src/ai/experiments/supervised/valence_regular_supervised_ml.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "TAyZCtgcISFO"
   },
   "source": [
    "## Testes de valência para o projeto final de IA369Y 2 Semestre 2018\n",
    "\n",
    "Passos para tratar os dados com valência, testar e escolher um classificador para utilizar no projeto final de IA369Y.\n",
    "\n",
    "1) Remover espaços duplos, quebras de linha, números e links do dataset e das frases a serem testadas.\n",
    "\n",
    "2) Remover stopwords e aplicar o stemmer.\n",
    "\n",
    "3) Treinar os classificadores.\n",
    "\n",
    "4) Realizar as predições com os classificadores.\n",
    "\n",
    "5) Avaliar as medidas obtidas com os classificadores."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:42.612664Z",
     "start_time": "2018-12-21T16:55:34.657019Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 632
    },
    "colab_type": "code",
    "id": "r2Wwyp_VHu32",
    "outputId": "3aef9901-ce0a-44de-df61-b6eeb50e676b"
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append('../..')\n",
    "\n",
    "import csv\n",
    "import codecs\n",
    "import copy\n",
    "import re\n",
    "from random import shuffle\n",
    "import time\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from nltk import word_tokenize\n",
    "import gensim\n",
    "\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
    "from sklearn.pipeline import make_pipeline\n",
    "from sklearn.calibration import CalibratedClassifierCV\n",
    "from sklearn.preprocessing import Normalizer, MinMaxScaler\n",
    "from sklearn.decomposition import TruncatedSVD, LatentDirichletAllocation\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.naive_bayes import MultinomialNB, GaussianNB, ComplementNB\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.svm import SVC, LinearSVC\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "from sklearn.externals import joblib\n",
    "\n",
    "from utils import tokenizer, load_six_emotions, load_3_emotions\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "np.warnings.filterwarnings('ignore')\n",
    "np.random.seed(12345)\n",
    "\n",
    "def highlight_max(data, color='green'):\n",
    "    attr = f'background-color: {color}; color: white; font-weight: bold;'\n",
    "    #remove % and cast to float\n",
    "    data = data.replace('%','', regex=True).astype(float)\n",
    "    if data.ndim == 1:  # Series from .apply(axis=0) or axis=1\n",
    "        is_max = data == data.max()\n",
    "        return [attr if v else '' for v in is_max]\n",
    "    else:  # from .apply(axis=None)\n",
    "        is_max = data == data.max().max()\n",
    "        return pd.DataFrame(np.where(is_max, attr, ''),\n",
    "                            index=data.index, columns=data.columns)\n",
    "\n",
    "\n",
    "def highlight_min(data, color='green'):\n",
    "    attr = f'background-color: {color}; color: white; font-weight: bold;'\n",
    "    #remove % and cast to float\n",
    "    data = data.replace('%','', regex=True).astype(float)\n",
    "    if data.ndim == 1:  # Series from .apply(axis=0) or axis=1\n",
    "        is_min = data == data.min()\n",
    "        return [attr if v else '' for v in is_min]\n",
    "    else:  # from .apply(axis=None)\n",
    "        is_min = data == data.min().min()\n",
    "        return pd.DataFrame(np.where(is_min, attr, ''),\n",
    "                            index=data.index, columns=data.columns)   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Matriz de Resultados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:42.643793Z",
     "start_time": "2018-12-21T16:55:42.615216Z"
    }
   },
   "outputs": [],
   "source": [
    "classf = {\n",
    "    'MultinomialNB': 0,\n",
    "    'ComplementNB': 0,\n",
    "    'LogisticRegression': 0,\n",
    "    'RandomForestClassifier': 0,\n",
    "    'KNeighborsClassifier': 0,\n",
    "    'MLPClassifier': 0,\n",
    "    'LinearSVC': 0,\n",
    "    'SVC': 0\n",
    "}\n",
    "\n",
    "matriz_resultados = {\n",
    "    'all': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),\n",
    "    },\n",
    "    'tweets': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),       \n",
    "    },\n",
    "    'titulo_noticias': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),\n",
    "    }\n",
    "}\n",
    "\n",
    "train_time = {\n",
    "    'all': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),\n",
    "    },\n",
    "    'tweets': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),       \n",
    "    },\n",
    "    'titulo_noticias': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),\n",
    "    }\n",
    "}\n",
    "\n",
    "predict_time = {\n",
    "    'all': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),\n",
    "    },\n",
    "    'tweets': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),       \n",
    "    },\n",
    "    'titulo_noticias': {\n",
    "        'tfidf': copy.deepcopy(classf),\n",
    "        'tfidf+lsa': copy.deepcopy(classf),\n",
    "        'tfidf+lda': copy.deepcopy(classf),\n",
    "        'count': copy.deepcopy(classf),\n",
    "        'count+lsa': copy.deepcopy(classf),\n",
    "        'count+lda': copy.deepcopy(classf),\n",
    "        'tfidf+count+w2c': copy.deepcopy(classf),\n",
    "    }\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "thdVOHWmKxaF"
   },
   "source": [
    "## Datasets\n",
    "\n",
    "Para validar, serão utilizados dois datasets.\n",
    "\n",
    "Os dos datasets foram obtidos do site minerando dados.\n",
    "\n",
    "O primeiro deles tem tweets de política de Minas Gerais com rótulos de valência: positivo, negativo e neutro. Foi feito um tratamento para eliminar tweets repetidos e dessa forma sobraram 3016 tweets.\n",
    "\n",
    "O segundo contém 2123 títulos de notícias com rótulos de valência: positivo, negativo e neutro."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:42.730468Z",
     "start_time": "2018-12-21T16:55:42.645522Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "colab_type": "code",
    "id": "tJiBeMzSTKlB",
    "outputId": "56408718-b65c-42be-dbc7-703dc96ea551"
   },
   "outputs": [],
   "source": [
    "path = '/ssd/programas/sentiment-analysis-2018-president-election/dataset/valencia/'\n",
    "\n",
    "#Carregando os datasets\n",
    "def carregar(filename):\n",
    "    frases = []\n",
    "    with open(filename, 'r', encoding='utf-8') as h:\n",
    "        reader = csv.reader(h, delimiter='|')\n",
    "        for row in reader:\n",
    "            frase = tokenizer(row[0]).strip()\n",
    "            valencia = row[1].upper()\n",
    "            if len(frase) > 5:\n",
    "                frases.append((valencia, frase))\n",
    "    return frases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:55.807072Z",
     "start_time": "2018-12-21T16:55:42.737175Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 71
    },
    "colab_type": "code",
    "id": "zD3gV2f0tsNL",
    "outputId": "5560031f-54b1-44a9-81e7-9e17f3f5417c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('NEUTRO', 'implic org govern dilm pression ditadur venezuel pag dív bilionári odebrecht'), ('NEUTRO', 'par fing estar melhor min ger brasil'), ('NEUTRO', 'govern min cham vagabund ladrã shoping paul'), ('POSITIVO', 'pmg veícul produt roub recuper'), ('NEGATIVO', 'campanh eleitoral ver torped internet cri polêm irã jov bombard mail mensag polít celul país milhã internaut milhã usar aparelh móvel')]\n",
      "--------------------\n",
      "[('NEGATIVO', 'ser aind gent apó ver'), ('POSITIVO', 'sinomarp mut vacin contr febr amarel realiz ipating'), ('NEGATIVO', 'calam financeir govern compr dois helicópter'), ('POSITIVO', 'timbet glob betim contag intensific vacin contr febr amarel nest sáb'), ('NEUTRO', 'petist pimentel')]\n",
      "--------------------\n",
      "[('NEGATIVO', 'dia descobr beij pesso esquec outr bobag'), ('POSITIVO', 'usimin lucr milhã tri melhor result desd notic uol econom'), ('NEGATIVO', 'petrobr mult val recomend mais notíc rad'), ('NEGATIVO', 'maior pens sensibil sent pensament hom vulg sent viv pens sab viv mim pens viv sent nã mais aliment pens'), ('POSITIVO', 'cientist usar twit question paranormal experient usar sit relacion test habil psíquic seguidor diz new scientist')]\n"
     ]
    }
   ],
   "source": [
    "#Carrega os datasets em separado\n",
    "tweets_mg = carregar(f'{path}tweets_mg_tratados.csv')\n",
    "shuffle(tweets_mg)\n",
    "titulo_noticias = carregar(f'{path}titulo_noticias.txt')\n",
    "shuffle(titulo_noticias)\n",
    "\n",
    "frases = tweets_mg + titulo_noticias\n",
    "shuffle(frases)\n",
    "\n",
    "print(frases[:5])\n",
    "print('-' * 20)\n",
    "print(tweets_mg[:5])\n",
    "print('-' * 20)\n",
    "print(titulo_noticias[:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:55.853562Z",
     "start_time": "2018-12-21T16:55:55.809458Z"
    }
   },
   "outputs": [],
   "source": [
    "def carregar(filename):\n",
    "    frases = []\n",
    "    with open(filename, 'r', encoding='utf-8') as h:\n",
    "        reader = csv.reader(h, delimiter='|')\n",
    "        for row in reader:\n",
    "            frase = ' '.join(row[0].strip().splitlines())\n",
    "            valencia = row[1].upper()\n",
    "            if len(frase) > 5:\n",
    "                frases.append((valencia, frase))\n",
    "    return frases\n",
    "\n",
    "titulo_noticias2 = carregar(f'{path}titulo_noticias.txt')\n",
    "only_neutro = []\n",
    "for sent, new in titulo_noticias2:\n",
    "    if sent == 'NEUTRO':\n",
    "        only_neutro.append(new)\n",
    "with open(f'{path}titulo_noticias_neutro.txt', 'w') as h:\n",
    "    for k in only_neutro:\n",
    "        h.write(f'{k}|NEUTRO\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:55.974390Z",
     "start_time": "2018-12-21T16:55:55.856081Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 139
    },
    "colab_type": "code",
    "id": "yJbG_ck5i8rI",
    "outputId": "8d2f6682-e474-43bc-8da5-2c1a6db52619"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['implic org govern dilm pression ditadur venezuel pag dív bilionári odebrecht', 'par fing estar melhor min ger brasil', 'govern min cham vagabund ladrã shoping paul', 'pmg veícul produt roub recuper', 'campanh eleitoral ver torped internet cri polêm irã jov bombard mail mensag polít celul país milhã internaut milhã usar aparelh móvel']\n",
      "['NEUTRO', 'NEUTRO', 'NEUTRO', 'POSITIVO', 'NEGATIVO']\n",
      "--------------------\n",
      "['ser aind gent apó ver', 'sinomarp mut vacin contr febr amarel realiz ipating', 'calam financeir govern compr dois helicópter', 'timbet glob betim contag intensific vacin contr febr amarel nest sáb', 'petist pimentel']\n",
      "['NEGATIVO', 'POSITIVO', 'NEGATIVO', 'POSITIVO', 'NEUTRO']\n",
      "--------------------\n",
      "['dia descobr beij pesso esquec outr bobag', 'usimin lucr milhã tri melhor result desd notic uol econom', 'petrobr mult val recomend mais notíc rad', 'maior pens sensibil sent pensament hom vulg sent viv pens sab viv mim pens viv sent nã mais aliment pens', 'cientist usar twit question paranormal experient usar sit relacion test habil psíquic seguidor diz new scientist']\n",
      "['NEGATIVO', 'POSITIVO', 'NEGATIVO', 'NEGATIVO', 'POSITIVO']\n"
     ]
    }
   ],
   "source": [
    "#all_datasets\n",
    "afrases = []\n",
    "avalencias = []\n",
    "for valencia, frase in frases:\n",
    "    afrases.append(frase)\n",
    "    avalencias.append(valencia)\n",
    "    \n",
    "print(afrases[:5])\n",
    "print(avalencias[:5])\n",
    "print('-' * 20)\n",
    "\n",
    "#tweets_mg\n",
    "atweets_mg = []\n",
    "aval_tweets_mg = []\n",
    "for valencia, frase in tweets_mg:\n",
    "    atweets_mg.append(frase)\n",
    "    aval_tweets_mg.append(valencia)\n",
    "\n",
    "print(atweets_mg[:5])\n",
    "print(aval_tweets_mg[:5])\n",
    "print('-' * 20)\n",
    "\n",
    "#titulo_noticias\n",
    "atitulo_noticias = []\n",
    "aval_titulo_noticias = []\n",
    "for valencia, frase in titulo_noticias:\n",
    "    atitulo_noticias.append(frase)\n",
    "    aval_titulo_noticias.append(valencia)\n",
    "\n",
    "print(atitulo_noticias[:5])\n",
    "print(aval_titulo_noticias[:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T01:16:59.018579Z",
     "start_time": "2018-11-29T01:16:58.927104Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "UicV1eOVjX2Q"
   },
   "outputs": [],
   "source": [
    "def run_ml_model(model, X_train, y_train, X_test, y_test):\n",
    "    start = time.time()\n",
    "    model.fit(X_train, y_train)\n",
    "    train_time = round(time.time() - start, 5)\n",
    "    accuracy = np.round(model.score(X_test, y_test) * 100, 2)\n",
    "    start = time.time()\n",
    "    y_pred = model.predict(X_test)\n",
    "    predict_time = round(time.time() - start, 5)\n",
    "    print(f'Modelo      : {model.__class__.__name__}')\n",
    "    print(f'Train Time  : {train_time}s')\n",
    "    print(f'Predict Time: {predict_time}s')\n",
    "    print(f'Acurácia    : {accuracy}%')\n",
    "    print(classification_report(y_test, y_pred))\n",
    "    print('Matrix de Confusão: ')\n",
    "    print(confusion_matrix(y_test, y_pred))\n",
    "    print('-' * 30)\n",
    "    print()\n",
    "    return accuracy, train_time, predict_time\n",
    "\n",
    "def split_data(X, y):\n",
    "    test_size = .33\n",
    "    random_state = 0\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=test_size, random_state=random_state)\n",
    "    return {\n",
    "        'X_train': X_train,\n",
    "        'y_train': y_train,\n",
    "        'X_test': X_test,\n",
    "        'y_test': y_test\n",
    "    }"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "k94TjXIfjjF6"
   },
   "source": [
    "## Classificadores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T01:16:59.104811Z",
     "start_time": "2018-11-29T01:16:59.023136Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "D-xa0I3PjexY"
   },
   "outputs": [],
   "source": [
    "classifiers = (\n",
    "    MultinomialNB(),\n",
    "    ComplementNB(),\n",
    "    LogisticRegression(multi_class='auto', solver='lbfgs'),\n",
    "    RandomForestClassifier(n_estimators=50, min_samples_split=8, random_state=10),\n",
    "    KNeighborsClassifier(n_neighbors=8, algorithm='auto'),\n",
    "    MLPClassifier(hidden_layer_sizes=(100, 25), max_iter=500, random_state=10),\n",
    "    LinearSVC(max_iter=150, random_state=10),\n",
    "    SVC(gamma='auto', max_iter=150),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "8T0sft8Ujwki"
   },
   "source": [
    "## TF-IDF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:18:50.929328Z",
     "start_time": "2018-11-29T00:18:50.259889Z"
    },
    "colab": {},
    "colab_type": "code",
    "id": "g9TiaAoGjvPj"
   },
   "outputs": [],
   "source": [
    "vec_tfidf = TfidfVectorizer(ngram_range=(1, 2))\n",
    "X_tfidf = vec_tfidf.fit_transform(afrases)\n",
    "\n",
    "vec_tfidf_tmg = TfidfVectorizer(ngram_range=(1, 2))\n",
    "X_tfidf_tmg = vec_tfidf_tmg.fit_transform(atweets_mg)\n",
    "\n",
    "vec_tfidf_tn = TfidfVectorizer(ngram_range=(1, 2))\n",
    "X_tfidf_tn = vec_tfidf_tn.fit_transform(atitulo_noticias)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:26:06.930332Z",
     "start_time": "2018-11-29T00:18:50.931034Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1037
    },
    "colab_type": "code",
    "id": "1keF94erj8UT",
    "outputId": "2c5c1870-cf5f-4f72-8f2d-c0b48bf05bcb",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "all_datasets\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.01453s\n",
      "Predict Time: 0.00138s\n",
      "Acurácia    : 58.86%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.54      0.60      0.57       790\n",
      "      NEUTRO       0.61      0.17      0.26       529\n",
      "    POSITIVO       0.62      0.79      0.69      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.59      0.52      0.51      2348\n",
