KG_Classification2_Homework.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Homework\n",
    "\n",
    "This week we are going to use a new data set which contains 1070 purchases where the customer either purchased Citrus Hill or Minute Maid Orange Juice. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Purchase</th>\n",
       "      <th>WeekofPurchase</th>\n",
       "      <th>StoreID</th>\n",
       "      <th>PriceCH</th>\n",
       "      <th>PriceMM</th>\n",
       "      <th>DiscCH</th>\n",
       "      <th>DiscMM</th>\n",
       "      <th>SpecialCH</th>\n",
       "      <th>SpecialMM</th>\n",
       "      <th>LoyalCH</th>\n",
       "      <th>SalePriceMM</th>\n",
       "      <th>SalePriceCH</th>\n",
       "      <th>PriceDiff</th>\n",
       "      <th>Store7</th>\n",
       "      <th>PctDiscMM</th>\n",
       "      <th>PctDiscCH</th>\n",
       "      <th>ListPriceDiff</th>\n",
       "      <th>STORE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>237</td>\n",
       "      <td>1</td>\n",
       "      <td>1.75</td>\n",
       "      <td>1.99</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.99</td>\n",
       "      <td>1.75</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.24</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>239</td>\n",
       "      <td>1</td>\n",
       "      <td>1.75</td>\n",
       "      <td>1.99</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>1.69</td>\n",
       "      <td>1.75</td>\n",
       "      <td>-0.06</td>\n",
       "      <td>0</td>\n",
       "      <td>0.150754</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.24</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>245</td>\n",
       "      <td>1</td>\n",
       "      <td>1.86</td>\n",
       "      <td>2.09</td>\n",
       "      <td>0.17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.680000</td>\n",
       "      <td>2.09</td>\n",
       "      <td>1.69</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.091398</td>\n",
       "      <td>0.23</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>227</td>\n",
       "      <td>1</td>\n",
       "      <td>1.69</td>\n",
       "      <td>1.69</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>1.69</td>\n",
       "      <td>1.69</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>228</td>\n",
       "      <td>7</td>\n",
       "      <td>1.69</td>\n",
       "      <td>1.69</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.956535</td>\n",
       "      <td>1.69</td>\n",
       "      <td>1.69</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Purchase  WeekofPurchase  StoreID  PriceCH  PriceMM  DiscCH  DiscMM  \\\n",
       "0         0             237        1     1.75     1.99    0.00     0.0   \n",
       "1         0             239        1     1.75     1.99    0.00     0.3   \n",
       "2         0             245        1     1.86     2.09    0.17     0.0   \n",
       "3         1             227        1     1.69     1.69    0.00     0.0   \n",
       "4         0             228        7     1.69     1.69    0.00     0.0   \n",
       "\n",
       "   SpecialCH  SpecialMM   LoyalCH  SalePriceMM  SalePriceCH  PriceDiff  \\\n",
       "0          0          0  0.500000         1.99         1.75       0.24   \n",
       "1          0          1  0.600000         1.69         1.75      -0.06   \n",
       "2          0          0  0.680000         2.09         1.69       0.40   \n",
       "3          0          0  0.400000         1.69         1.69       0.00   \n",
       "4          0          0  0.956535         1.69         1.69       0.00   \n",
       "\n",
       "   Store7  PctDiscMM  PctDiscCH  ListPriceDiff  STORE  \n",
       "0       0   0.000000   0.000000           0.24      1  \n",
       "1       0   0.150754   0.000000           0.24      1  \n",
       "2       0   0.000000   0.091398           0.23      1  \n",
       "3       0   0.000000   0.000000           0.00      1  \n",
       "4       1   0.000000   0.000000           0.00      0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "oj = pd.read_csv('data/OJ.csv')\n",
    "oj.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "oj.data = oj.iloc[:, 1:]\n",
    "oj.target = oj.iloc[:, 0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The features and the target variable are already prepared for you. What you are going to do is:\n",
    "\n",
    "**1. Split**\n",
    "\n",
