Decision+Trees+on+Amazon+reviews+data+set+%5BM%5D.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pandoc in /usr/local/lib/python3.6/site-packages\n",
      "Requirement already satisfied: ply in /usr/local/lib/python3.6/site-packages (from pandoc)\n",
      "\u001b[33mYou are using pip version 9.0.1, however version 10.0.1 is available.\n",
      "You should consider upgrading via the 'pip install --upgrade pip' command.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install --user pandoc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "from sklearn.grid_search import RandomizedSearchCV\n",
    "from scipy.stats import uniform\n",
    "from scipy.stats import norm\n",
    "\n",
    "from imblearn.over_sampling import SMOTE\n",
    "\n",
    "import sqlite3\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "import gensim\n",
    "\n",
    "from sklearn.metrics import classification_report, accuracy_score, confusion_matrix\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "con = sqlite3.connect('final.sqlite') # this is cleaned dataset\n",
    "final = pd.read_sql_query(\"\"\"\n",
    "SELECT Score, Text_not_included\n",
    "FROM reviews\n",
    "\"\"\", con)[:15000]\n",
    "\n",
    "for i, seq in enumerate(final['Text_not_included']):\n",
    "  final['Text_not_included'][i]=final['Text_not_included'][i].decode('UTF-8')\n",
    "X_train, X_test, y_train , y_test = train_test_split(final['Text_not_included'], final['Score'], test_size=0.2, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/site-packages/sklearn/feature_extraction/text.py:1089: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  if hasattr(X, 'dtype') and np.issubdtype(X.dtype, np.float):\n"
     ]
    }
   ],
   "source": [
    "tf_idf_vect=TfidfVectorizer(ngram_range=(1,2), min_df=10, dtype=float)\n",
    "tf_idf_vect.fit(X_train)\n",
    "tf_idf_train=tf_idf_vect.transform(X_train)\n",
    "tf_idf_test=tf_idf_vect.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sentences=[]\n",
    "for review in X_train:\n",
    "  sentence=[]\n",
    "  for word in review.split():\n",
    "    sentence.append(word)\n",
    "  sentences.append(sentence)\n",
    "\n",
    "w2vec_model=gensim.models.word2vec.Word2Vec(sentences, min_count=10)\n",
    "\n",
    "avg_w2vec_train=np.zeros(shape=(len(X_train), 100), dtype=float)\n",
    "\n",
    "for i, sentence in enumerate(sentences):\n",
    "  for word in sentence:\n",
    "    try:\n",
    "      avg_w2vec_train[i]+=w2vec_model.wv[word]\n",
    "    \n",
    "    except KeyError:\n",
    "      pass\n",
    "  \n",
    "  avg_w2vec_train[i]/=len(sentence)\n",
    "\n",
    "sentences=[]\n",
    "for review in X_test:\n",
    "  sentence=[]\n",
    "  for word in review.split():\n",
    "    sentence.append(word)\n",
    "  sentences.append(sentence)\n",
    "\n",
    "avg_w2vec_test=np.zeros(shape=(len(X_test), 100), dtype=float)\n",
    "\n",
    "for i, sentence in enumerate(sentences):\n",
    "  for word in sentence:\n",
    "    try:\n",
    "      avg_w2vec_test[i]+=w2vec_model.wv[word]\n",
    "    \n",
    "    except KeyError:\n",
    "      pass\n",
    "  \n",
    "  avg_w2vec_test[i]/=len(sentence)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sentences=[]\n",
    "for review in X_train:\n",
