SentiWordNet Lyrics Sentiment Analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to\n",
      "[nltk_data]     C:\\Users\\dtchmnt\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n",
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\dtchmnt\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "[nltk_data] Downloading package averaged_perceptron_tagger to\n",
      "[nltk_data]     C:\\Users\\dtchmnt\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package averaged_perceptron_tagger is already up-to-\n",
      "[nltk_data]       date!\n",
      "[nltk_data] Downloading package wordnet to\n",
      "[nltk_data]     C:\\Users\\dtchmnt\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n",
      "[nltk_data] Downloading package sentiwordnet to\n",
      "[nltk_data]     C:\\Users\\dtchmnt\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package sentiwordnet is already up-to-date!\n",
      "[nltk_data] Downloading package vader_lexicon to\n",
      "[nltk_data]     C:\\Users\\dtchmnt\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package vader_lexicon is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "%load ImportsDefinitions.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Τα NULL δεδομένα στο dataset μας είναι: 0\n",
      "Κάνουμε προεπισκόπηση των δεδομένων μας:\n",
      "Έχουμε 57650 διαφορετικά τραγούδια\n",
      "0    Look at her face, it's a wonderful face  \\nAnd...\n",
      "1    Take it easy with me, please  \\nTouch me gentl...\n",
      "2    I'll never know why I had to go  \\nWhy I had t...\n",
      "3    Making somebody happy is a question of give an...\n",
      "4    Making somebody happy is a question of give an...\n",
      "5    Well, you hoot and you holler and you make me ...\n",
      "6    Down in the street they're all singing and sho...\n",
      "7    Chiquitita, tell me what's wrong  \\nYou're enc...\n",
      "8    I was out with the morning sun  \\nCouldn't sle...\n",
      "9    I'm waitin' for you baby  \\nI'm sitting all al...\n",
      "Name: text, dtype: object\n"
     ]
    }
   ],
   "source": [
    "#pairnoume to dataset me 50000 stixous to opoio den periexei synaishtimatiko skor\n",
    "path='songdata.csv'\n",
    "data = pd.read_csv(path)\n",
    "all_texts=data.text\n",
    "text_samples1=data.text\n",
    "#ελέγχουμε αν εχουμε NULL δεδομένα\n",
    "print(\"Τα NULL δεδομένα στο dataset μας είναι: {}\".format(text_samples1.isnull().sum()))\n",
    "print(\"Κάνουμε προεπισκόπηση των δεδομένων μας:\")\n",
    "#text_samples=text_samples1.sample(100,random_state=1) # pairnoume ena tyxaio deigma apo ta sample mas\n",
    "text_samples=text_samples1\n",
    "\n",
    "print(\"Έχουμε {} διαφορετικά τραγούδια\".format(text_samples.count()))\n",
    "print(text_samples.head(10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "57650\n"
     ]
    }
   ],
   "source": [
    "#PREPROCESSING KAI YLOPOIISI  SYNAISTHIMATIKHS ANALYSIS me SENTIWORDNET\n",
    "#kanoume tokenize tis lekseis\n",
    "my_sentiments_class=SentimentAnalysis_Sentiwordnet(text_samples)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#σώζουμε τη λίστα με την συναισθηματική ανάλυση SentiWordNet. Με την αποκάτω γραμμή, μπορούμε να την φορτώσουμε, \n",
    "#σε περίπτωση που θέλουμε να γλιτώσουμε υπολογιστικο χρόνο και να δοκιμάσουμε διαφορετικά πράγματα στο σετ δεδομένων μας\n",
    "saveList(my_sentiments_class,\"my_sentiments_class.txt\")\n",
    "#my_sentiments_class=loadList(\"my_sentiments_class.txt.npy\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Έχουμε 47095 χαρούμενα τραγούδια\n",
      "Έχουμε 10555 στενάχωρα τραγούδια\n"
     ]
    }
   ],
   "source": [
    "sns.countplot(my_sentiments_class,label=\"Count\")\n",
    "plt.show()\n",
    "print(\"Έχουμε {} χαρούμενα τραγούδια\".format(my_sentiments_class.count(1)))\n",
    "print(\"Έχουμε {} στενάχωρα τραγούδια\".format(my_sentiments_class.count(0)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_s, X_test_s, y_train_s, y_test_s = train_test_split(text_samples, my_sentiments_class,test_size=0.2,random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Παρουσιάζουμε την επιτυχία των vectorizers για τη συναισθηματική κλάση από SentiWordNet ΧΩΡΙΣ ΤΟΚΕΝΙΖΕR\n",
      "Number of features: 73880\n",
      "to accuracy tou CountVectorizer me NB einai: 0.8073934933150149\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of features: 73880\n",
      "to accuracy tou TFIDF me NB einai 0.8149826513687691\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Παρουσιάζουμε την επιτυχία των vectorizers για τη συναισθηματική κλάση από SentiWordNet ΜΕ ΤΟΚΕΝΙΖΕR\n",
      "Number of features: 72672\n",
      "to accuracy tou CountVectorizer me NB einai: 0.8197527259160483\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of features: 72672\n",
      "to accuracy tou TFIDF me NB einai 0.8150043376967286\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#αξιολόγηση CountVect kai TFIDF gia synaisthimatikh analysh me SentiWordNet\n",
    "\n",
    "token_iser='0'\n",
    "scores=[]\n",
    "\n",
    "print(\"Παρουσιάζουμε την επιτυχία των vectorizers για τη συναισθηματική κλάση από SentiWordNet ΧΩΡΙΣ ΤΟΚΕΝΙΖΕR\")\n",
    "scores.append(countvect_test_simple(X_train_s,X_test_s,y_train_s,y_test_s,token_iser))\n",