    "Split the data set into two parts: training set and test set(with *random_state=0*, and *test_size=1.0/2*).\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "### your solution\n",
    "from sklearn import model_selection\n",
    "x_train, x_test, y_train, y_test = model_selection.train_test_split(oj.data, oj.target, test_size=1.0/2, random_state=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**2. Support vector machine**\n",
    "\n",
    "(1) Fit a svm model on the training set, report the training error and test error.(Just use the **svm.SVC** with default setting.)\n",
    "    \n",
    "(2) Change the value of parameter $C$ from $10^{-3}$ to $10^3$, make a plot to watch how the training error and test error varies. You can choose the value of $C$ from the array `np.logspace(-3, 3, 300)`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SVM Default Training error: 0.22616822429906547\n",
      "SVM Default Test error:     0.302803738317757\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFpCAYAAACxlXA1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11e3d5f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### your solution\n",
    "from sklearn import svm\n",
    "import numpy as np\n",
    "svm_model = svm.SVC()\n",
    "svm_model.fit(x_train, y_train)\n",
    "print(\"SVM Default Training error: {}\".format(1 - svm_model.score(x_train,y_train)))\n",
    "print(\"SVM Default Test error:     {}\".format(1 - svm_model.score(x_test,y_test)))\n",
    "\n",
    "C = np.logspace(start=-3, stop=3, num=300)\n",
    "\n",
    "svm_train_errors = []\n",
    "svm_test_errors = []\n",
    "\n",
    "for i in C: \n",
    "    svm_model.set_params(C=i)\n",
    "    svm_model.fit(x_train,y_train)\n",
    "    train_error = 1 - svm_model.score(x_train,y_train)\n",
    "    test_error = 1 - svm_model.score(x_test,y_test)\n",
    "    svm_train_errors.append(train_error)\n",
    "    svm_test_errors.append(test_error)\n",
    "    \n",
    "import matplotlib.pyplot as plt\n",
    "plt.plot(np.log10(C), svm_train_errors, c = \"purple\", label=\"train error\")\n",
    "plt.plot(np.log10(C), svm_test_errors, c = \"orange\", label=\"test error\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**3. Decision tree**\n",
    "\n",
    "(1) Fit a decision tree model on the training set with the default setting.\n",
    "    \n",
    "(2) Set the depth of the tree from 1 to 30. Look the varies of the training error and test error.\n",
    "    \n",
    "(3) Use the function **grid_search.GridSearchCV** to find the best parameters. What's the best parameters? What's the best score? What's the training error and test error of the best model. The possible combination of the parameters may be:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFpCAYAAACxlXA1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xd4VFX+x/H3SYEQegkdpENCJ6GJKII0CyCsbZW1A7Ku\nq65YVl3Z7qq7664/FVERXQVFEURBQQRFRJSEDqEXCTX0Ttr5/XGiUpNJMsmd8nk9Dw9k5t65Xy7D\nfOaee4qx1iIiIiKBL8LrAkRERMQ3Cm0REZEgodAWEREJEgptERGRIKHQFhERCRIKbRERkSCh0BYR\nEQkSPoW2MaafMWatMWaDMebR8zx/szFmuTFmhTFmgTGm7WnPbcl9fKkxJtmfxYuIiIQTk9/kKsaY\nSGAd0BtIAxYBN1lrV5+2zcVAqrX2gDGmPzDaWts597ktQJK1dm/x/BVERETCgy9X2p2ADdbaTdba\nDOBdYODpG1hrF1hrD+T+uBCo698yRUREJMqHbeoA2077OQ3onMf2dwKfnvazBWYbY7KBV6y1Y/M7\nYLVq1WyDBg18KE1ERCT4paSk7LXWxuW3nS+h7TNjzOW40L7ktIcvsdZuN8ZUBz43xqyx1s47z77D\ngGEA9evXJzlZt79FRCQ8GGO2+rKdL83j24F6p/1cN/exsw/YBngNGGit3ffj49ba7bm/7wGm4Jrb\nz2GtHWutTbLWJsXF5ftlQ0REJOz4EtqLgKbGmIbGmFLAjcC00zcwxtQHPgSGWmvXnfZ4WWNM+R//\nDPQBVvqreBERkXCSb/O4tTbLGHMvMBOIBMZZa1cZY0bkPj8G+ANQFXjJGAOQZa1NAmoAU3IfiwIm\nWGs/K5a/iYiISIjLd8iXF5KSkqzuaYuISLgwxqTkXuzmSTOiiYiIBAmFtoiISJBQaIuIiAQJhbaI\niEiQUGiLiIgECYW2iIhIkFBoi4iIBAmF9oXYHDi0xusqJBAc3QQZh7yuQvzlyAbIOuZ1FSKFotC+\nkNRnYXo8rH7G60rES8e2wfRW8Elz2DbF62qkqI5ugukJ8FlHF94iQUahfT45mbD2BYiMgaWPwKq/\ne12ReGX5E67VJaYGfD0Y5t8IJ9O9rkoKa+VfgAg4uRtmdoZdc7yuSKRAFNrns+1DOLEdur0LDW6G\nZb+HlX/1uiopaQeWwub/QfPfQr9kaP0nSPsQpreEH973urrgl3UM5vaHre+VzPGObIDNb0HTEdD3\neyhTE+b2gXUvlczxRfxAoX0+a/8D5ZpAnWugy5vQYKi74lrxJ68rk5JiLSx+CEpXgZaPQUQ0tH4S\n+qVA2Ytg/vXw9S/gxG6vKw1eSx6BnZ9B8r0l02dg5Z8hohQkPArlG0Ofb6FWf0j+NXx/j2thEwlw\nCu2z7VsEe7+F5r8BEwERkdDlDWh0G6x4CpaPdh/oEtp2fga7v4CWT0KpSj8/Xqm1+7Bv9zRs/8Td\nH938jt4TBbVrDqx/EWpfBaf2wup/FO/xDq+DLW9D03vcFTZAdAW4dCokPAIbxsCcPnBqX/HWIVJE\nCu2zrf0PRJV3If2jiEjo/Do0ugNW/tGFtz6kQ1dONix5GMo1dh/yZ4uIch/0/ZdA+Wbw7S0wbyAc\n31HytQajzMPw3R1QvilcMsndglr7b9fpr7is/DNElIb4h898PCLSfQHr+j/3ZX1mJzi4qvjqECki\nhfbpTuyEHyZBo9vdt/DTmQjo/Co0vst9ACx/QsEdqjaPh0Mrod3fIbLUhberGA+950