    "  sentence=[]\n",
    "  for word in review.split():\n",
    "    sentence.append(word)\n",
    "  sentences.append(sentence)\n",
    "  \n",
    "tf_idf_w2vec_train=np.zeros((len(X_train), 100), dtype=float)\n",
    "feat=tf_idf_vect.get_feature_names()\n",
    "for i, sentence in enumerate(sentences):\n",
    "  tf_idf_sum=0\n",
    "  for word in sentence:\n",
    "    try:\n",
    "      tf_idf_w2vec_train[i]+=w2vec_model.wv[word]*tf_idf_train[i, feat.index(word)]\n",
    "      tf_idf_sum+=tf_idf_train[i, feat.index(word)]  \n",
    "    except KeyError:\n",
    "      pass\n",
    "    except ValueError:\n",
    "      pass\n",
    "  tf_idf_w2vec_train[i]/=tf_idf_sum\n",
    "  \n",
    "sentences=[]\n",
    "for review in X_test:\n",
    "  sentence=[]\n",
    "  for word in review.split():\n",
    "    sentence.append(word)\n",
    "  sentences.append(sentence)\n",
    "  \n",
    "tf_idf_w2vec_test=np.zeros((len(X_test), 100), dtype=float)\n",
    "\n",
    "for i, sentence in enumerate(sentences):\n",
    "  tf_idf_sum=0\n",
    "  for word in sentence:\n",
    "    try:\n",
    "      tf_idf_w2vec_test[i]+=w2vec_model.wv[word]*tf_idf_test[i, feat.index(word)]\n",
    "      tf_idf_sum+=tf_idf_test[i, feat.index(word)]  \n",
    "    except KeyError:\n",
    "      pass\n",
    "    except ValueError:\n",
    "      pass\n",
    "  tf_idf_w2vec_test[i]/=tf_idf_sum"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Upsampling minority class\n",
    "over_sampler = SMOTE(ratio='minority')\n",
    "\n",
    "avg_w2vec_train_resampled, y_train_resampled = over_sampler.fit_sample(avg_w2vec_train, y_train)\n",
    "tf_idf_w2vec_train_resampled, y_train_resampled = over_sampler.fit_sample(tf_idf_w2vec_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classification using avg_word2Vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "param_grid={'max_depth' : [i for i in range(1,20,1)]}\n",
    "\n",
    "gscv=GridSearchCV(DecisionTreeClassifier(), param_grid, scoring='accuracy', cv=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GridSearchCV(cv=5, error_score='raise',\n",
      "       estimator=DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,\n",
      "            max_features=None, max_leaf_nodes=None,\n",
      "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "            min_samples_leaf=1, min_samples_split=2,\n",
      "            min_weight_fraction_leaf=0.0, presort=False, random_state=None,\n",
      "            splitter='best'),\n",
      "       fit_params={}, iid=True, n_jobs=1,\n",
      "       param_grid={'max_depth': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]},\n",
      "       pre_dispatch='2*n_jobs', refit=True, scoring='accuracy', verbose=0)\n"
     ]
    }
   ],
   "source": [
    "print(gscv.fit(avg_w2vec_train_resampled, y_train_resampled))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.34      0.61      0.44       505\n",
      "   positive       0.91      0.76      0.83      2495\n",
      "\n",
      "avg / total       0.81      0.74      0.76      3000\n",
      "\n",
      "[[ 308  593]\n",
      " [ 197 1902]]\n",
      "TPR = 0.7623246492985972\n",
      " TNR = 0.6099009900990099\n",
      " FPR = 0.3900990099009901\n",
      " FNR = 0.2376753507014028\n"
     ]
    }