    "scores.append(tfidf_test_simple(X_train_s,X_test_s,y_train_s,y_test_s,token_iser))\n",
    "\n",
    "token_iser='1'\n",
    "\n",
    "print(\"Παρουσιάζουμε την επιτυχία των vectorizers για τη συναισθηματική κλάση από SentiWordNet ΜΕ ΤΟΚΕΝΙΖΕR\")\n",
    "scores.append(countvect_test_simple(X_train_s,X_test_s,y_train_s,y_test_s,token_iser))\n",
    "scores.append(tfidf_test_simple(X_train_s,X_test_s,y_train_s,y_test_s,token_iser))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>params</th>\n",
       "      <th>scores</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cvec without tokenizer</td>\n",
       "      <td>0.807393</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>tvec without tokenizer</td>\n",
       "      <td>0.814983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cvec with tokenizer</td>\n",
       "      <td>0.819753</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>tvect with tokenizer</td>\n",
       "      <td>0.815004</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   params    scores\n",
       "0  cvec without tokenizer  0.807393\n",
       "1  tvec without tokenizer  0.814983\n",
       "2     cvec with tokenizer  0.819753\n",
       "3    tvect with tokenizer  0.815004"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# DataFrame Vectorizer Accuracy \n",
    "acc_df = pd.DataFrame()\n",
    "acc_df['params']= ['cvec without tokenizer','tvec without tokenizer','cvec with tokenizer','tvect with tokenizer']\n",
    "acc_df['scores']= scores\n",
    "acc_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of features: 5000\n",
      "To accuracy NB me max features 5000 είναι : [0.8189229  0.81656548 0.81630529 0.81568219] \n",
      "Number of features: 20000\n",
      "To accuracy NB me max features 20000 είναι : [0.81293903 0.80832611 0.80537728 0.8074421 ] \n",
      "Number of features: 50000\n",
      "To accuracy NB me max features 50000 είναι : [0.81805568 0.81162186 0.81040763 0.81342701] \n",
      "Number of features: 72672\n",
      "To accuracy NB me max features 100000 είναι : [0.82360593 0.82003469 0.8160451  0.81932518] \n",
      "Number of features: 72672\n",
      "To accuracy NB me max features None είναι : [0.82360593 0.82003469 0.8160451  0.81932518] \n",
      "Τα ποσοστα επιτυχιας του LR για το SentiWordNet με διαφορετικά max Features ειναι: [0.816868967027087, 0.8085211297282874, 0.8133780436108878, 0.8197527259160483, 0.8197527259160483]\n",
      "Number of features: 72650\n",
      "To accuracy NB me Max df: 0.25 είναι : [0.82360593 0.81760624 0.81084128 0.81811085] \n",
      "Number of features: 72670\n",
      "To accuracy NB me Max df: 0.5 είναι : [0.82586072 0.82072853 0.81760624 0.82192731] \n",
      "Number of features: 72672\n",
      "To accuracy NB me Max df: 0.75 είναι : [0.82360593 0.82003469 0.8160451  0.81932518] \n",
      "Number of features: 72672\n",
      "To accuracy NB me Max df: 1.0 είναι : [0.82360593 0.82003469 0.8160451  0.81932518] \n",
      "Τα ποσοστα επιτυχιας του LR για το SentiWordNet με διαφορετικά max DF ειναι: [0.8175410777302485, 0.82153070395903, 0.8197527259160483, 0.8197527259160483]\n",
      "Number of features: 72672\n",
      "To accuracy NB me 1-ngrams είναι : [0.82360593 0.82003469 0.8160451  0.81932518] \n",
      "Number of features: 1378205\n",
      "To accuracy NB me 2-ngrams είναι : [0.82135114 0.81838682 0.8169124  0.8177639 ] \n",
      "Number of features: 4464595\n",
      "To accuracy NB me 3-ngrams είναι : [0.80912323 0.80494363 0.80208153 0.80379912] \n",
      "Number of features: 8160494\n",
      "To accuracy NB me 4-ngrams είναι : [0.75544185 0.75550737 0.75221162 0.75219013] \n",
      "Τα ποσοστα επιτυχιας του LR για το SentiWordNet με διαφορετικά n-grams ειναι: [0.8197527259160483, 0.8186035648907053, 0.8049868750191295, 0.7538377438915187]\n"
     ]
    }
   ],
   "source": [
    "#parametropoihsh vectorizer\n",
    "ngrams=[1,2,3,4]\n",
    "mdf_params = [0.25, 0.5, 0.75, 1.0]\n",
    "maxfeat=[5000,20000,50000,100000,None]\n",
    "\n",
    "#θα υλοποιησουμε διαφορετικους count vectorizers με logreg για να βρουμε την καλυτερη επιδοση\n",
    "#θα ορισουμε τις λιστες στις οποιες θα αποθηκευσουμε τα αποτελεσματα\n",
    "ngram_scores=[]\n",
    "mdf_scores=[]\n",
    "maxfeat_scores=[]\n",
    "\n",
    "\n",
    "token_izer='1'\n",
    "\n",
    "for mf in maxfeat:\n",
    "    maxfeat_scores.append(countvect_test_maxfeat(X_train_s,X_test_s,y_train_s,y_test_s,token_izer,mf))\n",
    "print(\"Τα ποσοστα επιτυχιας του LR για το SentiWordNet με διαφορετικά max Features ειναι: {}\".format(maxfeat_scores))\n",
    "\n",
    "for mdf in mdf_params:\n",
    "    mdf_scores.append(countvect_test_maxdf(X_train_s,X_test_s,y_train_s,y_test_s,token_izer,mdf))\n",
    "print(\"Τα ποσοστα επιτυχιας του LR για το SentiWordNet με διαφορετικά max DF ειναι: {}\".format(mdf_scores))\n",
    "for n in ngrams:\n",
    "    ngram_scores.append(countvect_test_ngrams(X_train_s,X_test_s,y_train_s,y_test_s,token_izer,n))\n",
    "print(\"Τα ποσοστα επιτυχιας του LR για το SentiWordNet με διαφορετικά n-grams ειναι: {}\".format(ngram_scores))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   params    scores\n",
      "0       1  0.819753\n",
      "1       2  0.818604\n",
      "2       3  0.804987\n",
