQvUDXYmRiiV/eBS+Q8elf3gUdkPHpX94FHZDx6V/eA5Xst+jLU271ArHVeB3PHKGLPC\nWjtvsPMxEqnsB5fKf/Co7AePyn7wqOwHj8p+8KjsB89QL3s1rRQRERERERliFMiJiIiIiIgMMQrk\nDs/dg52BEUxlP7hU/oNHZT94VPaDR2U/eFT2g0dlP3iGdNmrj5yIiIiIiMgQoxo5ERERERGRIUaB\nnIiIiIiIyBCjQO4QjDFLjDGbjDEVxpibBjs/w5ExZocxZq0xZrUxZkV0WbYx5jljzJbo/6zocmOM\nuSu6P941xswd3NwPLcaYe4wxNcaYdd2WHXFZG2M+EV1/izHmE4PxWYaaPsr+28aYyuixv9oYc0G3\n126Olv0mY8x53ZbrnHSEjDElxpiXjDEbjDHrjTH/E12uYz/ODlL2OvbjzBiTaIxZZoxZEy3770SX\nlxtj3o4eww8YYzzR5QnR5xXR18u6pdXrPpHeHaTs/2yM2d7tuJ8dXa5zzgAzxjiNMauMMY9Hnw/P\n495aq78+/gAnsBUYC3iANcDUwc7XcPsDdgC5+y37EXBT9PFNwO3RxxcATwEGOAl4e7DzP5T+gIXA\nXGDd0ZY1kA1si/7Pij7OGuzPdrz/9VH23wa+3Mu6U6PnmwSgPHoecuqcdNRlXwTMjT5OAzZHy1jH\n/uCVvY79+Je9AVKjj93A29Hj+UHgyujy3wLXRx/fAPw2+vhK4IGD7ZPB/nzH899Byv7PwOW9rK9z\nzsDvgy8C9wGPR58Py+NeNXIHtwCosNZus9b6gfuBSwY5TyPFJcBfoo//Anyw2/K/2oi3gExjTNFg\nZHAosta+CjTst/hIy/o84DlrbYO1thF4DlgS/9wPbX2UfV8uAe631vqstduBCiLnI52TjoK1dq+1\n9p3o41ZgAzAaHftxd5Cy74uO/QESPX7bok/d0T8LLAb+GV2+/3Hf9X34J3CWMcbQ9z6RPhyk7Pui\nc84AMsYUAxcCf4g+NwzT416B3MGNBnZ3e76Hg/8AydGxwLPGmJXGmGujywqstXshciEA5EeXa58M\nvCMta+2DgXVjtCnNPV1N+1DZx0202cwcInfIdewfQ/uVPejYj7to87LVQA2RIGAr0GStDUZX6V6O\nsTKOvt4M5KCyPyr7l721tuu4vzV63P/UGJMQXabjfmD9DPgqEI4+z2GYHvcK5A7O9LJM8zUMvFOt\ntXOB84HPGWMWHmRd7ZNjp6+y1j4YOL8BxgGzgb3Aj6PLVfZxYIxJBf4FfMFa23KwVXtZpvLvh17K\nXsf+MWCtDVlrZwPFRGoTpvS2WvS/yn4A7V/2xpjpwM3AZGA+keaSX4uurrIfIMaYi4Aaa+3K7ot7\nWXVYHPcK5A5uD1DS7XkxUDVIeRm2rLVV0f81wL+J/NhUdzWZjP6via6ufTLwjrSstQ8GiLW2Ovpj\nHwZ+z/vNNlT2A8wY4yYSSPzdWvtwdLGO/WOgt7LXsX9sWWubgJeJ9L/KNMa4oi91L8dYGUdfzyDS\nHFxl3w/dyn5JtKmxtdb6gD+h4z4eTgU+YIzZQaQJ9mIiNXTD8rhXIHdwy4EJ0ZFuPEQ6QT42yHka\nVowxKcaYtK7HwLnAOiLl3DU60yeAR6OPHwP+KzrC00lAc1fTKDlqR1rWzwDnGmOyos2hzo0ukyO0\nX//OS4kc+xAp+yujo2mVAxOAZeicdFSi/R3+CGyw1v6k20s69uOsr7LXsR9/xpg8Y0xm9HEScDaR\nPoovAZdHV9v/uO/6PlwOvGittfS9T6QPfZT9xm43jgyRPlrdj3udcwaAtfZma22xtbaMyHniRWvt\nRxmmx73r0KuMXNbaoDHmRiJfGidwj7V2/SBna7gpAP4dOafhAu6z1j5tjFkOPGiMuQbYBVwRXf9J\nIqM7VQAdwCePfZaHLmPMP4BFQK4xZg/wLeA2jqCsrbUNxpjvEbmwAviutfZwB/EYsfoo+0UmMvy0\nJTJ663UA1tr1xpgHgfeAIPA5a20omo7OSUfuVODjwNponxWAr6Nj/1joq+yv0rEfd0XAX4wxTiI3\n7h+01j5ujHkPuN8Y831gFZFAm+j/e40xFURqJK6Eg+8T6VNfZf+iMSaPSLO91cBno+vrnBN/X2MY\nHvcmEnSKiIiIiIjIUKGmlSIiIiIiIkOMAjkREREREZEhRoGciIiIiIjIEKNATkREREREZIhRICci\nIiIiIjLEKJATEREREREZYhTIiYiIiIiIDDEK5ERERERERIYYBXIiIiIiIiJDjGuwM9Bdbm6uLSsr\nG+xsiIiIiIiIDIqVK1fWWWvzDrXecRXIlZWVsWLFisHOhoiIiIiIyKAwxuw8nPXUtFJERERERGSI\nUSAnIiIiIiIyxCiQG8L8wfBgZ0FERERERAaBArkh6rE1VUz8v6d4et3ewc6KiIiIiIgcYwrkhqgN\ne1sA2FrbPsg5ERERERGRY02B3BBl7WDnIP6stdS1+QY7GyIiIiIixx0FckOcMYOdg/j5w2vbmff9\n59lZr1pHEREREZHuFMgNUZbhXyX34sYaACobOwc5JyIiIiIixxcFckNVNI4bzk0sR0KwKiIiIiJy\nNBTIDXGh8PANdmJB6jBuPioiIiIicjTiHsgZY5YYYzYZYyqMMTfFe3sjRVeMEx7OVXJRRpGciIiI\niEgPcQ3kjDFO4FfA+cBU4CpjzNR4bnOkCEdr4sLDuUZusDMgIiIiInKcineN3AKgwlq7zVrrB+4H\nLonzNkeEULQm7niN42pavOysb+9foBl969KtdXgDoYHJmIiIiIjIMBDvQG40sLvb8z3RZTHGmGuN\nMSuMMStqa2vjdFb77QAAIABJREFUnJ3ho6tvXH27nz2NHQe8Zo+iyaU3EOKRVZXUtHr7lbf3qlpY\n8IMXOOOOl/nZC1v6lRbAXS9WcP+yXf1OJx5avAFqWvpXXgMtFLb88sUtNHcGBjsrIiIiIhIn8Q7k\neuvc1CPCsNbeba2dZ62dl5eXF+fsDB9dgdw/lu3itNtfYl1lM95AiOoWL1O/+TTffHQ9gVCYxnY/\nta0+3txaR2O7n+aOvi/uX95UwxceWM0dT2+KLXtpY00sUAyHLS3eQwcH3QPL6uajD3K6j1rZ3Bk8\n6nR6c7ByOBJn3vEyC37wAn98ffuApDcQnnuvmjuf3cxtT22IS9PbTn+IiprWo35/KGwPuPkgPXkD\nIW74+0p21PV/DsV2X5BfvVRBMBQegJzJYGrzDex5UEREhjZXnNPfA5R0e14MVMV5myPC/oOcXPSL\n1wGYVZyBLxjm/uW76PCH+Nc7ew54762XTmd8XioLyrN5ZXMtzZ0BPjBrFE3R4Ka+3c/O+nb+94HV\nvLOrCYC/fmoB/15Vyb9XVXL+9EJKs5OZWJDGxbNG4XE5qG7xYi00dwZYvqMhti1vMMS+Zi+7GztI\nT3QzqTAt9povGGJLdRsTC9IIW0swbAkEw7icBpfDwZrdzbF1A/tdhNa1+dhZ386ozCSykj14nJF7\nEg7H+/cOrLXc9UIFRRmJfHj++4fh0q31XPX7t5hSlM61C8u5YEYRCS7nke2AqPp2PwDfe/w9rjmt\nvMdrW6pbcTkdlOemxJa1+YK8trmWE8ZkkZ+eCMCHf7eUBJeDe685kaYOP5nJHipq2ghbS3FWEsme\nI/uadl3sNbYHmPO955hflsWV80txuxw8urqSG88cz9i81F7f+9a2etp9Qc6aUkC7L8jvX9vGtFEZ\nvLq5lmsXjqUkO5lbHlnLw+9Ucv+1JzEuL5W8tAQAGtr9JLodPfJrrcXsN2v9T5/bzC9fquDNmxYz\nKjPpiD7bwcz41jNcOLOI/LQEJhel43QYZhZnUJQR2YY/GKauzUdhemKP46S7QCiM2+lg7Z5mvv7v\ntSyZXsj1Z4w7YH1vIMTaymaKMhIpzko+IJ3Z332W86cX8cPLZhx2/v3BMFVNnRRnJfFGRR1Prt3H\nk2v38dKXF8WOodpWH82dfsbnpx0itYhgKMy0bz0DwJicZC6aOSr22kubakhyOzlpbE5smbWWNXua\nyUxyU9btuO2uttVHqzfQ5zF0qPx4g2E8TgdNHX4+/49VfHXJZE4YkxVbZ83uJkZlJlHZ1MmM0Rk4\nHQZ/MIzH5cAfDONymD73X1+W72igOCuJ7BQPzR0BOvwh8tISSPY4afMF6Yw+7zpWA6EwO+raGZ+f\nesDx2yUQCtPmDZKV4jnotr2BEL98sYLLTyiOlekjqyrZ3dBBeV5Kj31yMP9auYcvPbSGF750BuMO\ns+y7mqQnuo/u/Ha0fMEQG/e2MrkoLXZutdbS5guSlug+6nSttVjLEe//I0m/r/09UMJhyw+e3EBl\nUyfXLxrH9FEZvLWtngXl2bicI2Mg8Xjvx2PlzYo6Vu9p4typBYd9Th5KrLX8+uWtLJyQx4zijMHO\nzmFp7gywvqqZ6aMzSO/jXNPVam24fN/M0TTBO+zEjXEBm4GzgEpgOfD/rLXre1t/3rx5dsWKFXHL\nz1DW7gty9Z+WsXxHI4sm5ZGR5ObR1QMXEzsdJlbLd87UAtISXDy8qvKQ7zt5bA6fO3M8H/vj2we8\nlpPiiQU6XXJTE3A7DT+4bAbfeGQdexo7KclOYnfDwSf9vnbhWL5+wRQ6/EGWbq3nd69uY9n2hh7r\nJLgcnDY+l1PH53Lfsl1U1LTFXrt41ii+f8l0EtwO7nxmE3/Yrwbt8hOKCVtLfloiDhOp6QyELNkp\nHpwOw7lTC3hlcy37WrwUZSTx+cXj6fCH+PJDa2Jp7LjtQqqaOrn96Y0snJDHzQ+vxR8KM7M4A7fT\nwc76dura3i+PFI+TmcWZLN1WD8C8MVms2NnIzz4ymy88sDq23mVzR1OcmUSrL4g/GMYfDNMZCOEN\nhGjuDFDf5mdsXgpVTV5OGJPFvW/tBGBBWTbLdvQsoy6F6YlMKEilscNPhy+EPxSmptWHPxgJmP/3\n7ImEreXn3ZrGnjO1gDsvn8Ws7z7bI60LZxSxcV8LW2sjtUcfP2kMs0oyuXfpDt6tbGbm6Azy0hJw\nOgxnTS7gO/9ZT7s/xDcumsrepk7G5CSTkuDi6/9eS15aAgVpiTiMoSw3mVAYQuEwp4zPZV+zl85A\niHOmFgCQn5ZAZrKHsLX85uWt/OblrQd8zqlF6Xzjoqlsq2vjp89tpq7Nj8NE+paePSWfuz8+j0A4\nTCBk+elzm/nj69uZXJjGxn3v1zhOyE/lpvMn8/KmWv71zh6mjUpn+Y7G2OvnTSvgktmjWTQpD18g\nzA1/fye2T9d861wyktzc/PBadjd0MLc0k4mFaeys7yA90YXb6SAYtjyzfh+vbakDYHRmEv991ni+\n9q+1AOSmepg+OoN3djbS4o0E6Z9fPB6nw/D2tgYaO/yU5aQwqyQTY2Bfs5d/rtxDSXYym6tbY9/r\nu66awwdmRYIGay3lNz8JwEfmlfB6RR1JHicOA5urI9+b335sLm6ng621bZwxMZ+n1u1lb5OXB1ZE\nWstfOLOIgrRElu2oJxC0hG3k+zKvLItAyOIPhslLS+C08bmMz0/lrhe28PvXthG2kfPNLRdM4buP\nv8fZUwq44/KZBEJh2v0hzrzz5VjZ5qR4uH7ROL7/xAYmFaTR5gviD4X50NxigqEwhRmJeFwOFk3M\npzQnmQdX7GbD3hY6fCH2tXjJS0ugusUbK9v9pSW6aI2WaXluClfMK+b6M8bxwPLd3PTw2tj+T3A7\nSHI7GZ2ZRGl2Mv9eXRk7Z50+IXJs1rX5cDocTCmKXNCNz0/llgum8Pb2Bj76h7fJTfXws4/MoTEa\nwHZ54r9PY9qo9y+SXt9Sx9bayI2cCflphKxl+fYGfvlSBQB/uno+CyfmUd2tSXd+WgIuZyTQ/c3L\nW+nwB3l0dRX7Wrx4nA4+v3g8HpeDuWOyqGv1Mbs0k9+9so0n1u4lNcHF2NwUbjhzPM++t4+WziAt\nnQFafUEyktycPSUfj9OBx+Vg2fYGirOTmTcmi73NnSR7XCwoyz7ggvyuF7bwk+c2c/UpZXxm4VjC\nYcuN973Dmj3NLCjPZn5ZFjkpCXiDIQyG1EQX88uymFyY3ut+amj3s76qmU/+aTnBsCU31UOrN8hH\n5pfw8DuVzCnNJMHloKrJS3NngIL0BC6ZPZol0wtJ8jhjF3Qd/iBtviDBkI0cb74QYWvxuBz87pVt\n/OudPVx9Shl5aQk8s34flY2dLCjP5tZLZ5DdLWB/a1s99W1+CtITmFeW3SOvlU2dbK5u5ZVNtSS6\nnRRnJREMhXlzaz05qR6umFfCZb9+84DP6HQYHv3cqbHzT2cgRCAYpsMforrVy1Nr9+FyGl7eVBvb\n57/7+AnMKc3iP2uqCFvLlKL02L768kNrqGzq5JkvLMTpMLj7uGhdu6eZLz+0hkvnjuZTp5bz21e2\n8vGTxuB0GpZvb+C9qhYS3U5KspOZUJDKvUt3kpPiYe6YLK7/20pyUyNl7XIa5o3JYkF5do9g+MWN\n1fzlzZ0YA584pYxpo9L5+B+WsaO+nX9cexIJLgfhcKQlT7s/hNtpOH96EWFr2d3QQSBkueuFLSR5\nnKQlukhPdLNkeiF5aQmsq2xmQn4aW+vaCIUs4/Mjv2nNnQGmjYr87uysb+fSX79JUUYiD332ZOrb\n/Dz7XjWfOrWM2jYfDyzbTYI78t1JdDs5sTyHDn+QquZO/MEwM4szKc2O/EZ15w+GmfHtZ/AFw1ww\no5Abz5zAyl2NYC1hCy6nYWJBGq7otVUobClIT6QkO5mqpk78oTDrKpsZl5fK9NEZBENh/rJ0J9kp\nbjr9YUZlJvZ5o/MHT25g5c5G2n1BwhY+eWoZgVCY+jY/3/vgdJLcTt6oqGNdVQu1rT4mF6Zxwpgs\nxuenHnBTp/vNixZvAI/TEVtna20bZ/34FbJTPDz02ZNp9QbZ3dDB+PzU2A21qqZO2n1BgmFLY0eA\nURmJnDWlgPeqWqhu8bKpupXHVldx5xWzmDqq5/e7uTPApb9+g4UT8khLdLGuspk7rphFbmpCbB1v\nIES7L0ii20lKgotXNteS7HEyP/q921bbxrt7mgmFLWv2NPHXpZHrn5wUD184ZyK+buMrdIU79y/f\nxdbadj51ajkLJ+ayaFJ+r+U82IwxK6218w65XjwDuWhGLgB+BjiBe6y1t/a1rgK53nXdiT2U73xg\nGvPLsmnzBbnjmY09LjavOKGY+eXZpHhc/OS5TWytbefzi8fjD4Z5ZXMt6YnuPi/6u8wpzeS6hWOZ\nUJBGc2eA//rjsgOa+pw8Noc5pZl0+EP8ZemOAyYsP31Cbp8XVQczb0wW66ta6Oz2pSxIT+DGxRPY\nVtvGs+uryUtLYPXuptjrH5lXgstpeHR11VE3SZpdkhlLMz8tgQS3o8+gc+HEPF7dfGA/z7REF5ML\n06hv89PiDVLX5uOcqQWs2NFA40GaeBZnJbGnMbItYyDV48LjcuB0GGpafSS5ncwqySAchvVVzbT7\nQyS4HPiCh25CN7Egckd/c3UbDgMXzRxFYUYiO+vbeWZ9da/vyUx2k5+WELvQP1yjM5PISHLz3t6W\nw1q/64I9FLY0dvgP6/MciewUDw373WDYn9tpCIQOfm7sqiE6GKfD8NETS2M/LoeSkeTus29jotuB\nN3DosvA4HQTD4QMGQvrdx0/gvGmFAPzhtW18/4kNsdfSE120+oKMy0vtcQPkcGQmuzmpPIdVuxup\nbvHhchhcToPDGDr8PQcpyk31ML8sm6fW7TuibfRm//LfPwDvkuJx0r5fPj51ajkZSW52NXTQ4Q/y\n5tb6WLmfMi6HN7dGAvFzphbgdhp8gTDVrV7WVb5/DGcmuynKSKKlM4DbaZhdksmuhg68gTA76ttj\nn72vffrzK2fzP/ev5sIZRWSneNhR3866yuaDnhMA/vzJ+Tz7XjX3vf1+v+HUBBfnTitgxY5GdjW8\n32x54cQ83t5Wf9Dv0JJphTy9/v390f2m3uGYVZzBSeNyWDKtkG//5z2CoTDrq/r+rmclu/v8jKdP\nyAXoUV6t3iA76tsP+C3pTV/f21nFGbT6gmyrPfqmymdPKaCuzdfjNwbg6lPKSEt00ekPEbZwzxsH\nb2bfdSOpLCeZj59cxoodDT2+Dwcr/97OOfvvv77MKs7gzMn5lOemsHRrPWv2NJPgchzweQaKx+XA\nwICdv0dnJtHmCx5R3++zp+Tz/Iaa2PPS7OTY9+M/N57GDfetPORN5C7piS4mF6Wzpbr1gOPXGA7r\n+OzLdQvH0tQRiN0kOxyjMhLZ1+I9ogHvUjxO/t+Jpexr8bGlupXtde0YAwsn5LGzvoNN1a0YA9NG\npbNpX+shfwOPVGl2MoFQGAMYYwiGw1S3+A5YL3JDwuByOli58/3rWJfDEIx+4FnFGQe83tt6h+P0\nCbnce82JR/6B4uy4CeSOhAK5nqy13P3qNn741MbYsi+fO5E7n93c6/r//OzJB9wd7GqS2NfduC5t\nviAn3vp87ILn0jmjuXhWEbmpCXzpwTV8bclkzo7WhHTxBkL8+NlN/P617dywaBxfPndSjzuzM771\nDK2+ID/60Ey++q93ufOKWXxo7mjufnUbNa0+Thufy/zybKqaOglby6SCNGpafeSlRu7Ubq9r58K7\nIk1GuwKUq08pIyvZw/zyLGYWZ5K6312ybbVtbNjbyqySjFiTtw5/kIdW7CEYtviCISobO/l79CLo\nk6eW8X8XTsXpMLxRUUdJVjIFGQlU1LSRnuimJDuZLdWtbK1tZ9GkPBLdTipqWun0h8lIcvPVf63h\nrW3vB8CzSjL59Gnl7Gv20uIN8OnTxpKR3HdTooqaVm68b1XsAjQ3NYFpo9KZVJjGV86bRDBkCVlL\nktuJs1vZtvmCsTuvXWpbfeSmevAFw5z/89doigZB4/NT+e/FE8hMduMPhinLTelxpy8Utj3Sbmz3\n8/javbR6A8wYncHH/7gMiJyEu4LH2y6bydvbG3hy7V4+u2gc1S1e/MEwyR5n5A5yi5cpRemU5aTE\n8rhyZwNLt9ZTnpvKE2urGJ2ZxCdOKSMQsrT7gjy4YjefPm0spTnvN1W01vLse9XkpnrITklga00b\nTqcBCzvq2wmGLMZEfhTOnJRHWU5KpFaqxUswZFlX2YwlUquSkewmL3qnb3tdO69sro02B3XidhrS\nEt1s3NtCfnoi1y0cG2t20dTh561oDdu4vFRavAFavEGmjUonPdHNbU9t5PWKOhra/bEA8eMnjYnV\njKZ4nGQme/jyeRMJhyHZE2nOGAxHagWCIcve5k6yUzxMKEjjybV7+ceyXQfc8PjPjadRkp3Eo6ur\nyEtLIDPJTZsvSCAUaYI7Pj8SnCd7nBhjsNbyvcc3xC4sr1pQSnWLF4eBDXtbqWzq5LbLZnD+9KID\njtHKpk7+9Pp23tnVGGteffP5k2Pnoq0/uIDdDR34Q2FKs5NJdDux1uKL3tXu8u6eJm57aiPFWUmc\nN62QMybm4XI6+OfKPTy1di8vbKzh3KkFnDAmi7CF7XVt7G32MrEgjasWlPDWtgZSEpyUZifT7guR\nn55AaXYywbAlLcHFQyv38NjqKtZVNceahv/g0hlcMnsUy7Y3sKm6lU+eWkaCy0m7L0hNq4/S7OQe\nx3vXceYPhfnc31fx/IbIjYycFA8rv3FOj/Xq2nwYIhfiuamePpvhhcKWB5bvZtO+FryBcOwC7cHr\nTsZaS3qSm8mFaZz+o5diN2scBvLTEkn2OPnzJxfw3t5mfvFiBU6H4WMnjmF2aSbn/vRV0hNdtHiD\nTC1K5xOnjOEXL1bE0hibm8K00RmkJ7ooy0nhMwvHApGmjo+truKr/3qXZPf7ge3nF4/nS+dO4tN/\nWc6Gva2cODabn3x4Nh3+IE6HoakjEKkFjbYEeGLtXpZtb8AYWLXrwADAGFgcvbs9viCVspwUnMZQ\n3+6nrs3HDYvGkZOaQF1bJOCPHDeRpq9/fH07b1TU4XQYZpVkkuJ5/zgak5PCrJIMclISWFCezZbq\nNooyE2mMfn+7zgFF6Yls3NdKc2eAnfXtNHT4+VG0z/fozCTOmJTHlKJ03NEaqiSPE7fTgS8YIhS2\nzCnJwu0yeJwOOvwhalp9rNndxL1v7STB5SDZ44x9H762ZDK3P70x9rmT3E6CIUtGspsPzBpFUUYi\niyfnk5rgwuV04DDw42c3c+9bO8lO8fDil84gM/n9Wr5HV1fyzs5GkjwuxualUJSRiMfpwO1y4HIY\nNu5t5dK5o9lS3UarN4A/FOZLD66hpjVyIZya4OLWS6cTCkdqw3c2dFDb6iM1wcW9b+08IDh0GMhO\nSSBsLTcsGgdEaj5/HW3ZcMbEPDwuB9ecVo7LYWjzBWnqCLCnsYNg2OJ2OhiTk8w5UwuwNlJL+b8P\nrKbdF2JWSQYnjMkmFA5T2dTJk2sjgeaH5xWTkuBiTmkWmUnu2I1ZayOBUn56In9+cztOY8hPT2RP\nYwfv7GzioplFfP6sCQC8sKGaa/4SuU48a3I+xVlJ/GXpTs6eUsAFMwpJT3Tz65creGdXE5MK0kjy\nOPnI/BI27WvlvmW7DgiEk9xOLj+hmCfX7qUsN4U2b5CPnVTK7JIs/vbWTrJSPLT7gizb3kBGspuq\npk4CoTBjslNwOCLn1XWVzSS4nCycmMfYvBQMsKexk5pWX7SriCFs4fUttTy5dh/TR6ezYkcj9e1+\nktzOHjeov33xVOaUZlHd4u0zkHI7DWdMysPjjATij66uYsn0QmpbfayrbCbZ42JUZiIXzRyFwwFv\nVtTz+9e28ebWehLdDlI8rgNaS+WnJVDf7mdcXgrFWcm8uDESAC+enM8lsyMtOYIhi8tpaPeFSEt0\nRWqznQ7G5KTgcRm8gTB3v7qNlTsbsdZyxxWzyEr28OqWWm5/OvI7MKs4M/a9t9bS3BlgRnEGYRup\n5fzbWztjrZiC4cjv49wxWSS6nLT7g1Q2dlLT6sXlcMRavpw7tYBbLpyC2+kgO8VDKGzZXtdOXloC\nSdHzSNeZ2hiDw0Cyx0WLN8CN963CAH/51IJey3owKZAbBh5asZuv/PNdXA7DTedPZmZxJgvKs9nd\n0MGDK3bzixcreqx/JH0nehMMhRl/y1MAPP7505g+un9ton/1UgWvb6njT5+cf1R9NDr8Qf77H6t6\n3FFb/c1zevz4Ha2uC5Wj7RvXZdWuRm57aiOfPLWM0uwUphSlHXEfiweW74o1o3vyv08/oPnB0QiG\nwjgdZkD6e/xr5R6MgcvmFvc7reFsR1073/7PetIS3dxx+UwS3c5+7YeH39nD71/bjsNE7rrfuHj8\nUaWzo66dRdHmiskeJ+W5KYTClotmFnHj4glHlFbZTU9E0rztwiPORzyFw5bH1+7FEGnq25++N9Za\nnlq3j/LcFKYU9f+7CFDT6iXB5SQjqWfA3OqN9CkOhSM1gfs33+ouGArzqb+sYPWuRtIS3dz2oRmc\nPiEPbyBERU0bZbkpB9zY2l8obHEY6PCHaGj3Mzozqd/9lEJhy61PbKCm1cvUUencsGj8UaflD0a+\nL/sH2kfLWsu9b+1kybTCWJ/k/urqg3PKuFx8wcgFuMfpOOzvZjhsB6xvmC8Y4ul1+yLdCibkHXT/\nN7T78QfDtPmC+IIhynJSej3eXthQTWcgdNh9Nw9HU4cfbyDSFHog1Lf5cDkdB3yfDiUYCrOvxctd\nL2wh2eNixugMLpxZdMz7kPaWr64AI559NLvftA2HI83hnQ4TvQmcMOT6LHadP+eXZfer7+3xSoHc\nELdyZwMf+s1SoO/gpeuiKj3RRUaym2e/cEbs7sPROvPOl9le185bN581YCfd/lr4o5fY1dCBMVBx\n6wUD9iN/vHj83SpuvC/SZ+b1r53Z6+AZIkfLGwhx9Z+WUdfm58Pzirl24bijTuuljTXkpSX0+yaP\niIiI9O1wA7l4j1opR+lnz0cGmfjrpxYcsgbq3W+fN2DbvfeaBVQ1eY+bIA4izR/fqKjjpLE5wy6I\nA1g0KZ+vXzCZtEQ3owdwFEcRiIxYeP+1Jw9IWmdOPj47hYuIiIxECuSOQ7sbOnhtSx1jcpJZOPHY\nzq1XnJV83NUIffLUcj55avmhVxyiUhNc/aolEREREZGRR4HccejuV7cBcPP5Uw663s+vnE1p9vEV\ndImIiIiISPwpkDsOvbixhvREF0umFx50vUtmjz5GORIRERERkePJ8JjWfBipafVS2dTJdWeoqZ2I\niIiIiPROgdxx5v5lkfmG5u83H5yIiIiIiEgXBXLHkVDY8vi7VQDMLc0c5NyIiIiIiMjxSoHcceTF\njTVsrm7jktmjcDm1a0RE5P+3d+9Bdtb1HcffX3azm6u5B9JsIBEoiOjEkFI6mQoqIl4KdUqn6VRl\nsC2V4kyrdRB0pl5aptVe7DgdUawotlqkgCO1gELFOrYCBoRcuEiESG4QaK5kySa7++0f5wnupHvL\n7tn97dnzfs2c2ef8nie73/2e3z7JJ8/vPCtJUv9MCxPIXRueBeCTF59ZuBJJkiRJE5lBboJYt3UP\ntz60lSVzpjF72pTS5UiSJEmawAxyE8THbt8IwKd+67WFK5EkSZI00RnkJoCPfWsDP3lmDwtmtrP6\nlPmly5EkSZI0wRnkCtuyq5Mbf/RzZra3ct27VhIRpUuSJEmSNMEZ5Ap76oUDAHzxPav83XGSJEmS\nhsUgV9jazbsAOHnRjMKVSJIkSWoUBrnCNv9vJ4tnT2XRrKmlS5EkSZLUIAxyBV1z2zr+/ZHtLJrV\nXroUSZIkSQ3EIFdIb29y20PbaG89jk/4C8AlSZIkHQODXCHrtu2lq7uXP/+NM1ixdE7pciRJkiQ1\nEINcIV/64dMAvGbJ7MKVSJIkSWo0BrlCHt+xj+ULZvDaDq/GSZIkSTo2BrkCunt6eXLniyz0JieS\nJEmSRsAgV8CuzkMAvOG0RYUrkSRJktSIDHIFXH3regCWzZ9euBJJkiRJjcggV8DO/QcBOPe0hYUr\nkSRJktSIDHIFvHSoh7e/ZjHT21pLlyJJkiSpARnkCjh4uJdpbS2ly5AkSZLUoAxy4+ylQz1s2/MS\n06YY5CRJkiSNjEFunF3/g6cAmD+zrXAlkiRJkhqVQW6cHbnRyRXnnVy4EkmSJEmNalRBLiL+JiIe\nj4h1EfHNiJjTZ981EbEpIp6IiLeMvtTJYXfnIZbNn057q0srJUmSJI3MaK/I3Q2cmZmvBX4KXAMQ\nEWcAa4BXAxcCn4sIkwtwx/pnaWv1QqgkSZKkkRtVosjM72Zmd/X0PqCj2r4YuCkzuzLzaWATcPZo\nvtZkMau9lUWzppYuQ5IkSVIDq+elofcCd1bbS4AtffZtrcb+n4i4PCLWRsTa559/vo7lTEy9mZx+\nwqzSZUiSJElqYEP+RuqIuAc4oZ9dH83Mb1XHfBToBr525I/1c3z29/kz83rgeoBVq1b1e8xkcrgn\naW1xaaUkSZKkkRsyyGXm+YPtj4hLgXcAb8rMI0FsK7C0z2EdwPaRFjlZZCaHe3tpa+kv50qSJEnS\n8Iz2rpUXAh8GLsrMzj67bgfWRER7RCwHTgUeGM3Xmgx6epNMvCInSZIkaVSGvCI3hH8E2oG7IwLg\nvsx8X2ZujIibgUepLbm8MjN7Rvm1Gt7hntoFyykGOUmSJEmjMKogl5mnDLLvWuDa0Xz+yeZwby8A\nU1xaKUmSJGkUvDQ0jg53Hwlytl2SJEnSyJkoxpFLKyVJkiTVg4liHB3ucWmlJEmSpNEzyI2jXwQ5\n2y5JkiRp5EwU48illZIkSZLqwUQxjlxaKUmSJKkeDHLjyKWVkiRJkurBRDGOXFopSZIkqR5MFOOo\n26WVkiQjSwumAAAL4UlEQVRJkurAIDdOunt6ufTLDwAwpdW2S5IkSRq51tIFTHZX/MuD/HDTC+w/\n2A3AiqVzePUvvaJwVZIkSZIamUFujGQm/7Z2K3dueJaW44LLVi9j/ow2rnzDKUS4tFKSJEnSyBnk\nxshtD23jqlvXAXDvn53HifOnF65IkiRJ0mRhkBsDnYe6+as7H2PhrHbu+cC5zJ4+pXRJkiRJkiYR\n77oxBv72Oz/lhRcPcdnqZYY4SZIkSXVnkKuz5/d3ccN/P82vvXI+V5x7culyJEmSJE1CBrk627B9\nLwBrzl7qTU0kSZIkjQmDXB3tOnCID3zjYdpaj+P8Vx1fuhxJkiRJk5RBro6+8F8/Y0/nYd65Ygkz\n2r2PjCRJkqSxYZCrk8zk1oe2cVzAte88s3Q5kiRJkiYxg1ydfOaeJ3nhxS4++OZfprXFtkqSJEka\nOyaOOtiyq5PP/ueTTG9r4Q9f/8rS5UiSJEma5Axyo3S4p5ff+cKPAPjKZWfT3tpSuCJJkiRJk51B\nbpQe2bKH7XsPcuqimZy9fF7pciRJkiQ1AYPcKP3H+h0AXPeuswpXIkmSJKlZGORG6fn9Xcxoa+GU\nRTNLlyJJkiSpSRjkRmnLrk5WnjS3dBmSJEmSmohBbpSe2dXJ0nnTS5chSZIkqYkY5Ebh8Wf3sbvz\nMB1zp5UuRZIkSVITMciNwhPP7gdgxdI5hSuRJEmS1EwMcqOwY+9BAM5cMrtwJZIkSZKaiUFuhDZu\n38vn7t1Ex9xpvGLqlNLlSJIkSWoidQlyEfGhiMiIWFA9j4j4bERsioh1EbGyHl9nIrnqlnXsO9jN\neactLF2KJEmSpCYz6iAXEUuBNwPP9Bl+K3Bq9bgcuG60X2ciOdDVzcbt+3jzGcfzyYvOLF2OJEmS\npCZTjytynwGuArLP2MXAV7PmPmBORCyuw9eaEO5YvwOA1504h+OOi8LVSJIkSWo2owpyEXERsC0z\nHzlq1xJgS5/nW6uxSeHHm3cB8N7VywtXIkmSJKkZtQ51QETcA5zQz66PAh8BLujvj/Uzlv2MERGX\nU1t+yYknnjhUOcXtOnCIm9duZc70KUyd0lK6HEmSJElNaMggl5nn9zceEa8BlgOPRARAB/BQRJxN\n7Qrc0j6HdwDbB/j81wPXA6xatarfsDeRbNv9EgAfuuC0wpVIkiRJalYjXlqZmeszc1FmLsvMZdTC\n28rMfBa4HXhPdffKc4C9mbmjPiWXddfG2rdx+gmzClciSZIkqVkNeUVuhO4A3gZsAjqBy8bo64y7\nWx7cCsApi2YWrkSSJElSs6pbkKuuyh3ZTuDKen3uiaKru4fn9nXxvnNPZs70ttLlSJIkSWpSdfmF\n4M3ixv/ZDMCJ86aXLUSSJElSUzPIHYMtu2o3OrnkrI7ClUiSJElqZga5Y7Bj70ucfsIs2lptmyRJ\nkqRyTCTHYPuegyyePbV0GZIkSZKanEFumLq6e3h0xz4Wz5lWuhRJkiRJTc4gN0yfvusJAJbN90Yn\nkiRJksoyyA3T1t2dALz7nGVlC5EkSZLU9Axyw7S78zC/unwe09paSpciSZIkqckZ5IZp94FDzJvh\nLwGXJEmSVJ5BbhheOtTDkztfZK5BTpIkSdIEYJAbhs99fxMAS7xjpSRJkqQJwCA3DC+82AXAH/z6\n8sKVSJIkSZJBblgOdPWwfMEM2lu90YkkSZKk8gxyw9B5qJvp3q1SkiRJ0gRhkBuGA109BjlJkiRJ\nE4ZBbhhqV+RaS5chSZIkSYBBblgOHOphRrtX5CRJkiRNDAa5Ydh/8LBX5CRJkiRNGAa5IXzt/p/z\n3L4uZk01yEmSJEmaGAxyQ2hvbeGCM47n3eecVLoUSZIkSQLAy0xDuOSsDi45q6N0GZIkSZL0Mq/I\nSZIkSVKDMchJkiRJUoMxyEmSJElSgzHISZIkSVKDicwsXcPLIuJ54Oel6+jHAuCF0kU0KXtflv0v\nx96XY+/Lsffl2Pty7H05E7X3J2XmwqEOmlBBbqKKiLWZuap0Hc3I3pdl/8ux9+XY+3LsfTn2vhx7\nX06j996llZIkSZLUYAxykiRJktRgDHLDc33pApqYvS/L/pdj78ux9+XY+3LsfTn2vpyG7r3vkZMk\nSZKkBuMVOUmSJElqMAY5SZIkSWowBrkhRMSFEfFERGyKiKtL1zMZRcTmiFgfEQ9HxNpqbF5E3B0R\nT1Yf51bjERGfrV6PdRGxsmz1jSUiboiInRGxoc/YMfc6Ii6tjn8yIi4t8b00mgF6//GI2FbN/Ycj\n4m199l1T9f6JiHhLn3HPSccoIpZGxL0R8VhEbIyIP6nGnftjbJDeO/fHWERMjYgHIuKRqvefqMaX\nR8T91Rz+RkS0VePt1fNN1f5lfT5Xv6+J+jdI778SEU/3mfcrqnHPOXUWES0R8ZOI+Hb1fHLO+8z0\nMcADaAF+BrwSaAMeAc4oXddkewCbgQVHjX0auLravhr4VLX9NuBOIIBzgPtL199ID+D1wEpgw0h7\nDcwDnqo+zq2255b+3ib6Y4Defxz4UD/HnlGdb9qB5dV5qMVz0oh7vxhYWW3PAn5a9di5X673zv2x\n730AM6vtKcD91Xy+GVhTjX8euKLa/mPg89X2GuAbg70mpb+/ifwYpPdfAS7p53jPOfV/DT4IfB34\ndvV8Us57r8gN7mxgU2Y+lZmHgJuAiwvX1CwuBm6stm8EfrPP+Fez5j5gTkQsLlFgI8rMHwC7jho+\n1l6/Bbg7M3dl5m7gbuDCsa++sQ3Q+4FcDNyUmV2Z+TSwidr5yHPSCGTmjsx8qNreDzwGLMG5P+YG\n6f1AnPt1Us3fF6unU6pHAm8EbqnGj573R34ebgHeFBHBwK+JBjBI7wfiOaeOIqIDeDvwT9XzYJLO\ne4Pc4JYAW/o838rgfwFpZBL4bkQ8GBGXV2PHZ+YOqP1DAFhUjfua1N+x9trXoL7eXy2lueHI0j7s\n/Zipls28jtr/kDv3x9FRvQfn/pirlpc9DOykFgJ+BuzJzO7qkL59fLnH1f69wHzs/Ygc3fvMPDLv\nr63m/Wcior0ac97X1z8AVwG91fP5TNJ5b5AbXPQz5u9rqL/VmbkSeCtwZUS8fpBjfU3Gz0C99jWo\nn+uAk4EVwA7g76pxez8GImImcCvwp5m5b7BD+xmz/6PQT++d++MgM3sycwXQQe1qwqv6O6z6aO/r\n6OjeR8SZwDXA6cCvUFsu+eHqcHtfJxHxDmBnZj7Yd7ifQyfFvDfIDW4rsLTP8w5ge6FaJq3M3F59\n3Al8k9pfNs8dWTJZfdxZHe5rUn/H2mtfgzrJzOeqv+x7gS/yi2Ub9r7OImIKtSDxtcy8rRp27o+D\n/nrv3B9fmbkH+D6191/NiYjWalffPr7c42r/bGrLwe39KPTp/YXVUuPMzC7gyzjvx8Jq4KKI2Ext\nCfYbqV2hm5Tz3iA3uB8Dp1Z3ummj9ibI2wvXNKlExIyImHVkG7gA2ECtz0fuznQp8K1q+3bgPdUd\nns4B9h5ZGqURO9Zefwe4ICLmVsuhLqjGdIyOen/nO6nNfaj1fk11N63lwKnAA3hOGpHq/Q5fAh7L\nzL/vs8u5P8YG6r1zf+xFxMKImFNtTwPOp/YexXuBS6rDjp73R34eLgG+l5nJwK+JBjBA7x/v8x9H\nQe09Wn3nveecOsjMazKzIzOXUTtPfC8zf49JOu9bhz6keWVmd0S8n9oPTQtwQ2ZuLFzWZHM88M3a\nOY1W4OuZeVdE/Bi4OSJ+H3gG+O3q+Duo3d1pE9AJXDb+JTeuiPhX4DxgQURsBT4G/DXH0OvM3BUR\nf0HtH1YAn8zM4d7Eo2kN0Pvzonb76aR299Y/AsjMjRFxM/Ao0A1cmZk91efxnHTsVgPvBtZX71kB\n+AjO/fEwUO9/17k/5hYDN0ZEC7X/uL85M78dEY8CN0XEXwI/oRa0qT7+c0RsonZFYg0M/ppoQAP1\n/nsRsZDasr2HgfdVx3vOGXsfZhLO+6iFTkmSJElSo3BppSRJkiQ1GIOcJEmSJDUYg5wkSZIkNRiD\nnCRJkiQ1GIOcJEmSJDUYg5wkSZIkNRiDnCRJkiQ1mP8DYUgD4Gm8+BsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28702feae10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "wei,result = stocGradAscent1(np.array(dataArr),labelMat,numIter = 40)\n",
    "\n",
    "x0 = result.transpose()[0]\n",
    "x1 = result.transpose()[1]\n",
    "x2 = result.transpose()[2]\n",
    "y = np.arange(0,len(x0), 1)\n",
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(15,8))\n",
    "\n",
    "plt.subplot(311)\n",
    "plt.plot(y,x0)\n",
    "plt.subplot(312)\n",
    "plt.plot(y,x1)\n",
    "plt.subplot(313)\n",
    "plt.plot(y,x2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEKCAYAAAAFJbKyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCUWJQ0REQvn/ABDMn0DHjcUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28702f82898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotBestFit(wei)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EX : Horse Colic"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### by np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def classifyVector(inX,weights):\n",
    "    prob = sigmoid(sum(inX*weights))\n",
    "    if prob > 0.5: return 1\n",
    "    else: return 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def colicTest():\n",
    "    frTrain = open(\"./Data/CH5/horseColicTraining.txt\")\n",
    "    frTest = open(\"./Data/CH5/horseColicTest.txt\")\n",
    "    trainingSet = []; trainingLabels = []\n",
    "    for line in frTrain.readlines():\n",
    "        currLine = line.strip().split(\"\\t\")\n",
    "        lineArr = []\n",
    "        for i in range(21):\n",
    "            lineArr.append(float(currLine[i]))\n",
    "        trainingSet.append(lineArr)\n",
    "        trainingLabels.append(float(currLine[21]))\n",
    "\n",
    "    trainWeights, results = stocGradAscent1(np.array(trainingSet),trainingLabels,numIter = 50)\n",
    "    errorCount = 0; numTestVec = 0\n",
    "    for line in frTest.readlines():\n",
    "        numTestVec += 1\n",
    "        currLine = line.strip().split(\"\\t\")\n",
    "        lineArr = []\n",
    "        for i in range(21):\n",
    "            lineArr.append(float(currLine[i]))\n",
    "        if int(classifyVector(np.array(lineArr),trainWeights)) != int(currLine[21]):\n",
    "            errorCount += 1\n",
    "    errorRate = errorCount/numTestVec\n",
    "    logging.info(\"the error rate of this test is: %f\" , errorRate)\n",
    "    return errorRate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def multiTest():\n",
    "    numTests = 5; errorSum=0\n",
    "    for k in range(numTests):\n",
    "        errorSum +=colicTest()\n",
    "    logging.info(\"after %d iterations the average error rate is: %f\", numTests, errorSum/numTests)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2018-06-24 08:53:03 - [line:1] - INFO: \n"
     ]
    }
   ],
   "source": [
    "logging.info(\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:53:14 - [line:24] - INFO: the error rate of this test is: 0.447761\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:53:25 - [line:24] - INFO: the error rate of this test is: 0.537313\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:53:35 - [line:24] - INFO: the error rate of this test is: 0.343284\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:53:46 - [line:24] - INFO: the error rate of this test is: 0.417910\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:53:57 - [line:24] - INFO: the error rate of this test is: 0.343284\n",
      "2018-06-24 08:53:57 - [line:5] - INFO: after 5 iterations the average error rate is: 0.417910\n"
     ]
    }
   ],
   "source": [
    "multiTest()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### by pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_df = pd.read_table(\"./Data/CH5/horseColicTraining.txt\",header=None)\n",
    "test_df = pd.read_table(\"./Data/CH5/horseColicTest.txt\",header=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>38.5</td>\n",
       "      <td>66.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>8.4</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>39.2</td>\n",
       "      <td>88.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    0    1     2     3     4    5    6    7    8    9  ...    12   13   14  \\\n",
       "0  2.0  1.0  38.5  66.0  28.0  3.0  3.0  0.0  2.0  5.0 ...   0.0  0.0  0.0   \n",
       "1  1.0  1.0  39.2  88.0  20.0  0.0  0.0  4.0  1.0  3.0 ...   0.0  0.0  0.0   \n",
       "\n",
       "    15   16    17    18   19   20   21  \n",
       "0  3.0  5.0  45.0   8.4  0.0  0.0  0.0  \n",
       "1  4.0  2.0  50.0  85.0  2.0  2.0  0.0  \n",
       "\n",
       "[2 rows x 22 columns]"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>38.5</td>\n",
       "      <td>54</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5.9</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>42.0</td>\n",
       "      <td>6.3</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>37.6</td>\n",
       "      <td>48</td>\n",
       "      <td>36</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>44.0</td>\n",
       "      <td>6.3</td>\n",
       "      <td>1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0   1     2   3   4   5   6   7   8   9  ...  12  13   14  15  16    17  \\\n",
       "0   2   1  38.5  54  20   0   1   2   2   3 ...   2   2  5.9   0   2  42.0   \n",
       "1   2   1  37.6  48  36   0   0   1   1   0 ...   0   0  0.0   0   0  44.0   \n",
       "\n",
       "    18  19   20  21  \n",
       "0  6.3   0  0.0   1  \n",
       "1  6.3   1  5.0   1  \n",
       "\n",
       "[2 rows x 22 columns]"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def colicTest_df():\n",
    "    train_df = pd.read_table(\"./Data/CH5/horseColicTraining.txt\",header=None)\n",
    "    test_df = pd.read_table(\"./Data/CH5/horseColicTest.txt\",header=None)\n",
    "    trainingSet = train_df[list(range(21))].values\n",
    "    trainingLabels = train_df[[21]].values\n",
    "    trainWeights, results = stocGradAscent1(trainingSet,trainingLabels,numIter = 50)\n",
    "    errorCount = 0; numTestVec = 0\n",
    "    testSet = test_df[list(range(21))].values\n",
    "    testLabels = test_df[[21]].values.reshape(-1)\n",
    "    errorRate = sum(sigmoid(sum((testSet*trainWeights).T)).astype(\"int32\") != testLabels)/len(testLabels)\n",
    "    logging.info(\"the error rate of this test is: %f\" , errorRate)\n",
    "    return errorRate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100% (50 of 50) |########################| Elapsed Time: 0:00:11 Time:  0:00:11\n",
      "2018-06-24 08:54:08 - [line:11] - INFO: the error rate of this test is: 0.447761\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.44776119402985076"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "colicTest_df()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def multiTest_df():\n",
    "    numTests = 5; errorSum=0\n",
    "    for k in range(numTests):\n",
    "        errorSum +=colicTest_df()\n",
    "    logging.info(\"after %d iterations the average error rate is: %f\", numTests, errorSum/numTests)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:54:19 - [line:11] - INFO: the error rate of this test is: 0.492537\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:54:29 - [line:11] - INFO: the error rate of this test is: 0.402985\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:11 Time:  0:00:11\n",