P7f8Kuz929\n7k3j9b7Iz5JRcHybu/UUFQtt/+rO2fIni+d4h9bA1gnQ7F4oU+P82zS8Ba74CrKOw6yurhVFgkPm\nYTieVrhfQTj0L8rrAgLK+pchJ8s1jZ+PiYBOr4CJhFV/c72K2/4NjCnZOqX4ZB1z4VG1C9T7Rf7b\nR0RC/IOu/8N3d8LC22HrJPc+KVuv+OsNNjtnwYaxED8K4rq6x8peBM3vg9TnoMUDULmtf4+58k8Q\nWcYdMy/VOkO/RTBvEHw1wH1pi39Y/78DVU6WG5q7YjTkZBTuNaLKQ/tnocmwoPl3Vmj/KPskrB/j\n7rGVb3Lh7UwEdHzJ/b76abDZ0O4fQfMPLvlI/Zdrcbnk/YL9m1ZoCld8CetehKWPuqvuDv90LTN6\nbzgZB90Xmwrx0OasTp0tfw8bX3dX4T1n+e+Yh1bD1nch4WGIict/+9i6cMU8WHiH+3c8uNK1sEXG\n+K8mKboDy90X5AOLod4QqNWvEC9iYetEWDTCjQbp/BqUa+DvSv1Oof2jre/BqXRo8dv8tzURkPQi\nEOG+6dkc921NH86+2b/EtWq0+fOFmyu9cGI3pD4Dda+FuG4F399EuFaaOlfBd3fB98NcYLT9O1Tr\n5P96z5Z9Eja8Bnu+dF88KyRAxQSo0AKiyxX/8fOz+EH3hajPh+eGYKlK0OpJWPwA7JgJtfv655gr\n/gRRZaHFQ77vExUL3Sa6TofLn4Aj61yHtTK1/FOTFF52Bqz+O6z6K0RXcl+u6/vQInYhje90LT9L\nRsGMVu4CrOk97v9ygDI2AO+/JSUl2eTk5JI7oLXwWSLknIIrV/oevtZCyv2w7r/Q/H7o8C8Fd35y\nMuHT9nBoFcTWg8umQeV2XlflfH8PbHwNrloFFZoV7bVsjvswWP6E65Fcqy+0eurnJmF/yjrhjpX6\nDxeKsfXh5M4zhzCVbeACvGJCbpi3dPfko8v7v57z2f4JfHWNu6Jue4E5D7Iz3CyEUWWh3xJ366Eo\nDq6EGW3ckL0LHTM/26bAt0NdQFw6FaomFa0mKbz9S9zV9cFlcNFNkPhfiKnmn9c+9gN8dzfsmgXV\nL3NX3Xm1uBYDY0yKtTbfN5hCG2DP1zD7Uug4BpoOL9i+1roriLXPQ7P7IPF5BXdeUv8JSx5yH6Lr\nX4ZT+6HrW1B/iLd1HVrjvmk3GQEd/89/r5t5BNa/5O7XntoLNXtD66cKdyV/tqzjsOEVN9XuyV1Q\nvYd77Ro9XGAf2QiHV7sm4kOr3Relw2vOvP8XW+/nIK/U0rUylK5S9NpOd2q/O7elq0HfRRBZ+sLb\nbp0E39wAncdB49uLdtyvr4OdM2HglqL9nQ4sh3kD3Cxqbf/umtALykS5L25RZQpfR7jKPuU6/65+\nGkrHQacxUHeg/49jLWx6w32e52S4/krNflP0L48+UmgXxNe/gN1zYFCaaxorKGtd88qaf0LTX0PS\nCwru8zmeBp+0gOqXQ4+P4cQumHct7FsIrUe75lGvmqW+Ggi758KAjb7d+yyorGPuS0rqs3ByD9To\n6QK2+qUl+1o5WXB0c26Yr/o50A+nuub1shfBpdOgcpuC13UhC4a62wR9v4cq7fPe1lrXe/v4Nrhm\nfeH+P4IL2k/bQssnoO2fC/capzu5x31OpH9d+Neocw1c+pE+Gwpi3yJ3dX1oFTS8FRL/DaUqF+8x\nj2+H74fDjulQ7WLoMg4qNC/eY6LQ9t2xrTCtkbvn1b4IEzxY6+YpT33W3RNJ+r+Avi/iifnXw/aP\n4arVUK6heyz7pPsPsvkt11u763jXPFqS9syD2Ze5q/+Wvy/eY2Uddx0eU59xV27VL8sN3B75f5hn\nHnUTkpx+1d7qD1D9kqLXlZPtxil/cwNkHoKub0O9QUV/3W1T4etr3Zey1k/5ts+e+TC7O7T5C7R6\nvHDHnTfYTY4zcIv/PuRzsuHIWnfro6C2fejmd+jyJjT6lX/qCWXZJ2H5U7DmOYipBZ3GQp0rS+74\n1rrJeFJ+6/7PtvkztHiwWK+6fQ1trLUB9ysxMdGWmMWjrJ0Qae3RrUV/rZwca5c8au07WLvsyaK/\nXijZMdOdl+V/Ove5nBxrVz9n7TvG2hnt/fNv4aucbGs/7Wjth3WszTxWcsfNPG5t6vPWfljLnZdZ\n3a3dOdudi7NlHLZ25d+s/aCq23ZOX2v3LCieuo5td+fjHaxd8Zfz1+OrE+nWTq5u7Yx21mZnFGzf\nrwZZ+155a0/sLvhx9y/J/T/4VMH3LS7ZWdbO6mbtpIrWHkvzuprAtucbaz9u7v4NF95l7amD3tVy\nfId7L76DtZ91svbgqmI7FJBsfcjH8L4UzDoGG1519/HK1i/66xnj7oM0uNndZzy2teivGQqyT7n5\npcs1gYTzjJU1BuJ/B5d9Akc3wsyOkL6gZGrbOgn2L3JX2YVtii2MqDJupMKATZD4glsycs4V7gpz\n5yz3TT/jkFuV6qMGbtGaqp2hz0K4/LPi6dAGEFvbTTJy0S9dJ7oFv3Qd3Qoj+deQccD1WYiILti+\n7Z6G7OOw4o8FP+6K0RBdEVrcX/B9i0tEJHQZ7+6Vfne3JuA5n6zjkPIgfH6Je89dPssNtytV0bua\nytSC7h/CxRPdZ9On7d0cHR7OUx/eob35bcg86FZx8hdj3AeOMbCskE17oSb1OTiyHjq+mPd41zpX\nunm9o8rDF5fDpjeLt67sU7DsMajUFhrcUrzHupDIGGh+LwzY4IYRHtsKc/vCZ0kurJc/6Tqt9V0E\nPaa7CUCKW1QZuPht1+lq63vui8Tx7QV7ja2T3OyCrUe7oVMFVaE5NBnuOtodXuv7fvtTIO0jaPG7\nM+eMDwTlm7ghRTs/hU3jvK4mcGQedp/FM9q66WybjoCrVkKt3l5X5hgDDW50t/XqDnSf6zO7uH4T\nHgjf0LbWDdWq3N4/PXlPF1sXmj8AW95xHyLh7OhmWPUXd7+6Vp/8t6+YAH2/g7hLYOFtbqWtnOzi\nqW3d/8GxLW6MfQn1EL2gyBhoNhKu2eBGMeRkuF7g/VLcsLiSHmpkDLR81HWcOrzWfYnY+51v+57Y\nDckjoUrHc+f6LojWT0FkrOsr4qvlo909bF/mW/BCs1+7/gspD7hhRuEq4yBsesvNPDc5zg2rA+g1\nx01eVVJDEQsiprqbK/+SD+BEmmuh80D4dkTb+blbS7fLeGh0q/9fP+MQfNzEXWX0/CJ8e4x+NcD1\nzL96TcGGyuRkug+29S+65RO7TfRvM9mp/TCtsbtyvfwz/71uKDq40v07ntjhFs5pePOFt7UWvh4C\nO2ZA/8XuS1hRrPyra6a/Yh5U7573tvsWuQU/itKBrSQc3ezGj1fr4pqAi+uzIeOAu00QKB1iT+2H\n7dPc7GO7Pnf/x2PruS/09X/hzkeg1JqfjIO559Z//3a+dkQLkjNUDNb+x31zuujG4nn9UhXdZBq7\n57oPsHCU9rHrLd7qqYKPbY2IduOlO45x/8FndYHD6/1X26q/ul7S7Z7x32uGqkqt3HCtal3cimZL\nH71w68fWiZA2xfW2LWpgg5uLvEwdN7Y/vwuMFaOhVBU3j3kgK9cQ2j8Hu2a75v/isOMz+LAWzLrY\nDenzyql9bnrauf3hwxo/D99qdp/rnzFwCyT+C+IuDp7ABnfrxaMLsSA6S350eL0bg9dkRN4TPRRV\n0+Fu+cGlD7vxseEk6zik3Oc+uIvSIajpcOj5uRsnO6uz+6ArqqObXdN4o9v8Ox45lMVUc1eFTYa7\nta/nDXL3Ik93fIfrcFitqxse4w9RsdD2L7Dve3eP/EL2LnRfjuNHBWbT6tmaDHND9pY85N6P/rR9\nulsqtnxjOLoht/PU30vuM+hkupuhb04fF9Tf3eVusbR40H35G7AZOjznWrmCKagDRHiesXUvuCu5\n862V7E8R0a5T2qHVbqadcLLq7+5+cdKLBe85fLYaPdzqS2Vqw9x+brrR3V8W/l73st+7ldra+GHS\njXASWQo6vuzmINj5qbuKO7rJPWetG2+ffcLdcvJnH4EGQ6FSG1j6mOs8eD4rRrsZ15rd67/jFidj\n3K0GE+muPgsz9vt80j524+IrtXbLxl65CuoMcO/5WV3g4Ar/HOds1sKuOTCnL0yp6d4LRze5L1H9\nUtykRe3/AVU7hu+tQj8Jv9DOPOwCtP4NUKZm8R/vx8Unlv/BTY4RDg6vd5OHNLjZBa4/lGsEfRZA\nw6Gw+U3Xu3xqbRfgu77w/Spi3yI3O1eLByG2jn9qCyfGuM5Ul89097hndnJfoDa/BTs+cT3Oizpv\n+9kiIl1nwWOb3ZSwZ0tf4KYrjR8VGAuj+KpsPejwb9jzlVsdrqjSPoL5Q6BSO+g523XIK1MDur/v\nFtY49oNbY2HFn/w3ZMla1/o1+1KY0wsOrYCEx6D/EjejXbu/Q5UOCmo/Cr+OaGv+A4vvd800VTsW\nzzHOlv4tfH5xwWaFClbWuqvhfQvh6rXF88Uo86i70vvhfdcUmH3cXWXVvRbqX+e+KJzv6t5a+KIH\nHEp1Q6yiK/i/tnByZINbBOTIBnebqXIHtzxpcTV5zunreuwO2HjmLGdzesOBZTBwc8nPpldU1sJX\nV7u+L1cuL/wiFdumuBkHqyS6L1Tn67R5cq+7ZbV1ohvm2OWN/KeVzavunbPcWuV7F7h+BwmPQpO7\ntIxpIakj2vnkZLum8WpdSy6wwU2EUe8XborTEztL7rhe2DbZrZTT5i/F15IRXc6F8yWTYEi6G4JR\n8wrYOsGNCPiwpruPtuMzt3LUj7Z/7KYsbT1age0P5Zu4zkS1+rgFMbq8Ubz3KNs/63rtrvrbz4/t\nme+u9BIeCb7ABncF2mksRJR2QxwLc8vnh8m5gZ104cAG1y+h2wS3WtnJ3W4So2VPXPiWw/lYCzs+\ndfPDf9kPjv/gboEN2ODmG1BgF7vwutJO+9it1tPtXbjoBv+/fl6ObIDpCdDwNug8tmSPXVIyj8An\n8W7Bjb6LIKKEl2vPOuGaSbd9AGnTIOuIuyKrOxDqDXGLutgcN3FDUe+zy8+sdfeyS2JGuYW3w5YJ\nrhWnXAP4opfrjTxgU8nOaOdvm/8H3/4K2v8T4gvQie+H9+Gbm9xseZd/6vuX0YwDbkjl5jfdMq1d\n3sj7QsZa13l3xZ9ca0dsfTdPf6PbirczbxjRlfb5rP2Pa8apN7jkj12+CTS5Bza9DgdXlfzxS8LK\nP8GJ7ZD0UskHNriZvOoNcrN5DdnjVquqc41rOvzqGrcsZbt/KLD9zZiSC8w2f3adt5Y9Dru/cnMA\nJDwa3IENbka+ugNh+eNumVhfbH3PBXa1rm6ugYK0HpWq7BbnuWy6a72Y1cUN5cs+eeZ21rp75Z8l\nuf9Dp9Kh06vufnXT4QpsD4TPlfbBlTCjtZsbvOVj/n1tX53cCx83hrju0OMTb2ooLgdXwaft3EQ1\nnV/zupozZZ9yndVObIfGd6lTTLBb9rhrIi/fzLWmXLMxNNapPrELprd0w0R7z8/7i++WiW7MfLVu\n0GNG0TrgZRxyQ882vgYVWri1zKt1dmG98k9wYKnrCNrycdcRVF96i4WutIHsjGy+HP0lqyatcvey\nI2Og8d3eFRRTzTUp7ZjuOp6ECmvdtJXRFaDt015Xc67I0m5e8yZ3K7BDQcIjUDoOjqzLvcoOgcAG\n1wek40uw7ztY888Lb7f5bRfYcd2LHtjg7oF3ftWNw886Dp93g4+bwdeDXafPLuPd7YjGdyiwA0BI\nh3ZEVATrPl7HvD9MxW7+nxuCFFPN26Ka3efuBy1+yH9jM7225R3Xwavd370/vxL6oiu4oVJx3d0k\nJaGk/vWu0+ryP5z/Ntqmt9y97+qXuQVk/DnErVZv19+j2a/dbJFd34KrU13rmRe3u+S8Qjq0TYSh\n93O9adr0S0z2Cf+u5lVYUWXcMpAHFrsmrmCXcdA1rVXt5JqeRUpCw5uh97zQ661sTO6CGRVh4a1n\njqfeNN71MK/R0y1jWxy95aPLQ9ILP8+JoLAOOCEd2gANL6tH1wFL2Lq2EcezG3tdjtPgl25M67Lf\nn9vxI9gsf9J1Tun4sqYkFPGHmDjoNMatELj6H+6xjeNg4R1uaONlHwd/xzsptND/lE2bQtmy+/h2\nRmfm/Xme19U4JsKNOT3+A6x9wetqCm//YjdDVZN73KxHIuIf9QbDRb90HcGWPgrf3enmKr/0o9C5\nhy+FEvqhvfY/ULYhZdtfz6KXFrF/436vK3Jq9oTaV7rVpk7t87qagsvJhEX3uJnI2v7F62pEQk/S\nC1CqqrvartUPLlNgS6iHdnYGVEiA+Ifo8cdeRJaK5IvHvvC6qp+1e8YNWVkZhKG37HG38lLiC26Z\nOhHxr9JVoPsHbk71S6eE3v17KZTQDu3IUm72sWYjKV+rPBc/dDGr319N2sI0rytzKrWERnfA+hfh\nyEavq/Fd2jQ3JWvTe+Ci672uRiR0xXWD9s8osOUnoR3aZ7n4oYspV7Mcsx6aRcBMKtPmT2CiXae0\nYHB0C3x7q+tI1+FfXlcjIhJWwiq0S5UrRY8/9mDbN9tYM9XHqQKLW5laEP8Q/DAJ9n7ndTV5yz7l\nFibAuuX+9O1fRKREhVVoA7S/oz3V4qsx+5HZZGcWYkWd4hA/CmJquPHOgdICcD5LRrnFArq84aY1\nFBGREhV2oR0RFUHvZ3qzf/1+UsameF2OE10OWv8R0ue7+X4D0Q/vu6lgmz8A9a71uhoRkbAUdqEN\n0PSqpjTo0YCv/vgVpw4XYC3Z4tT4TqgQD0sfOXMN6EBweD0svBOqdoF2ATi3uIhImAjL0DbGTW96\nPP048/8x3+tynIgo6PBPtwjCqgAaApZ1AuZf5xYKuOQ91yNfREQ8EZahDVA7sTatf9mahf9ayOG0\nw16X49TuDw1/5ZYd3B8gTfcpv4WDy6Dr/6Bsfa+rEREJa2Eb2gA9/9oTm2OZ88Qcr0v5WeLzrlPa\nt7e63tpe2vw/2PgqJDzmlrYUERFPhXVoV2pQiU73dWLZW8vYtWyX1+U4pSpD59fg0CpYMdq7Og6t\nhu9HQPVL3VhyERHxXFiHNkD333cnplIMn4/63OtSfla7v+uYlvoM7F1Y8sfPOgZf/8L1ar94opbn\nExEJEGEf2mUql+HSJy9l0+eb2DBzg9fl/Kz9P6FMHbd+btaJkjuutfD9PXB4DVw8AWJrl9yxRUQk\nTz6FtjGmnzFmrTFmgzHm0fM8f7MxZrkxZoUxZoExpq2v+waCjiM7UqlhJT4f9Tk52Tlel+OUqghd\nxsHhtW7N6pKy8XXY8j9oPRpq9iq544qISL7yDW1jTCTwItAfSABuMsYknLXZZuAya21r4M/A2ALs\n67mo0lFc8fQV7Fmxh2VvLfO6nJ/VvAKajIA1/4I9JTA07cBSSL7Xrdvb8vHiP56IiBSIL1fanYAN\n1tpN1toM4F1g4OkbWGsXWGsP5P64EKjr676BIuG6BOp0rsPcJ+aSeTzT63J+1v5ZKHsRLLzd3Wsu\nLpmH4evroHRVuPhtiIgsvmOJiEih+BLadYBtp/2clvvYhdwJfFrIfT1jjKHPc304suMI3/77W6/L\n+Vl0OTfX99ENsPSx4jmGtfDdXXBsM3R7F2KqF89xRESkSPzaEc0YczkutB8pxL7DjDHJxpjk9PR0\nf5bls/qX1KfFoBZ8849vOLanGK9qC6pGD2h2n5v7e/eX/n/99S+5ucXb/g2qd/f/64uIiF/4Etrb\ngXqn/Vw397EzGGPaAK8BA621+wqyL4C1dqy1NslamxQXF+dL7cWi19O9yDyeyZd//NKzGs6r3d+g\nXBPXTJ55xH+vu306LH4Aal/tlggVEZGA5UtoLwKaGmMaGmNKATcC007fwBhTH/gQGGqtXVeQfQNN\ntebVSByeSMorKexdu9frcn4WVRa6jodjW2HJw0V/vYwD7gvAV1dD+ebQ9U0wYT8CUEQkoOX7KW2t\nzQLuBWYCqcAka+0qY8wIY8yI3M3+AFQFXjLGLDXGJOe1bzH8Pfyqx1M9iI6N5otHv/C6lDPFdYMW\nD8KGMbCzCJPBpH0M01u6aUpb/h76JUPpKv6rU0REioWx1npdwzmSkpJscnKypzXM+8s85j45l+FL\nh1OzbU1PazlD1gn4rIPrSX7lCjee21en9kHyfbB1AlRq7Tq4VUksvlpFRMQnxpgUa21SftupPfQC\nOo7sSGTpSFJeCZDVtn4UVQa6jIcT22HJ73zf74fJMD0BfpgErZ6CvskKbBGRIKPQvoAyVcrQ6oZW\nLH97ORlHM7wu50zVOkP8w272su0z8t725B6Yfz3M/4WbFrVfMrQZrXWxRUSCkEI7D4nDE8k4ksGK\niSu8LuVcrUdDxZbw/d2uU9nZrIWt77l712kfQZu/QN/voHLbc7cVEZGgoNDOQ92udaneunrgNZED\nRJZ2Pb5P7oaU+8987sQu+HoIfHMjlG0I/RZDq8chItqbWkVExC8U2nkwxpA4PJGdKTvZkbzD63LO\nVSXRzRG++S1Im+aurjf/z9273jED2j0DfRZApZZeVyoiIn6g0M5Hm1vaEB0bTfIr3vZmv6CWj0Pl\ndvD9MPjqGvj2V1AhHq5cBgmjtBa2iEgIUWjnI6ZiDK1uasXKCSs5eeik1+WcK7KU602esR92z4EO\n/4Yr5kGF5l5XJiIifqbQ9kHi8EQyj2ey4p0A7JAGrnNZ72/gqtXQ4n6t0CUiEqIU2j6onVSbWh1q\nkTwmmUCcjAaAqh2hXAOvqxARkWKk0PbBjx3S9qzYQ9rCNK/LERGRMKXQ9lGrm1pRqnypwBz+JSIi\nYUGh7aPS5UvT+ubWrHpvFScOnPC6HBERCUMK7QJIGpFE1skslr21zOtSREQkDCm0C6Bm25rU6VyH\nlDEpgdshTUREQpZCu4CSRiSxd81efvj6B69LERGRMKPQLqCW17ekdMXSJI8J0BnSREQkZCm0Cyg6\nNpq2t7YldXIqx/ce97ocEREJIwrtQkgankR2RjZLxy/1uhQREQkjCu1CiEuIo/4l9Ul5JQWbow5p\nIiJSMhTahZQ4IpH9G/azee5mr0sREZEwodAupIQhCZSpWoaUMZohTURESoZCu5CiYqJod1s71kxd\nw9FdR70uR0REwoBCuwgShyWSk5XDknFLvC5FRETCgEK7CKo2q0rDng1Z/OpicrJzvC5HRERCnEK7\niBKHJ3Jwy0E2ztrodSkiIhLiFNpF1GJQC8pWL6slO0VEpNgptIsoslQk7e5ox7pP1nE47bDX5YiI\nSAhTaPtB4t2J2BzL4tcXe12KiIiEMIW2H1RuVJnGfRqz5LUl5GSpQ5qIiBQPhbafJI1I4nDaYdbP\nWO91KSJ77FhLAAAgAElEQVQiEqIU2n7S7OpmlK9dXh3SRESk2Ci0/SQiKoL2d7Zn/afrObjloNfl\niIhICFJo+1GHuzpgjGHxa+qQJiIi/qfQ9qOK9SvS9MqmLHl9CdmZ2V6XIyIiIUah7WeJIxI5uuso\na6et9boUEREJMQptP2vSrwkV6lZg2ZvLvC5FRERCjELbzyIiI4gfEs/GWRvJOJrhdTkiIhJCFNrF\nIH5wPNmnstnw2QavSxERkRCi0C4G9brVIzYultQPU70uRUREQohCuxhEREbQfGBz1n2yjqxTWV6X\nIyIiIUKhXUziB8eTcSSDzXM2e12KiIiECIV2MWnYsyGlypdSE7mIiPiNQruYRJWOotnVzVj70Vpy\nsrXyl4iIFJ1CuxjFD47nePpxtn2zzetSREQkBCi0i1GTfk2ILB2pJnIREfELhXYxKlWuFE36NmHN\nlDVYa70uR0REgpxCu5i1GNyCQz8cYufinV6XIiIiQU6hXcyaX9McE2nURC4iIkWm0C5mZaqUoUGP\nBqz5cI3XpYiISJBTaJeA+MHx7F2zl/TUdK9LERGRIKbQLgHNBzYHYM0UXW2LiEjhKbRLQIU6Fajb\npa7ua4uISJEotEtIi8Et2Jmyk0M/HPK6FBERCVIK7RISf208AKlTdLUtIiKFo9AuIVWaVKF66+rq\nRS4iIoXmU2gbY/oZY9YaYzYYYx49z/MtjDHfGmNOGWMeOuu5LcaYFcaYpcaYZH8VHoziB8fzw/wf\nOLbnmNeliIhIEMo3tI0xkcCLQH8gAbjJGJNw1mb7gfuA5y7wMpdba9tZa5OKUmywa3FtC2yOZe20\ntV6XIiIiQciXK+1OwAZr7SZrbQbwLjDw9A2stXustYuAzGKoMWTUaFODyo0qqxe5iIgUii+hXQc4\nfW3JtNzHfGWB2caYFGPMsIIUF2qMMbQY3ILNX2zm5KGTXpcjIiJBpiQ6ol1irW2Ha17/tTHm0vNt\nZIwZZoxJNsYkp6eH7sxh8dfGk52RzfoZ670uRUREgowvob0dqHfaz3VzH/OJtXZ77u97gCm45vbz\nbTfWWptkrU2Ki4vz9eWDTt0udSlXs5x6kYuISIH5EtqLgKbGmIbGmFLAjcA0X17cGFPWGFP+xz8D\nfYCVhS02FJgIQ4trW7D+0/VknlAXABER8V2+oW2tzQLuBWYCqcAka+0qY8wIY8wIAGNMTWNMGvAg\n8IQxJs0YUwGoAcw3xiwDvgemW2s/K66/TLBocW0LMo9lsunzTV6XIiIiQSTKl42stTOAGWc9Nua0\nP+/CNZuf7TDQtigFhqIGPRoQUymG1A9TaT6gudfliIhIkNCMaB6IjI6k+YDmrPt4HdmZ2V6XIyIi\nQUKh7ZEW17bgxP4TbJ231etSREQkSCi0PdK4T2OiY6M10YqIiPhMoe2R6NhomvRvwtqpa7E51uty\nREQkCCi0PdTi2hYc2XGE7d/7POxdRETCmELbQ82uakZEdISayEVExCcKbQ/FVIqhUa9GrJmyBmvV\nRC4iInlTaHusxbUt2L9hP3tW7vG6FBERCXAKbY81H9gcDGoiFxGRfCm0PVauRjnqd6vPmilaQERE\nRPKm0A4ALQa3YPey3ezfuN/rUkREJIAptANA/LXxALraFhGRPCm0A0ClBpWo2b6mQltERPKk0A4Q\n8YPj2bZgG0d2HvG6FBERCVAK7QARPzi3iXyqrrZFROT8FNoBolp8Nao2q6omchERuSCFdoAwxtBi\ncAu2zN3CxlkbvS5HREQCkEI7gHT5bReqxVfjnf7vsPA/CzW1qYiInEGhHUDK1SzHnQvupPmA5sy8\nfyYf3/0x2RnZXpclIiIBQqEdYEqVK8X1k6+n+xPdWfL6Et7q9RbH9hzzuiwREQkACu0AZCIMPf/c\nkyETh7AjeQevdnqVXct2eV2WiIh4TKEdwFrd2Irbv76dnKwcxnUbR+oULSoiIhLOFNoBrnZSbe5e\ndDfVW1Zn0uBJzPvLPHVQExEJUwrtIFC+Vnlu++o22tzShrlPzmXyTZPJPJ7pdVkiIlLCorwuQHwT\nFRPFoLcGUb11dWY/Opv9G/Zz49QbqVC3gteliYhICdGVdhAxxtDt4W7cNO0m9q3dx6sdXyXtuzSv\nyxIRkRKi0A5Cza5uxp0L7yQ6Nprxl41n+dvLvS5JRERKgEI7SFVvWZ27vr+Lel3rMWXoFD5/5HNy\nsnO8LktERIqRQjuIxVaN5ZZZt5B0TxILnlnAJyM+8bokEREpRuqIFuQioyO56qWriKkUw/y/z6d+\nt/q0u62d12WJiEgx0JV2iLj8z5fT4PIGTB85nT0r93hdjoiIFAOFdoiIiIxgyIQhlK5Qmkm/mETG\n0QyvSxIRET9TaIeQcjXLMWTiEPav388nwz/RzGkiIiFGoR1iGl7ekB5/6sGKCStIGZvidTkiIuJH\nCu0Q1P2x7jTu25jPfvsZO5fs9LocERHxE4V2CDIRhsFvDya2WizvX/c+Jw+d9LokERHxA4V2iIqt\nFst1k67j0NZDTLtzmu5vi4iEAIV2CKt3cT16Pd2L1MmpfP/C916XIyIiRaTQDnFdH+xK84HNmfXQ\nLC0uIiIS5BTaIc4Yw8A3BlKhTgU+uP4DTuw/4XVJIiJSSArtMFCmchmue/86ju46ytRbp2JzdH9b\nRCQYKbTDRO2k2vT5Vx/WfbKOBc8t8LocEREpBIV2GOk4siMtr2/JF7//gq1fb/W6HBERKSCFdhgx\nxnDNq9dQuVFlJt84mWN7jnldkoiIFIBCO8yUrlCa696/jhP7T/DhLR+Sk53jdUkiIuIjhXYYqtm2\nJv1f6M+mzzfx9d++9rocERHxkUI7TLW/sz1thrbhy6e+ZNMXm7wuR0REfKDQDlPGGK56+Sri4uP4\n8JcfcurIKa9LEhGRfCi0w1ipsqW45rVrOLbnGCsmrPC6HBERyYdCO8zV7VKXGm1rkPJKihYVEREJ\ncArtMGeMIWlEEruW7GLHoh1elyMiInlQaAutf9ma6LLRJL+S7HUpIiKSB4W2ULpCaVrf3JqVE1dy\n8uBJr8sREZELUGgLAEnDk8g6kcXyt5d7XYqIiFyAQlsAqNWhFrU71iZ5TLI6pImIBCifQtsY088Y\ns9YYs8EY8+h5nm9hjPnWGHPKGPNQQfaVwJE4PJH0VelsW7DN61JEROQ88g1tY0wk8CLQH0gAbjLG\nJJy12X7gPuC5QuwrAaLVja0oXaE0KWNSvC5FRETOw5cr7U7ABmvtJmttBvAuMPD0Day1e6y1i4DM\ngu4rgaNU2VK0GdqGVe+v4vi+416XIyIiZ/EltOsAp7eXpuU+5oui7CseSByeSPapbJa9uczrUkRE\n5CwB0xHNGDPMGJNsjElOT0/3upywVaN1DepdXE8zpImIBCBfQns7UO+0n+vmPuYLn/e11o611iZZ\na5Pi4uJ8fHkpDonDE9m3bh9bvtzidSkiInIaX0J7EdDUGNPQGFMKuBGY5uPrF2Vf8UjCdQnEVI4h\n5RV1SBMRCSRR+W1grc0yxtwLzAQigXHW2lXGmBG5z48xxtQEkoEKQI4x5n4gwVp7+Hz7FtdfRvwj\nukw0bW9ty6IXF3FszzHKVi/rdUkiIgKYQLxvmZSUZJOTNQ+2l/au2cuL8S/S6+leXPLIJV6XIyIS\n0owxKdbapPy2C5iOaBJYqrWoxkWXXcTisYuxOYH3xU5EJBwptOWCEocncmDTATbN3uR1KSIigkJb\n8hA/OJ7YarHqkCYiEiAU2nJBUaWjaHd7O9Z8tIYjO454XY6ISNhTaEueEoclYrMtS8Yt8boUEZGw\np9CWPFVpUoVGVzQiZWwKOdk5XpcjIhLWFNqSr8QRiRzedpgNn23wuhQRkbCm0JZ8NR/QnHI1y2nJ\nThERjym0JV+R0ZG0v7M962es59APh7wuR0QkbCm0xScd7u6AtZbFry32uhQRkbCl0BafVLqoEk36\nNWHJ60vIyVKHNBERLyi0xWdJI5I4suMI6z5Z53UpIiJhSaEtPmt6ZVPK1ylP8hgt5iIi4gWFtvgs\nIiqCDnd3YOOsjRzYdMDrckREwo5CWwqkw50dMMaQ8qqGf4mIlDSFthRIhboVaHZNM5aOW0p2RrbX\n5YiIhBWFthRY4vBEju05xpqpa7wuRUQkrCi0pcAa92lMpQaVtGSniEgJU2hLgUVEug5pm+dsZt+6\nfV6XIyISNhTaUijt72hPRFQEsx+ZrclWRERKiEJbCqVczXJc8cwVrJm6hsk3TSY7U53SRESKW5TX\nBUjw6vpAVwBmPTgLm2MZMnEIkaUiPa5KRCR06UpbiqTrA13p+3xfUj9M5YMbPtAwMBGRYqTQliLr\n8tsu9PtvP9ZMXcP7171P1qksr0sSEQlJCm3xi86/6Uz//+vP2mlref8XCm4RkeKg0Ba/6fTrTlz1\n8lWs+2QdkwZPIuukgltExJ8U2uJXSSOSuPqVq1k/Yz3vXfuegltExI8U2uJ3icMSuebVa9gwcwPv\nDnyXzBOZXpckIhISFNpSLDrc1YEBrw9g4+cbeXfAu2QeV3CLiBSVQluKTfvb2zPwjYFs+mITE6+Z\nqOAWESkihbYUq3a3tmPQm4PY8uUWJlw1gYxjGV6XJCIStBTaUuzaDm3LoLcGsXXeViZcOYGMowpu\nEZHCUGhLiWhzcxuufftafpj/A+/0f4dTR055XZKISNBRaEuJaX1TawZPGMy2b7cx4coJmoBFRKSA\nFNpSolrd0Ipr/+euuOf9eZ7X5YiIBBWFtpS41je1pt1t7Zj/9Hy2L9rudTkiIkFDoS2e6PvvvpSr\nWY6pt07VrGkiIj5SaIsnYirFMOD1AexN3cvcp+Z6XY6ISFBQaItnmvRtQoe7O/Dtc9+y7dttXpcj\nIhLwFNriqT7P9aFC3QpMvXWqZkwTEcmHQls8VbpCaQaMG8D+9fv54vEvvC5HRCSgKbTFc416NSJp\nZBLf/ec7ts7b6nU5IiIBS6EtAaH3P3pTuWFlPrr9I01zKiJyAQptCQilypVi4BsDObD5ALMfne11\nOSIiAUmhLQHjoksvovNvO7PoxUVsnrPZ63JERAKOQlsCSq+/9qJK0yp8dMdHnDqsRUVERE6n0JaA\nEh0bzaDxgzi87TCzRs3yuhwRkYCi0JaAU+/ienT9XVcWj13MhpkbvC5HRCRgKLQlIF3+p8upFl+N\naXdO4+TBk16XIyISEBTaEpCiYqIY9OYgju46yswHZ3pdjohIQFBoS8Cq07EO3R7pxtI3lrLuk3Ve\nlyMi4jmFtgS0y/5wGdVbV+fjYR9zYv8Jr8sREfGUQlsCWlRp10x+PP04n973qdfliIh4SqEtAa9W\n+1p0f6I7K95ZQeqUVK/LERHxjEJbgkL333enZruaTB8xneN7j3tdjoiIJ3wKbWNMP2PMWmPMBmPM\no+d53hhj/pv7/HJjTIfTnttijFlhjFlqjEn2Z/ESPiKjIxn01iBOHDjB9JHTvS5HRMQT+Ya2MSYS\neBHoDyQANxljEs7arD/QNPfXMODls56/3FrbzlqbVPSSJVzVaF2DHqN7sPr91ayatMrrckRESpwv\nV9qdgA3W2k3W2gzgXWDgWdsMBN6yzkKgkjGmlp9rFaHbw92o3bE200dO5+juo16XIyJSonwJ7TrA\nttN+Tst9zNdtLDDbGJNijBlW2EJFACKiIhg0fhAZRzOYfs90rLVelyQiUmJKoiPaJdbadrgm9F8b\nYy4930bGmGHGmGRjTHJ6enoJlCXBKi4hjsv/fDlrpqxh5cSVXpcjIlJifAnt7UC9036um/uYT9tY\na3/8fQ8wBdfcfg5r7VhrbZK1NikuLs636iVsdX2wK3W71mXGvTM4suOI1+WIiJQIX0J7EdDUGNPQ\nGFMKuBGYdtY204Bf5fYi7wIcstbuNMaUNcaUBzDGlAX6ALo0kiKLiHTN5Fknsvhk+CdqJheRsJBv\naFtrs4B7gZlAKjDJWrvKGDPCGDMid7MZwCZgA/AqMDL38RrAfGPMMuB7YLq19jM//x0kTFVtVpVe\nf+/Fuk/WsezNZV6XIyJS7EwgXqEkJSXZ5GQN6Zb82RzL+B7j2b1sNyNXjaRC3QpelyQiUmDGmBRf\nhkVrRjQJaibCMPCNgeRk5TDtrmlqJheRkKbQlqBXpXEVrnjmCjbO3Mji1xZ7XY6ISLFRaEtI6HhP\nRxr2bMisB2dxcOtBr8sRESkWCm0JCSbCMOD1AQBMu2MaNkfN5CISehTaEjIqNahEn3/2YfOczSSP\nUUdGEQk9Cm0JKR3u7kDjPo35fNTnHNh0wOtyRET8SqEtIcUYwzWvXUNEVAQf3f6RmslFJKQotCXk\nVKxXkb7P92XrvK1898J3XpcjIuI3Cm0JSe1ua0fTq5ryxWNfsG/dPq/LERHxC4W2hCRjDNeMvYao\n0lFMvW0qOdk5XpckIlJkCm0JWeVrl6f/C/1J+zaNhf9e6HU5IiJFptCWkNb65ta0GNSCOU/MIT1V\n67SLSHBTaEtIM8Zw1ZirKFWuFFN/NZVTh095XZKISKEptCXklatRjgGvDWDnkp28fvHrGr8tIkFL\noS1hocWgFgz9fChHdx7l1Y6vsnnuZq9LEhEpMIW2hI2Glzfkru/vomyNsrzd520WvbzI65JERApE\noS1hpUrjKty18C4a923MjJEzmP7r6WRnZntdloiITxTaEnZKVyjNjR/dyMUPX0zyS8m83fdtju87\n7nVZIiL5UmhLWIqIjKD3P3oz6K1BbPtmG691fo301RoSJiKBTaEtYa3t0Lbc9tVtZBzN4LUur7Fu\n+jqvSxIRuSCFtoS9ul3qMix5GFWbVmXiNRP55tlvsFarg4lI4FFoiwAV6lbg9q9vp+V1LZn98Gym\n3jqVrJNZXpclInKGKK8LEAkU0bHRDHl3CNVbV2fuk3PZt24fN0y5gfK1yntdmogIoCttkTMYY7j0\niUu5fvL17Fmxh1c7vsqOlB1elyUiAii0Rc4rfnA8dyy4g4jICN7o/gZLxi3RfW4R8ZxCW+QCarat\nyd2L7qZu57pMu3MaE66cwKFth7wuS0TCmAnEq4ekpCSbnJx8xmOZmZmkpaVx8uRJj6oKTTExMdSt\nW5fo6GivSwlYNsey6KVFzH50NibC0Oe5PnS4uwPGGK9LE5EQYYxJsdYm5btdsIT25s2bKV++PFWr\nVtWHpZ9Ya9m3bx9HjhyhYcOGXpcT8A5sPsDHd33M5jmbadirIQNeG0ClBpW8LktEQoCvoR00zeMn\nT55UYPuZMYaqVauq9cJHlRtWZujsoVw15iq2f7edl1q9xPcvfo/NCbwvviISmoImtAEFdjHQOS0Y\nYwxJw5MYuWok9bvV59N7P+XNnm+yf8N+r0sTkTAQVKHtpYMHD/LSSy8Vat8rr7ySgwcP+rki8VLF\n+hW5+bObGfD6AHYt3cXLbV5m4fMLycnO8bo0EQlhCm0f5RXaWVl5z5w1Y8YMKlXy773Ps4+ZXw0F\n3U7yZ4yh/R3tGblqJA17NmTmAzMZf+l49q7d63VpIhKiFNo+evTRR9m4cSPt2rVj1KhRfPnll3Tv\n3p0BAwaQkJAAwKBBg0hMTKRly5aMHTv2p30bNGjA3r172bJlC/Hx8dx99920bNmSPn36cOLEiXOO\nlZ6ezpAhQ+jYsSMdO3bkm2++AWD06NEMHTqUbt26MXToUMaPH8+AAQPo2bMnvXr1wlrLqFGjaNWq\nFa1bt+a9994DOG+t4j8V6lTgpo9vYtBbg0hPTWdM2zF888w35GTpqltE/CsopzH97P7P2LV0l19f\ns2a7mvR7vt8Fn3/66adZuXIlS5cuBVwQLl68mJUrV/7U83rcuHFUqVKFEydO0LFjR4YMGULVqlXP\neJ3169czceJEXn31Va6//nomT57MLbfccsY2v/3tb3nggQe45JJL+OGHH+jbty+pqakArF69mvnz\n51OmTBnGjx/P4sWLWb58OVWqVGHy5MksXbqUZcuWsXfvXjp27Mill14KcE6t4l/GGNoObUujKxox\nY+QMZj8ym9UfrGbgGwOp3rK61+WJSIgIytAOFJ06dTojBP/73/8yZcoUALZt28b69evPCe2GDRvS\nrl07ABITE9myZcs5rzt79mxWr17908+HDx/m6NGjAAwYMIAyZcr89Fzv3r2pUqUKAPPnz+emm24i\nMjKSGjVqcNlll7Fo0SIqVKhwTq1SPMrXKs/1H17PqvdWMePeGYztMJZL/3Ap3R7uRmR0pNfliUiQ\nC8rQzuuKuCSVLVv2pz9/+eWXzJ49m2+//ZbY2Fh69Ohx3qFUpUuX/unPkZGR520ez