      "3       4  0.753838\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEHCAYAAABBW1qbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xl8lNXZ//HPJCEJWwBBWZQduVCCLEEBBUVU3AF99LHWLvZxr8Uq4lJ3rbtQtNal+09ra7UqLqgoLgVkUwIIEbgisiqo7Fsg+++PGYYBgQyQyZ1kvu/Xi1dz5tx35hsLc+U+97nPCZWXlyMiIgKQEnQAERGpPlQUREQkSkVBRESiVBRERCRKRUFERKLSgg5wMHJzczOAY4FVQGnAcUREaopUoCXwWU5OTmFsR40uCoQLwuSgQ4iI1FADgE9iX6jpRWEVQOfOnUlPTw86i4hIjVBUVER+fj5EPkNj1fSiUAqQnp5ORkZG0FlERGqaHwy760aziIhEqSiIiEiUioKIiESpKIiISJSKgoiIRNX02UfV2uK1m3lx1hJWbtpGs/oZXNSjHUe3aBx0LBGRvVJRSICS0jKuf/0znp2WT+x2FfdPmMf/9mjLXy86nnrp+k8vItVPwj6ZzCwFeBroDhQCl7v7opj+kcDFQBnwoLuPNbNGwAtAFpAOjHD3aYnKmCg3vDGTZ6bm77Hv5TnLKC4t55VLT6riVCIiFUvkPYVhQKa79wNuBUbv6DCzxsB1QD9gMPB4pGsE8KG7nwRcCjyVwHwJsXTdFp6Z6vs8Zuy85Xy2fE0VJao9ikpKmbrkeyb4Spat2xJ0HJFaKZFjGP2B8QDuPt3Mesf0bQWWAfUjf8oir48hfFWxI9v2BOZLiJdmLyWeHU5P/+MHtDukAfXqpFEvPZV66Wk7v66TFmnHvh573A+Pr58ePqZOau2bO1BWVs6jH+fxxKSFfL9l51+J0zq3ZPTQ3nTVfRqRSpPIopAFbIxpl5pZmruXRNorgPmEV+t7CMDdNwCYWQvCw0jXx/NGeXl5lZX5oH2+6Nu4jtu4vZjPV66v9PdPDUFmWgqZqSlkpoXITE2hbloKGZGvw32hXf93t+MrOiY9JUQoFKr07HtSXl7Ob2esZNzijT/om5C/in6Pv80fT21H5yaZVZJHpLZLZFHYBDSMaafEFIQzCS/b2j7Sfs/Mprj7p2bWDfg3MNLdJ8bzRtnZ2dVm7aOu6+ZC/rrA3r+0HLYWl7G1uKzigw9QKET0KqV+7JVMnVTq7vEKZufVT92Yq5zd+3ZeAaVSt04aKSkh3lu4knGLF+w1y9biMv4wfxNTrjshYT+vSG1TWFi411+mE1kUpgDnAi+bWV9gXkzfemAbUOju5Wa2AWhsZkcD/wEucvfPE5gtYS7q2Y67xlcc/aGze9K2SQO2FpWwrbiEgqJSCopLKCgqoaC4lIKiErYWlVBQXMK2aN/OY7ZGjisuTdyH/96Ul8PWSIbVFFZ8wgGqWyeV0rKKx+KmL1vD3JXrOaZVk4RlEUkWiSwKY4HTzGwqEAJ+YWYjgEXu/qaZnQpMN7Mywut5TwBeBzKBJ8wMYKO7D01gxkrXqVkWPz+2I8999tVejxlsrbjp5K6VMgRTXFq2x6KytWjXAhNbXLYWlvygyBQUl7LtB+eGj9leEsz+RduK43/featUFEQqQ8KKgruXAVfv9vLCmP67gbt3669RBWBvnrmgD8WlZfxr1pIf9J3epRX//umAShuTr5OaQp3UdLISOKReVlYeLjx7KTa7FJe99pXGXBXt+Zh4btDvzQdffsuw7NbUz6hTeT+4SBIKlR/Mv8SA5ebmtgOWVKd7CrHmrlzPC7mLWbmxgGYNMrm4ZzuOa9Osym7S1iTl5eVsLykNF4iYYvHwh/N4Ze7yuL5H03oZ/Kq/cW3/LjStX/3+PohUFzH3FNrn5OQsje1TUZBq7cvVm+j22Fv7de+kXnoqV/Q9khtOPJrWTeonMJ1IzbSvolD7JrVLrXLkoVk8d/EJpKXs+eqq1+GHcEWfI8lMS42+VlBUyhOTFtLpwbH837+nsuC7H05nFZE90wI8Uu1d1LMdRzVvxBOTFvD2gq8pKCrFDsvi//p04v+O60RGWir3ndmdJz9ZyNNT8tmwrQiAkrJynvvsK56f+RVDs1tzy6BsjmvTLOCfRqR60/CR1Cqbthfx52lfMmbSAlZt2vaD/pM7NefmQdmc1rml7u1I0tLwkSSNrMx0bjy5K1/dfh5/vLAvRzZruEv/x4u+48w/fcixY97h5TlLKS2r+uc8RKozFQWplTLSUrm875F8ccsQXvrZifQ64pBd+md/s46L/zGZox95kz9Ny6cwoGcxRKobFQWp1VJTUrige1s+vf4sxl95CoM6tdilf9GazVzzygw6PjCWUR9/wabtRQElFakeVBQkKYRCIU6zVky45jSm/fpMzuvWhthbCqs2beOWcbNo99vXuOOd2Xy3+Yf3I0SSgYqCJJ3j2jTjlUtPIu+mIfziuI67LDe+cXsxD32YR4f7x/KrV2ewZO3mAJOKVD0VBUlaXZo34i8XHc+i24Zxw0lHUT9mi9TtJaU8MzUfe/gNfvLCZOYmYJlzkepIRUGS3hGN6zNqSG+W3HE+95zenab1dk5vLi0r58XZS+k5ehzn/OUjJi/+LsCkIomnoiAS0bR+BncOPoYld5zH48N607pxvV36313wDQOfep8TnxzPuPlfUxbHst4iNY2Kgshu6mfUYfiAo/jytvP4+8XHc3TzRrv0T1m6mqF//Zieo8fxQu7iQPa0EEkUFQWRvaiTmsLPenfk85HnMvYXA+nbdtclMvK+3cDP/zUFe+h1nvpkIQVFJXv5TiI1h4qCSAVSUkIMyW7NJ8PP4ONfDuaMLq126V+2fivXjf2M9ve/xgMT5rK+IHG70YkkmoqCSJxCoRAndmzO21ecQu6Is7moRztSYh52WLO1kLvGf067+1/jpjdzWbmxIMC0IgdGRUHkAPQ4/BD+9dMBLLx1KFf160xG2s5/SlsKS/jdxPl0fGAsV748jfzVmwJMKrJ/VBREDkLHZg15+oI+LL79fG4Z1JWszJ3bgRaVlvHXGYs4+pE3+N/nJjJzxdoAk4rER0VBpBK0yKrLg2f3Yukd5/PQ2T1p3nDnptnl5fDq3OX0efwdBj87gQ/zV1GTl6yX2k1FQaQSNaqbzs2Dsll8+/k8fUEfOjRtsEv/h19+y+A/fkDfJ97l1bnLtHS3VDsJ23nNzFKAp4HuQCFwubsviukfCVwMlAEPuvvYmL7zgAvd/ceJyieSSJl1UrmqX2cuO64Tr8xdxqMffcHnMUtlzFyxlv99bhKdD81i5MlH85OcDmTEbCkqEpREXikMAzLdvR9wKzB6R4eZNQauA/oBg4HHY/qeAB5KcDaRKpGWmsKPerYnd8TZvH3FIE7q2HyX/vzVm7jy5ekc+eDrjJk4n83biwNKKhKWyA/e/sB4AHefDvSO6dsKLAPqR/7EXkNPBa5JYC6RKhcKhTijy+F89MvBfDL8DIZ0PWKX/m82FjDyzVza3/8ad4+fw+ot2wNKKskuYcNHQBawMaZdamZp7r7jsc8VwHwglfCVAQDu/pKZDdyfN4rsNSpSI6QDd3TP4sftOvKP+WsYv3QjpZH7zuu3FXH/hHk89lEeQzs24ZKjDqFl/fRA80pySWRR2ATEbpCbElMQzgRaAu0j7ffMbIq7f3ogb5SdnU1GRkbFB4pUIznAhYNg+fqtjJk4n7/M+JKCovC2oIWl5bycv45XF63n4p7tuenko8lu2STYwFJrFBYW7vWX6UQOH00BzgIws77AvJi+9cA2oNDdtwMbgMYJzCJSbbVpUp8xw45lye3nc+dpx9Ck7s4rg9Kycl7IXUz3UeMY+tePmbZ0dYBJJRkksiiMBbab2VRgDHCDmY0wsyHuPhn4DJhuZtOAfGBCArOIVHvNGmRyzxndWXrn+YweksPhjXZdunvc/K/p/+R4Tn7qPd5d8I2edZCECNXkv1i5ubntgCUaPpLaqKiklH/OWsJjH32B72GpjGNaNuHmQV25sHtb0lI1WU/iFzN81D4nJ2dpbJ/+JolUU+lpqfziuE7k3TyEVy49iWNbN92lf+6q9fzkn5/Q5eE3eGaqs61YS3fLwVNREKnmUlJCnNetDdN+fSYTrj6VUzu33KV/ybot/OrVT+lw/1ge/nAeG7YVBZRUagMVBZEaIhQKMejIlrx31al8ev1ZXNC9LTErd/P9lu3c/s4c2t//Gr8ZN4tVm7R0t+w/FQWRGiindVNe+tmJLLhlKJf37UR6zD2FTduLefTjL+j4wFiueWU6i9bsfenu8vJy3bCWXehGs0gtsHJjAU9MWsCz0/LZUrjrvYWUUIgLurfh5pOz6XnEIZSVlfPCrMU8OyWfz1asJSUEAzocxrX9uzAsuzWh2MsPqZX2daNZRUGkFllfUMizU/N5YvICVm/54baggzu3pKi0jP9+9d0ezx8+oAtjhvZWYajlNPtIJEk0qZfBb07txpI7zucP5x9Hu0Pq79L/fv6qvRYEgCcnL+SlOUsTnFKqMxUFkVqobp00rjnB8FuH8fyPTyC7RfwLBvx+0sIEJpPqTkVBpBZLS03hkpwOzBl5Dn+8sE9c58xYvoaCIj3zkKxUFESSQCgUon/75hUfGFFcqh3hkpWKgkiSaHdIAxpl1qnwuA5NG5AVx3FSO6koiCSJzDqpXHpcxwqPu7pfZ80+SmIqCiJJ5K7B3enWcu83nZvWz+BXA7pUYSKpblQURJJI47rpfPzLwVx9fGfqp/9wj621WwuZqj0bkpqKgkiSaVIvg6f+pw/f3H0BU687g2m/PpMeh+/c1e3GN2ZSWqYbzclKRUEkSTXMrEOftodyXJtmPD7s2Ojrn69cz//77KsAk0mQVBREhAEdmnP+MW2i7TvfncPm7cUBJpKgqCiICACPnNMrutrqd5u38/BHe97YXWo3FQURAaBD04b8+sSjou0xE+ezdN2WABNJEFQURCTqtlOzObRBeMXhwpIyfvP2rIATSVX74Zy0SmJmKcDTQHegELjc3RfF9I8ELgbKgAfdfayZ1QVeAA4DNgM/d3fNjxOpIlmZ6dx7Rg9++coMAF6es4zh/b/n+PaHBZxMqkoirxSGAZnu3g+4FRi9o8PMGgPXAf2AwcDjka5rgHnuPgB4HrgjgflEZA8uO67TLg+4jXhjJmVlNXffFdk/iSwK/YHxAO4+Hegd07cVWAbUj/wp2/0c4F3g1ATmE5E9SEtNYdSQnf9cP1uxln/NXhJgIqlKCRs+ArKAjTHtUjNLc/cda/KuAOYDqcBDezhnM9AonjeK7CAkIpWkCdC/VQM+WRm+0XzT2Bm0K1lL3TTdhqztElkUNgENY9opMQXhTKAl0D7Sfs/Mpux2TkNgQzxvpO04RSrfn1p34pjH3qKkrJzvt5Xw0YZ07jq9e9CxpBLEbMf5AxWWfTNLN7Pbzex5M8sys7vMLD2O950CnBX5Hn2BeTF964FtQKG7byf84d849hzChWNyHO8jIglghzXilydYtP3Yf7/gm40FASaSqhDPteBThMf9ewElQCfgb3GcNxbYbmZTgTHADWY2wsyGuPtk4DNguplNA/KBCcAzQFcz+wS4Erh3f38gEak8dw4+hiZ1w78DFhSVcvs7swNOJIkWKi/f96wCM5vl7r3MbLa79zSzEOEZQtlVE3HvcnNz2wFLNHwkkjhPTl7A9a/PjLZnXH8WvVs3DTCRHKyY4aP2OTk5S2P74rlSKI8MF+2oHs1ivhaRWu7q4w07NCvaHvH6Z1T0y6TUXPEUhceBD4AWZvY4MJPwcJCIJIE6qSk8NiQn2p6ydDWvzF0eYCJJpHiKwrvA1cADwGLgXHeP556CiNQSZx11OKd2bhlt3zoul+3FpQEmkkSJZ0rqZHc/ivAzBSKShEKhEKOH5NBz9NuUlZezdN1Wfj95ATcPCvzWolSyeIrC52b2U+BTwtNIAXB3XT+KJJHslk24ou+R/HFaPgAPfpDHz4/tSPOGdQNOJpUpnuGjPsB9hJefmBj5898EZhKRaureM7qTlVkHgM2Fxdw1fk7AiaSyVXil4O7tKzpGRJLDoQ0yuf3UbtwyLryk9t9mfMW1J3ThmFZNKjhTaop4nmg+1MxeMrM1ZrbBzF4zs+ZVEU5Eqp/hA7rQoWkDAMrKyxn55kxNUa1F4hk++iPhp487AO2A6cBfE5hJRKqxjLRUHjln5xTVD7/8lre++DrARFKZ4ikKHdx9lLtvcvcN7v4o0DbRwUSk+jqvW2tO6rhzwODmt3IpKtEU1dog3ieaW+9omFkboDhxkUSkuguFQowakkMoFG5/uWYzz0zNDzaUVIp4isKdwDQze9XMXgOmRV4TkSTW64im/Lx3x2j7vvfnsnZrYYCJpDJUWBTcfRzQk/DKqH8Herj724kOJiLV3/1n9aB+engS44ZtRdz73ucBJ5KDFc/so5OB1yOFIB+YYWbHJzyZiFR7LbPqccugrtH2s9PyWfDdxn2cIdVdPMNHo4GrANzdCW+C80QiQ4lIzTFi4NG0blwPgNKycm56KzfgRHIw4ikKme4e3bfN3RcCdRIXSURqkrp10njo7F7R9rsLvuG9hSsDTCQHI56isNDMHjGzbDPramb3Ex5GEhEB4Ec929G3bbNoe+SbMykpLQswkRyoeIrCZYS343wR+AfQALgikaFEpGYJhUKMHto72p7/3Ub+POPLABPJgYpn9tF6d/+Vu3cDTgNucHfdSRKRXfRteyg/6tku2r5n/Ods2FYUXCA5IHstCpE1j14xs4FmlhJ5RmEZsMjMjq66iCJSUzx0di8y01IBWLO1kAcmzAs4keyvfV0pPEl4682ZwIVAL6AV8BM0+0hE9qBNk/rcOHDn74xPfrKQRWs2BZhI9te+ls4+2t1/BGBmZwIvu/smwk83t6roG5tZCvA00B0oBC5390WRvh6E937eoS8wDHDgOSBE+KrkSncv2O+fSkQCc/Ogrvzt00Ws2rSN4tIybhk3i1cvHRh0LInTvq4UYtfCHQR8ENOuF8f3HkZ4Oms/4FbCzzsA4O5z3H2guw8EngJec/fxwGPAs+4+gPBGPiPi+SFEpPpokFGH+8/sGW2/Pm8FHy/6NsBEsj/2VRSWmdlFZvYLwkXgvwBm9hPgizi+d3/Cu7Xh7tOB3rsfYGb1gXuB6yIvHQ28G/l6SuR7iEgN87PeHeh1xCHR9sg3ZlJapimqNcG+ho+uJbyXQnPgx+5eZGa/A84l/FRzRbKA2FlKpWaW5u4lMa9dBvzH3ddE2nOAIYSHkIYQngpboby8vIoPEpEqdWWXhlz99ToA5qxcz32vfMSQjtqhrbrba1Fw9xX88MP/t8BId4+n5G8CGsa0U3YrCACXABfEtG8E/mBmFwMfAmuIQ3Z2NhkZGfEcKiJVJAd477uJjJ23HIC/zN/AyCEn0TBTCyIErbCwcK+/TMfz8FpU5JmFeK8BpxApKmbWF9hlbpqZNQIyIsVnh9OAe939DKAMmLA/+USkennknF6kp4Y/Zr7dvI1HPtJVfXW3X0VhP40FtpvZVGAMcIOZjTCzIZH+zsDS3c5x4G9mNgUw4M8JzCciCdaxWUOuG9Al2v7dxPksW7clwERSkVBFG26bWQt3r5ZTB3Jzc9sBSzR8JFJ9bdxWhD38Oqu3hDfguahHO/710wEBp0puMcNH7XNycpbG9sVzpTDJzN42swvNLD0RAUWk9mpUN517Tu8Rbb80ZylTl3wfYCLZl3jWPuoMPAycTnjF1D+Y2Q+ml4qI7M3lfTqR3aJxtH3jmzMpK9v3KIUEI657Cu4+GRgO3AMMBV4zs9zIDWQRkX1KS01h1JCcaPvT5Wt5cfaSABPJ3sSzHecpZvYcsAgYAFzk7m2AS4FXEhtPRGqL06wVZx11eLR929uzKSjafZa6BC2eK4W7gY+AI939CnefCuDu84BRiQwnIrXLY+fmkJYSAuDrjQX8buL8gBPJ7uIpCmcDDdy9wMwON7P7zKwegLs/XsG5IiJRXZo34poTLNp+5KM8vtmoNS+rk3iKwj8JL5kNsDlyzj8SlkhEarW7Bh9Dk7rhiYwFRaXc8c7sgBNJrHiKQlt3vx3A3Te5+x1Ax8TGEpHa6pB6Gdw1+Jho+/mZi8ldsTbARBIrnqJQbmbddjTMrAtQnLhIIlLbXXOCYYdmRdsj3phJRQ/SStWIpyiMBCaY2Uwzmwm8R3jhOhGRA1InNYVHY6aofrLke16duzzARLJDPA+vfQC0Aa4C/g/o7O6TEh1MRGq3s486nFOObBFt3zpuFtuLSwNMJBDfcwpHEt4R7ZfA9cCfzExFQUQOSigUYvTQ3qSEwlNUl6zbwpOTFwacSuIZPnoR2AD0JLwJThtA69+KyEHr1rIJl/ftFG0/8ME8vtu8LcBEEk9RSHf3uwlvrTmL8B4JJyU0lYgkjXtP705WZOOdzYXF3D3+84ATJbd4ikKBmWUA+UCOu6uMi0ilOaxhXW47JTrBkb/OWMS8VesDTJTc4ikKLwBvAW8Dw83sXeCbhKYSkaRy3Yld6NC0AQBl5eXcqCmqgYlrPwXgf9x9NTAQ+BNwXiJDiUhyyUhL5eFzekXbH375LePmfx1gouSVFscxL7n7UQDu/jWg/6dEpNKd360NJ3Y4jEmLwxvw3PzWLE63VqSnpQacLLnEUxTmm9ldwAwgej9BzyqISGUKhUKMGtKbPk+8Q3k55K/exLNT87nuxKOCjpZU4ikKhwAnR/7sUA4MSkgiEUlaOa2b8rPeHXnus68AuO/9uVyS04Gm9bUHe1WpsCi4+8kVHSMiUlnuP7MHr3y+jK1FJazfVsR973/OE+cdF3SspFFhUTCzjwlfGezC3fd5pWBmKcDTQHegELjc3RdF+noAsXsx9AWGAfMJL8sdAtYBP3Z3LbYukkRaNarHzYO6Rp9XeGZqPtccb3Rp3ijgZMkhntlH9wD3Rv48CHwBfBLHecOATHfvB9wKjN7R4e5z3H2guw8EngJec/fxwA2Eb2yfGHmfy+L/UUSkthhx0tEc0ageAKVl5dz0Vm7AiZJHPAviTYz584G7DwdOj+N79yf8FDTuPh3ovfsBZlafcLG5LvLSHKBJ5OsstES3SFKql57GQzFTVN9Z8A3v+8oAEyWPeIaP2sQ0Q0BXoGkc3zsL2BjTLjWzNHeP3an7MuA/7r4m0v4aeNjMfgxkEL5KqVBenpZiEqltOpeXk920Lnlrw5Mer33pE144s0N0j2dJjHhmH02M+bocWA0Mj+O8TUDDmHbKbgUB4BLggpj2Y8Cl7v6emZ0NPE94j+h9ys7OJiNDsxNEaptnm7Wl/5PjAVi8sZDZxVlcfXzngFPVfIWFhXv9ZTqe4aP2hPdQaA8YMMjd343jfacQXjwPM+sLzIvtNLNGQIa7r4h5eT07ry5WsnMoSUSSUL92h3JRj3bR9t3j57BxW1FwgZJAPPspXEh4dVQIL5u90MyGxvG9xwLbzWwqMAa4wcxGmNmQSH9nYOlu5wwHHjSzicATwLVxvI+I1GIPnd2TzMhTzWu2FvLAB/MqOEMORqiiRafMbC5wmrt/F2kfBrzv7j2qIN8+5ebmtgOWaPhIpHa7893ZPPhBeLijTmoKX9w8hI7NGlZwluxNzPBR+5ycnKWxffHup/Ddjoa7f0/4hrOISJW4ZVA2LRrWBaC4tIxbxs2q4Aw5UPHcaP7EzF4E/kn4RvOPgGkJTSUiEqNBRh3uP6sHl78U/ugZO285E7/6jpM6Ng84We0Tz5XCtUAucBXhKaQz2flcgYhIlfh57470PPyQaPvGN2ZSWlYWYKLaKZ6iUAfY5u7nEr4R3JT4rjBERCpNSkqI0UN3PgM7+5t1PD9zcYCJaqd4isK/gFaRrzdHzvlHwhKJiOzFSR2bM6xb62j7jnfmsKVQCx9UpniKQlt3vx3A3Te5+x1Ax8TGEhHZs0fPyaFOavij69vN23j0oy8CTlS7xFMUys0suqu2mXVBaxKJSEA6NmvIdQO6RNuj/zuf5eu3BpiodomnKIwEJpjZTDP7DHgPGJHYWCIie3f7qd1oFtl4Z3tJKb95W1NUK0s8y1x8QPhJ5muAtwgvPxHPMhciIgnRqG4695zRPdr+9+ylTFu6OsBEtUc8y1y0J7y89TjgdsIFoV1iY4mI7NsVfY6ka4udG+/c+MZMysr2vUKDVGyvRcHMzjOz8cCnhKeh/gRY5e73xSx1LSISiLTUFEYN2TlFdcbyNfx7ztLgAtUS+7pSeBXYAPRz9yvdfQKgJ0VEpNoYbK0486jDo+3b3p5FQdHuK/TL/thXUTgG+IbwMhfTzezX6KE1EalmRp2bQ2pk450VGwoYM3F+wIlqtr0WBXfPc/cbgSOAh4GTgeZm9raZnVVVAUVE9qVL80ZcE7PxzsMf5bFyY0GAiWq2eGYflbj76+4+jHCB+Ah4KOHJRETidNfg7jSpmw5AQVEpd7w7J+BENVc8zylEuftqdx/t7t0rPlpEpGo0rZ/BnYOPibafn/kVs75eG2Cimmu/ioKISHV1zfGd6XxoFgDl5eEpqhVtIiY/pKIgIrVCeloqj57bK9qetPh7Xpu3PMBENZOKgojUGuccfQSnHNki2r513CwKS0oDTFTzqCiISK0RCoUYNaQ3KaHwFNXFa7fw5OSFAaeqWVQURKRWOaZVEy7r0ynavn/CPL7fvC3ARDVLwh5GM7MU4GmgO1AIXO7uiyJ9PYDHYw7vCwwDzgB6RF5rAWxw976JyigitdO9Z3Tn37OXsrmwmM2Fxdz93uc8c4E+SuKRyCuFYUCmu/cDbgVG7+hw9znuPtDdBwJPAa+5+3h3vz7y2mnARuCKBOYTkVqqecO63HZqdrT9l+mLmLdqfYCJao5EFoX+wHgAd58O9N79ADOrT3gF1ut26xoOvO/u8xKYT0RqsesGHEX7QxoAUFZerimqcUrkWkZZhH/b36HUzNLcPXa1qsuA/8Suumpm6cBVwHHxvlFeXt7BZhWRWujKoxvzm0+2APDhl9/yxFsTGXB4w4BTVW+JLAqbgNj/+im7FQT+Wf3cAAAMVklEQVSAS4ALdnvtVGCSu28kTtnZ2WRkZBxYShGptXr1Kuedle8zefH3APxp/kauPfvE6B7PyaqwsHCvv0wn8r/MFOAsADPrC+wyFGRmjYAMd1+x23mnop3dRKQS7JiiuoOv3sSzUz3ARNVfIovCWGC7mU0FxgA3mNkIMxsS6e8MLN3DeQYsTmAuEUkivVs35We9O0Tb9743l3UFhQEmqt5CNfnGS25ubjtgiYaPRGRfVm4swB5+nYKi8NPNwwd04fFhxwacKjgxw0ftc3Jylsb2JffAmogkhVaN6nHzyTunqD4zxfHv475tmVRUFEQkKdw48GiOaFQPgJKycm56KzfgRNWTioKIJIV66Wk8eHbPaPvt+d8wwVcGmKh6UlEQkaRxcc/2HNemabQ98s1cSkrLAkxU/agoiEjSSEkJMTpmimretxv466eLAkxU/agoiEhSOb79YVzUo120fff4OWzcVhRcoGpGRUFEks5DZ/ckMy0VgNVbCnnwAy2ztoOKgogknbaHNOCGk46Ktn8/eSGL124OMFH1oaIgIknplkHZtGhYF4Ci0jJuGTcr4ETVg4qCiCSlhpl1+O2ZPaLt1+YuZ9JX3wWYqHpQURCRpPXzYzvQo1WTaPvGN2dSVlZzl/6pDCoKIpK0UlNSGD105xTVWV+v4/mZyb0ep4qCiCS1gZ1aMDS7dbR9x7uz2VJYHGCiYKkoiEjSe/TcXtGNd1Zt2sZjH38RcKLgqCiISNLr1CyL4f27RNujPp7P8vVbA0wUHBUFERHg9tO60ax+eF+W7SWl3PZ2ck5RVVEQEQEa103nnjO6R9svzl7KjGWrA0wUDBUFEZGIK/ocSdcWjaLtEW/MpCbvTnkgVBRERCLSUlN47NydU1SnL1vDv2cvDS5QAFQURERinN6lFWd0aRVt/+btWWwrLgkwUdVSURAR2c2oIb1JTQkBsGJDAWMmLgg4UdVJS9Q3NrMU4GmgO1AIXO7uiyJ9PYDHYw7vCwwDJgPPAO2BdGC4u3+aqIwiIntyVPNGXN2vM09NcQAe/jCPS4/tSKvIHs+1WSKvFIYBme7eD7gVGL2jw93nuPtAdx8IPAW85u7jgZuAPHcfAFwBWALziYjs1d2nd6dx3XQAthaVcOe7cwJOVDUSWRT6A+MB3H060Hv3A8ysPnAvcF3kpdOBIjN7D7gTeC+B+URE9qpp/QzuPK1btP3czK+Y9fXaABNVjYQNHwFZwMaYdqmZpbl77B2by4D/uPuaSLsZ0MTdTzeznwGjgJ9V9EZ5eXmVlVlEJKpv3XJaN0xnxeYiysvhqn/+l2dOaUsoFAo6WsIksihsAhrGtFN2KwgAlwAXxLTXAm9Gvn6L8LBThbKzs8nIyDjQnCIie/X7es057+//BWDW9wUsTz+M849pE2yog1RYWLjXX6YTOXw0BTgLwMz6ArtsgmpmjYAMd18R8/InO84BTgSSd1UqEakWzu16BIM6tYi2bxmXS2FJaYCJEiuRRWEssN3MpgJjgBvMbISZDYn0dwaW7nbOg0BPM5sG3Ej4xrOISGBCoRCjh/YmJTJktHjtFv4weWHAqRInVJMf4c7NzW0HLNHwkYgk2lX/mcZfpi8CICuzDvm/GcahDTIDTnVgYoaP2ufk5CyN7dPDayIicbjvjB40zKgDwKbtxdw9/vOAEyWGioKISByaN6zLb07Jjrb/PP1L8latDzBRYqgoiIjE6dcnHkW7Q+oDUFZezsg3c2vdKqoqCiIiccqsk8rD5+RE2xPyV/HuwpUBJqp8KgoiIvvhgmPa0L/9YdH2yDdmUlxaFmCiyqWiICKyH0KhEKOG7Lxa8NWb+OPU/AATVS4VBRGR/XRsm2b8tHeHaPve9z9nXUFhgIkqTyKXuRARqbUeOKsnr85dRkFRKesKirj4+Un0a3cYjerWYWh2azo0bVjxN6mGdKUgInIADm9Uj5sGdo22P/jyW347YS4j38yl80Ov85MXJrO1sDjAhAdGRUFE5ABlRR5m2115Obw4eyn/+/wkyspq1pRVFQURkQOwtbCY+ybM3ecx4xeu5MMvV1VRosqhoiAicgDe+OJrNm6veHjo/332VRWkqTwqCiIiB2DF+q2Velx1oaIgInIAmtRLj/O4mrWCs4qCiMgBGNK1NXVSK/4IvaB72ypIU3lUFEREDkCLrLpc0ffIfR5jh2ZxoYqCiEhyGD0kh4t7tttjX5fDsnjnylPIrJNataEOkp5oFhE5QOlpqfzjkv4MH9CFv3/6FUvWbaFx3XTO79aG87q1Jj2tZhUEUFEQETkooVCIPm0PpU/bQ4OOUik0fCQiIlEqCiIiEpWw4SMzSwGeBroDhcDl7r4o0tcDeDzm8L7AMOBTIB/Ii7w+1t2fSFRGERHZVSLvKQwDMt29n5n1BUYDQwHcfQ4wEMDMLgRWuvt4MzsVeNHdhycwl4iI7EUii0J/YDyAu083s967H2Bm9YF7gRMjL+UAvcxsIvA9cJ2716zVpEREarBEFoUsYGNMu9TM0ty9JOa1y4D/uPuaSHshkOvuH5jZJcCTwAUVvVFeXl5Fh4iISBwSWRQ2AbFbD6XsVhAALmHXD/2PgILI12OB+yp4j1SAzp07k54e3zokIiLJrqioiPz8fIh8hsZKZFGYApwLvBy5pzAvttPMGgEZ7r4i5uW/AK8CLwOnALkVvEdLYMcPJyIi+6clsMva3oksCmOB08xsKhACfmFmI4BF7v4m0BlYuts5twJ/M7NfAluByyt4j8+AAcAqoLQSs4uI1GaphAvCZ7t3hMrLa9ZWcSIikjh6eE1ERKJUFEREJEpFQUREolQUREQkSkVBRESiVBRERCRKm+xIjRBZOLEZsMrdXw86T01nZkMIL0Oz3d1nBJ2nJjOz/wGKI89f1Xi6UkgQMxtiZg8FnaM2MLPrgGsJL53ygpkNCjhSjWZmw4FbgF7Aw2a2793npSL9gfvMrGfQQSqDikLiHAH82MwuDTpITWZm9YDBwK/c/Z/AbUCnYFPVXGZWFzgDuNrdxwBfAE3NrHbsJVmFzCwU+XI9sBa40cwam1mrAGMdNBWFBDCzBsCxwGOEl/c4K+BINVkpMBvY8eh9a/T39mAUA9OAbZHfbC8Ffg7MM7MmQQargXYUhc3ARcAywuu13RlTMGoc/eNKAHffAvyV8AJ/jwN3mFnXYFPVTO5eCIxy9y8iLx0JLAAws+sDC1ZDRVYqfjSyC+J3wAB3vwZ4g/AVhMTJ3csiX24GziT8C8s24Dt3r7HrB6koJIi7f+Lu2919LPAC8LR+Ezsw7h67L8dywr/lvgi0DyhSjebuRZH/XQmsMLPGQDvCC0vK/isGbgc+IDxCMNjMTtz3KdWXZh9VjWcJj4P3A94JOEtNdzYwBLjX3f8edJiazMz6Az8GugH/dvf/BpuoxvoX8I27fwBgZue7+7cBZzpgWiW1iphZqrtree+DYGapwDOEP8A+CjpPTWdmaYQ3wmoVMzwnB8HMUoDymjx8pKIgNcoetnQVkUqkoiAiIlG60SwiIlEqCiIiEqWiICIiUSoKIiISpaIgIiJRenhNZDdmNpDwwnsFwFHAPMIPeV0NDAc2AAuBr9z9HjNbDcwEWhJ+ovVpIBtoDswFLo58/XrkvK7ALGAq4bWHmgDnufsCMxsFnAaUAa+7+72J/4lFdtKVgsieHQ/8inBRaANcQ3j57hxgAOE1mHZoBjzi7j0IP7Ve5O79CD/F3hjYsSDiMcAjQHfgBKBd5LgXgSvNrC1wprvv6D/azDIT+lOK7EZXCiJ7lufuXwOY2QLCH+7j3H1T5LUXCf+Gv8MMAHefZGZrzexaoAvh4tEgcsy37j47cv7XwIeR15cRXsfpG8LrOk0BxgG3uPv2BP6MIj+gKwWRPYv9MC4nPGS0138v7r4Nojua/ZPw0NPfgUnsXGK5aLfTdnkyO/Kkdh/gTqApMM3MOh/4jyCy/1QUROJTCpxlZllmlg78Dzv3eIh1KvByZLG+DcDJQGo8bxDZ32AiMMndRwLzAauM8CLxUlEQic8W4PeEN6iZTHgN/W17OO7PwMVmNg/4DzCFOJf4jgwtTQPyzGwW4aLw7sFHF4mf1j4SiUNkGOfsyBaWmNkbwF/c/a1gk4lULt1oFonPMuBYM8sjPGz0HuGbwSK1iq4UREQkSvcUREQkSkVBRESiVBRERCRKRUFERKJUFEREJEpFQUREov4/vQ58uhTpns0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ngrams=[1,2,3,4]\n",
    "ngram_params = ['cvec gram_1','cvec gram_2','cvec gram_3','cvec gram_4']\n",
    "ngram_df = pd.DataFrame({'params':ngrams, 'scores':ngram_scores}, index=[0,1,2,3])\n",
    "print(ngram_df)\n",
    "# adding cvec score with default params\n",
    "print(\"Τα αποτελέσματα για τα διαφορετικά n-grams είναι\")\n",
    "#ngram_df = ngram_df.append(acc_df.iloc[:1,:])\n",
    "# plot scores on graph\n",
    "sns.pointplot(x='params', y='scores', data =ngram_df)\n",
    "plt.ylabel('Accuracy Score')\n",
    "plt.xlabel('ngrams')\n",
    "plt.xticks(rotation=40)\n",