      "weighted avg       0.59      0.59      0.56      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[476  36 278]\n",
      " [207  89 233]\n",
      " [191  21 817]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.01132s\n",
      "Predict Time: 0.00073s\n",
      "Acurácia    : 63.12%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.65      0.61       790\n",
      "      NEUTRO       0.55      0.51      0.53       529\n",
      "    POSITIVO       0.73      0.68      0.70      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.62      0.61      0.61      2348\n",
      "weighted avg       0.64      0.63      0.63      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[513 129 148]\n",
      " [143 272 114]\n",
      " [240  92 697]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 6.29359s\n",
      "Predict Time: 0.00242s\n",
      "Acurácia    : 62.69%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.55      0.69      0.61       790\n",
      "      NEUTRO       0.57      0.34      0.43       529\n",
      "    POSITIVO       0.72      0.72      0.72      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.61      0.59      0.59      2348\n",
      "weighted avg       0.63      0.63      0.62      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[547  84 159]\n",
      " [210 182 137]\n",
      " [233  53 743]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 8.89272s\n",
      "Predict Time: 0.13253s\n",
      "Acurácia    : 62.61%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.56      0.70      0.62       790\n",
      "      NEUTRO       0.55      0.51      0.53       529\n",
      "    POSITIVO       0.76      0.62      0.68      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.62      0.61      0.61      2348\n",
      "weighted avg       0.64      0.63      0.63      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[555 129 106]\n",
      " [156 272 101]\n",
      " [289  97 643]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00243s\n",
      "Predict Time: 0.42257s\n",
      "Acurácia    : 57.84%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.51      0.64      0.56       790\n",
      "      NEUTRO       0.50      0.38      0.43       529\n",
      "    POSITIVO       0.69      0.64      0.66      1029\n",
      "\n",
      "   micro avg       0.58      0.58      0.58      2348\n",
      "   macro avg       0.56      0.55      0.55      2348\n",
      "weighted avg       0.58      0.58      0.58      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[502 106 182]\n",
      " [210 200 119]\n",
      " [276  97 656]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 500.15232s\n",
      "Predict Time: 0.02753s\n",
      "Acurácia    : 61.97%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.58      0.59      0.59       790\n",
      "      NEUTRO       0.54      0.42      0.47       529\n",
      "    POSITIVO       0.68      0.74      0.71      1029\n",
      "\n",
      "   micro avg       0.62      0.62      0.62      2348\n",
      "   macro avg       0.60      0.59      0.59      2348\n",
      "weighted avg       0.61      0.62      0.61      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[468 112 210]\n",
      " [152 222 155]\n",
      " [187  77 765]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.22786s\n",
      "Predict Time: 0.00103s\n",
      "Acurácia    : 64.78%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.69      0.63       790\n",
      "      NEUTRO       0.59      0.45      0.51       529\n",
      "    POSITIVO       0.74      0.72      0.73      1029\n",
      "\n",
      "   micro avg       0.65      0.65      0.65      2348\n",
      "   macro avg       0.64      0.62      0.62      2348\n",
      "weighted avg       0.65      0.65      0.65      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[548  97 145]\n",
      " [182 237 110]\n",
      " [224  69 736]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.8579s\n",
      "Predict Time: 0.50209s\n",
      "Acurácia    : 48.17%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.52      0.17      0.26       790\n",
      "      NEUTRO       0.51      0.11      0.18       529\n",
      "    POSITIVO       0.48      0.91      0.63      1029\n",
      "\n",
      "   micro avg       0.48      0.48      0.48      2348\n",
      "   macro avg       0.50      0.40      0.35      2348\n",
      "weighted avg       0.50      0.48      0.40      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[135  32 623]\n",
      " [ 60  58 411]\n",
      " [ 67  24 938]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "tweets_mg\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00563s\n",
      "Predict Time: 0.00054s\n",
      "Acurácia    : 61.98%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.13      0.01      0.02       201\n",
      "      NEUTRO       0.54      0.73      0.62       362\n",
      "    POSITIVO       0.72      0.81      0.76       426\n",
      "\n",
      "   micro avg       0.62      0.62      0.62       989\n",
      "   macro avg       0.46      0.52      0.47       989\n",
      "weighted avg       0.53      0.62      0.56       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  2 153  46]\n",
      " [  9 265  88]\n",
      " [  4  76 346]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.00645s\n",
      "Predict Time: 0.00053s\n",
      "Acurácia    : 59.25%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.29      0.32      0.31       201\n",
      "      NEUTRO       0.54      0.54      0.54       362\n",
      "    POSITIVO       0.81      0.76      0.78       426\n",
      "\n",
      "   micro avg       0.59      0.59      0.59       989\n",
      "   macro avg       0.55      0.54      0.54       989\n",
      "weighted avg       0.60      0.59      0.60       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 65 108  28]\n",
      " [115 197  50]\n",
      " [ 44  58 324]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 2.41728s\n",
      "Predict Time: 0.00165s\n",
      "Acurácia    : 63.4%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.20      0.04      0.07       201\n",
      "      NEUTRO       0.52      0.82      0.64       362\n",
      "    POSITIVO       0.84      0.76      0.80       426\n",
      "\n",
      "   micro avg       0.63      0.63      0.63       989\n",
      "   macro avg       0.52      0.54      0.50       989\n",
      "weighted avg       0.59      0.63      0.59       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  8 172  21]\n",
      " [ 25 297  40]\n",
      " [  7  97 322]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 3.62224s\n",
      "Predict Time: 0.05584s\n",
      "Acurácia    : 63.4%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.27      0.06      0.10       201\n",
      "      NEUTRO       0.52      0.85      0.64       362\n",
      "    POSITIVO       0.89      0.72      0.79       426\n",
      "\n",
      "   micro avg       0.63      0.63      0.63       989\n",
      "   macro avg       0.56      0.54      0.51       989\n",
      "weighted avg       0.63      0.63      0.60       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 13 175  13]\n",
      " [ 29 309  24]\n",
      " [  6 115 305]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00186s\n",
      "Predict Time: 0.14588s\n",
      "Acurácia    : 61.17%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.35      0.32      0.34       201\n",
      "      NEUTRO       0.53      0.60      0.56       362\n",
      "    POSITIVO       0.80      0.76      0.78       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.56      0.56      0.56       989\n",
      "weighted avg       0.61      0.61      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 64 111  26]\n",
      " [ 93 216  53]\n",
      " [ 24  77 325]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 41.46804s\n",
      "Predict Time: 0.01048s\n",
      "Acurácia    : 61.68%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.35      0.22      0.27       201\n",
      "      NEUTRO       0.55      0.62      0.58       362\n",
      "    POSITIVO       0.76      0.80      0.78       426\n",
      "\n",
      "   micro avg       0.62      0.62      0.62       989\n",
      "   macro avg       0.55      0.55      0.54       989\n",
      "weighted avg       0.60      0.62      0.60       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 44 119  38]\n",
      " [ 64 226  72]\n",
      " [ 18  68 340]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.11271s\n",
      "Predict Time: 0.00067s\n",
      "Acurácia    : 63.9%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.35      0.16      0.22       201\n",
      "      NEUTRO       0.54      0.75      0.63       362\n",
      "    POSITIVO       0.84      0.77      0.80       426\n",
      "\n",
      "   micro avg       0.64      0.64      0.64       989\n",
      "   macro avg       0.58      0.56      0.55       989\n",
      "weighted avg       0.63      0.64      0.62       989\n",
      "\n",
      "Matrix de Confusão: \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 32 149  20]\n",
      " [ 48 273  41]\n",
      " [ 12  87 327]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.60397s\n",
      "Predict Time: 0.34286s\n",
      "Acurácia    : 53.69%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.33      0.22      0.26       201\n",
      "      NEUTRO       0.52      0.35      0.42       362\n",
      "    POSITIVO       0.59      0.85      0.69       426\n",
      "\n",
      "   micro avg       0.54      0.54      0.54       989\n",
      "   macro avg       0.48      0.47      0.46       989\n",
      "weighted avg       0.51      0.54      0.51       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 44  69  88]\n",
      " [ 68 127 167]\n",
      " [ 20  46 360]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "titulo_noticias\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.01788s\n",
      "Predict Time: 0.00189s\n",
      "Acurácia    : 65.34%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.65      0.84      0.73       635\n",
      "      NEUTRO       1.00      0.01      0.02       130\n",
      "    POSITIVO       0.66      0.60      0.63       594\n",
      "\n",
      "   micro avg       0.65      0.65      0.65      1359\n",
      "   macro avg       0.77      0.48      0.46      1359\n",
      "weighted avg       0.69      0.65      0.62      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[533   0 102]\n",
      " [ 51   1  78]\n",
      " [240   0 354]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.0225s\n",
      "Predict Time: 0.00187s\n",
      "Acurácia    : 65.19%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.70      0.71      0.70       635\n",
      "      NEUTRO       0.42      0.32      0.37       130\n",
      "    POSITIVO       0.64      0.66      0.65       594\n",
      "\n",
      "   micro avg       0.65      0.65      0.65      1359\n",
      "   macro avg       0.59      0.57      0.57      1359\n",
      "weighted avg       0.65      0.65      0.65      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[449  26 160]\n",
      " [ 28  42  60]\n",
      " [168  31 395]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 6.15197s\n",
      "Predict Time: 0.00667s\n",
      "Acurácia    : 65.86%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.65      0.82      0.72       635\n",
      "      NEUTRO       0.50      0.04      0.07       130\n",
      "    POSITIVO       0.68      0.62      0.65       594\n",
      "\n",
      "   micro avg       0.66      0.66      0.66      1359\n",
      "   macro avg       0.61      0.49      0.48      1359\n",
      "weighted avg       0.65      0.66      0.63      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[521   0 114]\n",
      " [ 66   5  59]\n",
      " [220   5 369]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 11.44937s\n",
      "Predict Time: 0.19044s\n",
      "Acurácia    : 62.62%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.61      0.83      0.70       635\n",
      "      NEUTRO       0.48      0.09      0.15       130\n",
      "    POSITIVO       0.67      0.53      0.59       594\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      1359\n",
      "   macro avg       0.58      0.48      0.48      1359\n",
      "weighted avg       0.62      0.63      0.60      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[524   4 107]\n",
      " [ 67  12  51]\n",
      " [270   9 315]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00397s\n",
      "Predict Time: 0.44656s\n",
      "Acurácia    : 60.49%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.62      0.72      0.67       635\n",
      "      NEUTRO       0.38      0.25      0.30       130\n",
      "    POSITIVO       0.61      0.56      0.58       594\n",
      "\n",
      "   micro avg       0.60      0.60      0.60      1359\n",
      "   macro avg       0.54      0.51      0.52      1359\n",
      "weighted avg       0.60      0.60      0.60      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[458  25 152]\n",
      " [ 40  32  58]\n",
      " [235  27 332]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 149.8301s\n",
      "Predict Time: 0.01744s\n",
      "Acurácia    : 66.0%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.70      0.70      0.70       635\n",
      "      NEUTRO       0.49      0.35      0.41       130\n",
      "    POSITIVO       0.65      0.68      0.66       594\n",
      "\n",
      "   micro avg       0.66      0.66      0.66      1359\n",
      "   macro avg       0.61      0.58      0.59      1359\n",
      "weighted avg       0.66      0.66      0.66      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[445  23 167]\n",
      " [ 29  46  55]\n",
      " [163  25 406]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.17626s\n",
      "Predict Time: 0.00096s\n",
      "Acurácia    : 68.65%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.69      0.78      0.74       635\n",
      "      NEUTRO       0.57      0.25      0.34       130\n",
      "    POSITIVO       0.69      0.68      0.68       594\n",
      "\n",
      "   micro avg       0.69      0.69      0.69      1359\n",
      "   macro avg       0.65      0.57      0.59      1359\n",
      "weighted avg       0.68      0.69      0.68      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[496  12 127]\n",
      " [ 41  32  57]\n",
      " [177  12 405]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.83014s\n",
      "Predict Time: 0.51707s\n",
      "Acurácia    : 51.29%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.77      0.22      0.34       635\n",
      "      NEUTRO       0.27      0.28      0.28       130\n",
      "    POSITIVO       0.50      0.88      0.64       594\n",
      "\n",
      "   micro avg       0.51      0.51      0.51      1359\n",
      "   macro avg       0.51      0.46      0.42      1359\n",
      "weighted avg       0.61      0.51      0.46      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[139  57 439]\n",
      " [ 10  37  83]\n",
      " [ 31  42 521]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(\"\\nall_datasets\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_tfidf, avalencias))\n",
    "        matriz_resultados['all']['tfidf'][classifier.__class__.__name__] = acc\n",
    "        train_time['all']['tfidf'][classifier.__class__.__name__] = tt\n",
    "        predict_time['all']['tfidf'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "print(\"\\ntweets_mg\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_tfidf_tmg, aval_tweets_mg))\n",
    "        matriz_resultados['tweets']['tfidf'][classifier.__class__.__name__] = acc\n",
    "        train_time['tweets']['tfidf'][classifier.__class__.__name__] = tt\n",
    "        predict_time['tweets']['tfidf'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "print(\"\\ntitulo_noticias\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_tfidf_tn, aval_titulo_noticias))\n",