8nJYeHChcTE\nxOR5zPP97EutUryMMbS6sRUNezbk0998ytwn5rLmwzUMfGMgNdrU8Lo8EQliuqfto/Lly3PkyJEL\nPn/o0CEqV65MbGwsa9asYeHChYU+Vp8+fXjhhRd++vnHJvn8dO/enffee4/s7GzS09OZN28enTp1\nKnQdUjRlq5flF+/9gus+uI7DaYcZmziWL0d/SXZGtteliUiQUmj7qGrVqnTr1o1WrVoxatSoc57v\n168fWVlZxMfH8+ijj9KlS5dCH+u///0vycnJtGnThoSEBMaMGePTftdeey1t2rShbdu29OzZk2ee\neYaaNWsWug7xj4QhCYxcPZKWN7Tkqz9+xasdX2Xn4p1elyUiQShopjFNTU0lPj7eo4pCm85tyVk7\nbS2fjPiEY3uO0e2Rblz2h8uIKh2Ud6lExI9CbhpTkVDQfEBzRq4aSduhbZn/t/mM7TCW7d9v97os\nEQkSCm2RElamchkGvjGQX874JacOn+L1rq/z+cOfk3ki0+vSRCTAKbRFPNK0f1PuWXkP7e9sz4Jn\nF/BKu1fYtmCb12WJSABTaIt4KKZiDNeMvYahnw8l61QW4y4Zx2cPfEbGsQyvSxORAKQeMCIBoNEV\njbhnxT188dgXfPf8d6z7eB2JwxOp3rI6cQlxVKxfEROhFdlEwp1CWyRAlC5fmiv/70oSrktg+j3T\nmf3w7J+eiy4bTVx8HHEJcVRLqPZTmFdqUElhLhJGFNo+OnjwIBMmTGDkyJGF2v/5559n2LBhxMbG\n+rkyCTUNLmvAr1f/muP7jrM3dS/pq9PZs2oPe1fvZdPsTSx7a9lP20aViTojzOMS4qjesjqVGlYi\nIlJ3v0RCjcZp+2jLli1cffXVrFy5slD7N2jQgOTkZKpVq1ao/bOysoiKirrgz77udz5en1spmJMH\nT5K+Ov2cX4e3Hf5pm8jSkVRr4UI8LiGOuJbu9yqNqxARpTAXCTS+jtPWlbaPTl+as3fv3jz77LM8\n++yzTJo0iVOnTnHttdfyxz/+kWPHjnH99deTlpZGdnY2Tz75JLt372bHjh1cfvnlVKtWjblz557x\n2ikpKTz44IMcPXqUatWqMX78eGrVqkWPHj1o167dTwuBrFixgpiYGJYsWUK3bt144oknuOOOO9i0\naROxsbGMHTuWNm3aMHr0aDZu3MimTZuoX78+EydO9OisSXGIqRRDvYvrUe/iemc8furwKdJT00lf\nlf7T79sWbGPlxJ+/aEaWiqRqs6pnBHlcQhxVmlQhspQWNBEJdMEZ2in3wwHf5uP2WeV2kPj8BZ8+\ne2nOWbNmsX79er7//nustQwYMIB58+aRnp5O7dq1mT59OuDmJK9YsSL/+te/mDt37jlX2pmZmfzm\nN7/ho48+Ii4ujvfee4/HH3+ccePGAZCRkcGPrQ633XYbaWlpLFiwgMjISH7zm9/Qvn17pk6dypw5\nc/jVr371U32nL+Ep4aF0hdLU7VyXup3rnvF4xtEM9q45s5l9R/IOVr2/CnIb2iKiIqjSpAoxlc5d\npEZEztX5/s60uqFViR83OEM7AMyaNYtZs2bRvn17AI4ePcr69evp3r07v/vd73jkkUe4+uqr6d69\ne56vs3btWlauXEnv3r0ByM7OplatWj89f8MNN5yx/XXXXUdkpLsimj9/PpMnTwagZ8+e7Nu3j8OH\nXRPp2Ut4SvgqVa4UtZNqUzup9hmPZx7P/CnM01enszd1L5nHNcGLiC+8apkKztDO44q4pFhreeyx\nxxg+fPg5zy1evJgZM2bwxBNP0KtXL/7whz/k+TotW7bk22+/Pe/zWopTikt0bDS1OtSiVoda+W8s\nIgFBPVJ8dPbSnH379mXcuHEcPXoUgO3bt7Nnzx527NhBbGwst9xyC6NGjWLx4sXn3f9HzZs3Jz09\n/afQzszMZNWqVT7V1L17d9555x3AreddrVo1KlSoUKS/p4iIBK7gvNL2wOlLc/bv359nn32W1NRU\nunbtCkC5cuV4++232bBhA6NGjSIiIoLo6GhefvllAIYNG0a/fv2oXbv2GR3RSpUqxQcffMB9993H\noUOHyMrK4v7776dly5b51jR69GjuuOMO2rRpQ2xsLG+++Wbx/OVFRCQgaMiX6NyKiHhMS3OKiIiE\nGJ9C2xjTzxiz1hizwRjz6HmeN8aY/+Y+v9wY08HXfUVERMQ3+Ya2MSYSeBHoDyQANxljEs7arD/Q\nNPfXMODlAuwrIiIiPvDlSrsTsMFau8lamwG8Cww8a5uBwFvWWQhUMsbU8nFfnwXi/fdgp3MqIhI8\nfAntOsC2035Oy33Ml2182dcnMTEx7Nu3TyHjR9Za9u3bR0yMZsESEQkGATPkyxgzDNe0Tv369c95\nvm7duqSlpZGenl7SpYW0mJgY6tatm/+GIiLiOV9Ceztw+soEdXMf82WbaB/2BcBaOxYYC27I19nP\nR0dH07BhQx/KFRERCU2+NI8vApoaYxoaY0oBNwLTztpmGvCr3F7kXYBD1tqdPu4rIiIiPsj3Stta\nm2WMuReYCUQC46y1q4wxI3KfHwPMAK4ENgDHgdvz2rdY/iYiIiIhLmhmRBMREQlVvs6IFpChbYxJ\nB7b68SWrAXv9+HqhRucnfzpHedP5yZ/OUd7C/fxcZK2Ny2+jgAxtfzPGJPvyDSZc6fzkT+cobzo/\n+QteyVoAAALnSURBVNM5ypvOj28097iIiEiQUGiLiIgEiXAJ7bFeFxDgdH7yp3OUN52f/Okc5U3n\nxwdhcU9bREQkFITLlbaIiEjQC+nQ1lre+TPGbDHGrDDGLDXGaHA8YIwZZ4zZY4xZedpjVYwxnxtj\n1uf+XtnLGr10gfMz2hizPfd9tNQYc6WXNXrJGFPPGDPXGLPaGLPKGPPb3Mf1HsqVxznS+ygfIds8\nnruW9zqgN251sUXATdba1Z4WFmCMMVuAJGttOI+PPIMx5lLgKG652Va5jz0D7LfWPp37BbCytfYR\nL+v0ygXOz2jgqLX2OS9rCwS5yxLXstYuNsaUB1KAQcBt6D0E5HmOrkfvozyF8pW2X9fylvBhrZ0H\n7D/r4YHAm7l/fhP3AROWLnB+JJe1dqe1dnHun48AqbglifUeypXHOZJ8hHJo+20t7xBngdnGmJTc\n5VHl/GrkLoIDsAuo4WUxAeo3xpjluc3nYdv0ezpjTAOgPfAdeg+d11nnCPQ+ylMoh7b45hJrbTug\nP/Dr3KZPyYN195RC875S4b0MNALaATuBf3pbjveMMeWAycD91trDpz+n95BznnOk91E+Qjm0fVkH\nPOxZa7fn/r4HmIK7rSDn2p17H+7H+3F7PK4noFhrd1trs621OcCrhPn7yBgTjQujd6y1H+Y+rPfQ\nac53jvQ+yl8oh7bW8s6HMaZsbicQjDFlgT7Ayrz3ClvTgFtz/3wr8JGHtQScH8Mo17WE8fvIGGOA\n14FUa+2/TntK76FcFzpHeh/lL2R7jwPkDhd4np/X8v6rxyUFFGNMI9zVNbi11SfoHIExZiLQA7fq\n0G7gKWAqMAmoj1uB7nprbVh2xrrA+emBa9K0wBZg+Gn3b8OKMeYS4GtgBZCT+/Dvcfds9R4iz3N0\nE3of5SmkQ1tERCSUhHLzuIiISEhRaIuIiAQJhbaIiEiQUGiLiIgECYW2iIhIkFBoi4iIBAmFtoiI\nSJBQaIuIiASJ/wcGDfkAwU4o1QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11ec10400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Grid best params: {'criterion': 'entropy', 'max_depth': 3}\n",
      "Grid best score:  0.8149532710280374\n",
      "Best model training error: 0.17570093457943925\n",
      "Best model test error:     0.22056074766355138\n"
     ]
    }
   ],
   "source": [
    "from sklearn import tree\n",
    "tree_model = tree.DecisionTreeClassifier()\n",
    "\n",
    "tree_model.fit(x_train,y_train)\n",
    "\n",
    "tree_train_errors = []\n",
    "tree_test_errors = []\n",
    "\n",
    "for i in range(1,30):\n",
    "    tree_model.set_params(max_depth=i)\n",
    "    tree_model.fit(x_train,y_train)\n",