    "plt.show()\n",
    "\n",
    "mdf_params = [0.25, 0.5, 0.75, 1.0]\n",
    "maxdf_df = pd.DataFrame({'params':mdf_params, 'scores':mdf_scores}, index=[0,1,2,3])\n",
    "# adding cvec score with default params\n",
    "#ngram_df = ngram_df.append(acc_df.iloc[:1,:])\n",
    "print(\"Τα αποτελέσματα για τα διαφορετικά max df είναι\")\n",
    "# plot scores on graph\n",
    "sns.pointplot(x='params', y='scores', data =maxdf_df)\n",
    "plt.ylabel('Accuracy Score')\n",
    "plt.xlabel(mdf_params)\n",
    "plt.xticks(rotation=40)\n",
    "plt.show()\n",
    "\n",
    "maxfeat=['None',5000,20000,50000,100000]\n",
    "maxfeat_df = pd.DataFrame({'params':maxfeat, 'scores':maxfeat_scores}, index=[0,1,2,3,4])\n",
    "# adding cvec score with default params\n",
    "print(\"Τα αποτελέσματα για τα διαφορετικά max features είναι\")\n",
    "\n",
    "#ngram_df = ngram_df.append(acc_df.iloc[:1,:])\n",
    "# plot scores on graph\n",
    "sns.pointplot(x='params', y='scores', data =maxfeat_df)\n",
    "plt.ylabel('Accuracy Score')\n",
    "plt.xlabel(maxfeat)\n",
    "plt.xticks(rotation=40)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8197527259160483\n",
      "0.8197527259160483\n",
      "0.82153070395903\n"
     ]
    }
   ],
   "source": [
    "print(max(maxfeat_scores))#Βλέπουμε ότι το καλύτερο score είναι με max_df=0.5\n",
    "print(max(ngram_scores))\n",
    "print(max(mdf_scores))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#δοκιμαζουμε με τον καλυτερο vectorizer διαφορετικους classifiers\n",
    "# initalise the vectoriser \n",
    "cvec = CountVectorizer(tokenizer=tokenizer_preproccessor,max_df=0.5)\n",
    "# fit the training data on the model\n",
    "cvec.fit(X_train_s)\n",
    "\n",
    "#transform training data into sparse matrix\n",
    "X_train_cvec = cvec.transform(X_train_s)\n",
    "X_test_cvec=cvec.transform(X_test_s)\n",
    "#print(X_train_tvec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Η επιτυχία του  LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
      "          intercept_scaling=1, max_iter=1000, multi_class='warn',\n",
      "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
      "          tol=0.0001, verbose=0, warm_start=False) είναι:  0.9219645168437871\n",
      "Η επιτυχία του  MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True) είναι:  0.8163479606514983\n",
      "Η επιτυχία του  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') είναι:  0.8231138904164318\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py:246: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Η επιτυχία του  RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "            max_depth=None, max_features='auto', 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, n_estimators=10, n_jobs=None,\n",
      "            oob_score=False, random_state=None, verbose=0,\n",
      "            warm_start=False) είναι:  0.8338897836317578\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Η επιτυχία του  LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
      "     intercept_scaling=1, loss='squared_hinge', max_iter=2000,\n",
      "     multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
      "     verbose=0) είναι:  0.9037728750037083\n",
      "                    params    scores\n",
      "0     Logistisc Regression  0.921965\n",
      "1  Multinomial Naive Bayes  0.816348\n",
      "2            Decision Tree  0.823114\n",
      "3            Random Forest  0.833890\n",
      "4               Linear SVC  0.903773\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    }
   ],
   "source": [
    "#δοκιμαζουμε διαφορετικούς classifiers\n",
    "classifier_finder(X_train_cvec,X_test_cvec,y_train_s,y_test_s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 10 folds for each of 10 candidates, totalling 100 fits\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=0.01, penalty=l1, score=0.9080858443529157, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.6s remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=0.01, penalty=l1, score=0.9067851723390419, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    1.4s remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=0.01, penalty=l1, score=0.9080858443529157, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    2.2s remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=0.01, penalty=l1, score=0.9052677216561891, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    2.9s remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=0.01, penalty=l1, score=0.9030999349663993, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l1, score=0.9056603773584906, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l1, score=0.8974192149208414, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l1, score=0.9080459770114943, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l1, score=0.9052266319670353, total=   0.6s\n",