   ],
   "source": [
    "predictions = gscv.best_estimator_.predict(avg_w2vec_test)\n",
    "print(classification_report(y_test, predictions))\n",
    "print(confusion_matrix(y_test, predictions).T)\n",
    "tn, fp, fn, tp = confusion_matrix(y_test, predictions).ravel()\n",
    "\n",
    "print(\"TPR = {}\\n TNR = {}\\n FPR = {}\\n FNR = {}\".format(tp/(fn+tp), tn/(tn+fp), fp/(tn+fp), fn/(fn+tp)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=16,\n",
       "            max_features=None, max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, presort=False, random_state=None,\n",
       "            splitter='best')"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gscv.best_estimator_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classification using tfidf weighted word2vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GridSearchCV(cv=5, error_score='raise',\n",
      "       estimator=DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,\n",
      "            max_features=None, max_leaf_nodes=None,\n",
      "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "            min_samples_leaf=1, min_samples_split=2,\n",
      "            min_weight_fraction_leaf=0.0, presort=False, random_state=None,\n",
      "            splitter='best'),\n",
      "       fit_params={}, iid=True, n_jobs=1,\n",
      "       param_grid={'max_depth': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]},\n",
      "       pre_dispatch='2*n_jobs', refit=True, scoring='accuracy', verbose=0)\n"
     ]
    }
   ],
   "source": [
    "print(gscv.fit(tf_idf_w2vec_train_resampled, y_train_resampled))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "   negative       0.32      0.57      0.41       505\n",
      "   positive       0.90      0.76      0.82      2495\n",
      "\n",
      "avg / total       0.80      0.73      0.75      3000\n",
      "\n",
      "[[ 289  606]\n",
      " [ 216 1889]]\n",
      "TPR = 0.7571142284569138\n",
      " TNR = 0.5722772277227722\n",
      " FPR = 0.4277227722772277\n",
      " FNR = 0.24288577154308616\n"
     ]
    }
   ],
   "source": [
    "predictions = gscv.best_estimator_.predict(tf_idf_w2vec_test)\n",
    "print(classification_report(y_test, predictions))\n",
    "print(confusion_matrix(y_test, predictions).T)\n",
    "tn, fp, fn, tp = confusion_matrix(y_test, predictions).ravel()\n",
    "\n",
    "print(\"TPR = {}\\n TNR = {}\\n FPR = {}\\n FNR = {}\".format(tp/(fn+tp), tn/(tn+fp), fp/(tn+fp), fn/(fn+tp)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=18,\n",
       "            max_features=None, max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, presort=False, random_state=None,\n",
       "            splitter='best')"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gscv.best_estimator_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Effect of changing depth"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "depths=range(1,30,1)\n",
    "\n",
    "accuracy=np.zeros((2,30))\n",
    "\n",
    "for d in depths:\n",
    "    clf_w2v=DecisionTreeClassifier(max_depth=d)\n",
    "    clf_w2v.fit(avg_w2vec_train_resampled, y_train_resampled)\n",
    "    \n",
    "    clf_tfidfw2v=DecisionTreeClassifier(max_depth=d)\n",
    "    clf_tfidfw2v.fit(tf_idf_w2vec_train_resampled, y_train_resampled)\n",