    "        matriz_resultados['titulo_noticias']['tfidf'][classifier.__class__.__name__] = acc\n",
    "        train_time['titulo_noticias']['tfidf'][classifier.__class__.__name__] = tt\n",
    "        predict_time['titulo_noticias']['tfidf'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass      "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Ny241RlukicC"
   },
   "source": [
    "## LSA (usando TF-IDF)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:26:51.815552Z",
     "start_time": "2018-11-29T00:26:06.932372Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1054
    },
    "colab_type": "code",
    "id": "C2B7FluMklZB",
    "outputId": "b867c791-ef17-4df6-9f27-6fb0870dacd1",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "all_datasets\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.0242s\n",
      "Predict Time: 0.0012s\n",
      "Acurácia    : 43.91%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.00      0.00      0.00       790\n",
      "      NEUTRO       1.00      0.00      0.01       529\n",
      "    POSITIVO       0.44      1.00      0.61      1029\n",
      "\n",
      "   micro avg       0.44      0.44      0.44      2348\n",
      "   macro avg       0.48      0.33      0.21      2348\n",
      "weighted avg       0.42      0.44      0.27      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[   0    0  790]\n",
      " [   0    2  527]\n",
      " [   0    0 1029]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.02366s\n",
      "Predict Time: 0.00117s\n",
      "Acurácia    : 54.64%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.53      0.45      0.49       790\n",
      "      NEUTRO       0.47      0.57      0.52       529\n",
      "    POSITIVO       0.61      0.60      0.60      1029\n",
      "\n",
      "   micro avg       0.55      0.55      0.55      2348\n",
      "   macro avg       0.54      0.54      0.54      2348\n",
      "weighted avg       0.55      0.55      0.55      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[359 161 270]\n",
      " [ 91 302 136]\n",
      " [227 180 622]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 1.06266s\n",
      "Predict Time: 0.00126s\n",
      "Acurácia    : 58.73%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.52      0.63      0.57       790\n",
      "      NEUTRO       0.54      0.37      0.44       529\n",
      "    POSITIVO       0.67      0.66      0.67      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.58      0.56      0.56      2348\n",
      "weighted avg       0.59      0.59      0.58      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[497  93 200]\n",
      " [198 198 133]\n",
      " [267  78 684]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 6.65577s\n",
      "Predict Time: 0.07394s\n",
      "Acurácia    : 57.75%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.51      0.57      0.54       790\n",
      "      NEUTRO       0.56      0.44      0.49       529\n",
      "    POSITIVO       0.64      0.66      0.65      1029\n",
      "\n",
      "   micro avg       0.58      0.58      0.58      2348\n",
      "   macro avg       0.57      0.55      0.56      2348\n",
      "weighted avg       0.58      0.58      0.58      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[448 100 242]\n",
      " [164 233 132]\n",
      " [272  82 675]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.04872s\n",
      "Predict Time: 7.79778s\n",
      "Acurácia    : 54.98%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.48      0.64      0.55       790\n",
      "      NEUTRO       0.48      0.46      0.47       529\n",
      "    POSITIVO       0.69      0.52      0.60      1029\n",
      "\n",
      "   micro avg       0.55      0.55      0.55      2348\n",
      "   macro avg       0.55      0.54      0.54      2348\n",
      "weighted avg       0.57      0.55      0.55      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[509 131 150]\n",
      " [198 245  86]\n",
      " [360 132 537]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 8.38732s\n",
      "Predict Time: 0.00986s\n",
      "Acurácia    : 57.37%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.49      0.66      0.56       790\n",
      "      NEUTRO       0.53      0.38      0.45       529\n",
      "    POSITIVO       0.69      0.60      0.65      1029\n",
      "\n",
      "   micro avg       0.57      0.57      0.57      2348\n",
      "   macro avg       0.57      0.55      0.55      2348\n",
      "weighted avg       0.59      0.57      0.57      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[524  98 168]\n",
      " [221 203 105]\n",
      " [328  81 620]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 2.21711s\n",
      "Predict Time: 0.00967s\n",
      "Acurácia    : 58.99%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.53      0.59      0.56       790\n",
      "      NEUTRO       0.55      0.42      0.48       529\n",
      "    POSITIVO       0.66      0.68      0.67      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.58      0.56      0.57      2348\n",
      "weighted avg       0.59      0.59      0.59      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[467 101 222]\n",
      " [167 222 140]\n",
      " [251  82 696]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 1.80184s\n",
      "Predict Time: 1.08235s\n",
      "Acurácia    : 46.25%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.38      0.18      0.25       790\n",
      "      NEUTRO       0.45      0.13      0.21       529\n",
      "    POSITIVO       0.48      0.84      0.61      1029\n",
      "\n",
      "   micro avg       0.46      0.46      0.46      2348\n",
      "   macro avg       0.44      0.39      0.36      2348\n",
      "weighted avg       0.44      0.46      0.40      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[146  52 592]\n",
      " [110  71 348]\n",
      " [124  36 869]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "tweets_mg\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.01163s\n",
      "Predict Time: 0.01165s\n",
      "Acurácia    : 50.56%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.00      0.00      0.00       201\n",
      "      NEUTRO       0.54      0.33      0.41       362\n",
      "    POSITIVO       0.49      0.89      0.64       426\n",
      "\n",
      "   micro avg       0.51      0.51      0.51       989\n",
      "   macro avg       0.35      0.41      0.35       989\n",
      "weighted avg       0.41      0.51      0.42       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  0  55 146]\n",
      " [  0 119 243]\n",
      " [  0  45 381]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.02089s\n",
      "Predict Time: 0.00066s\n",
      "Acurácia    : 57.84%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.25      0.01      0.02       201\n",
      "      NEUTRO       0.49      0.79      0.60       362\n",
      "    POSITIVO       0.72      0.66      0.69       426\n",
      "\n",
      "   micro avg       0.58      0.58      0.58       989\n",
      "   macro avg       0.49      0.49      0.44       989\n",
      "weighted avg       0.54      0.58      0.52       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  2 160  39]\n",
      " [  3 287  72]\n",
      " [  3 140 283]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.74969s\n",
      "Predict Time: 0.0147s\n",
      "Acurácia    : 63.4%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.40      0.08      0.14       201\n",
      "      NEUTRO       0.53      0.83      0.65       362\n",
      "    POSITIVO       0.80      0.73      0.77       426\n",
      "\n",
      "   micro avg       0.63      0.63      0.63       989\n",
      "   macro avg       0.58      0.55      0.52       989\n",
      "weighted avg       0.62      0.63      0.60       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 17 157  27]\n",
      " [ 14 299  49]\n",
      " [ 11 104 311]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 2.3482s\n",
      "Predict Time: 0.03586s\n",
      "Acurácia    : 61.27%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.31      0.15      0.21       201\n",
      "      NEUTRO       0.53      0.72      0.61       362\n",
      "    POSITIVO       0.80      0.73      0.77       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.54      0.54      0.53       989\n",
      "weighted avg       0.60      0.61      0.59       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 31 138  32]\n",
      " [ 53 262  47]\n",
      " [ 17  96 313]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.02979s\n",
      "Predict Time: 1.00287s\n",
      "Acurácia    : 60.47%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.34      0.29      0.31       201\n",
      "      NEUTRO       0.52      0.72      0.61       362\n",
      "    POSITIVO       0.88      0.65      0.75       426\n",
      "\n",
      "   micro avg       0.60      0.60      0.60       989\n",
      "   macro avg       0.58      0.56      0.56       989\n",
      "weighted avg       0.64      0.60      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 59 130  12]\n",
      " [ 75 262  25]\n",
      " [ 40 109 277]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 6.01657s\n",
      "Predict Time: 0.00564s\n",
      "Acurácia    : 60.26%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.50      0.01      0.02       201\n",
      "      NEUTRO       0.52      0.72      0.61       362\n",
      "    POSITIVO       0.68      0.78      0.73       426\n",
      "\n",
      "   micro avg       0.60      0.60      0.60       989\n",
      "   macro avg       0.57      0.50      0.45       989\n",
      "weighted avg       0.59      0.60      0.54       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  2 146  53]\n",
      " [  1 260 101]\n",
      " [  1  91 334]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.82166s\n",
      "Predict Time: 0.00076s\n",
      "Acurácia    : 64.31%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.45      0.06      0.11       201\n",
      "      NEUTRO       0.54      0.86      0.66       362\n",
      "    POSITIVO       0.81      0.73      0.77       426\n",
      "\n",
      "   micro avg       0.64      0.64      0.64       989\n",
      "   macro avg       0.60      0.55      0.52       989\n",
      "weighted avg       0.64      0.64      0.60       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 13 159  29]\n",
      " [  8 312  42]\n",
      " [  8 107 311]]\n",
      "------------------------------\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo      : SVC\n",
      "Train Time  : 0.76614s\n",
      "Predict Time: 0.38243s\n",
      "Acurácia    : 46.01%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.28      0.11      0.16       201\n",
      "      NEUTRO       0.48      0.11      0.18       362\n",
      "    POSITIVO       0.48      0.92      0.63       426\n",
      "\n",
      "   micro avg       0.46      0.46      0.46       989\n",
      "   macro avg       0.41      0.38      0.32       989\n",
      "weighted avg       0.44      0.46      0.37       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 22  28 151]\n",
      " [ 39  40 283]\n",
      " [ 17  16 393]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "titulo_noticias\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.01477s\n",
      "Predict Time: 0.00198s\n",
      "Acurácia    : 46.73%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.47      1.00      0.64       635\n",
      "      NEUTRO       0.00      0.00      0.00       130\n",
      "    POSITIVO       0.00      0.00      0.00       594\n",
      "\n",
      "   micro avg       0.47      0.47      0.47      1359\n",
      "   macro avg       0.16      0.33      0.21      1359\n",
      "weighted avg       0.22      0.47      0.30      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[635   0   0]\n",
      " [130   0   0]\n",
      " [594   0   0]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.01448s\n",
      "Predict Time: 0.00082s\n",
      "Acurácia    : 57.17%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.61      0.63      0.62       635\n",
      "      NEUTRO       0.32      0.34      0.33       130\n",
      "    POSITIVO       0.59      0.56      0.57       594\n",
      "\n",
      "   micro avg       0.57      0.57      0.57      1359\n",
      "   macro avg       0.51      0.51      0.51      1359\n",
      "weighted avg       0.57      0.57      0.57      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[401  46 188]\n",
      " [ 41  44  45]\n",
      " [214  48 332]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.65222s\n",
      "Predict Time: 0.00164s\n",
      "Acurácia    : 63.72%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.65      0.74      0.69       635\n",
      "      NEUTRO       0.47      0.17      0.25       130\n",
      "    POSITIVO       0.64      0.63      0.63       594\n",
      "\n",
      "   micro avg       0.64      0.64      0.64      1359\n",
      "   macro avg       0.58      0.51      0.52      1359\n",
      "weighted avg       0.63      0.64      0.62      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[468   7 160]\n",
      " [ 53  22  55]\n",
      " [200  18 376]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 3.48411s\n",
      "Predict Time: 0.03911s\n",
      "Acurácia    : 60.41%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.61      0.73      0.67       635\n",
      "      NEUTRO       0.46      0.10      0.16       130\n",
      "    POSITIVO       0.60      0.58      0.59       594\n",
      "\n",
      "   micro avg       0.60      0.60      0.60      1359\n",
      "   macro avg       0.56      0.47      0.47      1359\n",
      "weighted avg       0.59      0.60      0.58      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[463   5 167]\n",
      " [ 52  13  65]\n",
      " [239  10 345]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.02049s\n",
      "Predict Time: 1.30266s\n",
      "Acurácia    : 54.97%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.56      0.80      0.66       635\n",
      "      NEUTRO       0.24      0.17      0.20       130\n",
      "    POSITIVO       0.59      0.37      0.45       594\n",
      "\n",
      "   micro avg       0.55      0.55      0.55      1359\n",
      "   macro avg       0.47      0.44      0.44      1359\n",
      "weighted avg       0.54      0.55      0.53      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[507  22 106]\n",
      " [ 61  22  47]\n",
      " [330  46 218]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 9.2845s\n",
      "Predict Time: 0.00498s\n",
      "Acurácia    : 61.52%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.62      0.77      0.69       635\n",
      "      NEUTRO       0.35      0.32      0.33       130\n",
      "    POSITIVO       0.67      0.51      0.58       594\n",
      "\n",
      "   micro avg       0.62      0.62      0.62      1359\n",
      "   macro avg       0.55      0.53      0.53      1359\n",
      "weighted avg       0.62      0.62      0.61      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[492  30 113]\n",
      " [ 52  41  37]\n",
      " [246  45 303]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 1.18968s\n",
      "Predict Time: 0.00096s\n",
      "Acurácia    : 62.99%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.63      0.75      0.69       635\n",
      "      NEUTRO       0.44      0.16      0.24       130\n",
      "    POSITIVO       0.64      0.60      0.62       594\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      1359\n",
      "   macro avg       0.57      0.51      0.51      1359\n",
      "weighted avg       0.62      0.63      0.62      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[478   9 148]\n",
      " [ 59  21  50]\n",
      " [219  18 357]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 1.0472s\n",
      "Predict Time: 0.55322s\n",
      "Acurácia    : 45.33%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.55      0.14      0.22       635\n",