      "2018-06-24 08:54:41 - [line:11] - INFO: the error rate of this test is: 0.492537\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:10 Time:  0:00:10\n",
      "2018-06-24 08:54:51 - [line:11] - INFO: the error rate of this test is: 0.552239\n",
      "100% (50 of 50) |########################| Elapsed Time: 0:00:11 Time:  0:00:11\n",
      "2018-06-24 08:55:03 - [line:11] - INFO: the error rate of this test is: 0.567164\n",
      "2018-06-24 08:55:03 - [line:5] - INFO: after 5 iterations the average error rate is: 0.501493\n"
     ]
    }
   ],
   "source": [
    "multiTest_df()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sklearn"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def loadDataSet():\n",
    "    dataMat = [];labelMat = []\n",
    "    with open(\"./Data/testSet.txt\") as fr:\n",
    "        for line in fr.readlines():\n",
    "            lineArr = line.strip().split()\n",
    "            dataMat.append([1, float(lineArr[0]), float(lineArr[1])])\n",
    "            labelMat.append(int(lineArr[2]))\n",
    "    return dataMat, labelMat    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dataMat, labelMat = loadDataSet()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = np.array(dataMat)\n",
    "y = np.array(labelMat)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### GD"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "lr = LogisticRegression()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='ovr', n_jobs=1,\n",
       "          penalty='l2', random_state=None, solver='liblinear', tol=0.0001,\n",
       "          verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 2.45317293,  0.51690909, -0.71377635]])"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2.45317293])"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr.intercept_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Decision Boundary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 4.662541,  7.067335,  6.866805,  6.920334,  3.16702 ,  5.749003,\n",
       "        1.589505,  6.151823,  2.690988,  0.444165,  5.754399,  3.230619,\n",
       "       -0.55754 , -1.67873 ,  2.672471, -1.618087,  3.523953,  3.989286,\n",
       "        8.294984, -2.406051,  4.951851,  3.297303,  6.105155,  0.548467,\n",
       "        2.676045, -2.166125,  5.921938,  5.027866,  6.835839,  7.718645,\n",
       "        4.7473  ,  4.119083,  1.960539, -1.466018,  6.476801,  0.468339,\n",
       "        2.874846, -0.027255,  2.171228,  5.019981,  4.375691, -3.181888,\n",
       "        8.401766,  1.688274, -1.733869,  3.849075,  5.468166,  0.059736,\n",
       "       -0.7153  ,  0.922311,  1.22053 ,  2.556666,  0.143632])"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# for line\n",
    "clf_x = np.arange(-4,4.1,0.1)\n",
    "clf_y = (-lr.coef_[0][0]-lr.coef_[0][1]*clf_x - lr.intercept_)/lr.coef_[0][2]\n",
    "Class1 = X[[np.where(y==1)],1:][0][0]\n",
    "Class0 = X[[np.where(y==0)],1:][0][0]\n",
    "Class1[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#  np.c_ np.r_ 矩阵拼接\n",
    "# np.c_[np.ones(xx.ravel().shape),xx.ravel(), yy.ravel()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# for decision boundary\n",
    "xx, yy = np.meshgrid(np.arange(-4, 4.1, 0.1), np.arange(-5, 20, 0.1))\n",
    "Z = lr.predict(np.c_[np.ones(xx.ravel().shape),xx.ravel(), yy.ravel()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAEzCAYAAAAlyyi6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28702bbd780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(6,5))\n",
    "\n",
    "Z = Z.reshape(xx.shape)\n",
    "# for line\n",
    "# plt.contour(xx,yy,Z)\n",
    "plt.plot(clf_x,clf_y,c=\"r\")\n",
    "# for mesh\n",
    "plt.pcolormesh(xx, yy, Z, cmap=plt.cm.Vega20c_r, alpha=0.6)\n",
    "# for data\n",
    "plt.scatter(x=Class0[:,0],y=Class0[:,1])\n",
    "plt.scatter(x=Class1[:,0],y=Class1[:,1],marker=\"s\")\n",
    "plt.ylim(-5,20)\n",
    "plt.xlim(-4,4)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SGD"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import SGDClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# SGD_clf = SGDClassifier()\n",
    "SGD_clf = SGDClassifier(max_iter=50)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.0001, average=False, class_weight=None, epsilon=0.1,\n",
       "       eta0=0.0, fit_intercept=True, l1_ratio=0.15,\n",
       "       learning_rate='optimal', loss='hinge', max_iter=50, n_iter=None,\n",
       "       n_jobs=1, penalty='l2', power_t=0.5, random_state=None,\n",
       "       shuffle=True, tol=None, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SGD_clf.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 85.01416903,  31.98983997, -51.43373729]])"
      ]
     },
     "execution_count": 138,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SGD_clf.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([345.08727735])"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SGD_clf.intercept_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Decision Boundary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 4.662541,  7.067335,  6.866805,  6.920334,  3.16702 ,  5.749003,\n",
       "        1.589505,  6.151823,  2.690988,  0.444165,  5.754399,  3.230619,\n",
       "       -0.55754 , -1.67873 ,  2.672471, -1.618087,  3.523953,  3.989286,\n",
       "        8.294984, -2.406051,  4.951851,  3.297303,  6.105155,  0.548467,\n",
       "        2.676045, -2.166125,  5.921938,  5.027866,  6.835839,  7.718645,\n",
       "        4.7473  ,  4.119083,  1.960539, -1.466018,  6.476801,  0.468339,\n",
       "        2.874846, -0.027255,  2.171228,  5.019981,  4.375691, -3.181888,\n",
       "        8.401766,  1.688274, -1.733869,  3.849075,  5.468166,  0.059736,\n",
       "       -0.7153  ,  0.922311,  1.22053 ,  2.556666,  0.143632])"
      ]
     },
     "execution_count": 140,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# for line\n",
    "clf_x = np.arange(-4,4.1,0.1)\n",
    "clf_y = (-SGD_clf.coef_[0][0]-SGD_clf.coef_[0][1]*clf_x - SGD_clf.intercept_)/SGD_clf.coef_[0][2]\n",
    "Class1 = X[[np.where(y==1)],1:][0][0]\n",
    "Class0 = X[[np.where(y==0)],1:][0][0]\n",
    "Class1[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# for decision boundary\n",
    "xx, yy = np.meshgrid(np.arange(-4, 4.1, 0.1), np.arange(-5, 20, 0.1))\n",
    "Z = SGD_clf.predict(np.c_[np.ones(xx.ravel().shape),xx.ravel(), yy.ravel()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAEzCAYAAAAlyyi6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3XlwXNd1JvDvoLuxNrhgbYAA2CAJ\nbZYpRaS1RJ6hZJmx7HjkpSpxNCMlcqwwNRnXOFVxJk5clWQqyYxrLDueqmwjy0smSpR4KpHichJb\njKIlckWyRFqUSAIEesHSIAUQG4l9vfNHL+jlvV5f91v6+1WxyNd8uO+gGzj9+i7nilIKRETkHDVm\nB0BERMZiYicichgmdiIih2FiJyJyGCZ2IiKHYWInInKYvBO7iPSKyIsiMigiF0Tkc7HHW0TktIiM\nxP7eX75wiYgoF8l3HruIdAHoUkqdFZFmAGcAfBzAYwDmlFJfEpEvANivlPqNcgVMRETZ5X3HrpS6\nopQ6G/v3IoBBAAcAfAzAn8dO+3NEkz0REZkk7zv2lC8S8QN4BcCtAMaVUvuS/m9eKcXuGCIik7gL\n/QIR8QL4WwC/qpS6LiL5ft0pAKcAoKmp6dhNN91U6KWJiKramTNnZpRS7bnOK+iOXUQ8AL4H4AdK\nqa/GHrsE4D6l1JVYP/xLSqkbs7Vz2223qeeffz5xvLq6ioaGBsOOy9FmJa5hlzYZt/3bZNz2axMA\nfD7fGaXUceSQ9x27RG/NvwFgMJ7UY74L4BcAfCn299/naksphdXV1cTxxsZGyv+XelyONitxDbu0\nybjt3ybjtl+bhSikK+ZeAI8CeEdE3oo99luIJvTviMhnAIwD+JmioyEiopLlndiVUq8C0OtQf8CY\ncIiIqFRceUpE5DBM7EREDsPETkTkMEzsREQOw8ROROQwTOxERA7DxE5E5DAF14oxgohkLJU1+rgc\nbVbiGnZpk3Hbv03Gbb8282VKYmdJAXu3ybjt3ybjtl+bhWBXDBGRwzCxExE5DBM7EZHDMLETETkM\nEzsRkcMwsRMROQwTOxGRwzCxExE5DBM7EZHDMLETETkMa8VY7Bp2aZNx279Nxm2/NvPFWjEWuoZd\n2mTc9m+TcduvzUKwK4aIyGGY2ImIHIaJnYjIYZjYiYgchomdiMhhmNiJiByGiZ2IyGGY2ImIHIaJ\nnYjIYZjYiYgchomdiMhhWATMYtewS5uM2/5tMm77tZkvFgGz0DXs0ibjtn+bjNt+bRaCXTFERA7D\nxE5E5DBM7EREDsPETkTkMEzsREQOw8ROROQwTOxERA7DxE5E5DBM7EREDsPETkTkMKwVY7Fr2KVN\nxm3/Nhm3/drMF2vFWOgadmmTcdu/TcZtvzYLkXdXjIh8U0SmReR80mO/KyKTIvJW7M9Hio6EiIgM\nUUgf+7cBPKjx+B8qpW6P/flHY8IiIqJi5Z3YlVKvAJgrYyxERGQAI/rYPysiPw/gTQC/ppSa1zpJ\nRE4BOAUA3d3dBlyWyHwvBK/j22fmML20iQ6vB4/f1YETBxvNDouqXKnTHf8UwGEAtwO4AuAreicq\npZ5USh1XSh1vbW0t8bJE5js9soCvvTqNqaVNKABTS5t44uXLeCF43ezQqMqVlNiVUlNKqW2l1A6A\nrwO405iwiKzvqdensb6tUh5b31L41puzJkVEFFVSYheRrqTDTwA4r3cukdNML21qPn51eavCkRCl\nyruPXUSeAXAfgDYRiQD4HQD3icjtABSAUQC/XIYYiSypw+vBlEZyb28yZXkIUULeP4FKqYc1Hv6G\ngbEQ2crjd3XgiZcup3TH1LkFnz7OMSQyF28tiIp0cmAftra2OCuGLIeJnagEDxzegw/f3JbyWClL\nwYmMwCJgFruGXdp0WtzfH5zFn/5wElOLG+hsrsV/vvcAHry51TJxl6NNxm2/NvPFImAWuoZd2nRa\n3KdHFvDEy5exvhXtK393cQP/459HsbG5kdGt4vTnwupt2jVuo9rIF+uxU9V76vXpRFKPW99SeOr1\naZMiIioN+9ipKp0eWcBTr09jOrZqVIvePPVyY5kCKhXv2KnqxLteprIkdSA6T73SWKaAjMDETlVH\nq+slXZ1b8PhdHRWKaBfLFJAR2BVDVSdbF4sAie6PkwP7Kj51kWUKyAhM7FR19EoBdHo9+Iuf9aO2\nttaEqKJYpoCMwK4YqjqP39WBOrekPGZW10u6x+/qQJ0rMzaWKaBC8DaAqs7JgX0AkJgVY2bXi1Zs\n5SpTkDwTqL3JjV+6uzPxXJCzMLFTVTo5sC+RyM3setFSjjIFLwSv42s/3B00nl7ewhMvXwYAJncH\nYlcMURX41puzXIRVRVgrxmLXsEubjNteberNqple2kw5z2pxm3UNq7aZL9aKsdA17NIm47Zfm+1N\nbkxrJPcOryfxu2jFuO36fJtdK4Z97EQmqPRA5qePt6b0sQOlzwRKL33w2LGWjLEBMgcTO1GFmTGQ\n+cDhPXC73Xjq9Wi5ghpJ7WMv9Lrx0gfxVbJTS5v42qvTcLvdHIy1AA6eElWYWQOZJwf2Jebw78Qu\nH69Fc3pkoaC2NEsfbHMw1ip4x05UouRulXy6JLINZJZbthLFJw76825HL1azKmJSKt6xE5XgheD1\nlEqR8S6JbHfAeuUBKlFN0qiErBerGRUxKRMTO1EJNLtVcnRJfPp4q2klDYxKyJqlD1zWKMtA7Iqh\nGC43L47WFEIAmoW84pIHMiv9fD9+V0fKNoBAcW8qWqUPHjvWwp8Zi2BiJy43L0GNIDEQmf54NvGS\nBgAqWtYgfs2vvzaFq8tbJdXJSS99YHadHdrFxE5ZZ2kwsWenldTjj3/q6WFLfgI6ObAPJw42Wq5G\nDhmHfexk6iwNu+vIUic9PqAa/wRU6JRComIxsZOpszTydXpkAY/8TRj3/9kFfOrpYcskSa2BUC0s\nuEWVxCJgFruGGW3+0t2d+MrLV7C2tZN4rN5dg195f0/iPDPj/v7gbEp8U0ub+MrLV4ATXXjo6F5D\nrlHs8YdvbkNjQyP+9IeTmFrcQGdzLd5d1O5rTi64ZYWfCyv87Fn1GlZtM18sAmaha5jV5omDjcCJ\nroxZGif8jVhdXTU97j95NZLypgMAa1s7+PprUyn91mY93yf8jTjhH0g89rN/cSlrwa1KxPlPgzOJ\nGSvNddEP5ovrO1k37rDDz7NVfmdYBIxsodyzNOIFo+J1SnZUdI/R+IyMbOy2wXM5Cm4VIr2Oy/X1\n3TfFeAmBrXs7WLDLwdjHTmUXTzTxud2F1inR6+u36gbPDxzeg8+f6Ean1wNBdID18ye6KzYrRquO\nS7L1LYVvvTlbkVjIHNb8zSBbOT2ykDEvOvmjfrZEk0+dEr1FNVbe4NmseepAfrOZrPpph4zBO3Yq\nyemRBTzx8mVML28laqU88fJlvBC8njgnV6LJ9f8nB/bh8ye60dHkhiDahfP5E9144PAeA74D58ln\nNpNVP+2QMfjqUkn0qgV+683ZRB9uh9eTdYl9PolIa1ENVzpqe/yuDjzx0mXdT0lW/7RDpWNip5Lk\nM7CZLdFUclCxEnJ1S1VCeh2XfGfFkHMwsVNJ9O7Gkz/qJycavVkxydPzzE4+xRZEi3dLxT/BTC1t\n4g9emMQfIP8ZQEbRquPCTzvVg4mdSpLvwGa2RKO1zZpZU/JKKYim1S0VF/+eAPBumcqOg6dUEiMG\nNjW3WTNpSl4p29blGgRmWQGqFN6xU8lKHdg0cgFSqXXlSymIlmuQOFs7rIdPupRCzewsPIFA3l/C\nWjEWu4Zd2jTyGnr1VTq8noLaPD2ykFJTZnp5C195+QpqPbV48ObWvOLWS86dzbU567z8yvt78D//\neSyj/EF6O7W1tQXHnet7N+LYLm3aNe6c5yiFprk5uEdG4B4eBgYHURcKwT0ygpq5uYy2smGtGAtd\nwy5txo+z3WUW0uZn7mzPmDVT5xY8dqyloJ+Tr782pVlT5k9ejeCEvzGv5+axYy2a5QA+c2d7zjov\nJ/yN2DjRha+/NqVZKybeTqFx5/O9W/HnpBxt2jXulMe2tuCemIBnaAi1o6PwBALwBAJoCwTgWllJ\nnL+5dy8W/UewdOKDWPYfxpL/EPC5X8poVwu7Yqgo6TNAStl1SWubtWJmxRTbjZL+BvXgDXvx2vhy\nUd0iyd1Sye1m26mI9fAdan0dnnAYnkAAtYEAtgcvwRMKwTsWRs3m7mu71t6Bxf4jOHv3RxDq6EfE\n58dklx8XNuvR01LcQDsTOxVFb2FSsbsupc+aAaJ3OIX0PTfX1aQUvIrLtgBKaxbM94evJWq7lFIO\nIF5WIFcb7U1u3WqQZH2ytIS6oSE0jo2hNnb3XTM8grrIBGQn+vOoRHC1tQshXz8iJ96HyS4/zjZ2\nYeOGG7HS4AUAROZX0bM/qWtmflXrcnlhYqei6N1NGnmXWcjUw9MjC1jeyEzq7prsC6CssC2g2dUg\nixWv2Jm8mbWTK0bWzM8nuk22Lg7BFQjBGw6iYepK4pwdtwfLvQdxoa0f4fd8ABGfH5Guflzu6EVo\nWaUk7sj8KnqKrLeeCxM7FUVvkNHIu8xCku5Tr09DawV9o0eyJmgrdIM8cHgP3G63rWbFaK09+Nqr\n03C73ZaOOyel4JqaQuPQEBpi/d+1gQBcIwF45nan36576hDx+XGx7ygidz2Ec95urAzciKm2A9h2\nuTPvvgFgufg78ELlndhF5JsAPgpgWil1a+yxFgB/A8APYBTAzyql5o0Pk6xGb2GSkXeZhSRdvUS8\nqNE1k8wq3SBmVoMshubag20bbYC+vQ13JBLtNrkwiJqRELyjITSFg/AsLyVO22zegyX/Ibx98z0I\nd0b7vye6+vEW9uBAa1PiPM1EbqJC7ti/DeCPAPzfpMe+AOAFpdSXROQLsePfMC48sqr4L2857zIL\nSbrFfoKwazeI2SrRFWeIjQ2snx/C8nAYTaMheMMBtI0G0TA2Ctf6euK0+T2tmPD5MXH8Q4j4/Hir\nqQtrAzdhYU8LIJKRuFUJ/d+VkHdiV0q9IiL+tIc/BuC+2L//HMBLYGKvGuW+yywk6WoVGqtz5U7Q\n5e4GsUJRsHKoRFdcIWRlBZ5gMNp1cnEodgceRGNkHDXb24nzplq7cK61FzPv/2S0/9vnx48afNjb\nnfpzEplfRc9e69yBF6rUPvZOpdQVAFBKXRER3uaQYQpJulpTJh871pJXgs73DarQJK1VFMwp29IV\n+0ZaqsXpOWxcCsMbCsI7GoQ3HIRvNIj6K5cT52zVuPBuew/O+/yYuPkE3m7uxurhGzDZ2Yf1uoaM\nu+/F+VXs1bqYjVVs8FRETgE4BQDd3d2VuizZXCGfCrQKjRmlmCSdT616uyrljTQnpeC6ejW6hP6d\ni5BAGN5wNIn752YSp617ajHZeRBn22/C7PGPJu7A33S3wte+W6vIav3flVBqYp8Ska7Y3XoXAN0K\nR0qpJwE8CQBHjx7V35CRyIKKSdJmbcKdzyeL9KmKxXQRlfxGurODleA4rg2H4Q0H4A2H0DkWQGMo\nBM/i7g5cy/VNmPT58daNd+H83h4s9Q8g0uXH1RYfdmpcGYl7y+L935VQamL/LoBfAPCl2N9/X3JE\nVDStX2hbzFCwAd3BwuUtfOrpYc3nWq8fWiT6WpXjtcnnk0UlyyQvrG1jbnYJjZEJeMNBtIyNwBMK\nwzsaRNNoGP613SS80LwfgbY+XL39AUR8/Yh0HcQb9T7UHeyNPmmozrvvYhQy3fEZRAdK20QkAuB3\nEE3o3xGRzwAYB/AzebZly8JAligUpHP8/cHZlEJSU0ubiUJSJ/yNVfVclOMaeoXKgNTnOrnYmF5R\nsB2FnMXJio3zGz8a0fxk8e0zc/jkHb0AED1Ho0xy8jmFXDNO1tbgDQbhHhmBa3gY7uFhNJwbxLF3\nJ+Da2n2Du9behZnewxj60HGcbepC423vxUzvIazu2Y+z4wu4oy/6hucGUJd0HJd+rPVYqcdWbfPZ\njBa1FTIr5mGd/3og3zaS2mIRMIPb/JNXI7qFpO7p9ls27kq2Wco1PnNne8a8/WTpxcbipRD0Kj1m\nK05WSpxTOm8+00ubWF1dxemRBd03qPg5ua4h16+jNhhE/dAQ6sPhxGwU98QEREWfH1VTg+WePkRa\ne/HqLfci0hWdA/5GXQdafLsVKyPzq+hpbgAWACwsIJLWjZLrOJ9znNJmIbjy1CFsM6/YpuLdJnqV\nG4Hd5zq9FIKecrw22bYqjHfT6EnezhBKwTUzg/rx8egy+pERqEvDqAuFUH91dyhtu7YWy31+LNx4\nK16/7SQCnf2Y9Plxub0XW57ajK6TlflVtBjzrVIWTOwOYbV5xU4Ur9z46HdGsz7XWqUQtJTjtcm2\nVaHm1n1KoXvxKm6Zn8CnmxfQ+oVvJe7AXQsLidO2Ghsx3nEQo4ePY/Le2OyTBh/k8CHs1LgAsP/b\nSpjYHULvF/ruviY88jdh2yyQMWsnIb0Su1pylVPIZ9ZLuVa3Jn+ySHnND9Ti/z17FidnJ3BkdgJH\nZsZxZDaCI7MTaNpcS3z9xt59WOo/gqX7P4RXan0IdhxCxOfH7L52RBbWMotYxZI6WQsTu0No/ULf\n3deE7w9fs80CGSNrvJdy3VwbT6eXU0h/I9ArhVAjgFIo6xuWrK/jI1tXcH/dINR4CN43g/B+O4im\n8VH8S1IN8CveVgRae/GdoycRbu/Dzk234I0GH5p6d9eY8A7cvpjYHSR979FPPT1sqwUyRtd4N+K6\nJw76Nb8mW611vVIIRtR4j5OlJdROTCSqD+4MXkJtKITGyd0a4DsimGrrxlCnH0N334FIlx8/qu9A\noKUHi3XRAlYiQF9LE1q99ZifX0VTtouSbTCxO5hZC2SKZdYAsNHXNbL+TM3cXHTg8vwgEAijKbYC\ns2H63cQ5O24PLnf0YLDzECbe+wAivugmDttHjmCjth7A7t333NIa1hZWge0deFw16N7XgFZvfVHf\nJ1kXE7uDZZshYUVmDQDnc91C+uCBAgukKQXXlSvwjIzAEwxi++IQ3MHoJg61C7tVsFfrGjDZeRAX\n/Lfj4m0HcD22AvPdtgMYv76Z2f9dm5mwW731aPXWs5vF4az5G06GyDZDwooqUeO9mOsW2geva3sb\n7okJbLx9EevDITTFilh5x0JwLy8nTltqbMZEVz/OvOffYcLnx6TPjzcafPD4D0LV1ADQ6v/mtFba\nxcTuYLozJCw6K6YSNd7zuW76xtOF9sFfW1zF2shwovrgvrEA6oMhNI2H4Upa9DO7tw1jPj8uvfck\n5v0DiSJWF7YyNzF+d34VPbGkTpQLE7vDpQ+oAsZWPTSaWTsJZRsM1d2daW4RG2ffxmqsiFVTOFoD\nvHdiHK6daA3wHRG8u68TI92HMPHvjyHi8+PHTV1YH7gRK43NAIzdxJjy9/T0x9GoYs/11O7jK9KA\n+zzPmBOUQUxJ7KwVY/82qynuAdcamsdCODI7gYGk+d8915NWYLrcmOs+iIneIzh/8wm4b30PZnoP\nYfZAP/7r+Q+jW53HMZzfbXQRWF9uxJePvQjAnNoldqmPUq5rNE5pv4E2qugbrVXiTGZ4rRgjsVaM\nvdt0ZNxKwTU9DdfQENyxTYw9gQD2j4zg+dndTYzX3LUItvTgbO8tGH7vJ3B+Xx9G2g/h3Y4ebLui\nv04pd+DT67t3hWnqdlZwdjx3fRStx6xYy8SucWuxS5x62BVD1WVnJ7GJsScQ2J2BMhqCZ2kxcdqm\ntxlL/kN458a7Ee7sx4V9B/BGQydGvW1wu92JaYKcXUJWxMROjpBei/7UHfvxAK6icWwskcTl0jDq\nR0fhWt9dQr+wpwUTnX5E7jgZ3cTY2421gRsxv6c1YxPj/QCWmcjJBpjYydZkdRVnXjqHSz84g0en\nJzAwO44jMxM4uHAFnp3dTYxXuroRaOnD6L0fj5WQPYiIz4+hjdrM+d823sSYCGBiJ5uQ69dRPzSE\nxqT+75rhEdRdnsRBpfBJAFtSg7H9XQi09uIHN9yDYFsfam69FZOdBzU3MQYAbHAGSrVakQbN8Y8V\nsf8bOxM7WYdSqJmZQW0scW9dHIIrEJ1CWD9zNXFatAZ4Py523YjwTzyIf3W1IdDag9H9B7DhTl2l\nekef+YuxltGAJjgzgdjZIx3PAdApdmbzKadM7FR5OztYCY3j2vAovKEAvKPRTYwbQiHUXr+WOG21\nvhETPj/ePvI+RN7fj7e8XVg7cgOm27pTNjF+JzKPze3MnYpqYvtkmu3+2mcyu3ts2k+fMvcbSMz/\ndsLcbydhYqfy2dqCe2wMtYEAXBcGUROILeRJ28T4mnc/gm29uHr0fkx0RXfgifj8OKe8KSsw9RJi\n974GjM8tQyUtDhUBmuqc+eM9u7SGmcVVTF9fSRTyqhS9qZt6j5M5nPmTTxUla2tYf2cQK8NheMNB\nNIWDaB8LomFsDDVJmxhf3d+JkM+PyXv+A97ZcwBLh29AxOfHondfSR+H49UJJ+ZWsKNUItmtbube\nxchuZpfWUt7ENrd3MD63DG+dB4A9PwWQ8ZjYKW+yuAhPIICd84PYGY72fTeNxmqAxzLNttRgqv0A\n3mzpxewH7knUP/lRfWfmJsYGdke0euuxuqkyujyc5vLCasonEyC6ecfyujVLMZM5mNgpw9KVGWwM\nhRJFrNrHgqgPhVJqgG+63Ljc2YcR3wAiR0/iXHM3Vo7cgCsdvdj01HET4zLRGksAgJ30bE9VjYm9\nCi2sbWN2YQ310+/CGw6iZTyAmmAI3lgRK39aDfDR1l5M+W9H5O7YJsZNXcChQ9hx7f742HlA0E48\nrhpLDxSTNbAImMWuYXib29twh8PwTkzANTwM9/Aw3Gcv4rbLYdStriROW2nei9newxi964N4y9uF\n+qO3Yqb3MK63+XA2ch139O1DYHoRb47OYWl9C96ryzjub8GRjuZEG1YsmuSUwlfxx/bUu/CvI1ex\nvbN7h+6qEfTsb6xI3OtXG1G3s4J06zWNmoWzionDSs+31dpkEbBqK3y1vg5POIzaoSHUjo2hdmQk\nupAnHIYkfc1aeweutfbh3F0/jYjPj4nYJg5NPV3RqSRIuvteAxC5jsj8asag3dL6Fl4enkZ4ZjlR\nMyWZFYomVfIahUwDLDXOnv2NmgPFZ8cXDLuG3vHD7c+mHOca07DTa2j1NgvBrhibSd7E2BMMQg0O\noTaYuomxEsFqdw8W/Iew9L73419cnQh09CPi82OlwZvxCzk/v4qmHB/l9QbtLi+scs9MGDsN8MWN\nh9GkUVJ2ZboBj3Q8VzUDxVQ8JvYK6Xv6HtRsZX6E3XE3YvyRf8t4vGZ+Hp5AAA2XLqE+HE6sxnRf\nvpz0tR5caT+AId8hTNz6AUR8fpxp7ML2wEBiE2PAmP5vvUE7vcftTm8ThmU04FE8V9Zra61SBThX\nnPLHxF4hWkkdSqFmbgn1r76a2MgYl4ZRFwyibn4ucdp2XT2W+g9h7ugx/Nv7PopgZ/Tu+932Hozl\nuYlxqfQG7TwuZ27XppdE9ZIukZUwsVfC9jYwvwNc3Qau7gAzO7G/t4F1wIdHAQCbzXsw2t6HsZt+\nMlaB0I83G3xw9ftN38RYb3VnJVc9ElF+mNiNtLEBT1L1wZ2hYXgCQTSNh4H19d3zvAK01QBHPUCb\nC799w5cR8fVjYU8LIgtrKYn7skU2MY73o19eWMXm9k5i0M7I/vXZpbWU9us9LnA1JVHhmNiLICsr\n8IRC2H77ArZGwvCOhqI1UCLjkO3dGuBXW7sw4vNj8v0/gYcanwXaa4A2F9CQOlB5vvN4pb+ForR6\n68u2a9DqxiauLm6mLJXf3N7B7NKaLQZnnVwC1ioS4x5TqY9XYtzDbpjYs6i5di1x9715YRCuYLQW\nSuOVycQ52zUuXGnvwXmfH4P+n8S1/gFEfH5MdvYhuIJEAnxo6h/M+jZsYXl9K2PWDWCfWTfxErCA\nRndZgTNWWOZXG8c98sfErhRcMzPYPncBy7H6J95QEM1jQdTOzCRO2/DUYrLzIIYO3ILJYx+JzUDx\nYefIALZiNcAzfqFXdn/geEeXnd6SeLNm3ei9XssV6BpKL/MLcGUvFaZqEvvCyiZWBifQFOs22T8W\nRF0oBG84CM/i9cR5y/VNiHT144d9xzH/7w4j4utHpMuPH8tedLd6U9qMzK+iJ21jBz2PdDynuaCD\nv6xRNSKayd2sWTd6mzBE5lfRY0pERPlzXmLf3IRnfBybb1/E+kg0cXtHg+gKhVC3sbuJ8VzTPkx2\n9yNy+wcR8fnxY29XdBPjvW0ZmxgDwA4XgJRVU50bS+ubGd0x5Z51o7kYaCp6x57cvWIlen3N8QVM\nRPatFQPAHQzCPTICdfEi6kMhuEdG4AqHIZu7UwCvtXdhpucQBu99CPKe92Cm9zBmeg/hhwuSUofh\nyvgC7ujbh/6ka9i1nkSxbSbXgllZ30ipBVOJuPfUu3Zr0dS50eqtw8lbfIZeI/1Ya4UnEO3PjZ9r\ntdewMY+YrRh3ycdpb2RluYbF23RMrRi5fh01Q0NwjY0lBjL3Do+gbjKSqAGuamqwcqAX1/oPY+me\nE1jyH8azm20YbvNjrb4JQNpH6gVr1H6wUpvZasGkb1hRrrh79jfgBt/elMdKqX9Sau2Ns+MLKV+T\n751yJV7DbDEbdQ2r/TxnG/cw8ufE6LiNbCNf1uiKiW1i7AkEUBsIoGZ4GPWhUHQJ/dTub1F0E2M/\nBn0DCN/2IUR8BxHp6scbnjZ0tu9NaTIyv4qeevZf5ytbLZj9TdaflVIJ3BbOXFrjVADHPbSY0xWz\nuIg93/hG4g7cFQjAs7D7jrvd0IDF/iOYPXY3lvuP4J9UO4Y7DmVsYhy3yf7vklm9Fgz7laOenv64\n2SGQDZiS2D2jo2j5/d/Hxt59WOo/gnPvPYFwpx8Tvugy+nNozmsTYzKO1WvB8G45qtq+XyqOKYl9\nsqMPj/3an+F6834ApS/ooNJlqwXjxE2h45y0GMiOMVN5mJLY1+obE0mdrCFbLRgn1/pOXwxk50+H\n5e6SsuPU0GpljcFTsoRy1oJxgmpfPcw68fbBxE6UJ64eJruwxsgYUQ56d8XVcrccp/f9VqKGDdmH\nIXfsIjIKYBHANoAtpZQ96tCyRqCQAAANFUlEQVSSbfBuOYo1bCgfRnbF3K+Umsl9GhFZRcrerkBi\nncCKNOA+zzPGtF3law/MwD52IocoJkkXsj6g0KmhXHtgHqMSuwLwvIgoAP9HKfVktpMba12GF9ip\nRBEfuxQK4nNh7jXMajNbcbCe/Q3aX5+lsFb61zyK51KOz8YK5wHAHVox5Vm0y67PdzUUAbtXKXVZ\nRDoAnBaRIaXUK8kniMgpAKcAoKHFV/biTka3aZdCQdX6XOTzsd+KcZerzUK/vpivKaVoVbHFyqz6\nfDuyCJhS6nLs72kReRbAnQBeSTvnSQBPAsD+gzc7dymjg6R8tE9KmFbcY5If+4l2lTzdUUSaRKQ5\n/m8APwXgfKntkvm4xySRPRlxx94J4FkRibf3V0qp7xvQLhGVWTlX01b7Sl0zlZzYlVIhALcZEAtV\nEdYdMV4xiTR9gxAji/Fx7YF5ON2RTMG6I8YrZ5Ime2FJAXIElhwg2sU7dtKVbY9Jq+HHfqJdTOyk\nyy51Sew0LZOoEtgVQ7bHaZlEqZjYiYgcxpSuGNaKsX+bJV/DyDoiebaldZz3NRzYJuO2X5uVrhVT\nkJWNbdaKsXGb5ahlkqzYOiLZ2tI73wrPhV1fQ8Zd2TYLwcFToiLpFR7joC2ZjX3sZHtmbRfHQVuy\nKt6xk+3ZZVomUaUwsZMpCt2NRwvrzVhPxi5OAF8TEzCxkynur32m5FWirDeTqpz7l+aLdfGtgX3s\nRA7BpEpxTOxERTJr0JYoF3bFUNXQ7JNH6r6ohdAqPAZw0JbMxzt2qhrsk6dqwTt2i+PMD31GzKyJ\n+7upD+0emPj82n3RE7fDswbWirHYNdIf0+o6AKJ3mfHzquW5SD9+FM+lHJ8dX9h9TrS+PktNmXSF\nPL9GPheNOq93E1ZztrF+tRF1OysZX7te04ie/Q0VeQ2/2vdSynHO18SA42pq09K1Yoiq3a+fuT81\nCRfwpqPny8deTPw7OaECAJJq5pDzsQiYha6h95ies+MLVfVcVLKIEpDf81ts3Fp31vnEU8g1Cjm2\nS5t2jduoNvLFwVOqGpyGSNWCXTFUNdJXuwJpg6ZEDsHEbnFGzvygTHZ5fkv5tKE306bY+ftkfUzs\nFmdETRXSZ8Xn987a5wxd9MRSA9WHiZ2oDHIV5OJ8byonJnaiMsh1l5zcBZLxKaHE2TxEnBVDROQw\nvGMnw1ihHjgRMbGTgThIZ7yUN8ukWS2F1I5hf371YWInsjAjNszWKi9shdk/VD4sAmaxa9ilTa1r\nZKt3kk8RKie9hqUW5EocZ3lO7fJc2PU1tGKbli4Cxlox9m6zmBo3donbqDYfbn825Tj9brnQNrUY\n+TtkRBtOew2t1mYh2BVDlmNEvzJRNeN0RzKM3mBcoYN0RvQrE1Uz3rGTYbjoxnh6M1pYqZKyYWIn\n0mGF4lnx62jOaqlIBGRH7Ioh0sF5+WRXTOxERA7DxE6WozfYyn5lovywj50sh/3KhXtx42E0TaV1\nEU1F3yS5mUb1YWKnnJg0rE9vKijHA6oTEzvlZPekobfgKVfVSRbPIrtirRiLXcOSbeZZr8RycceO\nG9M/bcQ0qtWsdVu+2vcSzo4vpPx//PiOCsRdUJt5vEa2/Nkz6RpWbZO1YlijwtA29MRfRyvGnc/3\nYZe4S32NrBq3Fa9h1TYLwa4YC7PCAhkish9DEruIPAjgfwNwAXhKKfUlI9qtdlwgYw477gS1jAbN\nsRCOB1SnkhO7iLgA/DGAkwAiAN4Qke8qpS6W2jYZr5hPAdWWNOz4hnp/7TPcSIMSjLhjvxNAQCkV\nAgAR+WsAHwPAxG5BxSQtuycNzm6hamNEYj8AYCLpOALgrvSTROQUgFMA0NDiM+CyRPnRW/AEgFUn\nyZGMKCkgGo+pjAeUelIpdVwpdbzOmzkNiIiIjGFEYo8A6E067gFw2YB2q55RG1cQUXUxoivmDQAD\nItIPYBLAzwH4jwa0W/W4u7w52CdPdldyYldKbYnIZwH8ANHpjt9USl0oOTIyVMYUvjRMWru4ExTZ\nnSHz2JVS/wjgH41oi8ojW1K/s/Y5fgpIYsd57ETJWI+dKI0d57ETJWMRMItdo1xtZisSla0QViHX\nsMNzkc81Sn2uKhWnHa5RjjbtGrcRbbAIGIsPZX2s3NewYptGFNeq5ufCCm3aNW6j2sgXu2KIiByG\nib1KcE48UfVg2d4qwSl8+eM8drI7JnaiNHwTJLtjYifTZJsvzo1EiIrHPnYyDeeLE5WHLe7YX9x4\nGE3pGxJP8c6OiEiLLRK71u49AO/s7OzFjYfL0iZvAIjYFUMm0XuzLkebvAGgasPETkTkMPaoFZOl\ndkf8PNaoqGybJV8jy2u6XtOo+7ra4efELm0ybvu16ahaMdmcHV9gjYoKt2nENbJ5uP1ZQON1tcPP\niV3aZNz2a7MQthg8XUaDZv8pVwI6U8r89qS78GU04FFwEJQoF1sk9vtrn+H2cA6T7c1ab7Az14Ar\nbwCIomyR2Ml5sr1Z/93Uhwxvk6iaMLFTTpwfTmQvnO5IOXF+OJG9MLETETkMEztZjt5g5zLYX06U\nD/axk+XE++3TBz8j86voMSsoIhvhHTsRkcMwsVNOel0gnB9OZE3siqGcOD+cyF7sUQSsiONytGmX\nQkFmPhe/fuZ+1O2sRB9IKgewIg34at9Llo27Gttk3PZr01FFwKxQxMcuhYLMfi4SST1No1otuPAb\nX8Pytcm47ddmIdjHTkTkMEzsREQOw8ROROQwTOxERA7DxE6GYjkAIvNxHjsZiuUAiMzHO3YiIodh\nYicichgmdiIih2FiJyJyGNaKsdg17NIm47Z/m4zbfm2yVgxrVJStTcZt/zYZt/3aLAS7YoiIHIaJ\nnYjIYZjYiYgchomdiMhhmNiJiBympMQuIr8rIpMi8lbsz0eMCoyIiIpjxHTHP1RKPWFAO0REZAB2\nxRAROYwRd+yfFZGfB/AmgF9TSs3n+gKuPLV/m4zb/m0ybvu1me/KU1FKZT9B5J8B+DT+64sAXgMw\nA0AB+D0AXUqpX9Rp5xSAU7HDWwGczzNGM7Uh+v1ZHeM0jh1iBBin0ewS541KqeZcJ+VM7PkSET+A\n7ymlbs3j3DeVUscNuXAZMU5j2SFOO8QIME6jOS3OUmfFdCUdfgL2uAsnInK0UvvY/5eI3I5oV8wo\ngF8uOSIiIipJSYldKfVokV/6ZCnXrSDGaSw7xGmHGAHGaTRHxWlYHzsREVkD57ETETmM6YldRD4v\nIkpE2syORYuI/J6IvB0rmfC8iHSbHVM6EfmyiAzF4nxWRDIn0FqAiPyMiFwQkR0RsdwMBBF5UEQu\niUhARL5gdjxaROSbIjItIpaeqCAivSLyoogMxl7zz5kdkxYRqReRH4nIuVic/93smPSIiEtEfiwi\n38t1rqmJXUR6AZwEMG5mHDl8WSl1VCl1O4DvAfhtswPScBrArUqpowCGAfymyfHoOQ/gkwBeMTuQ\ndCLiAvDHAD4M4BYAD4vILeZGpenbAB40O4g8bCG6YPFmAHcD+C8WfT7XAXxAKXUbgNsBPCgid5sc\nk57PARjM50Sz79j/EMB/Q3RWjSUppa4nHTbBgrEqpZ5XSm3FDl8D0GNmPHqUUoNKqUtmx6HjTgAB\npVRIKbUB4K8BfMzkmDIopV4BMGd2HLkopa4opc7G/r2IaEI6YG5UmVTUUuzQE/tjud9xEekB8NMA\nnsrnfNMSu4g8BGBSKXXOrBjyJSJ/ICITAP4TrHnHnuwXAfyT2UHY0AEAE0nHEVgwEdlRbPHiTwB4\n3dxItMW6ON4CMA3gtFLKinF+DdGb4J18Ti7rZtY5yhH8FoCfKuf185UtTqXU3yulvgjgiyLymwA+\nC+B3KhogcscYO+eLiH4E/stKxpYsnzgtSjQes9ydm92IiBfA3wL41bRPv5ahlNoGcHtsbOpZEblV\nKWWZMQwR+SiAaaXUGRG5L5+vKWtiV0p9UOtxEXkvgH4A50QEiHYdnBWRO5VS75YzJi16cWr4KwD/\nABMSe64YReQXAHwUwAPKxDmsBTyXVhMB0Jt03APgskmxOIKIeBBN6n+plPo7s+PJRSm1ICIvITqG\nYZnEDuBeAA/F9ruoB7BHRJ5WSj2i9wWmdMUopd5RSnUopfxKKT+iv1R3mJHUcxGRgaTDhwAMmRWL\nHhF5EMBvAHhIKbVidjw29QaAARHpF5FaAD8H4Lsmx2RbEr1j+waAQaXUV82OR4+ItMdnkYlIA4AP\nwmK/40qp31RK9cRy5c8B+JdsSR0wf/DUDr4kIudF5G1Eu46sOG3rjwA0Azgdm5b5Z2YHpEVEPiEi\nEQD3APgHEfmB2THFxQafPwvgB4gO9H1HKXXB3KgyicgzAP4NwI0iEhGRz5gdk457ATwK4AMW32Gt\nC8CLsd/vNxDtY885ndDquPKUiMhheMdOROQwTOxERA7DxE5E5DBM7EREDsPETkTkMEzsREQOw8RO\nROQwTOxERA7z/wFWFvICrNxakwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x28702bc4cc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(6,5))\n",
    "\n",
    "Z = Z.reshape(xx.shape)\n",
    "# for line\n",
    "# plt.contour(xx,yy,Z)\n",
    "plt.plot(clf_x,clf_y,c=\"r\")\n",
    "# for mesh\n",
    "plt.pcolormesh(xx, yy, Z, cmap=plt.cm.Vega20c_r, alpha=0.6)\n",
    "# for data\n",
    "plt.scatter(x=Class0[:,0],y=Class0[:,1])\n",
    "plt.scatter(x=Class1[:,0],y=Class1[:,1],marker=\"s\")\n",
    "plt.ylim(-5,20)\n",
    "plt.xlim(-4,4)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "以上，\n",
    "- 参数默认，效果不如GD,每次的效果随机，因为默认的max_iter_是5(v0.19)，0.21将默认1000\n",
    "- 调整max_iter就收敛了\n",
    "- 或者调整tol=1e-8一样会收敛\n",
    "- learning_rate可以选optimal，这样会是变系数的类似改进SGD"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EX : Horse Colic "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>38.5</td>\n",
       "      <td>66.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>8.4</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>39.2</td>\n",
       "      <td>88.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>38.3</td>\n",
       "      <td>40.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>33.0</td>\n",
       "      <td>6.7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>39.1</td>\n",
       "      <td>164.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>7.2</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.3</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>37.3</td>\n",
       "      <td>104.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>7.4</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    0    1     2      3     4    5    6    7    8    9  ...    12   13   14  \\\n",
       "0  2.0  1.0  38.5   66.0  28.0  3.0  3.0  0.0  2.0  5.0 ...   0.0  0.0  0.0   \n",
       "1  1.0  1.0  39.2   88.0  20.0  0.0  0.0  4.0  1.0  3.0 ...   0.0  0.0  0.0   \n",
       "2  2.0  1.0  38.3   40.0  24.0  1.0  1.0  3.0  1.0  3.0 ...   0.0  0.0  0.0   \n",
       "3  1.0  9.0  39.1  164.0  84.0  4.0  1.0  6.0  2.0  2.0 ...   1.0  2.0  5.0   \n",
       "4  2.0  1.0  37.3  104.0  35.0  0.0  0.0  6.0  2.0  0.0 ...   0.0  0.0  0.0   \n",
       "\n",
       "    15   16    17    18   19   20   21  \n",
       "0  3.0  5.0  45.0   8.4  0.0  0.0  0.0  \n",
       "1  4.0  2.0  50.0  85.0  2.0  2.0  0.0  \n",
       "2  1.0  1.0  33.0   6.7  0.0  0.0  1.0  \n",
       "3  3.0  0.0  48.0   7.2  3.0  5.3  0.0  \n",
       "4  0.0  0.0  74.0   7.4  0.0  0.0  0.0  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = pd.read_table(\"./Data/CH5/horseColicTraining.txt\",header=None)\n",
    "train_df.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>38.5</td>\n",
       "      <td>54</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>5.9</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>42.0</td>\n",
       "      <td>6.3</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>37.6</td>\n",
       "      <td>48</td>\n",
       "      <td>36</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>44.0</td>\n",
       "      <td>6.3</td>\n",
       "      <td>1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>37.7</td>\n",
       "      <td>44</td>\n",
       "      <td>28</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>45.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>37.0</td>\n",
       "      <td>56</td>\n",
       "      <td>24</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>38.0</td>\n",
       "      <td>42</td>\n",
       "      <td>12</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>37.0</td>\n",
       "      <td>5.8</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0   1     2   3   4   5   6   7   8   9  ...  12  13   14  15  16    17  \\\n",
       "0   2   1  38.5  54  20   0   1   2   2   3 ...   2   2  5.9   0   2  42.0   \n",
       "1   2   1  37.6  48  36   0   0   1   1   0 ...   0   0  0.0   0   0  44.0   \n",
       "2   1   1  37.7  44  28   0   4   3   2   5 ...   1   1  0.0   3   5  45.0   \n",
       "3   1   1  37.0  56  24   3   1   4   2   4 ...   1   1  0.0   0   0  35.0   \n",
       "4   2   1  38.0  42  12   3   0   3   1   1 ...   0   0  0.0   0   2  37.0   \n",
       "\n",
       "     18  19   20  21  \n",
       "0   6.3   0  0.0   1  \n",
       "1   6.3   1  5.0   1  \n",
       "2  70.0   3  2.0   1  \n",
       "3  61.0   3  2.0   0  \n",
       "4   5.8   0  0.0   1  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df = pd.read_table(\"./Data/CH5/horseColicTest.txt\",header=None)\n",
    "test_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train = train_df.values[:,:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_train = train_df.values[:,-1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import SGDClassifier\n",
    "SGD_clf = SGDClassifier(loss = \"log\",max_iter=50,random_state=2018)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.0001, average=False, class_weight=None, epsilon=0.1,\n",
       "       eta0=0.0, fit_intercept=True, l1_ratio=0.15,\n",
       "       learning_rate='optimal', loss='log', max_iter=50, n_iter=None,\n",
       "       n_jobs=1, penalty='l2', power_t=0.5, random_state=2018,\n",
       "       shuffle=True, tol=None, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SGD_clf.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_test = test_df.values[:,:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_test = test_df.values[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = SGD_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "print(\"error rate is %1.2f\" % (100*np.mean(y_test==y_pred)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7164179104477612"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(y_test==y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(67,)"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 381.46407344,   53.87496098,  103.85488041,  -42.75562163,\n",
       "          74.58070964, -167.59856095, -113.06092488, -494.82865602,\n",
       "          -2.12150406, -314.25609958,  340.89460268, -138.74972483,\n",
       "         456.14132845,  -93.39288322, -167.04011532,  232.73237231,\n",
       "        -153.11825359,  -21.64036846,   31.71965708, -195.16459814]])"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SGD_clf.coef_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "以上\n",
    "- 固定随机种子，结果会不变\n",
    "- 不固定，变化范围不小\n",
    "- 20维数据，不好像之前的数据那样可视化决策边界\n",
    "- 参考[sklearn文档](http://scikit-learn.org/stable/modules/sgd.html#sgd)SGD对输入数据敏感"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler,StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sss = StandardScaler()\n",
    "mms = MinMaxScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.00000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "      <td>299.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.397993</td>\n",
       "      <td>1.64214</td>\n",
       "      <td>30.509365</td>\n",
       "      <td>66.220736</td>\n",
       "      <td>24.551839</td>\n",
       "      <td>1.906355</td>\n",
       "      <td>1.555184</td>\n",
       "      <td>2.404682</td>\n",
       "      <td>1.167224</td>\n",
       "      <td>2.404682</td>\n",
       "      <td>...</td>\n",
       "      <td>1.143813</td>\n",
       "      <td>1.023411</td>\n",
       "      <td>0.834448</td>\n",
       "      <td>1.816054</td>\n",
       "      <td>2.230769</td>\n",
       "      <td>41.816054</td>\n",
       "      <td>21.595318</td>\n",
       "      <td>0.913043</td>\n",
       "      <td>1.026756</td>\n",
       "      <td>0.595318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.490305</td>\n",
       "      <td>2.17730</td>\n",
       "      <td>15.326224</td>\n",
       "      <td>33.741787</td>\n",
       "      <td>19.922496</td>\n",
       "      <td>1.315088</td>\n",
       "      <td>1.250288</td>\n",
       "      <td>1.817175</td>\n",
       "      <td>0.606560</td>\n",
       "      <td>1.644631</td>\n",
       "      <td>...</td>\n",
       "      <td>0.987873</td>\n",
       "      <td>0.998045</td>\n",
       "      <td>1.982029</td>\n",
       "      <td>1.657688</td>\n",
       "      <td>2.144978</td>\n",
       "      <td>16.934178</td>\n",
       "      <td>26.905928</td>\n",
       "      <td>1.149466</td>\n",
       "      <td>1.836639</td>\n",
       "      <td>0.491653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>37.200000</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>37.000000</td>\n",
       "      <td>6.200000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>38.000000</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>22.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>44.000000</td>\n",
       "      <td>7.200000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>38.500000</td>\n",
       "      <td>88.000000</td>\n",
       "      <td>34.500000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>52.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>9.00000</td>\n",
       "      <td>40.800000</td>\n",
       "      <td>184.000000</td>\n",
       "      <td>96.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>7.500000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>75.000000</td>\n",
       "      <td>89.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>10.100000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               0          1           2           3           4           5   \\\n",
       "count  299.000000  299.00000  299.000000  299.000000  299.000000  299.000000   \n",
       "mean     1.397993    1.64214   30.509365   66.220736   24.551839    1.906355   \n",
       "std      0.490305    2.17730   15.326224   33.741787   19.922496    1.315088   \n",
       "min      1.000000    1.00000    0.000000    0.000000    0.000000    0.000000   \n",
       "25%      1.000000    1.00000   37.200000   48.000000   12.000000    1.000000   \n",
       "50%      1.000000    1.00000   38.000000   60.000000   22.000000    2.000000   \n",
       "75%      2.000000    1.00000   38.500000   88.000000   34.500000    3.000000   \n",
       "max      2.000000    9.00000   40.800000  184.000000   96.000000    4.000000   \n",
       "\n",
       "               6           7           8           9      ...              12  \\\n",
       "count  299.000000  299.000000  299.000000  299.000000     ...      299.000000   \n",
       "mean     1.555184    2.404682    1.167224    2.404682     ...        1.143813   \n",
       "std      1.250288    1.817175    0.606560    1.644631     ...        0.987873   \n",
       "min      0.000000    0.000000    0.000000    0.000000     ...        0.000000   \n",
       "25%      1.000000    1.000000    1.000000    1.000000     ...        0.000000   \n",
       "50%      1.000000    2.000000    1.000000    2.000000     ...        1.000000   \n",
       "75%      3.000000    4.000000    2.000000    4.000000     ...        2.000000   \n",
       "max      4.000000    6.000000    3.000000    5.000000     ...        3.000000   \n",
       "\n",
       "               13          14          15          16          17          18  \\\n",
       "count  299.000000  299.000000  299.000000  299.000000  299.000000  299.000000   \n",
       "mean     1.023411    0.834448    1.816054    2.230769   41.816054   21.595318   \n",
       "std      0.998045    1.982029    1.657688    2.144978   16.934178   26.905928   \n",
       "min      0.000000    0.000000    0.000000    0.000000    0.000000    0.000000   \n",
       "25%      0.000000    0.000000    0.000000    0.000000   37.000000    6.200000   \n",
       "50%      1.000000    0.000000    1.000000    2.000000   44.000000    7.200000   \n",
       "75%      1.000000    0.000000    4.000000    5.000000   50.000000   52.000000   \n",