    "    \n",
    "    predictions_w2vec=clf_w2v.predict(avg_w2vec_test)\n",
    "    predictions_tfidfword2vec=clf_tfidfw2v.predict(tf_idf_w2vec_test)\n",
    "    \n",
    "    acc_w2v=accuracy_score(y_test, predictions_w2vec)\n",
    "    acc_tfidfw2v=accuracy_score(y_test, predictions_tfidfword2vec)\n",
    "    \n",
    "    accuracy[0,d-1]=acc_w2v\n",
    "    accuracy[1,d-1]=acc_tfidfw2v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f304d524438>]"
      ]
     },
     "execution_count": 140,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD8CAYAAACb4nSYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl8W+Wd7/HPz/IWL3G8ZXX22BDWQExaSklYCqQb0Clb\nlxl6py3lFminndtO6e1tO5SZ2zu3M92GLkybue10aKDQljCkBQqFAAXqBAIhgcQhhMRJHDtxYkuO\nJVnSc/+Q5CiOF9lyYuvo+369/LJ0dM7Rcyz46cnvPM/vMeccIiKSG/LGuwEiInLyKOiLiOQQBX0R\nkRyioC8ikkMU9EVEcoiCvohIDlHQFxHJIQr6IiI5REFfRCSH5Kezk5mtBL4L+ICfOOe+2e/1bwMX\nJ56WAFOdc1MSr0WBTYnXdjnnrhzqvWpqaty8efPSvgAREYENGzYccM7VDrffsEHfzHzAXcBlQAvQ\nZGZrnHNbkvs45z6Xsv9twDkpp+hxzi1Jt+Hz5s1j/fr16e4uIiKAmb2Vzn7ppHeWAdudczucc2Fg\nNXDVEPt/CPhlOm8uIiInVzpBfxawO+V5S2LbccxsLjAfeCJlc7GZrTez583s6lG3VEREMpZWTn8E\nbgDud85FU7bNdc7tMbMFwBNmtsk590bqQWZ2E3ATwJw5c8a4SSIikpROT38PMDvleV1i20BuoF9q\nxzm3J/F7B/Akx+b7k/vc7ZxrdM411tYOex9CRERGKZ2g3wTUm9l8MyskHtjX9N/JzE4FKoHnUrZV\nmllR4nENcAGwpf+xIiJycgyb3nHORczsVuAR4kM2VznnNpvZHcB651zyC+AGYLU7dlWWxcCPzSxG\n/Avmm6mjfkRE5OSyibZyVmNjo9OQTRGRkTGzDc65xuH2y8kZubGY476m3YQjsfFuiojISZWTQX9j\ny2G++MArPLWtfbybIiJyUuVk0O/s6QXgQCA0zi0RETm5cjLoB4IRADq6w+PcEhGRkys3g34oHvTV\n0xeRXJObQV89fRHJUTkZ9P0hBX0RyU05GfSTPf0DAQV9EcktuRn0Q/HROx3dyumLSG4Z6yqbWSGQ\nkt5xzmFm49wiEfEK5xx7O4Ns2+9n+/4AAO8+czp1lSXj3LK4nAz6/kR6pzfq8IciTC4uGOcWiUi2\nicUcew730NzmZ9v+AM37A2xv87O9LUB3OHrMvv+w9jUa51Zy1TmzeO+ZM6gqLRynVudo0E/29AEO\nBsIK+iJpcM7x5LZ2joSiXHH6NPJ93s0OhyJRDgTCtPtDtHUFaQ+EaPcf/dnb2cP2tgDB3qOlXKaW\nF9EwrZxrG2dTP62M+qnl1E8tIxCKsOblvTy4cQ//67ev8vdrNrO8oZarlszkstOmUVJ4csNwTgZ9\nfzBCaaGP7nCUju4Q82tKx7tJkqO6gr08uHEvBXlGbXlR3091aRGF+SMPquFIjM6eXjp7woCxsLY0\n4/RlLOZ4dMt+vvd4M1v2dQEwp6qEWy9exAfOnUVBlgf/g4EQD27cyxOvt9HaFaTdH+qbtd9fVWkh\ntWVFTKso5sPLqmmYVkb9tDIW1ZZTUTJw57GytJBbLl7Epy9ayOutfn67cQ8PJd5vUoGPy0+fxtVL\nZvHO+pqT8rfMyaAfCEaYW13Kln1dHNQIHhkHkWiM1U27+fZj2zg4yNDhypICppYXH/NlUDGpgO5Q\nhMM9vRw+EubwkV4OH+mlM/G8f1qhYVoZ1yyt4+pzZjG1vHhEbYzFHL/f3Mr3Hm/m9VY/86pL+Na1\nZ1MxqYDvPd7MFx94he890cwtFy/ig+fWjepLKt129PRG6Q5HOBKKEoxEmVNVklEPuTca48mt7fxq\n/W6eeL2NSMxx6vR4z/wdC6upLSti6uTE370s/hlUlxVmFJTNjMUzJrN4xmT+7opTadrZwYMv72Xt\npn08uHEvlSUFfPDcOr7yvtNG/R5ptSMXSyuf8bVHuLC+ht+92sr//osz+dAyLdEoJ8+6be3c+fAW\ntu0PsGx+FV9+z2Jqy4uOSR+0+0O0B4K0dYX6Ugtt/hDhSIz8PGNKSSFTSgqYMqmAKSUFVEzq97yk\nkM6eXn7zYgsv7jqML8+4qKGWaxvruOTUaUMG6GjM8btX9/H9x7ezdb+fBbWl3HbJIt5/1sy+lI5z\njj9ubeO7f2jm5ZZOZk2ZxKcvXsg1S+soyvel9XdwzvFGe4CmnYfY8NYh2v0hjoQjdIei8d/hKN2h\nCEf6fZEB5OcZp8+q4Ly5lZw3v4rGuZVUlxUN+56vt3Zx//oWfrtxDwcCYWrKCvnAObO4ZulsTple\nnla7x1o4EmPdtnYefHkvhb48/vm6s0d1nnRLK+dc0I/FHAu+vJZPrVjAj5/awReuOIVbLl50wt5P\nJGl7m587H36NJ7e2M6eqhC+/51SuOH162ukX5xyhSIyi/LwRpWy2twV44MUWfv1iC/u7QlSWFHDV\nkllcs7SOM2ZV9O0XjTn+65W9fP+J7WxvC7CwtpTPXFrP+86aiS9v4PdzzvHUtna++3gzL+06zMyK\nYv77RQu5tnE2xQXHBv9wJMbmvZ007eygaech1u/s4NCReBqlurSQuqoSSgt9lBTmU1qU+F3oo6To\n2N8Fvjxeb+2iaechNu4+3FcifUFtKefNreK8+VWcN6+SOVUlmBmHj4R5cONe7t/QwqY9nRT4jEtP\nncY1S+tYcUrthEpPZTKaUEF/EP5gL2d+