      "      NEUTRO       0.25      0.13      0.17       130\n",
      "    POSITIVO       0.45      0.86      0.59       594\n",
      "\n",
      "   micro avg       0.45      0.45      0.45      1359\n",
      "   macro avg       0.42      0.38      0.33      1359\n",
      "weighted avg       0.48      0.45      0.38      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 86  28 521]\n",
      " [ 12  17 101]\n",
      " [ 57  24 513]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "vec_tfidf = TfidfVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_tfidf_svd = vec_tfidf.fit_transform(afrases)\n",
    "\n",
    "#all_datasets\n",
    "svd = TruncatedSVD(100, random_state=10)\n",
    "normalizer = MinMaxScaler(copy=False)\n",
    "lsa = make_pipeline(svd, normalizer)\n",
    "X_svd = lsa.fit_transform(X_tfidf_svd)\n",
    "\n",
    "print(\"\\nall_datasets\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_svd, avalencias))\n",
    "        matriz_resultados['all']['tfidf+lsa'][classifier.__class__.__name__] = acc\n",
    "        train_time['all']['tfidf+lsa'][classifier.__class__.__name__] = tt\n",
    "        predict_time['all']['tfidf+lsa'][classifier.__class__.__name__] = pt\n",
    "    except Exception as e:\n",
    "        print(e)\n",
    "\n",
    "#tweets_mg\n",
    "vec_tfidf_tmg = TfidfVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_tfidf_svd_tmg = vec_tfidf_tmg.fit_transform(atweets_mg)\n",
    "X_svd = lsa.fit_transform(X_tfidf_svd_tmg)\n",
    "\n",
    "print(\"\\ntweets_mg\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_svd, aval_tweets_mg))\n",
    "        matriz_resultados['tweets']['tfidf+lsa'][classifier.__class__.__name__] = acc\n",
    "        train_time['tweets']['tfidf+lsa'][classifier.__class__.__name__] = tt\n",
    "        predict_time['tweets']['tfidf+lsa'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "\n",
    "#titulo_noticias\n",
    "vec_tfidf_tn = TfidfVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_tfidf_svd_tn = vec_tfidf_tn.fit_transform(atitulo_noticias)\n",
    "X_svd = lsa.fit_transform(X_tfidf_svd_tn)\n",
    "\n",
    "print(\"\\ntitulo_noticias\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_svd, aval_titulo_noticias))\n",
    "        matriz_resultados['titulo_noticias']['tfidf+lsa'][classifier.__class__.__name__] = acc\n",
    "        train_time['titulo_noticias']['tfidf+lsa'][classifier.__class__.__name__] = tt\n",
    "        predict_time['titulo_noticias']['tfidf+lsa'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "6gV9r8SSlNUd"
   },
   "source": [
    "## LDA (usando TF-IDF)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:28:25.677529Z",
     "start_time": "2018-11-29T00:26:51.817359Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1037
    },
    "colab_type": "code",
    "id": "hYzcOnavlS6o",
    "outputId": "c3065879-7b29-48ea-ca7a-001316da7f0a",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "all_datasets\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.0186s\n",
      "Predict Time: 0.00125s\n",
      "Acurácia    : 48.47%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.42      0.47      0.44       790\n",
      "      NEUTRO       0.57      0.12      0.19       529\n",
      "    POSITIVO       0.52      0.69      0.59      1029\n",
      "\n",
      "   micro avg       0.48      0.48      0.48      2348\n",
      "   macro avg       0.50      0.42      0.41      2348\n",
      "weighted avg       0.50      0.48      0.45      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[370  23 397]\n",
      " [216  62 251]\n",
      " [299  24 706]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.02311s\n",
      "Predict Time: 0.00117s\n",
      "Acurácia    : 50.0%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.44      0.52      0.48       790\n",
      "      NEUTRO       0.43      0.48      0.45       529\n",
      "    POSITIVO       0.62      0.49      0.55      1029\n",
      "\n",
      "   micro avg       0.50      0.50      0.50      2348\n",
      "   macro avg       0.50      0.50      0.49      2348\n",
      "weighted avg       0.52      0.50      0.50      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[414 175 201]\n",
      " [163 255 111]\n",
      " [360 164 505]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.73606s\n",
      "Predict Time: 0.00124s\n",
      "Acurácia    : 50.13%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.43      0.51      0.47       790\n",
      "      NEUTRO       0.50      0.32      0.39       529\n",
      "    POSITIVO       0.56      0.59      0.58      1029\n",
      "\n",
      "   micro avg       0.50      0.50      0.50      2348\n",
      "   macro avg       0.50      0.47      0.48      2348\n",
      "weighted avg       0.50      0.50      0.50      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[401  84 305]\n",
      " [187 167 175]\n",
      " [336  84 609]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 8.03179s\n",
      "Predict Time: 0.0438s\n",
      "Acurácia    : 50.21%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.43      0.47      0.45       790\n",
      "      NEUTRO       0.48      0.33      0.39       529\n",
      "    POSITIVO       0.57      0.62      0.59      1029\n",
      "\n",
      "   micro avg       0.50      0.50      0.50      2348\n",
      "   macro avg       0.49      0.47      0.48      2348\n",
      "weighted avg       0.50      0.50      0.50      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[371 110 309]\n",
      " [179 175 175]\n",
      " [316  80 633]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.02413s\n",
      "Predict Time: 2.84385s\n",
      "Acurácia    : 45.36%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.41      0.51      0.46       790\n",
      "      NEUTRO       0.32      0.39      0.35       529\n",
      "    POSITIVO       0.62      0.45      0.52      1029\n",
      "\n",
      "   micro avg       0.45      0.45      0.45      2348\n",
      "   macro avg       0.45      0.45      0.44      2348\n",
      "weighted avg       0.48      0.45      0.46      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[401 207 182]\n",
      " [223 205 101]\n",
      " [345 225 459]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 32.79967s\n",
      "Predict Time: 0.01163s\n",
      "Acurácia    : 48.94%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.43      0.43      0.43       790\n",
      "      NEUTRO       0.40      0.39      0.40       529\n",
      "    POSITIVO       0.58      0.59      0.58      1029\n",
      "\n",
      "   micro avg       0.49      0.49      0.49      2348\n",
      "   macro avg       0.47      0.47      0.47      2348\n",
      "weighted avg       0.49      0.49      0.49      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[336 162 292]\n",
      " [172 208 149]\n",
      " [270 154 605]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.79302s\n",
      "Predict Time: 0.00147s\n",
      "Acurácia    : 50.13%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.44      0.51      0.47       790\n",
      "      NEUTRO       0.49      0.33      0.39       529\n",
      "    POSITIVO       0.56      0.58      0.57      1029\n",
      "\n",
      "   micro avg       0.50      0.50      0.50      2348\n",
      "   macro avg       0.50      0.47      0.48      2348\n",
      "weighted avg       0.50      0.50      0.50      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[403  89 298]\n",
      " [186 173 170]\n",
      " [336  92 601]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 1.81846s\n",
      "Predict Time: 1.0698s\n",
      "Acurácia    : 41.23%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.36      0.14      0.20       790\n",
      "      NEUTRO       0.31      0.16      0.21       529\n",
      "    POSITIVO       0.44      0.75      0.55      1029\n",
      "\n",
      "   micro avg       0.41      0.41      0.41      2348\n",
      "   macro avg       0.37      0.35      0.32      2348\n",
      "weighted avg       0.38      0.41      0.36      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[113  68 609]\n",
      " [ 67  84 378]\n",
      " [136 122 771]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "tweets_mg\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00858s\n",
      "Predict Time: 0.00087s\n",
      "Acurácia    : 55.71%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.00      0.00      0.00       201\n",
      "      NEUTRO       0.49      0.66      0.56       362\n",
      "    POSITIVO       0.62      0.73      0.67       426\n",
      "\n",
      "   micro avg       0.56      0.56      0.56       989\n",
      "   macro avg       0.37      0.46      0.41       989\n",
      "weighted avg       0.45      0.56      0.50       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  0 136  65]\n",
      " [  0 240 122]\n",
      " [  0 115 311]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.01123s\n",
      "Predict Time: 0.00069s\n",
      "Acurácia    : 52.07%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.25      0.23      0.24       201\n",
      "      NEUTRO       0.49      0.53      0.51       362\n",
      "    POSITIVO       0.67      0.65      0.66       426\n",
      "\n",
      "   micro avg       0.52      0.52      0.52       989\n",
      "   macro avg       0.47      0.47      0.47       989\n",
      "weighted avg       0.52      0.52      0.52       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 46 108  47]\n",
      " [ 79 193  90]\n",
      " [ 57  93 276]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.30723s\n",
      "Predict Time: 0.00077s\n",
      "Acurácia    : 55.31%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.32      0.04      0.07       201\n",
      "      NEUTRO       0.49      0.68      0.57       362\n",
      "    POSITIVO       0.64      0.69      0.66       426\n",
      "\n",
      "   micro avg       0.55      0.55      0.55       989\n",
      "   macro avg       0.48      0.47      0.43       989\n",
      "weighted avg       0.52      0.55      0.51       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  8 133  60]\n",
      " [  9 245 108]\n",
      " [  8 124 294]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 2.51137s\n",
      "Predict Time: 0.04095s\n",
      "Acurácia    : 53.79%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.30      0.16      0.21       201\n",
      "      NEUTRO       0.48      0.58      0.53       362\n",
      "    POSITIVO       0.65      0.68      0.66       426\n",
      "\n",
      "   micro avg       0.54      0.54      0.54       989\n",
      "   macro avg       0.48      0.47      0.47       989\n",
      "weighted avg       0.52      0.54      0.52       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 32 111  58]\n",
      " [ 53 211  98]\n",
      " [ 22 115 289]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.01879s\n",
      "Predict Time: 1.03105s\n",
      "Acurácia    : 53.59%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.24      0.16      0.19       201\n",
      "      NEUTRO       0.47      0.69      0.56       362\n",
      "    POSITIVO       0.75      0.58      0.66       426\n",
      "\n",
      "   micro avg       0.54      0.54      0.54       989\n",
      "   macro avg       0.49      0.48      0.47       989\n",
      "weighted avg       0.55      0.54      0.53       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 32 140  29]\n",
      " [ 58 249  55]\n",
      " [ 41 136 249]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 12.12654s\n",
      "Predict Time: 0.0057s\n",
      "Acurácia    : 53.69%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.29      0.27      0.28       201\n",
      "      NEUTRO       0.48      0.57      0.52       362\n",
      "    POSITIVO       0.74      0.63      0.68       426\n",
      "\n",
      "   micro avg       0.54      0.54      0.54       989\n",
      "   macro avg       0.50      0.49      0.49       989\n",
      "weighted avg       0.55      0.54      0.54       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 55 115  31]\n",
      " [ 89 208  65]\n",
      " [ 44 114 268]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.31729s\n",
      "Predict Time: 0.00181s\n",
      "Acurácia    : 55.71%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.30      0.06      0.11       201\n",
      "      NEUTRO       0.49      0.68      0.57       362\n",
      "    POSITIVO       0.65      0.68      0.67       426\n",
      "\n",
      "   micro avg       0.56      0.56      0.56       989\n",
      "   macro avg       0.48      0.48      0.45       989\n",
      "weighted avg       0.52      0.56      0.52       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 13 132  56]\n",
      " [ 15 247 100]\n",
      " [ 15 120 291]]\n",
      "------------------------------\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo      : SVC\n",
      "Train Time  : 0.76851s\n",
      "Predict Time: 0.43126s\n",
      "Acurácia    : 46.21%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.24      0.14      0.18       201\n",
      "      NEUTRO       0.49      0.18      0.26       362\n",
      "    POSITIVO       0.49      0.85      0.62       426\n",
      "\n",
      "   micro avg       0.46      0.46      0.46       989\n",
      "   macro avg       0.41      0.39      0.36       989\n",
      "weighted avg       0.44      0.46      0.40       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 29  37 135]\n",
      " [ 59  65 238]\n",
      " [ 33  30 363]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "titulo_noticias\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.0108s\n",
      "Predict Time: 0.00099s\n",
      "Acurácia    : 53.57%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.54      0.76      0.63       635\n",
      "      NEUTRO       0.00      0.00      0.00       130\n",
      "    POSITIVO       0.53      0.41      0.46       594\n",
      "\n",
      "   micro avg       0.54      0.54      0.54      1359\n",
      "   macro avg       0.36      0.39      0.36      1359\n",
      "weighted avg       0.48      0.54      0.50      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[484   0 151]\n",
      " [ 68   0  62]\n",
      " [350   0 244]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.01521s\n",
      "Predict Time: 0.0008s\n",
      "Acurácia    : 48.64%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.53      0.55       635\n",
      "      NEUTRO       0.18      0.30      0.22       130\n",
      "    POSITIVO       0.52      0.48      0.50       594\n",
      "\n",
      "   micro avg       0.49      0.49      0.49      1359\n",
      "   macro avg       0.42      0.44      0.42      1359\n",
      "weighted avg       0.51      0.49      0.50      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[339  89 207]\n",
      " [ 39  39  52]\n",
      " [221  90 283]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.39617s\n",
      "Predict Time: 0.00103s\n",
      "Acurácia    : 53.79%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.55      0.67      0.61       635\n",
      "      NEUTRO       0.40      0.08      0.13       130\n",
      "    POSITIVO       0.53      0.49      0.51       594\n",
      "\n",
      "   micro avg       0.54      0.54      0.54      1359\n",
      "   macro avg       0.49      0.41      0.41      1359\n",
      "weighted avg       0.53      0.54      0.52      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[427   6 202]\n",
      " [ 58  10  62]\n",
      " [291   9 294]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 4.32926s\n",
      "Predict Time: 0.0536s\n",
      "Acurácia    : 52.24%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.54      0.63      0.58       635\n",
      "      NEUTRO       0.24      0.06      0.10       130\n",
      "    POSITIVO       0.51      0.51      0.51       594\n",
      "\n",
      "   micro avg       0.52      0.52      0.52      1359\n",