      "[CV] C=0.01, penalty=l1 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l1, score=0.9008891780524831, total=   0.6s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9219596791675699, total=   3.6s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9239106871883807, total=   3.9s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9236939085194017, total=   3.5s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9197918924777801, total=   3.1s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9239106871883807, total=   3.1s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9245283018867925, total=   4.0s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9173715029277814, total=   3.5s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9195402298850575, total=   2.8s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9260464107568858, total=   2.9s\n",
      "[CV] C=0.01, penalty=l2 ..............................................\n",
      "[CV] ..... C=0.01, penalty=l2, score=0.9199739752765127, total=   3.4s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9343160632993713, total=   1.1s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ....... C=0.1, penalty=l1, score=0.937134185996098, total=   1.0s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9343160632993713, total=   1.2s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9312811619336657, total=   1.1s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9356167353132452, total=   1.0s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9329863370201692, total=   1.2s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9303838646714379, total=   0.9s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9277813923227066, total=   1.2s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9373237909347213, total=   1.1s\n",
      "[CV] C=0.1, penalty=l1 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l1, score=0.9271307742355237, total=   1.6s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l2, score=0.9304140472577498, total=   8.4s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ....... C=0.1, penalty=l2, score=0.928679817905918, total=   9.2s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l2, score=0.9297637112508129, total=   8.1s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ....... C=0.1, penalty=l2, score=0.928679817905918, total=   9.8s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l2, score=0.9314979406026447, total=   7.6s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l2, score=0.9334200824116243, total=   8.7s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l2, score=0.9253957926697028, total=   8.9s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ....... C=0.1, penalty=l2, score=0.927564519626979, total=   8.5s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l2, score=0.9301669919757103, total=   7.1s\n",
      "[CV] C=0.1, penalty=l2 ...............................................\n",
      "[CV] ...... C=0.1, penalty=l2, score=0.9292995011927998, total=   8.0s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9280294818989812, total=   2.1s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9258616952091914, total=   2.2s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9258616952091914, total=   2.2s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9262952525471494, total=   2.5s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9278127032300022, total=   2.3s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9301669919757103, total=   2.3s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9253957926697028, total=   2.7s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9262632834526133, total=   2.4s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ....... C=1.0, penalty=l1, score=0.929733246584255, total=   2.3s\n",
      "[CV] C=1.0, penalty=l1 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l1, score=0.9258295380611581, total=   2.5s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=1.0, penalty=l2, score=0.9256449165402124, total=  16.5s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ....... C=1.0, penalty=l2, score=0.929330153912855, total=  12.2s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=1.0, penalty=l2, score=0.9252113592022545, total=  18.1s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=1.0, penalty=l2, score=0.9226100151745068, total=  15.5s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=1.0, penalty=l2, score=0.9271623672230652, total=  13.8s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l2, score=0.9240945564953372, total=  26.5s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ........ C=1.0, penalty=l2, score=0.92474517458252, total=  16.7s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=1.0, penalty=l2, score=0.9240945564953372, total=  16.7s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l2, score=0.9245283018867925, total=  25.6s\n",
      "[CV] C=1.0, penalty=l2 ...............................................\n",
      "[CV] ...... C=1.0, penalty=l2, score=0.9219258295380611, total=  22.0s\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l1, score=0.9113375243876003, total=26.6min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=10.0, penalty=l1, score=0.905484500325168, total=19.9min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n",