/VH+53sW850/bOP68+bw1fef2H9OSW7r6A7znT9s4z9f\n2EVJgY/bLl3Eje+Yl3aPeKxEojGe2X6AX21o4bHN+wlHYyyeMZlrltZRMamAHzy5nR3t3TRMK+O2\nS+p5z5kzBg32/TnneLr5AN99vJkNbx1i+uRibl6xgAW1Zazf2cGfd3awcffhvhuf86pLOG9eFefN\nq6JxXiXza0Z37yEUifLqnk6adh6i6c0O1r91qC8fX1texKLaMja8dYhwNMZpMyZzbWMdVy2ZNa6j\nZ06UdIN+zuX0kyN3yorzqS4r4qAmaEmagr1ReqMxyory0wpQoUiUn//pLb73RDNHwlE+vGwOf/Ou\n+rTSECdCvi+Pi06ZykWnTOXwkTAPvRzv/X7jv+IrmJ46vZwffORcVp4+nbw0g32SmbG8oZYL62v4\n0xsH+e4fmvn6Q/Hz5hmcPrOCDy+by3nzKlk6r3LE9xcGU5TvY+ncKpbOreLmFQuJxRzb2wPxf028\n2cHrrX4+8vY5XLO0jtNnVgx/whyQe0E/MUa/rCifqtJC1d+RYe06eIRVz77Jfet3cyQcTeTUC6iY\nVBDPrU8qoKKkgCnJvHpJAWbGT57ewVsHj7CioZavvHcx9dPGJ2c8kCklhfzl+fP4y/PnsW2/nwP+\nEG9fUD3iYN+fmXHBohouWFTDhrcO0ROOsmTOFMqKTk6oycszGqaV0zCtnI+8be5Jec9sk3NB35/a\n0y8tZF9ncJxbJBPVhrc6+MnTb/LI5lbyzHj/2TNZPKM8PmKmp5fOI70c7gnT2hXk9VY/nT29x8wB\nqZ9axv/7b+dx0SlTx/EqhpcMkmNt6dzKMT+nZC7ngn6yp19elE91WSGv7u0c5xbJRBKJxnh0y37+\n7ekdvLTrMJOL8/nUioXceP48plcMn5IIR2J0BXvxByPMrpzk6QlMkp1yL+gnemLlxQVUlRap/o4A\n8f8u7mvazapn36TlUA9zqkr4+ytP55qldZSOIDVRmJ9HTVkRNeOUtxcZTu4F/eCx6R3V38ltrZ1B\n/v3ZN7nnz7vwByM0zq3kK+89jctOm5b2yBWRbJJzQb8vp59I74Dq7+Si7lCEHz31Bnev20FvNMa7\nz5zBJ945n3PmKA8t3pZzQb//6B1A9XdySCzmuP/FFr71yFba/CHef/ZMvnD5Kcypnhhlb0VOtJwL\n+v5gLyWFPnx5RnVpPO+q+ju54bk3DnLnw1vYvLeLJbOn8MOPLtUIE8k5ORf0A6FI35jhvvSOxup7\n2s4D3fzj2td4dMt+ZlYU890blnDl2TN1815yUs4FfX8oQllx/LKPpncU9L2os6eX7z/ezM+e20mB\nL4//cXkDn7hwwXE1YURySc4F/UAwQnmip19c4KO00Kf0jgc45+iNOnqjMcKRGGte3st3/rCNwz29\nXLd0Nn97eQNTJ4/N1H+RbJZ7QT+lpw9QVVaoBdInsDZ/kHXbDvDUtnZe29dFbzRGbyRGOBHgj/4c\nXzjw/AXVfOV9i1VzRSRFWkHfzFYC3wV8wE+cc9/s9/q3gYsTT0uAqc65KYnXbgS+knjtTufcz8ai\n4aMVCEaoKTs6UqO6tEg5/QmkNxrjpV2HeWpbG09ubWfz3vhKTTVlRZw7ZwolidK6Bfl5FPryKMzP\no8Bn8W2++LYCn7GgtowL62uUtxfpZ9igb2Y+4C7gMqAFaDKzNc65Lcl9nHOfS9n/NuCcxOMq4GtA\nI+CADYljD43pVYxA/Ebu0TH5Y1l/JxSJ0tYVYnaVhv+NxL7OHp7a2s5T29p5pvkA/lAEX56xdG4l\nX7jiFFY01HLajMkZFwMTkfR6+suA7c65HQBmthq4CtgyyP4fIh7oAa4AHnPOdSSOfQxYCfwyk0Zn\nwh/spTw1vVNa2NebzNQvnt/FP/3+dV748qVMKfFeve6x9qc3DnDHQ1t4vdUPwIyKYt571gwuOqWW\ndyyq0YQ5kRMgnaA/C9id8rwFeNtAO5rZXGA+8MQQx84a4LibgJsA5sw5cUsXOueOGbIJ9NXUH4v6\nOzvaA4Qi8YUq3nfWzEyb62m/27SPz67eSF3lJL78nlNZ0TCVhmllSseInGBjXQLwBuB+59zxi1oO\nwTl3t3Ou0TnXWFtbO8ZNOqqnN0rMcUxPP7X+TqZaE2mip7cdyPhcXvbLP+/ilnte5My6Cn796Xdw\n0/KFnDK9XAFf5CRIJ+jvAWanPK9LbBvIDRybuhnJsSdcarG1pL6x+mMwbDN5b2BdczsTbRnKicA5\nx11/3M7tv97E8oZa/uPjy5QGEznJ0gn6TUC9mc03s0LigX1N/53M7FSgEnguZfMjwOVmVmlmlcDl\niW3jIrXYWtLRWbmZD9ts7QpSVpTPvs4g29sCGZ/PS2Ixx50Pv8b/fWQrVy2Zyb/9VSMlhTk3Ylhk\n3A0b9J1zEeBW4sH6NeA+59xmM7vDzK5M2fUGYLVL6eImbuB+g/gXRxNwR/Km7njoW0DlmPTO2NTf\nCfZG6egO8/6zZwCwrlkpnqTeaIz/8auX+ekzb/Kxd8zj29ctoUCLi4iMi7S6Ws65tcDaftu+2u/5\n1wc5dhWwapTtG1P+vgqbR0eFVJWNTSmG/V3x1M7SuVW88GYH67a18/F3zs/onF7QE45y6z0v8vjr\nbfztZQ3ceski5e5FxlFOdbcCoV6gX3qndGyKriXz+TMqilleX8sLbx4k2Dui+9me09nTy1+teoEn\ntrZx59VncNul9Qr4IuMsp4K+f4D0zljV30mO3JleUczyhhqCvTHW7xy3OWjjrq0ryPU/fo6Nuw/z\n/Q+dw0ffPne8myQi5FjQDwxwIxfGpv5Osqc/fXIxb19QTYHPeLq5PaNzZqu3DnZzzY+eY1fHEf79\nY8s0Z0FkAsmtoJ/o6fdf6LpqDOrvtHb2MLk4n9KifEoK82mcW8VT23Ir6Dvn+P2r+/jgD5/DH+zl\nnk++nXfW14x3s0QkRW4F/VCEovx4ka5UNaWFGad39nUGmVExqe/58oZaXm/109aVeV2f+9bvZtUz\nb2Z8nhPp1T2d3HD389z8ixepKi3gVzefz5LZU8a7WSLST04NlPaHIsfk85PGov5Oa1eQ6RVH67Uv\nb6jh//wenm4+wAeX1o36vEfCEb7x0Bb8oQjVZYVcteS4Khbjan9XkH/6/VZ+/VILVSWF/MMHzuD6\nxtnka0imyISUU0E/EIwcl8+HeE4/0/o7+zqDnDZjct/zxdMnU1NWyLrm9oyC/tpNrfhDEeZUlfCl\nBzaxeMZkGqaVj/p8Y6UnHOXudTv40VNvEI05blq+gFsuXqQiaSITXE51x/ovoJJUU1qUUf2dcCTG\ngUDomJ5+Xp5xYX0tzzQfIBYbfUmGe5t2saCmlF/dfD6lRfnc/IsNfTekx0Ms5vjNSy1c8s9P8u0/\nbOPiU2v5w+dXcPu7Fyvgi2SB3Ar6wQjlRccHpkzr77T5gzgXH6Of6sL6Gg52h9myb3Spo+1tAZp2\nHuL682YzbXIx3//QOew80M3f3f/KuNT2adrZwQd+8Cyfu/dlasuL+NXN5/ODjyxlTrXWDxDJFjmV\n3vGHItRVTjpue1XZ0Qla82pKR3zeo2P0jz13cuTKuuZ2zpg18iX7frV+N/l5xl+cG08Pnb+wmi+u\nPJVv/u51lj5byV+fhBm/nUd6eXp7Ow+9vJdHNu9n+uRi/uW6s7l6ySwtaiKShXIq6AdCvZQXHZ8P\nr+mrvzO6sfqps3FTTS0vZvGMyazb1s6nL1o0onOGIzEeeLGFSxdPpba8qG/7p5YvYMNbh/jHta9x\nVl0FjfOqRtXmwcRijk17OnlqW3wlq5d2HSLmoGJSAZ97VwM3LV/ApELfmL6niJw8uRX0gwPn9DOt\nv5M6G7e/5Q01rHrmTbpDkePmBwzlidf3cyAQ5obzjl1Uxsz41rVnc+W/PsMt97zIw5+5kJqyokHO\nkp4DgRBPN7fz1NZ21jUfoKM7jBmcNauCWy+pZ0VDLWfXVWhEjogH5EzQd87hH2T0Tqb1d/Z1Bikt\n9FE+wLlX1Nfy46d28PyOg1y6eFra57y3aTfTJxezvOH4RWUqJhXww48s5QM/eJbb7nmJ//j4shEH\n5K5gLz//004e3bKfTXs6cS7+d1jRUMuKhlourK+hOsMvExGZeHIm6IciMSIxN2BPP9P6O61dPUyv\nKB5wuOfSeZVMKvCxblt72kF/7+EentrWzi0XL8I3SN78tJmTufPqM/jC/a/wL49t44srT03r3KFI\nlP98fhfff6KZQ0d6WTq3ks+/q4EVp9RyxswK5elFPC5ngn5fsbVBUiyZ1N/pPxs3VVG+j7cvqOLp\nEdTXv39DCzEH1zXOHnK/axtn8+KuQ/zgyTc4d04l7zpt8C+VWMzx0Ct7+dajW9nd0cMFi6r50srF\nnFk38hvMIpK9ciZJ21dsbYCePmRWf6e1MzhgPj/pwvpadhzoZnfHkWHPFYs57lu/mwsWVTO7avih\nkF97/+mcMWsyn7tvI7sODnz+Z5oPcOVdz/DZ1RspLyrg53+9jF98/G0K+CI5KHeC/gALqKSqHmX9\nnUg0Rps/dNzInVTJvHw6vf0/vXGQlkM9XN/vBu5gigt8/PAjS8kz4+ZfbDimhv+rezr5y5++wEd/\n+gKHj/TyneuX8F+3vZPlDbWqay+So3Im6PsHWEAlVXVp4ahG7xwIhInG3JA9/YW1pcysKGZdGlU3\nVzftYkpJAZcPkarpb3ZVCd++/my27Oviaw9uZnfHEf5m9Uu87/vPsGlPJ19572Ie/9sVXH2OxtaL\n5LqcyekPtD5uqnhOPzzi+jv7OnuA48fopzIzljfU8vCmfUSisUFH2hzqDvPo5v18+G1zKC4Y2Vj4\nS06dxq0XL+Jf/7idB15sId9nfPqihdx80UKVRxCRPrkT9AdZQCWpurSQcDSGPxQZUZDsG6M/eeAb\nuUnLG2pZ3bSbl1sOs3TuwBOqfvPSHsLRGNefN/QN3MF87rIGWruCFPjy+Oyl9UP+60NEclPOBf3B\nevrViVm5HYHwiIL+YLNx+7tgYQ15Bk9tOzBg0HfOcW/Tbs6ePYXFKdU6R8KXF5+4JSIymNzJ6QeH\nGb1TNroJWq1dQYry85hSMvQXRUVJAWfPnjLoEoobdx9m634/1w8zTFNEJBM5E/QDoQiFvjyK8gfO\nlffNyh1h/Z34GP2BJ2b1d2F9LS/vPkznkd7jXrtv/W4mFfh4/9kzRvT+IiIjkTtBf5C6O0nJkgMj\nHcHT2tmTdu58RUMNMQfPvnHs0M3uUIQ1G/fyvrNmUK6briJyAuVO0A8NXHcnabT1d4aajdvf2XVT\nKC/OP27o5sOv7KM7HB31DVwRkXTlTND3B3uHDPrFBT5KCn0j6unHYo79XUPPxk2V78vjgoU1PN18\n4JhFUFY37WJhbSlL51am/d4iIqORQ0F/6PQOxFfQGklO/2B3mN6oG3bkTqrlDbXsOdzDG+3dADTv\n9/PirsPccN4czZIVkRMuZ4J+IBQZtNhaUnXZyOrvHB2jn37QvzC5mlYixXNvU3x1rA+cOyvtc4iI\njFZOBf3hevojLcVwdDZuejl9iJdMWFBTytPN7YQjMX790h4uO21axguhiIikI62gb2YrzWyrmW03\nsy8Nss91ZrbFzDab2T0p26NmtjHxs2asGj5SgUEWUElVNcKia61dg6+YNZQL62t4fkcHazfto6M7\nrBu4InLSDDsj18x8wF3AZUAL0GRma5xzW1L2qQduBy5wzh0ys6kpp+hxzi0Z43aPmD+dnv4I6+/s\n6wxS4LO+kT/pWt5Qy8+ee4s7H36NmRXFXFh//OpYIiInQjo9/WXAdufcDudcGFgNXNVvn08Cdznn\nDgE459rGtpmZCUWihCOx4XP6ifo7yZINw2ntDDJtcvGIK1e+fUE1BT7jQCDENY2zB10dS0RkrKUT\n9GcBu1OetyS2pWoAGszsWTN73sxWprxWbGbrE9uvzrC9o9IditeYHz69E8+rp5vi2dfZM6KRO0ml\nRfksnVuJGVy7tG7Ex4uIjNZYFVzLB+qBi4A6YJ2ZnemcOwzMdc7tMbMFwBNmtsk590bqwWZ2E3AT\nwJw56S0eMhJHyyoPPdu1OqX+zrya0mHP29oZ5My6KaNq02cvbeC1fV1prY4lIjJW0unp7wFS7zTW\nJbalagHWOOd6nXNvAtuIfwngnNuT+L0DeBI4p/8bOOfuds41Oucaa2vHPr/dt4BKGqN3IL1SDM65\nvro7o3H+wmr++p3zR3WsiMhopRP0m4B6M5tvZoXADUD/UTi/Jd7Lx8xqiKd7dphZpZkVpWy/ANjC\nSRYYZlH0pKoRFF07fKSXUCQ2ojH6IiLjbdj0jnMuYma3Ao8APmCVc26zmd0BrHfOrUm8drmZbQGi\nwBeccwfN7B3Aj80sRvwL5pupo35OluEWRU9K1tRPZ4JWunX0RUQmkrRy+s65tcDaftu+mvLYAZ9P\n/KTu8yfgzMybmZnhVs1KmlSYfv2d1q74xCytTiUi2SQnZuQOt4BKqnTr7xzt6ac/G1dEZLzlVNAv\nLxq+Vn11aWFa6Z3WziC+PKO2XOUTRCR75ETQD4R68eUZxQXDX251WVFa6Z19nUGmlhdpYpWIZJXc\nCPqJujvplFZIt/5Oa2f6dfRFRCaKnAj6/mFWzUqVrLSZusjJQEY7G1dEZDzlRNAPBCOUp3ETF+Kz\ncoerv5OcmDV9sm7iikh2yY2gP4KefrL+zlB5fX8owpFwVD19Eck6uRP00+3pJ2blHhgir9+3YpaC\nvohkmdwI+sHIsMXWkpJF14bq6Ws2rohkq5wI+iO5kVvVV3Rt8AlarZ2ajSsi2Skngv6IbuQmcvpD\npXf2dQYxg6nlCvoikl08H/Qj0Rg9vdG0e/rp1N9p7QxSU1ZEYb7n/3wi4jGej1rprpqVqioxVn8w\nmdTRFxEZT54P+ukuoJKqurSQA0MUXWvtDKqOvohkJe8H/TQXUEk1fE9fs3FFJDt5Puinu4BKqqGK\nrnWHInQFI0xXSWURyULeD/rB9BZQSVWdKLo2UP2d1i6N0ReR7OX5oO9P9PTTHbIJ8fTOYPV3NBtX\nRLKZ54P+0Z5+ejNyIZ7egYFn5Wo2rohkM+8H/VGO3oGBJ2glZ+NO0+gdEclC3g/6wQhmUFLgS/uY\no6UYBu7pV5UWUjyC84mITBSeD/rJujt5I1jW8GjRtePH6muMvohkM88H/UAwMqIx+nC0/s5AC6Rr\nNq6IZDPvB/0R1NJPmlToY1KBb8C1clu7tDauiGSv3Aj6I+zpQzzF0z+nH+yN0tEdVk9fRLKW54O+\nPxihLM0FVFJVlxYel97Z35Uco6/ZuCKSnTwf9AOhkef0IT6C52C/omvJMfoz1dMXkSzl+aDvD/aO\nKr1TVXp8/R3NxhWRbOf5oB8IjvxGLkBNWTy9k1p/Z5+CvohkubSCvpmtNLOtZrbdzL40yD7XmdkW\nM9tsZvekbL/RzJoTPzeOVcPTEY05usPpr5qVqqq0kHDk2Po7rZ09VEwqoKRw5OcTEZkIho1eZuYD\n7gIuA1qAJjNb45zbkrJPPXA7cIFz7pCZTU1srwK+BjQCDtiQOPbQ2F/K8brDIy+2lpQ6K7c8cSNY\nY/RFJNul09NfBmx3zu1wzoWB1cBV/fb5JHBXMpg759oS268AHnPOdSReewxYOTZNH95oyion1ZQd\nP0FLY/RFJNulE/RnAbtTnrcktqVqABrM7Fkze97MVo7g2BNmNAuoJCV7+qkTtNTTF5FsN1bJ6Xyg\nHrgIqAPWmdmZ6R5sZjcBNwHMmTNnjJqUslTiKMbpH03vxIdthiMxDgRCTJ+sMfoikr3S6envAWan\nPK9LbEvVAqxxzvU6594EthH/EkjnWJxzdzvnGp1zjbW1tSNp/5D6evqjnJELR9M7bf4gzqmOvohk\nt3SCfhNQb2bzzawQuAFY02+f3xLv5WNmNcTTPTuAR4DLzazSzCqByxPbTopAcPQ3cksK84+pv6Mx\n+iLiBcNGQ+dcxMxuJR6sfcAq59xmM7sDWO+cW8PR4L4FiAJfcM4dBDCzbxD/4gC4wznXcSIuZCB9\nC6iMoqcP8RRPcoKWVswSES9IKxo659YCa/tt+2rKYwd8PvHT/9hVwKrMmjk6yZz+aG7kwtEJWqCe\nvoh4g6dn5CZz+qWjnEyVWn9nX2eQsqL8Ud0UFhGZKLwd9IMRSgt9+Eawalaq1Po7rV096uWLSNbz\ndND3j7LuTlJ1Sv0djdEXES/wdNAf7QIqSdWJ+jvd4ajWxhURT/B00PeHRreASlJyglZbV5A2f0g9\nfRHJep4O+oFg76gWUElKTtDa2uonGnNaMUtEsp63g37G6Z140bXNe7sAjdEXkezn7aCf4Y3cZHrn\n1b2dgMbUMBBMAAAKc0lEQVToi0j283TQ94cioyrBkJRM76inLyJe4dmg75wb9aLoScn6O+3+EMUF\neVRM0sQsEclung36R8JRnBt9CYakZIpnRsUkzEY3yUtEZKLwbNA/WlY5s955MsWjMfoi4gWeDfqZ\nFltLqu7r6Svoi0j282zQT/b0M8npQ7z+Dmjkjoh4g3eD/lj19MvU0xcR7/Bs0PcHM1tAJSl5I1ez\ncUXEC7wb9DNYHzeVcvoi4iWZRcQJLJP1cVNdccZ0Dh/p5bQZk8eiWSIi48q7QT+5alaGPf3JxQV8\ncvmCsWiSiMi482x6JxCKUFyQR4HPs5coIjJino2I/mAk44lZIiJe49mgHwhFmJxhPl9ExGu8G/SD\nvRmP0RcR8RrvBv0MF1AREfEizwb9eE5fQV9EJJVng34glNmqWSIiXuTpoJ9psTUREa/xZNB3zmW8\nPq6IiBd5MugHe2NEYk7j9EVE+vFk0PeHEhU21dMXETlGWkHfzFaa2VYz225mXxrg9Y+ZWbuZbUz8\nfCLltWjK9jVj2fjB9BVbU05fROQYw0ZFM/MBdwGXAS1Ak5mtcc5t6bfrvc65Wwc4RY9zbknmTU1f\nYIzKKouIeE06Pf1lwHbn3A7nXBhYDVx1YpuVmbFaNUtExGvSCfqzgN0pz1sS2/r7oJm9Ymb3m9ns\nlO3FZrbezJ43s6sHegMzuymxz/r29vb0Wz+I5AIqmdbSFxHxmrG6kfsQMM85dxbwGPCzlNfmOuca\ngQ8D3zGzhf0Pds7d7ZxrdM411tbWZtyYozl9jd4REUmVTtDfA6T23OsS2/o45w4650KJpz8Blqa8\ntifxewfwJHBOBu1NS19OXz19EZFjpBP0m4B6M5tvZoXADcAxo3DMbEbK0yuB1xLbK82sKPG4BrgA\n6H8DeMwdXTXLd6LfSkQkqwzbFXbORczsVuARwAescs5tNrM7gPXOuTXAZ8zsSiACdAAfSxy+GPix\nmcWIf8F8c4BRP2POH4xQmJ9HUb6CvohIqrTyH865tcDaftu+mvL4duD2AY77E3Bmhm0csUCoV2P0\nRUQG4MkZuaq7IyIyMG8GfS2gIiIyIE8G/S4toCIiMiBPBv1AMKKJWSIiA/Bm0Fd6R0RkQN4N+urp\ni4gcx5tBPxjRAioiIgPwXNAPRaKEozHl9EVEBuC5oN9XbE1BX0TkON4L+lpARURkUJ4L+v6ggr6I\nyGA8F/RVVllEZHDeC/paQEVEZFDeC/rq6YuIDMpzQd+vG7kiIoPyXNDXkE0RkcF5Luj7g73k5xlF\n+Z67NBGRjHkuMibr7pjZeDdFRGTC8V7QVy19EZFBeS7o