      "   macro avg       0.43      0.40      0.40      1359\n",
      "weighted avg       0.50      0.52      0.50      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[400  12 223]\n",
      " [ 60   8  62]\n",
      " [279  13 302]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.02605s\n",
      "Predict Time: 1.53888s\n",
      "Acurácia    : 50.4%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.52      0.70      0.59       635\n",
      "      NEUTRO       0.28      0.15      0.20       130\n",
      "    POSITIVO       0.52      0.37      0.43       594\n",
      "\n",
      "   micro avg       0.50      0.50      0.50      1359\n",
      "   macro avg       0.44      0.41      0.41      1359\n",
      "weighted avg       0.49      0.50      0.49      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[444  27 164]\n",
      " [ 67  20  43]\n",
      " [349  24 221]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 17.96285s\n",
      "Predict Time: 0.007s\n",
      "Acurácia    : 48.86%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.55      0.54      0.55       635\n",
      "      NEUTRO       0.15      0.18      0.17       130\n",
      "    POSITIVO       0.51      0.50      0.50       594\n",
      "\n",
      "   micro avg       0.49      0.49      0.49      1359\n",
      "   macro avg       0.41      0.41      0.41      1359\n",
      "weighted avg       0.50      0.49      0.49      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[345  70 220]\n",
      " [ 45  24  61]\n",
      " [234  65 295]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.42384s\n",
      "Predict Time: 0.00084s\n",
      "Acurácia    : 53.42%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.55      0.66      0.60       635\n",
      "      NEUTRO       0.33      0.09      0.14       130\n",
      "    POSITIVO       0.52      0.50      0.51       594\n",
      "\n",
      "   micro avg       0.53      0.53      0.53      1359\n",
      "   macro avg       0.47      0.42      0.42      1359\n",
      "weighted avg       0.52      0.53      0.52      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[419  10 206]\n",
      " [ 57  12  61]\n",
      " [285  14 295]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 1.04952s\n",
      "Predict Time: 0.58414s\n",
      "Acurácia    : 45.18%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.49      0.36      0.42       635\n",
      "      NEUTRO       0.08      0.04      0.05       130\n",
      "    POSITIVO       0.46      0.64      0.53       594\n",
      "\n",
      "   micro avg       0.45      0.45      0.45      1359\n",
      "   macro avg       0.34      0.35      0.33      1359\n",
      "weighted avg       0.44      0.45      0.43      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[231  35 369]\n",
      " [ 43   5  82]\n",
      " [194  22 378]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#all_datasets\n",
    "vec_tfidf = TfidfVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_tfidf_svd = vec_tfidf.fit_transform(afrases)\n",
    "\n",
    "lda = LatentDirichletAllocation(n_components=100, max_iter=25, random_state=10, n_jobs=5)\n",
    "normalizer = MinMaxScaler(copy=False)\n",
    "lda = make_pipeline(lda, normalizer)\n",
    "X_lda = lda.fit_transform(X_tfidf_svd)\n",
    "\n",
    "print(\"\\nall_datasets\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_lda, avalencias))\n",
    "        matriz_resultados['all']['tfidf+lda'][classifier.__class__.__name__] = acc\n",
    "        train_time['all']['tfidf+lda'][classifier.__class__.__name__] = tt\n",
    "        predict_time['all']['tfidf+lda'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "\n",
    "#tweets_mg\n",
    "vec_tfidf_tmg = TfidfVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_tfidf_svd_tmg = vec_tfidf_tmg.fit_transform(atweets_mg)\n",
    "X_lda = lda.fit_transform(X_tfidf_svd_tmg)\n",
    "\n",
    "print(\"\\ntweets_mg\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_lda, aval_tweets_mg))\n",
    "        matriz_resultados['tweets']['tfidf+lda'][classifier.__class__.__name__] = acc\n",
    "        train_time['tweets']['tfidf+lda'][classifier.__class__.__name__] = tt\n",
    "        predict_time['tweets']['tfidf+lda'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "\n",
    "#titulo_noticias\n",
    "vec_tfidf_tn = TfidfVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_tfidf_svd_tn = vec_tfidf_tn.fit_transform(atitulo_noticias)\n",
    "X_lda = lda.fit_transform(X_tfidf_svd_tn)\n",
    "\n",
    "print(\"\\ntitulo_noticias\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_lda, aval_titulo_noticias))\n",
    "        matriz_resultados['titulo_noticias']['tfidf+lda'][classifier.__class__.__name__] = acc\n",
    "        train_time['titulo_noticias']['tfidf+lda'][classifier.__class__.__name__] = tt\n",
    "        predict_time['titulo_noticias']['tfidf+lda'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "xyJKv6JUlhox"
   },
   "source": [
    "## Count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:35:52.891169Z",
     "start_time": "2018-11-29T00:28:25.679702Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1037
    },
    "colab_type": "code",
    "id": "bohbg-sHlgop",
    "outputId": "336629b8-a138-423d-9fd2-39e6aa3dae24",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "all_datasets\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.0262s\n",
      "Predict Time: 0.00281s\n",
      "Acurácia    : 63.2%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.65      0.61       790\n",
      "      NEUTRO       0.55      0.51      0.53       529\n",
      "    POSITIVO       0.74      0.68      0.71      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.62      0.61      0.61      2348\n",
      "weighted avg       0.64      0.63      0.63      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[516 138 136]\n",
      " [147 270 112]\n",
      " [244  87 698]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.035s\n",
      "Predict Time: 0.00291s\n",
      "Acurácia    : 63.12%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.60      0.61      0.60       790\n",
      "      NEUTRO       0.50      0.65      0.56       529\n",
      "    POSITIVO       0.77      0.64      0.70      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.62      0.63      0.62      2348\n",
      "weighted avg       0.65      0.63      0.64      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[480 197 113]\n",
      " [101 344  84]\n",
      " [219 152 658]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 10.64482s\n",
      "Predict Time: 0.00464s\n",
      "Acurácia    : 64.18%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.58      0.65      0.61       790\n",
      "      NEUTRO       0.57      0.48      0.52       529\n",
      "    POSITIVO       0.73      0.72      0.72      1029\n",
      "\n",
      "   micro avg       0.64      0.64      0.64      2348\n",
      "   macro avg       0.63      0.62      0.62      2348\n",
      "weighted avg       0.64      0.64      0.64      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[515 113 162]\n",
      " [165 254 110]\n",
      " [213  78 738]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 21.97293s\n",
      "Predict Time: 0.15753s\n",
      "Acurácia    : 62.82%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.64      0.60       790\n",
      "      NEUTRO       0.56      0.54      0.55       529\n",
      "    POSITIVO       0.72      0.67      0.69      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.62      0.61      0.61      2348\n",
      "weighted avg       0.63      0.63      0.63      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[506 126 158]\n",
      " [138 284 107]\n",
      " [244 100 685]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00272s\n",
      "Predict Time: 0.62637s\n",
      "Acurácia    : 34.11%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.39      0.07      0.12       790\n",
      "      NEUTRO       0.24      0.82      0.37       529\n",
      "    POSITIVO       0.76      0.31      0.43      1029\n",
      "\n",
      "   micro avg       0.34      0.34      0.34      2348\n",
      "   macro avg       0.46      0.40      0.31      2348\n",
      "weighted avg       0.52      0.34      0.31      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 54 677  59]\n",
      " [ 54 433  42]\n",
      " [ 30 685 314]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 420.74058s\n",
      "Predict Time: 0.01651s\n",
      "Acurácia    : 61.75%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.56      0.63      0.59       790\n",
      "      NEUTRO       0.53      0.44      0.48       529\n",
      "    POSITIVO       0.70      0.70      0.70      1029\n",
      "\n",
      "   micro avg       0.62      0.62      0.62      2348\n",
      "   macro avg       0.60      0.59      0.59      2348\n",
      "weighted avg       0.62      0.62      0.62      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[498 117 175]\n",
      " [168 235 126]\n",
      " [219  93 717]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.33628s\n",
      "Predict Time: 0.00056s\n",
      "Acurácia    : 62.73%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.62      0.59       790\n",
      "      NEUTRO       0.53      0.48      0.50       529\n",
      "    POSITIVO       0.73      0.71      0.72      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.61      0.60      0.60      2348\n",
      "weighted avg       0.63      0.63      0.63      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[491 131 168]\n",
      " [169 254 106]\n",
      " [204  97 728]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.43936s\n",
      "Predict Time: 0.30641s\n",
      "Acurácia    : 48.04%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.56      0.13      0.20       790\n",
      "      NEUTRO       0.51      0.18      0.27       529\n",
      "    POSITIVO       0.47      0.91      0.62      1029\n",
      "\n",
      "   micro avg       0.48      0.48      0.48      2348\n",
      "   macro avg       0.51      0.40      0.36      2348\n",
      "weighted avg       0.51      0.48      0.40      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 99  55 636]\n",
      " [ 20  96 413]\n",
      " [ 58  38 933]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "tweets_mg\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00594s\n",
      "Predict Time: 0.00087s\n",
      "Acurácia    : 60.57%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.30      0.28      0.29       201\n",
      "      NEUTRO       0.54      0.62      0.58       362\n",
      "    POSITIVO       0.83      0.75      0.79       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.56      0.55      0.55       989\n",
      "weighted avg       0.62      0.61      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 57 124  20]\n",
      " [ 95 223  44]\n",
      " [ 41  66 319]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.0062s\n",
      "Predict Time: 0.00041s\n",
      "Acurácia    : 59.56%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.30      0.38      0.34       201\n",
      "      NEUTRO       0.55      0.53      0.54       362\n",
      "    POSITIVO       0.83      0.75      0.79       426\n",
      "\n",
      "   micro avg       0.60      0.60      0.60       989\n",
      "   macro avg       0.56      0.55      0.56       989\n",
      "weighted avg       0.62      0.60      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 77 104  20]\n",
      " [126 191  45]\n",
      " [ 53  52 321]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 2.64707s\n",
      "Predict Time: 0.00103s\n",
      "Acurácia    : 63.3%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.35      0.21      0.27       201\n",
      "      NEUTRO       0.54      0.71      0.61       362\n",
      "    POSITIVO       0.83      0.77      0.80       426\n",
      "\n",
      "   micro avg       0.63      0.63      0.63       989\n",
      "   macro avg       0.57      0.56      0.56       989\n",
      "weighted avg       0.63      0.63      0.62       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 43 136  22]\n",
      " [ 61 256  45]\n",
      " [ 19  80 327]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 1.71502s\n",
      "Predict Time: 0.03709s\n",
      "Acurácia    : 62.59%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.29      0.08      0.12       201\n",
      "      NEUTRO       0.51      0.82      0.63       362\n",
      "    POSITIVO       0.86      0.72      0.78       426\n",
      "\n",
      "   micro avg       0.63      0.63      0.63       989\n",
      "   macro avg       0.55      0.54      0.51       989\n",
      "weighted avg       0.62      0.63      0.59       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 16 170  15]\n",
      " [ 31 298  33]\n",
      " [  9 112 305]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00149s\n",
      "Predict Time: 0.10617s\n",
      "Acurácia    : 50.76%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.32      0.19      0.24       201\n",
      "      NEUTRO       0.43      0.82      0.56       362\n",
      "    POSITIVO       0.95      0.39      0.55       426\n",
      "\n",
      "   micro avg       0.51      0.51      0.51       989\n",
      "   macro avg       0.57      0.47      0.45       989\n",
      "weighted avg       0.63      0.51      0.49       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 38 162   1]\n",
      " [ 57 298   7]\n",
      " [ 22 238 166]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 33.42182s\n",
      "Predict Time: 0.0262s\n",
      "Acurácia    : 61.07%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.36      0.30      0.33       201\n",
      "      NEUTRO       0.55      0.59      0.57       362\n",
      "    POSITIVO       0.77      0.77      0.77       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.56      0.55      0.55       989\n",
      "weighted avg       0.60      0.61      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 60 109  32]\n",
      " [ 79 214  69]\n",
      " [ 28  68 330]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.1838s\n",
      "Predict Time: 0.00035s\n",
      "Acurácia    : 61.48%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.36      0.28      0.31       201\n",
      "      NEUTRO       0.53      0.64      0.58       362\n",
      "    POSITIVO       0.80      0.76      0.78       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.56      0.56      0.56       989\n",
      "weighted avg       0.61      0.61      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 56 121  24]\n",
      " [ 76 230  56]\n",
      " [ 23  81 322]]\n",
      "------------------------------\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo      : SVC\n",
      "Train Time  : 0.25913s\n",
      "Predict Time: 0.18555s\n",
      "Acurácia    : 48.63%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.34      0.10      0.16       201\n",
      "      NEUTRO       0.54      0.17      0.26       362\n",
      "    POSITIVO       0.49      0.94      0.64       426\n",
      "\n",
      "   micro avg       0.49      0.49      0.49       989\n",
      "   macro avg       0.46      0.40      0.35       989\n",
      "weighted avg       0.48      0.49      0.40       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 21  32 148]\n",
      " [ 33  61 268]\n",
      " [  7  20 399]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "titulo_noticias\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00879s\n",
      "Predict Time: 0.00106s\n",
      "Acurácia    : 65.56%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.71      0.67      0.69       635\n",
      "      NEUTRO       0.48      0.22      0.31       130\n",
      "    POSITIVO       0.62      0.73      0.67       594\n",
      "\n",
      "   micro avg       0.66      0.66      0.66      1359\n",