       "max      3.000000    7.500000    4.000000    5.000000   75.000000   89.000000   \n",
       "\n",
       "               19          20          21  \n",
       "count  299.000000  299.000000  299.000000  \n",
       "mean     0.913043    1.026756    0.595318  \n",
       "std      1.149466    1.836639    0.491653  \n",
       "min      0.000000    0.000000    0.000000  \n",
       "25%      0.000000    0.000000    0.000000  \n",
       "50%      0.000000    0.000000    1.000000  \n",
       "75%      2.000000    2.000000    1.000000  \n",
       "max      3.000000   10.100000    1.000000  \n",
       "\n",
       "[8 rows x 22 columns]"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>67.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.492537</td>\n",
       "      <td>1.477612</td>\n",
       "      <td>32.894030</td>\n",
       "      <td>63.970149</td>\n",
       "      <td>24.835821</td>\n",
       "      <td>2.104478</td>\n",
       "      <td>1.373134</td>\n",
       "      <td>2.701493</td>\n",
       "      <td>1.164179</td>\n",
       "      <td>2.417910</td>\n",
       "      <td>...</td>\n",
       "      <td>1.014925</td>\n",
       "      <td>1.059701</td>\n",
       "      <td>1.386567</td>\n",
       "      <td>1.805970</td>\n",
       "      <td>2.373134</td>\n",
       "      <td>38.304478</td>\n",
       "      <td>22.695522</td>\n",
       "      <td>1.208955</td>\n",
       "      <td>1.255224</td>\n",
       "      <td>0.701493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.503718</td>\n",
       "      <td>1.909769</td>\n",
       "      <td>13.067805</td>\n",
       "      <td>27.570279</td>\n",
       "      <td>20.322101</td>\n",
       "      <td>1.244799</td>\n",
       "      <td>1.191330</td>\n",
       "      <td>1.517723</td>\n",
       "      <td>0.566676</td>\n",
       "      <td>1.509056</td>\n",
       "      <td>...</td>\n",
       "      <td>0.961258</td>\n",
       "      <td>1.071399</td>\n",
       "      <td>2.652837</td>\n",
       "      <td>1.707726</td>\n",
       "      <td>2.102205</td>\n",
       "      <td>17.627608</td>\n",
       "      <td>28.364507</td>\n",
       "      <td>1.200219</td>\n",
       "      <td>1.822193</td>\n",
       "      <td>0.461057</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>37.500000</td>\n",
       "      <td>44.500000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>35.000000</td>\n",
       "      <td>5.850000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>38.000000</td>\n",
       "      <td>60.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>41.000000</td>\n",
       "      <td>6.800000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>38.300000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>36.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>2.500000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>47.000000</td>\n",
       "      <td>56.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>39.700000</td>\n",
       "      <td>150.000000</td>\n",
       "      <td>96.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>8.500000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>74.000000</td>\n",
       "      <td>79.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              0          1          2           3          4          5   \\\n",
       "count  67.000000  67.000000  67.000000   67.000000  67.000000  67.000000   \n",
       "mean    1.492537   1.477612  32.894030   63.970149  24.835821   2.104478   \n",
       "std     0.503718   1.909769  13.067805   27.570279  20.322101   1.244799   \n",
       "min     1.000000   1.000000   0.000000    0.000000   0.000000   0.000000   \n",
       "25%     1.000000   1.000000  37.500000   44.500000  12.000000   1.000000   \n",
       "50%     1.000000   1.000000  38.000000   60.000000  20.000000   2.000000   \n",
       "75%     2.000000   1.000000  38.300000   80.000000  36.000000   3.000000   \n",
       "max     2.000000   9.000000  39.700000  150.000000  96.000000   4.000000   \n",
       "\n",
       "              6          7          8          9     ...             12  \\\n",
       "count  67.000000  67.000000  67.000000  67.000000    ...      67.000000   \n",
       "mean    1.373134   2.701493   1.164179   2.417910    ...       1.014925   \n",
       "std     1.191330   1.517723   0.566676   1.509056    ...       0.961258   \n",
       "min     0.000000   0.000000   0.000000   0.000000    ...       0.000000   \n",
       "25%     1.000000   1.000000   1.000000   1.000000    ...       0.000000   \n",
       "50%     1.000000   3.000000   1.000000   2.000000    ...       1.000000   \n",
       "75%     2.500000   3.000000   1.500000   3.000000    ...       2.000000   \n",
       "max     4.000000   6.000000   2.000000   5.000000    ...       3.000000   \n",
       "\n",
       "              13         14         15         16         17         18  \\\n",
       "count  67.000000  67.000000  67.000000  67.000000  67.000000  67.000000   \n",
       "mean    1.059701   1.386567   1.805970   2.373134  38.304478  22.695522   \n",
       "std     1.071399   2.652837   1.707726   2.102205  17.627608  28.364507   \n",
       "min     0.000000   0.000000   0.000000   0.000000   0.000000   0.000000   \n",
       "25%     0.000000   0.000000   0.000000   0.000000  35.000000   5.850000   \n",
       "50%     1.000000   0.000000   1.000000   3.000000  41.000000   6.800000   \n",
       "75%     2.000000   0.000000   4.000000   4.000000  47.000000  56.000000   \n",
       "max     3.000000   8.500000   4.000000   5.000000  74.000000  79.000000   \n",
       "\n",
       "              19         20         21  \n",
       "count  67.000000  67.000000  67.000000  \n",
       "mean    1.208955   1.255224   0.701493  \n",
       "std     1.200219   1.822193   0.461057  \n",
       "min     0.000000   0.000000   0.000000  \n",
       "25%     0.000000   0.000000   0.000000  \n",
       "50%     1.000000   0.000000   1.000000  \n",
       "75%     2.000000   2.000000   1.000000  \n",
       "max     3.000000   8.000000   1.000000  \n",
       "\n",
       "[8 rows x 22 columns]"
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "all_data = pd.concat([train_df,test_df],ignore_index=True).reindex()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21]"
      ]
     },
     "execution_count": 163,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_data.columns.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for idx in all_data.columns.tolist()[:-1]:\n",
    "    all_data[[idx]] = mms.fit_transform(all_data[[idx]].values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(299, 21)"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train = all_data[:train_df.shape[0]].values[:,:-1]\n",
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_train =  all_data[:train_df.shape[0]].values[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_test = all_data[train_df.shape[0]:].values[:,:-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_test =  all_data[train_df.shape[0]:].values[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.        , 0.        , 0.94362745, ..., 0.07078652, 0.        ,\n",
       "        0.        ],\n",
       "       [1.        , 0.        , 0.92156863, ..., 0.07078652, 0.33333333,\n",
       "        0.4950495 ],\n",
       "       [0.        , 0.        , 0.92401961, ..., 0.78651685, 1.        ,\n",
       "        0.1980198 ],\n",
       "       ...,\n",
       "       [0.        , 0.        , 0.93137255, ..., 0.73033708, 1.        ,\n",
       "        0.1980198 ],\n",
       "       [1.        , 0.        , 0.93137255, ..., 0.06516854, 0.        ,\n",
       "        0.        ],\n",
       "       [1.        , 0.        , 0.92156863, ..., 0.06741573, 0.        ,\n",
       "        0.        ]])"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "SGD_clf = SGDClassifier(loss = \"log\",max_iter=1500,random_state=2018)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "SGD_clf.fit(X_train,y_train)\n",
    "y_pred = SGD_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7313432835820896"
      ]
     },
     "execution_count": 172,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(y_pred==y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 用了没有处理之前的模型参数，只处理数据，居然变差了...\n",
    "- 改变迭代次数，正确率能提升\n",
    "- 用mms和sss差别不大"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "TODO:\n",
    "1. 用原始数据集，数据处理也做一下\n",
    "1. 考虑数据可视化，分析下"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}