    "    train_error = 1 - tree_model.score(x_train,y_train)\n",
    "    test_error = 1 - tree_model.score(x_test,y_test)\n",
    "    tree_train_errors.append(train_error)\n",
    "    tree_test_errors.append(test_error)\n",
    "    \n",
    "plt.plot(tree_train_errors, c = \"purple\", label=\"train error\")\n",
    "plt.plot(tree_test_errors, c = \"orange\", label=\"test error\")\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "grid_para_tree = {'criterion': ['gini', 'entropy'], 'max_depth': range(1, 31)}\n",
    "\n",
    "grid_search_tree = GridSearchCV(tree_model, grid_para_tree, cv=5, scoring='accuracy')\n",
    "grid_search_tree.fit(x_train,y_train)\n",
    "\n",
    "print(\"Grid best params: {}\".format(grid_search_tree.best_params_))\n",
    "print(\"Grid best score:  {}\".format(grid_search_tree.best_score_))\n",
    "\n",
    "best_model = grid_search_tree.best_estimator_\n",
    "best_model.fit(x_train,y_train)\n",
    "\n",
    "print(\"Best model training error: {}\".format(1 - best_model.score(x_train,y_train)))\n",
    "print(\"Best model test error:     {}\".format(1 - best_model.score(x_test,y_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(4) What are the first 5 important features?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('WeekofPurchase', 0.024060621158382475),\n",
       " ('SpecialCH', 0.02832348525941605),\n",
       " ('LoyalCH', 0.79940581662347732),\n",
       " ('SalePriceCH', 0.033513255939795532),\n",
       " ('PriceDiff', 0.1146968210189287)]"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### your solution\n",
    "all_features = list(zip(x_train.columns[:], best_model.feature_importances_))\n",
    "[x for x in all_features if x[1] > 0.0] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**4. Random Forest.**\n",
    "\n",
    "(1) Fit a random forest on the training set. Report the training error and test error.\n",
    "    \n",
    "(2) Use the function **grid_search.GridSearchCV** to find the best parameters. What's the best parameters? What's the best score? What's the training error and test error of the best model. The possible combination of the parameters may be:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest training error: 0.024299065420560706\n",
      "Random Forest test error:     0.22242990654205608\n",
      "Best Params: {'criterion': 'gini', 'max_depth': 4, 'n_estimators': 50}\n",
      "Best Score:  0.8168224299065421\n",
      "RandomForest Best Model training error: 0.14392523364485976\n",
      "RandomForest Best Model test error: 0.19626168224299068\n"
     ]
    }
   ],
   "source": [
    "grid_para_forest = {\n",
    "    'criterion': ['gini', 'entropy'],\n",
    "    'max_depth': range(1, 31),\n",
    "    'n_estimators': range(10, 110, 10)\n",
    "}\n",
    "\n",
    "from sklearn import ensemble\n",
    "randomForest = ensemble.RandomForestClassifier()\n",
    "randomForest.fit(x_train,y_train)\n",
    "print(\"Random Forest training error: {}\".format(1 - randomForest.score(x_train,y_train)))\n",
    "print(\"Random Forest test error:     {}\".format(1 - randomForest.score(x_test,y_test)))\n",
    "\n",
    "randomGridSearch = GridSearchCV(randomForest, grid_para_forest, cv=5, scoring='accuracy')\n",
    "randomGridSearch.fit(x_train,y_train)\n",
    "\n",
    "print(\"Best Params: {}\".format(randomGridSearch.best_params_))\n",
    "print(\"Best Score:  {}\".format(randomGridSearch.best_score_))\n",
    "\n",
    "randomForestBestModel = randomGridSearch.best_estimator_\n",
    "randomForestBestModel.fit(x_train,y_train)\n",
    "print(\"RandomForest Best Model training error: {}\".format(1 - randomForestBestModel.score(x_train,y_train)))\n",
    "print(\"RandomForest Best Model test error: {}\".format(1 - randomForestBestModel.score(x_test,y_test)))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(3) What's the first 5 important features?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('StoreID', 0.058713539676419166),\n",
       " ('LoyalCH', 0.56978919146419926),\n",
       " ('SalePriceMM', 0.046894434682285428),\n",
       " ('PriceDiff', 0.068992304463941972),\n",
       " ('ListPriceDiff', 0.046274454942400498)]"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### your solution\n",
    "all_features = list(zip(x_train.columns[:], randomForestBestModel.feature_importances_))\n",
    "[x for x in all_features if x[1] > 0.045] "
   ]
  },
  {
   "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}