      "   macro avg       0.61      0.54      0.56      1359\n",
      "weighted avg       0.65      0.66      0.65      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[427  15 193]\n",
      " [ 31  29  70]\n",
      " [143  16 435]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.01115s\n",
      "Predict Time: 0.00092s\n",
      "Acurácia    : 60.49%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.73      0.60      0.66       635\n",
      "      NEUTRO       0.25      0.52      0.34       130\n",
      "    POSITIVO       0.65      0.63      0.64       594\n",
      "\n",
      "   micro avg       0.60      0.60      0.60      1359\n",
      "   macro avg       0.55      0.58      0.55      1359\n",
      "weighted avg       0.65      0.60      0.62      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[378  99 158]\n",
      " [ 20  67  43]\n",
      " [119  98 377]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 6.46886s\n",
      "Predict Time: 0.00307s\n",
      "Acurácia    : 66.96%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.69      0.74      0.72       635\n",
      "      NEUTRO       0.50      0.26      0.34       130\n",
      "    POSITIVO       0.67      0.68      0.67       594\n",
      "\n",
      "   micro avg       0.67      0.67      0.67      1359\n",
      "   macro avg       0.62      0.56      0.58      1359\n",
      "weighted avg       0.66      0.67      0.66      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[473  17 145]\n",
      " [ 38  34  58]\n",
      " [174  17 403]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 5.72836s\n",
      "Predict Time: 0.06809s\n",
      "Acurácia    : 63.8%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.61      0.83      0.70       635\n",
      "      NEUTRO       0.57      0.13      0.21       130\n",
      "    POSITIVO       0.69      0.54      0.61       594\n",
      "\n",
      "   micro avg       0.64      0.64      0.64      1359\n",
      "   macro avg       0.62      0.50      0.51      1359\n",
      "weighted avg       0.64      0.64      0.62      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[528   5 102]\n",
      " [ 73  17  40]\n",
      " [264   8 322]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00167s\n",
      "Predict Time: 0.15371s\n",
      "Acurácia    : 47.76%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.51      0.77      0.62       635\n",
      "      NEUTRO       0.15      0.20      0.17       130\n",
      "    POSITIVO       0.59      0.22      0.32       594\n",
      "\n",
      "   micro avg       0.48      0.48      0.48      1359\n",
      "   macro avg       0.42      0.40      0.37      1359\n",
      "weighted avg       0.51      0.48      0.45      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[491  71  73]\n",
      " [ 86  26  18]\n",
      " [381  81 132]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 79.28158s\n",
      "Predict Time: 0.02133s\n",
      "Acurácia    : 67.18%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.70      0.74      0.72       635\n",
      "      NEUTRO       0.47      0.36      0.41       130\n",
      "    POSITIVO       0.68      0.67      0.67       594\n",
      "\n",
      "   micro avg       0.67      0.67      0.67      1359\n",
      "   macro avg       0.61      0.59      0.60      1359\n",
      "weighted avg       0.67      0.67      0.67      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[469  25 141]\n",
      " [ 33  47  50]\n",
      " [168  29 397]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.25415s\n",
      "Predict Time: 0.00052s\n",
      "Acurácia    : 65.86%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.68      0.73      0.71       635\n",
      "      NEUTRO       0.45      0.32      0.37       130\n",
      "    POSITIVO       0.66      0.65      0.66       594\n",
      "\n",
      "   micro avg       0.66      0.66      0.66      1359\n",
      "   macro avg       0.60      0.57      0.58      1359\n",
      "weighted avg       0.65      0.66      0.65      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[466  26 143]\n",
      " [ 35  41  54]\n",
      " [182  24 388]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.33907s\n",
      "Predict Time: 0.18923s\n",
      "Acurácia    : 45.18%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.65      0.09      0.15       635\n",
      "      NEUTRO       0.18      0.08      0.12       130\n",
      "    POSITIVO       0.45      0.92      0.61       594\n",
      "\n",
      "   micro avg       0.45      0.45      0.45      1359\n",
      "   macro avg       0.43      0.36      0.29      1359\n",
      "weighted avg       0.52      0.45      0.35      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 55  30 550]\n",
      " [  4  11 115]\n",
      " [ 26  20 548]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#all_datasets\n",
    "vec_count = CountVectorizer(ngram_range=(1, 2))\n",
    "X_count = vec_count.fit_transform(afrases)\n",
    "\n",
    "print(\"\\nall_datasets\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_count, avalencias))\n",
    "        matriz_resultados['all']['count'][classifier.__class__.__name__] = acc\n",
    "        train_time['all']['count'][classifier.__class__.__name__] = tt\n",
    "        predict_time['all']['count'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "      \n",
    "#tweets_mg\n",
    "vec_count = CountVectorizer(ngram_range=(1, 2))\n",
    "X_count_tmg = vec_count.fit_transform(atweets_mg)\n",
    "\n",
    "print(\"\\ntweets_mg\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_count_tmg, aval_tweets_mg))\n",
    "        matriz_resultados['tweets']['count'][classifier.__class__.__name__] = acc\n",
    "        train_time['tweets']['count'][classifier.__class__.__name__] = tt\n",
    "        predict_time['tweets']['count'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "      \n",
    "      \n",
    "#titulo_noticias\n",
    "vec_count = CountVectorizer(ngram_range=(1, 2))\n",
    "X_count_tn = vec_count.fit_transform(atitulo_noticias)\n",
    "\n",
    "print(\"\\ntitulo_noticias\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_count_tn, aval_titulo_noticias))\n",
    "        matriz_resultados['titulo_noticias']['count'][classifier.__class__.__name__] = acc\n",
    "        train_time['titulo_noticias']['count'][classifier.__class__.__name__] = tt\n",
    "        predict_time['titulo_noticias']['count'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "fyGZBGpklwRd"
   },
   "source": [
    "## LSA (usando Count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:36:30.738714Z",
     "start_time": "2018-11-29T00:35:52.893873Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1037
    },
    "colab_type": "code",
    "id": "uUzc5Dr9lx9S",
    "outputId": "e40ec931-ab0c-474b-b065-2c543c487ada",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "all_datasets\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.02481s\n",
      "Predict Time: 0.00102s\n",
      "Acurácia    : 43.82%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.00      0.00      0.00       790\n",
      "      NEUTRO       0.33      0.00      0.00       529\n",
      "    POSITIVO       0.44      1.00      0.61      1029\n",
      "\n",
      "   micro avg       0.44      0.44      0.44      2348\n",
      "   macro avg       0.26      0.33      0.20      2348\n",
      "weighted avg       0.27      0.44      0.27      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[   0    1  789]\n",
      " [   0    1  528]\n",
      " [   0    1 1028]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.01517s\n",
      "Predict Time: 0.00157s\n",
      "Acurácia    : 50.6%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.61      0.18      0.27       790\n",
      "      NEUTRO       0.43      0.57      0.49       529\n",
      "    POSITIVO       0.53      0.73      0.61      1029\n",
      "\n",
      "   micro avg       0.51      0.51      0.51      2348\n",
      "   macro avg       0.52      0.49      0.46      2348\n",
      "weighted avg       0.53      0.51      0.47      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[139 183 468]\n",
      " [ 23 299 207]\n",
      " [ 66 213 750]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.90832s\n",
      "Predict Time: 0.00079s\n",
      "Acurácia    : 59.28%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.53      0.60      0.56       790\n",
      "      NEUTRO       0.54      0.42      0.47       529\n",
      "    POSITIVO       0.67      0.67      0.67      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.58      0.57      0.57      2348\n",
      "weighted avg       0.59      0.59      0.59      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[474 110 206]\n",
      " [168 224 137]\n",
      " [254  81 694]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 2.45774s\n",
      "Predict Time: 0.03132s\n",
      "Acurácia    : 59.07%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.52      0.59      0.55       790\n",
      "      NEUTRO       0.55      0.44      0.49       529\n",
      "    POSITIVO       0.67      0.67      0.67      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.58      0.57      0.57      2348\n",
      "weighted avg       0.59      0.59      0.59      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[466 112 212]\n",
      " [165 231 133]\n",
      " [259  80 690]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.02668s\n",
      "Predict Time: 4.58188s\n",
      "Acurácia    : 54.43%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.48      0.61      0.54       790\n",
      "      NEUTRO       0.46      0.53      0.49       529\n",
      "    POSITIVO       0.71      0.50      0.58      1029\n",
      "\n",
      "   micro avg       0.54      0.54      0.54      2348\n",
      "   macro avg       0.55      0.55      0.54      2348\n",
      "weighted avg       0.57      0.54      0.55      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[485 167 138]\n",
      " [172 281  76]\n",
      " [354 163 512]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 5.18085s\n",
      "Predict Time: 0.00675s\n",
      "Acurácia    : 59.75%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.52      0.63      0.57       790\n",
      "      NEUTRO       0.54      0.50      0.52       529\n",
      "    POSITIVO       0.71      0.62      0.66      1029\n",
      "\n",
      "   micro avg       0.60      0.60      0.60      2348\n",
      "   macro avg       0.59      0.59      0.58      2348\n",
      "weighted avg       0.61      0.60      0.60      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[498 130 162]\n",
      " [160 267 102]\n",
      " [294  97 638]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.77893s\n",
      "Predict Time: 0.00067s\n",
      "Acurácia    : 58.9%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.53      0.62      0.57       790\n",
      "      NEUTRO       0.53      0.43      0.48       529\n",
      "    POSITIVO       0.67      0.65      0.66      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.58      0.57      0.57      2348\n",
      "weighted avg       0.59      0.59      0.59      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[486 113 191]\n",
      " [168 228 133]\n",
      " [271  89 669]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.61788s\n",
      "Predict Time: 0.27102s\n",
      "Acurácia    : 44.85%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.38      0.05      0.09       790\n",
      "      NEUTRO       0.41      0.08      0.13       529\n",
      "    POSITIVO       0.45      0.94      0.61      1029\n",
      "\n",
      "   micro avg       0.45      0.45      0.45      2348\n",
      "   macro avg       0.41      0.36      0.28      2348\n",
      "weighted avg       0.42      0.45      0.33      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 41  31 718]\n",
      " [ 39  42 448]\n",
      " [ 29  30 970]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "tweets_mg\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00694s\n",
      "Predict Time: 0.00047s\n",
      "Acurácia    : 53.08%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.00      0.00      0.00       201\n",
      "      NEUTRO       0.54      0.46      0.50       362\n",
      "    POSITIVO       0.53      0.84      0.65       426\n",
      "\n",
      "   micro avg       0.53      0.53      0.53       989\n",
      "   macro avg       0.36      0.43      0.38       989\n",
      "weighted avg       0.42      0.53      0.46       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  0  75 126]\n",
      " [  0 167 195]\n",
      " [  0  68 358]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.007s\n",
      "Predict Time: 0.00043s\n",
      "Acurácia    : 57.13%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.43      0.01      0.03       201\n",
      "      NEUTRO       0.49      0.73      0.59       362\n",
      "    POSITIVO       0.68      0.70      0.69       426\n",
      "\n",
      "   micro avg       0.57      0.57      0.57       989\n",
      "   macro avg       0.53      0.48      0.43       989\n",
      "weighted avg       0.56      0.57      0.52       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  3 149  49]\n",
      " [  3 265  94]\n",
      " [  1 128 297]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.34512s\n",
      "Predict Time: 0.00186s\n",
      "Acurácia    : 61.98%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.28      0.04      0.07       201\n",
      "      NEUTRO       0.52      0.81      0.63       362\n",
      "    POSITIVO       0.80      0.73      0.76       426\n",
      "\n",
      "   micro avg       0.62      0.62      0.62       989\n",
      "   macro avg       0.53      0.53      0.49       989\n",
      "weighted avg       0.59      0.62      0.57       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  8 165  28]\n",
      " [ 16 294  52]\n",
      " [  5 110 311]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 0.79983s\n",
      "Predict Time: 0.01153s\n",
      "Acurácia    : 60.97%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.27      0.13      0.18       201\n",
      "      NEUTRO       0.52      0.72      0.60       362\n",
      "    POSITIVO       0.81      0.74      0.77       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.53      0.53      0.52       989\n",
      "weighted avg       0.59      0.61      0.59       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 27 146  28]\n",
      " [ 55 260  47]\n",
      " [ 18  92 316]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00686s\n",
      "Predict Time: 0.56981s\n",
      "Acurácia    : 58.04%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.27      0.22      0.24       201\n",
      "      NEUTRO       0.50      0.72      0.59       362\n",
      "    POSITIVO       0.90      0.63      0.74       426\n",
      "\n",
      "   micro avg       0.58      0.58      0.58       989\n",
      "   macro avg       0.55      0.52      0.52       989\n",
      "weighted avg       0.62      0.58      0.58       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 44 147  10]\n",
      " [ 79 262  21]\n",
      " [ 42 116 268]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 2.52885s\n",
      "Predict Time: 0.00231s\n",
      "Acurácia    : 64.21%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.29      0.02      0.04       201\n",
      "      NEUTRO       0.52      0.91      0.66       362\n",
      "    POSITIVO       0.88      0.71      0.79       426\n",
      "\n",
      "   micro avg       0.64      0.64      0.64       989\n",
      "   macro avg       0.56      0.55      0.49       989\n",
      "weighted avg       0.63      0.64      0.59       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  4 182  15]\n",
      " [  8 328  26]\n",
      " [  2 121 303]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.47325s\n",
      "Predict Time: 0.00059s\n",