      "[CV] ..... C=10.0, penalty=l1, score=0.9117710817255582, total=   5.3s\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l1, score=0.9074355083459787, total=20.0min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l1, score=0.9130717537394321, total=19.5min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l1, score=0.9112990674474084, total=19.9min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=10.0, penalty=l1, score=0.912600303621774, total=19.8min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l1, score=0.9052266319670353, total=19.9min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l1, score=0.9171546302320538, total=24.9min\n",
      "[CV] C=10.0, penalty=l1 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l1, score=0.9052266319670353, total=28.2min\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9256449165402124, total=  20.0s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ...... C=10.0, penalty=l2, score=0.929546932581834, total=  15.7s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9247778018642966, total=  22.1s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9202254498157382, total=  21.0s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9221764578365489, total=  25.1s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9262632834526133, total=  20.6s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9243114291910648, total=  18.7s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9201908479722403, total=  23.6s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9284320104098894, total=  19.4s\n",
      "[CV] C=10.0, penalty=l2 ..............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=10.0, penalty=l2, score=0.9238776837996097, total=  23.3s\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l1, score=0.8887925428137872, total=30.0min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=100.0, penalty=l1, score=0.882722740082376, total=30.0min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l1, score=0.8900932148276609, total=30.5min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l1, score=0.8913938868415348, total=33.2min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l1, score=0.8859744201170605, total=26.0min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l1, score=0.8887443070917371, total=26.2min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=100.0, penalty=l1, score=0.890479288657558, total=24.3min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l1, score=0.8811537627412709, total=26.0min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l1, score=0.8893949251789199, total=34.5min\n",
      "[CV] C=100.0, penalty=l1 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=100.0, penalty=l1, score=0.891563652136196, total=35.0min\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9256449165402124, total=  20.8s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ..... C=100.0, penalty=l2, score=0.928896596574897, total=  17.3s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9243442445263386, total=  22.9s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9204422284847171, total=  23.1s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9219596791675699, total=  27.8s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9238776837996097, total=  29.7s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9223595749295164, total=  22.4s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9223595749295164, total=  22.7s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9277813923227066, total=  24.6s\n",
      "[CV] C=100.0, penalty=l2 .............................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:922: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n",
      "[Parallel(n_jobs=1)]: Done 100 out of 100 | elapsed: 508.0min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .... C=100.0, penalty=l2, score=0.9234439384081544, total=  28.3s\n",
      "tuned hpyerparameters :(best parameters)  {'C': 0.1, 'penalty': 'l1'}\n",
      "accuracy : 0.932827406764961\n"
     ]
    }
   ],
   "source": [
    "# Grid search me cross validation\n",
    "grid={\"C\":np.logspace(-2,2,5), \"penalty\":[\"l1\",\"l2\"]}\n",
    "logreg=LogisticRegression()\n",
    "logreg_cv=GridSearchCV(logreg,grid,cv=10,verbose=5)\n",
    "logreg_cv.fit(X_train_cvec,y_train_s)\n",
    "print(\"tuned hpyerparameters :(best parameters) \",logreg_cv.best_params_)\n",
    "print(\"accuracy :\",logreg_cv.best_score_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\dtchmnt\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=0.1, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l1', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_clf_sentiwordnet = LogisticRegression(**logreg_cv.best_params_)\n",
    "#best_clf_sentiwordnet=LogisticRegression(C=0.1,penalty='l1')\n",
    "best_clf_sentiwordnet.fit(X_train_cvec,y_train_s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved successfully!\n",
      "Saved successfully!\n",
      "Saved successfully!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "['sentiwordnet_cvec.pkl']"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#σωζουμε τα test sets, τον best_clf_sentiwordnet και τον vectorizer\n",
    "saveList(X_test_s,\"X_test_sentiwordnet.npy\")\n",
    "saveList(y_test_s,\"y_test_sentiwordnet.npy\")\n",
    "\n",
    "\n",
    "saveList(X_test_cvec,\"X_test_sentiwordnet_dtm\")\n",
    "joblib.dump(best_clf_sentiwordnet, 'sentiwordnet_clf.pkl')\n",
    "joblib.dump(cvec, 'sentiwordnet_cvec.pkl')"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}