+1VWWURkUJ4L+oFghMnFGqMvIjIQ7wV9\n1dIXERmUN4O+0jsiIgPyXND3B9XTFxEZjOeCfiDUS7l6+iIiA/JU0O+Nxgj2xpTeEREZhKeCfreK\nrYmIDMlTQV8LqIiIDM1TQT9ZVlnF1kREBpZW0DezlWa21cy2m9mXBnj9Y2bWbmYbEz+fSHntRjNr\nTvzcOJaN7+9oT1+Ts0REBjJsl9jMfMBdwGVAC9BkZmucc1v67Xqvc+7WfsdWAV8DGgEHbEgce2hM\nWt9PINQLKKcvIjKYdHr6y4DtzrkdzrkwsBq4Ks3zXwE85pzrSAT6x4CVo2vq8JTTFxEZWjpBfxaw\nO+V5S2Jbfx80s1fM7H4zmz3CY8dEMqc/WT19EZEBjdWN3IeAec65s4j35n82koPN7CYzW29m69vb\n20fdiOSqWUrviIgMLJ2gvweYnfK8LrGtj3PuoHMulHj6E2Bpuscmjr/bOdfonGusra1Nt+3HCYQi\n5BlMKvCN+hwiIl6WTtBvAurNbL6ZFQI3AGtSdzCzGSlPrwReSzx+BLjczCrNrBK4PLHthPAnFlDR\nqlkiIgMbNg/inIuY2a3Eg7UPWOWc22xmdwDrnXNrgM+Y2ZVABOgAPpY4tsPMvkH8iwPgDudcxwm4\nDiDe0y9XLX0RkUGllfx2zq0F1vbb9tWUx7cDtw9y7CpgVQZtTJuWShQRGZrnZuTqJq6IyOA8FfS1\nPq6IyNA8FfQDwV719EVEhuCtoB+KaAEVEZEheCro+3UjV0RkSJ4J+tGY40g4qvSOiMgQPBP0k3V3\n1NMXERmcZ4I+Dt531gwappWPd0tERCYsz3SLK0oK+NcPnzvezRARmdC809MXEZFhKeiLiOQQBX0R\nkRyioC8ikkMU9EVEcoiCvohIDlHQFxHJIQr6IiI5xJxz492GY5hZO/BWv801wIFxaM6JpuvKPl69\nNq9eF3j32vpf11znXO1wB024oD8QM1vvnGsc73aMNV1X9vHqtXn1usC71zba61J6R0Qkhyjoi4jk\nkGwJ+nePdwNOEF1X9vHqtXn1usC71zaq68qKnL6IiIyNbOnpi4jIGJjQQd/MVprZVjPbbmZfGu/2\njCUz22lmm8xso5mtH+/2jJaZrTKzNjN7NWVblZk9ZmbNid+V49nG0Rrk2r5uZnsSn9tGM3vPeLZx\nNMxstpn90cy2mNlmM/tsYntWf25DXFdWf2ZmVmxmfzazlxPX9feJ7fPN7IVEfLzXzArTOt9ETe+Y\nmQ/YBlwGtABNwIecc1vGtWFjxMx2Ao3OuaweP2xmy4EA8HPn3BmJbf8EdDjnvpn4sq50zv3deLZz\nNAa5tq8DAefct8azbZkwsxnADOfci2ZWDmwArgY+RhZ/bkNc13Vk8WdmZgaUOucCZlYAPAN8Fvg8\n8Gvn3Goz+xHwsnPuh8OdbyL39JcB251zO5xzYWA1cNU4t0n6cc6tAzr6bb4K+Fni8c+I/4+XdQa5\ntqznnNvnnHsx8dgPvAbMIss/tyGuK6u5uEDiaUHixwGXAPcntqf9eU3koD8L2J3yvAUPfIApHPCo\nmW0ws5vGuzFjbJpzbl/icSswbTwbcwLcamavJNI/WZUC6c/M5gHnAC/goc+t33VBln9mZuYzs41A\nG/AY8AZw2DkXSeySdnycyEHf697pnDsXeDdwSyKV4Dkunj+cmDnE0fkhsBBYAuwD/nl8mzN6ZlYG\nPAD8jXOuK/W1bP7cBriurP/MnHNR59wSoI54FuTU0Z5rIgf9PcDslOd1iW2e4Jzbk/jdBvyG+Afp\nFfsT+dVknrVtnNszZpxz+xP/A8aAfyNLP7dEbvgB4D+dc79ObM76z22g6/LKZwbgnDsM/BE4H5hi\nZvmJl9KOjxM56DcB9Yk71IXADcCacW7TmDCz0sSNJsysFLgceHXoo7LKGuDGxOMbgQfHsS1jKhkU\nEz5AFn5uiRuDPwVec879S8pLWf25DXZd2f6ZmVmtmU1JPJ5EfHDLa8SD/zWJ3dL+vCbs6B2AxNCq\n7wA+YJVz7h/GuUljwswWEO/dA+QD92TrtZnZL4GLiFf82w98DfgtcB8wh3jF1Oucc1l3Q3SQa7uI\neJrAATuBT6XkwbOCmb0TeBrYBMQSm79MPP+dtZ/bENf1IbL4MzOzs4jfqPUR76jf55y7IxFHVgNV\nwEvAR51zoWHPN5GDvoiIjK2JnN4REZExpqAvIpJDFPRFRHKIgr6ISA5R0BcRySEK+iIiOURBX0Qk\nhyjoi4jkkP8PF9h5NeLzj9MAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f304d1915f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(depths, accuracy[1][:29])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f3047c5e898>]"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD8CAYAAACb4nSYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f304cd7f358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(depths, accuracy[0][:29])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusion\n",
    "\n",
    "Ignoring a few irregularities, accuracy in general increases with depth upto some value of d and then starts decreasing<br><br>\n",
    "\n",
    "The ideal values of depth for both vector representations have been found<br>\n",
    "tfidf weighted : 18<br>\n",
    "avg word2vec : 16<br><br>\n",
    "\n",
    "Decisions trees provide a significant boost in training speed as opposed to SVMs and also offer better TNR"
   ]
  }
 ],
 "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}