      "Acurácia    : 62.49%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.30      0.05      0.09       201\n",
      "      NEUTRO       0.52      0.83      0.64       362\n",
      "    POSITIVO       0.81      0.72      0.77       426\n",
      "\n",
      "   micro avg       0.62      0.62      0.62       989\n",
      "   macro avg       0.54      0.53      0.50       989\n",
      "weighted avg       0.60      0.62      0.58       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 11 162  28]\n",
      " [ 20 299  43]\n",
      " [  6 112 308]]\n",
      "------------------------------\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo      : SVC\n",
      "Train Time  : 0.34286s\n",
      "Predict Time: 0.16097s\n",
      "Acurácia    : 44.49%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.29      0.10      0.15       201\n",
      "      NEUTRO       0.45      0.08      0.14       362\n",
      "    POSITIVO       0.46      0.92      0.61       426\n",
      "\n",
      "   micro avg       0.44      0.44      0.44       989\n",
      "   macro avg       0.40      0.37      0.30       989\n",
      "weighted avg       0.42      0.44      0.34       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 20  12 169]\n",
      " [ 37  29 296]\n",
      " [ 11  24 391]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "titulos_noticias\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00921s\n",
      "Predict Time: 0.00057s\n",
      "Acurácia    : 46.73%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.47      1.00      0.64       635\n",
      "      NEUTRO       0.00      0.00      0.00       130\n",
      "    POSITIVO       0.00      0.00      0.00       594\n",
      "\n",
      "   micro avg       0.47      0.47      0.47      1359\n",
      "   macro avg       0.16      0.33      0.21      1359\n",
      "weighted avg       0.22      0.47      0.30      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[635   0   0]\n",
      " [130   0   0]\n",
      " [594   0   0]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.00869s\n",
      "Predict Time: 0.00051s\n",
      "Acurácia    : 48.64%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.65      0.54      0.59       635\n",
      "      NEUTRO       0.18      0.66      0.28       130\n",
      "    POSITIVO       0.65      0.39      0.48       594\n",
      "\n",
      "   micro avg       0.49      0.49      0.49      1359\n",
      "   macro avg       0.49      0.53      0.45      1359\n",
      "weighted avg       0.61      0.49      0.52      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[346 191  98]\n",
      " [ 20  86  24]\n",
      " [166 199 229]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.41299s\n",
      "Predict Time: 0.00128s\n",
      "Acurácia    : 60.71%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.60      0.74      0.66       635\n",
      "      NEUTRO       0.33      0.05      0.08       130\n",
      "    POSITIVO       0.62      0.59      0.61       594\n",
      "\n",
      "   micro avg       0.61      0.61      0.61      1359\n",
      "   macro avg       0.52      0.46      0.45      1359\n",
      "weighted avg       0.59      0.61      0.58      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[468   4 163]\n",
      " [ 72   6  52]\n",
      " [235   8 351]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 1.08461s\n",
      "Predict Time: 0.01543s\n",
      "Acurácia    : 60.71%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.61      0.71      0.66       635\n",
      "      NEUTRO       0.61      0.17      0.27       130\n",
      "    POSITIVO       0.60      0.59      0.60       594\n",
      "\n",
      "   micro avg       0.61      0.61      0.61      1359\n",
      "   macro avg       0.61      0.49      0.51      1359\n",
      "weighted avg       0.61      0.61      0.59      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[450   3 182]\n",
      " [ 54  22  54]\n",
      " [230  11 353]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.01027s\n",
      "Predict Time: 0.94733s\n",
      "Acurácia    : 54.89%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.76      0.65       635\n",
      "      NEUTRO       0.28      0.29      0.29       130\n",
      "    POSITIVO       0.60      0.38      0.46       594\n",
      "\n",
      "   micro avg       0.55      0.55      0.55      1359\n",
      "   macro avg       0.48      0.48      0.47      1359\n",
      "weighted avg       0.55      0.55      0.53      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[482  35 118]\n",
      " [ 57  38  35]\n",
      " [305  63 226]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 4.4502s\n",
      "Predict Time: 0.00337s\n",
      "Acurácia    : 58.65%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.75      0.44      0.56       635\n",
      "      NEUTRO       0.32      0.11      0.16       130\n",
      "    POSITIVO       0.53      0.85      0.65       594\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      1359\n",
      "   macro avg       0.53      0.47      0.46      1359\n",
      "weighted avg       0.61      0.59      0.56      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[281  13 341]\n",
      " [ 18  14  98]\n",
      " [ 75  17 502]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.55289s\n",
      "Predict Time: 0.00048s\n",
      "Acurácia    : 61.59%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.61      0.75      0.67       635\n",
      "      NEUTRO       0.30      0.05      0.08       130\n",
      "    POSITIVO       0.64      0.60      0.62       594\n",
      "\n",
      "   micro avg       0.62      0.62      0.62      1359\n",
      "   macro avg       0.52      0.46      0.46      1359\n",
      "weighted avg       0.59      0.62      0.59      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[475   4 156]\n",
      " [ 78   6  46]\n",
      " [228  10 356]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.40392s\n",
      "Predict Time: 0.19878s\n",
      "Acurácia    : 46.14%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.63      0.12      0.20       635\n",
      "      NEUTRO       0.07      0.01      0.01       130\n",
      "    POSITIVO       0.45      0.92      0.60       594\n",
      "\n",
      "   micro avg       0.46      0.46      0.46      1359\n",
      "   macro avg       0.38      0.35      0.27      1359\n",
      "weighted avg       0.50      0.46      0.36      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 77   9 549]\n",
      " [  5   1 124]\n",
      " [ 40   5 549]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#all_datasets\n",
    "vec_tfidf = CountVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_count_lsa = vec_tfidf.fit_transform(afrases)\n",
    "\n",
    "svd = TruncatedSVD(100, random_state=10)\n",
    "normalizer = MinMaxScaler(copy=False)\n",
    "lda = make_pipeline(svd, normalizer)\n",
    "X_svd = lda.fit_transform(X_count_lsa)\n",
    "\n",
    "print(\"\\nall_datasets\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_svd, avalencias))\n",
    "        matriz_resultados['all']['count+lsa'][classifier.__class__.__name__] = acc\n",
    "        train_time['all']['count+lsa'][classifier.__class__.__name__] = tt\n",
    "        predict_time['all']['count+lsa'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "      \n",
    "\n",
    "#tweets_mg\n",
    "vec_tfidf = CountVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_count_lsa_tmg = vec_count.fit_transform(atweets_mg)\n",
    "X_svd = lda.fit_transform(X_count_lsa_tmg)\n",
    "\n",
    "print(\"\\ntweets_mg\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_svd, aval_tweets_mg))\n",
    "        matriz_resultados['tweets']['count+lsa'][classifier.__class__.__name__] = acc\n",
    "        train_time['tweets']['count+lsa'][classifier.__class__.__name__] = tt\n",
    "        predict_time['tweets']['count+lsa'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "      \n",
    "      \n",
    "#titulos_noticias\n",
    "vec_tfidf = CountVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_count_lsa_tn = vec_count.fit_transform(atitulo_noticias)\n",
    "X_svd = lda.fit_transform(X_count_lsa_tn)\n",
    "\n",
    "print(\"\\ntitulos_noticias\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_svd, aval_titulo_noticias))\n",
    "        matriz_resultados['titulo_noticias']['count+lsa'][classifier.__class__.__name__] = acc\n",
    "        train_time['titulo_noticias']['count+lsa'][classifier.__class__.__name__] = tt\n",
    "        predict_time['titulo_noticias']['count+lsa'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "fIN0NwXTmFva"
   },
   "source": [
    "## LDA (usando Count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:38:22.900280Z",
     "start_time": "2018-11-29T00:36:30.741146Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1037
    },
    "colab_type": "code",
    "id": "MoMB5nY4mJwM",
    "outputId": "5eaf4044-4f8d-46a9-f699-71840fcd854d",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "all_datasets\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00936s\n",
      "Predict Time: 0.0009s\n",
      "Acurácia    : 50.55%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.44      0.51      0.47       790\n",
      "      NEUTRO       0.53      0.18      0.27       529\n",
      "    POSITIVO       0.55      0.67      0.60      1029\n",
      "\n",
      "   micro avg       0.51      0.51      0.51      2348\n",
      "   macro avg       0.51      0.45      0.45      2348\n",
      "weighted avg       0.51      0.51      0.48      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[399  45 346]\n",
      " [206  96 227]\n",
      " [296  41 692]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.01651s\n",
      "Predict Time: 0.00087s\n",
      "Acurácia    : 53.88%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.49      0.55      0.52       790\n",
      "      NEUTRO       0.47      0.56      0.51       529\n",
      "    POSITIVO       0.65      0.52      0.57      1029\n",
      "\n",
      "   micro avg       0.54      0.54      0.54      2348\n",
      "   macro avg       0.54      0.54      0.54      2348\n",
      "weighted avg       0.55      0.54      0.54      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[437 173 180]\n",
      " [121 298 110]\n",
      " [338 161 530]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.56125s\n",
      "Predict Time: 0.0016s\n",
      "Acurácia    : 54.22%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.49      0.56      0.52       790\n",
      "      NEUTRO       0.52      0.40      0.45       529\n",
      "    POSITIVO       0.60      0.60      0.60      1029\n",
      "\n",
      "   micro avg       0.54      0.54      0.54      2348\n",
      "   macro avg       0.54      0.52      0.52      2348\n",
      "weighted avg       0.54      0.54      0.54      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[443  98 249]\n",
      " [156 211 162]\n",
      " [312  98 619]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 1.15241s\n",
      "Predict Time: 0.03065s\n",
      "Acurácia    : 53.32%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.47      0.56      0.51       790\n",
      "      NEUTRO       0.49      0.40      0.44       529\n",
      "    POSITIVO       0.61      0.58      0.60      1029\n",
      "\n",
      "   micro avg       0.53      0.53      0.53      2348\n",
      "   macro avg       0.53      0.51      0.52      2348\n",
      "weighted avg       0.54      0.53      0.53      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[440 115 235]\n",
      " [173 212 144]\n",
      " [327 102 600]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.01475s\n",
      "Predict Time: 2.04595s\n",
      "Acurácia    : 50.68%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.46      0.54      0.50       790\n",
      "      NEUTRO       0.41      0.48      0.45       529\n",
      "    POSITIVO       0.63      0.50      0.56      1029\n",
      "\n",
      "   micro avg       0.51      0.51      0.51      2348\n",
      "   macro avg       0.50      0.51      0.50      2348\n",
      "weighted avg       0.52      0.51      0.51      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[423 174 193]\n",
      " [166 255 108]\n",
      " [329 188 512]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 17.66538s\n",
      "Predict Time: 0.00676s\n",
      "Acurácia    : 50.77%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.46      0.48      0.47       790\n",
      "      NEUTRO       0.42      0.43      0.43       529\n",
      "    POSITIVO       0.60      0.57      0.58      1029\n",
      "\n",
      "   micro avg       0.51      0.51      0.51      2348\n",
      "   macro avg       0.49      0.49      0.49      2348\n",
      "weighted avg       0.51      0.51      0.51      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[376 157 257]\n",
      " [163 227 139]\n",
      " [286 154 589]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.31192s\n",
      "Predict Time: 0.00079s\n",
      "Acurácia    : 54.05%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.49      0.56      0.52       790\n",
      "      NEUTRO       0.50      0.42      0.45       529\n",
      "    POSITIVO       0.61      0.59      0.60      1029\n",
      "\n",
      "   micro avg       0.54      0.54      0.54      2348\n",
      "   macro avg       0.53      0.52      0.52      2348\n",
      "weighted avg       0.54      0.54      0.54      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[441 113 236]\n",
      " [153 220 156]\n",
      " [310 111 608]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.62617s\n",
      "Predict Time: 0.38366s\n",
      "Acurácia    : 46.64%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.48      0.13      0.21       790\n",
      "      NEUTRO       0.39      0.22      0.28       529\n",
      "    POSITIVO       0.48      0.85      0.61      1029\n",
      "\n",
      "   micro avg       0.47      0.47      0.47      2348\n",
      "   macro avg       0.45      0.40      0.37      2348\n",
      "weighted avg       0.46      0.47      0.40      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[106  96 588]\n",
      " [ 45 117 367]\n",
      " [ 70  87 872]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "tweets_mg\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00333s\n",
      "Predict Time: 0.00041s\n",
      "Acurácia    : 53.69%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.33      0.01      0.02       201\n",
      "      NEUTRO       0.48      0.63      0.55       362\n",
      "    POSITIVO       0.59      0.71      0.64       426\n",
      "\n",
      "   micro avg       0.54      0.54      0.54       989\n",
      "   macro avg       0.47      0.45      0.40       989\n",
      "weighted avg       0.50      0.54      0.48       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  2 123  76]\n",
      " [  2 228 132]\n",
      " [  2 123 301]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.00678s\n",
      "Predict Time: 0.00047s\n",
      "Acurácia    : 50.35%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.25      0.24      0.25       201\n",
      "      NEUTRO       0.46      0.51      0.48       362\n",
      "    POSITIVO       0.67      0.62      0.64       426\n",
      "\n",
      "   micro avg       0.50      0.50      0.50       989\n",
      "   macro avg       0.46      0.46      0.46       989\n",
      "weighted avg       0.51      0.50      0.51       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 49 106  46]\n",
      " [ 93 185  84]\n",
      " [ 52 110 264]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.17787s\n",
      "Predict Time: 0.00045s\n",
      "Acurácia    : 54.5%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.29      0.02      0.04       201\n",
      "      NEUTRO       0.48      0.67      0.56       362\n",
      "    POSITIVO       0.62      0.69      0.65       426\n",
      "\n",
      "   micro avg       0.54      0.54      0.54       989\n",
      "   macro avg       0.46      0.46      0.42       989\n",
      "weighted avg       0.50      0.54      0.49       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  4 131  66]\n",
      " [  4 243 115]\n",
      " [  6 128 292]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 0.41222s\n",
      "Predict Time: 0.01528s\n",
      "Acurácia    : 53.19%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.25      0.13      0.17       201\n",
      "      NEUTRO       0.48      0.62      0.54       362\n",
      "    POSITIVO       0.66      0.65      0.65       426\n",
      "\n",
      "   micro avg       0.53      0.53      0.53       989\n",
      "   macro avg       0.46      0.47      0.45       989\n",
      "weighted avg       0.51      0.53      0.51       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 26 123  52]\n",
      " [ 46 225  91]\n",
      " [ 32 119 275]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00604s\n",
      "Predict Time: 0.26528s\n",
      "Acurácia    : 48.33%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.22      0.15      0.18       201\n",
      "      NEUTRO       0.44      0.60      0.51       362\n",
      "    POSITIVO       0.65      0.54      0.59       426\n",
      "\n",
      "   micro avg       0.48      0.48      0.48       989\n",
      "   macro avg       0.44      0.43      0.42       989\n",
      "weighted avg       0.48      0.48      0.48       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 30 124  47]\n",
      " [ 67 217  78]\n",
      " [ 40 155 231]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 8.12206s\n",
      "Predict Time: 0.00311s\n",
      "Acurácia    : 53.19%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.26      0.10      0.15       201\n",
      "      NEUTRO       0.47      0.68      0.55       362\n",
      "    POSITIVO       0.68      0.61      0.64       426\n",
      "\n",
      "   micro avg       0.53      0.53      0.53       989\n",
      "   macro avg       0.47      0.46      0.45       989\n",
      "weighted avg       0.52      0.53      0.51       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 21 139  41]\n",
      " [ 35 245  82]\n",
      " [ 24 142 260]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.17695s\n",
      "Predict Time: 0.01041s\n",
      "Acurácia    : 54.8%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.32      0.04      0.08       201\n",
      "      NEUTRO       0.48      0.67      0.56       362\n",
      "    POSITIVO       0.64      0.68      0.66       426\n",
      "\n",
      "   micro avg       0.55      0.55      0.55       989\n",
      "   macro avg       0.48      0.47      0.43       989\n",
      "weighted avg       0.52      0.55      0.50       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[  9 133  59]\n",
      " [ 12 244 106]\n",
      " [  7 130 289]]\n",
      "------------------------------\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo      : SVC\n",
      "Train Time  : 0.32485s\n",
      "Predict Time: 0.15401s\n",
      "Acurácia    : 44.29%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.26      0.17      0.21       201\n",
      "      NEUTRO       0.50      0.04      0.08       362\n",
      "    POSITIVO       0.47      0.91      0.62       426\n",
      "\n",
      "   micro avg       0.44      0.44      0.44       989\n",
      "   macro avg       0.41      0.37      0.30       989\n",
      "weighted avg       0.44      0.44      0.34       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 34   9 158]\n",
      " [ 65  16 281]\n",
      " [ 31   7 388]]\n",
      "------------------------------\n",
      "\n",
      "\n",
      "titulo_noticias\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.00855s\n",
      "Predict Time: 0.00631s\n",
      "Acurácia    : 47.46%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.48      0.65      0.55       635\n",
      "      NEUTRO       0.00      0.00      0.00       130\n",
      "    POSITIVO       0.46      0.39      0.42       594\n",
      "\n",
      "   micro avg       0.47      0.47      0.47      1359\n",
      "   macro avg       0.31      0.35      0.33      1359\n",
      "weighted avg       0.43      0.47      0.44      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[412   0 223]\n",
      " [ 83   0  47]\n",
      " [361   0 233]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.00805s\n",
      "Predict Time: 0.00056s\n",
      "Acurácia    : 39.74%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.50      0.39      0.43       635\n",
      "      NEUTRO       0.12      0.28      0.17       130\n",
      "    POSITIVO       0.46      0.43      0.45       594\n",
      "\n",
      "   micro avg       0.40      0.40      0.40      1359\n",
      "   macro avg       0.36      0.37      0.35      1359\n",
      "weighted avg       0.45      0.40      0.42      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[245 140 250]\n",
      " [ 46  37  47]\n",
      " [203 133 258]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 0.2256s\n",
      "Predict Time: 0.00056s\n",
      "Acurácia    : 47.31%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.48      0.62      0.54       635\n",
      "      NEUTRO       0.00      0.00      0.00       130\n",
      "    POSITIVO       0.46      0.42      0.44       594\n",
      "\n",
      "   micro avg       0.47      0.47      0.47      1359\n",
      "   macro avg       0.31      0.35      0.33      1359\n",
      "weighted avg       0.43      0.47      0.45      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[395   2 238]\n",
      " [ 77   0  53]\n",
      " [343   3 248]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 0.56448s\n",
      "Predict Time: 0.0211s\n",
      "Acurácia    : 48.93%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.52      0.59      0.55       635\n",
      "      NEUTRO       0.25      0.13      0.17       130\n",
      "    POSITIVO       0.48      0.46      0.47       594\n",
      "\n",
      "   micro avg       0.49      0.49      0.49      1359\n",
      "   macro avg       0.42      0.39      0.40      1359\n",
      "weighted avg       0.48      0.49      0.48      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[377  24 234]\n",
      " [ 59  17  54]\n",
      " [295  28 271]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.00959s\n",
      "Predict Time: 0.43912s\n",
      "Acurácia    : 47.83%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.50      0.67      0.57       635\n",
      "      NEUTRO       0.19      0.08      0.11       130\n",
      "    POSITIVO       0.48      0.36      0.41       594\n",
      "\n",
      "   micro avg       0.48      0.48      0.48      1359\n",
      "   macro avg       0.39      0.37      0.36      1359\n",
      "weighted avg       0.46      0.48      0.46      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[427  16 192]\n",
      " [ 78  10  42]\n",
      " [355  26 213]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 13.7403s\n",
      "Predict Time: 0.00434s\n",
      "Acurácia    : 46.43%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.50      0.51      0.50       635\n",
      "      NEUTRO       0.21      0.13      0.16       130\n",
      "    POSITIVO       0.46      0.49      0.48       594\n",
      "\n",
      "   micro avg       0.46      0.46      0.46      1359\n",
      "   macro avg       0.39      0.38      0.38      1359\n",
      "weighted avg       0.45      0.46      0.46      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[323  29 283]\n",
      " [ 58  17  55]\n",
      " [268  35 291]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.19812s\n",
      "Predict Time: 0.00047s\n",
      "Acurácia    : 47.53%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.49      0.62      0.55       635\n",
      "      NEUTRO       0.00      0.00      0.00       130\n",
      "    POSITIVO       0.47      0.43      0.45       594\n",
      "\n",
      "   micro avg       0.48      0.48      0.48      1359\n",
      "   macro avg       0.32      0.35      0.33      1359\n",
      "weighted avg       0.43      0.48      0.45      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[393   5 237]\n",
      " [ 77   0  53]\n",
      " [335   6 253]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 0.36003s\n",
      "Predict Time: 0.18362s\n",
      "Acurácia    : 43.93%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.43      0.10      0.16       635\n",
      "      NEUTRO       0.00      0.00      0.00       130\n",
      "    POSITIVO       0.44      0.90      0.59       594\n",
      "\n",
      "   micro avg       0.44      0.44      0.44      1359\n",
      "   macro avg       0.29      0.33      0.25      1359\n",
      "weighted avg       0.39      0.44      0.33      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 64   0 571]\n",
      " [ 24   0 106]\n",
      " [ 61   0 533]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#all_datasets\n",
    "vec_tfidf = CountVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_count_lda = vec_tfidf.fit_transform(afrases)\n",
    "\n",
    "lda = LatentDirichletAllocation(n_components=100, max_iter=25, random_state=10, n_jobs=5)\n",
    "normalizer = MinMaxScaler(copy=False)\n",
    "lda = make_pipeline(lda, normalizer)\n",
    "X_lda = lda.fit_transform(X_count_lda)\n",
    "\n",
    "\n",
    "print(\"\\nall_datasets\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_lda, avalencias))\n",
    "        matriz_resultados['all']['count+lda'][classifier.__class__.__name__] = acc\n",
    "        train_time['all']['count+lda'][classifier.__class__.__name__] = tt\n",
    "        predict_time['all']['count+lda'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "\n",
    "\n",
    "#tweets_mg\n",
    "vec_tfidf = CountVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_count_lda_tmg = vec_count.fit_transform(atweets_mg)\n",
    "X_lda = lda.fit_transform(X_count_lda_tmg)\n",
    "\n",
    "print(\"\\ntweets_mg\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_lda, aval_tweets_mg))\n",
    "        matriz_resultados['tweets']['count+lda'][classifier.__class__.__name__] = acc\n",
    "        train_time['tweets']['count+lda'][classifier.__class__.__name__] = tt\n",
    "        predict_time['tweets']['count+lda'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass\n",
    "      \n",
    "      \n",
    "#titulo_noticias\n",
    "vec_tfidf = CountVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_count_lda_tn = vec_count.fit_transform(atitulo_noticias)\n",
    "X_lda = lda.fit_transform(X_count_lda_tn)\n",
    "\n",
    "print(\"\\ntitulo_noticias\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X_lda, aval_titulo_noticias))\n",
    "        matriz_resultados['titulo_noticias']['count+lda'][classifier.__class__.__name__] = acc\n",
    "        train_time['titulo_noticias']['count+lda'][classifier.__class__.__name__] = tt\n",
    "        predict_time['titulo_noticias']['count+lda'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "qRkHm_VjdG36"
   },
   "source": [
    "## Count + TF-IDF + Word2Vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:40:04.890350Z",
     "start_time": "2018-11-29T00:38:22.907570Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "xgvYhs55dLsM",
    "outputId": "2c9cd2ce-e399-4381-ae30-ecfca12bc0ef"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(32746561, 35424500)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#all_datasets\n",
    "# Count\n",
    "vec_count = CountVectorizer()\n",
    "X_count = vec_count.fit_transform(afrases)\n",
    "weights_count = pd.DataFrame(np.round(X_count.toarray().T, 8), index=vec_count.get_feature_names())\n",
    "\n",
    "# TF-IDF\n",
    "vec_tfidf = TfidfVectorizer()\n",
    "X_tfidf = vec_tfidf.fit_transform(afrases)\n",
    "weights_tfidf = pd.DataFrame(np.round(X_tfidf.toarray().T, 8), index=vec_tfidf.get_feature_names())\n",
    "\n",
    "# Word2Vec preprocessing\n",
    "frases_w2v = []\n",
    "for frase in afrases:\n",
    "    bigram = []\n",
    "    p_frase = word_tokenize(frase)\n",
    "    for m, palavra in enumerate(p_frase):\n",
    "        next_p = None\n",
    "        try:\n",
    "            next_p = p_frase[m+1]\n",
    "        except:\n",
    "            pass\n",
    "        bigram += [f'{palavra}']\n",
    "#         if next_p:\n",
    "#             bigram += [f'{palavra} {next_p}']\n",
    "    frases_w2v += [bigram]\n",
    "\n",
    "# Word2Vec\n",
    "model = gensim.models.Word2Vec(\n",
    "    sentences=frases_w2v,\n",
    "    sg=1,\n",
    "    hs=1,\n",
    "    size=1,\n",
    "    window=5,\n",
    "    min_count=1,\n",
    "    seed=0,\n",
    "    workers=10)\n",
    "model.train(frases_w2v, total_examples=len(frases_w2v), epochs=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:51:49.044980Z",
     "start_time": "2018-11-29T00:40:04.893395Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 306
    },
    "colab_type": "code",
    "id": "07BlXxa5h1_c",
    "outputId": "061a7282-3c97-4685-dbb1-9aa8162bc34f",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "all_datasets\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 21.4986s\n",
      "Predict Time: 0.08318s\n",
      "Acurácia    : 64.22%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.58      0.67      0.62       790\n",
      "      NEUTRO       0.57      0.43      0.49       529\n",
      "    POSITIVO       0.73      0.73      0.73      1029\n",
      "\n",
      "   micro avg       0.64      0.64      0.64      2348\n",
      "   macro avg       0.62      0.61      0.61      2348\n",
      "weighted avg       0.64      0.64      0.64      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[532 106 152]\n",
      " [174 229 126]\n",
      " [213  69 747]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 28.98565s\n",
      "Predict Time: 0.29381s\n",
      "Acurácia    : 62.95%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.57      0.65      0.61       790\n",
      "      NEUTRO       0.56      0.51      0.53       529\n",
      "    POSITIVO       0.72      0.67      0.69      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.62      0.61      0.61      2348\n",
      "weighted avg       0.63      0.63      0.63      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[516 124 150]\n",
      " [137 271 121]\n",
      " [248  90 691]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 3.27147s\n",
      "Predict Time: 269.98029s\n",
      "Acurácia    : 59.41%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.51      0.63      0.56       790\n",
      "      NEUTRO       0.53      0.48      0.50       529\n",
      "    POSITIVO       0.72      0.63      0.67      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.59      0.58      0.58      2348\n",
      "weighted avg       0.61      0.59      0.60      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[496 126 168]\n",
      " [189 254  86]\n",
      " [287  97 645]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 66.92027s\n",
      "Predict Time: 0.04415s\n",
      "Acurácia    : 59.28%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.54      0.63      0.58       790\n",
      "      NEUTRO       0.47      0.40      0.43       529\n",
      "    POSITIVO       0.70      0.67      0.68      1029\n",
      "\n",
      "   micro avg       0.59      0.59      0.59      2348\n",
      "   macro avg       0.57      0.56      0.56      2348\n",
      "weighted avg       0.59      0.59      0.59      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[495 131 164]\n",
      " [190 211 128]\n",
      " [233 110 686]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.08492s\n",
      "Predict Time: 0.02392s\n",
      "Acurácia    : 62.52%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.56      0.63      0.59       790\n",
      "      NEUTRO       0.54      0.48      0.50       529\n",
      "    POSITIVO       0.73      0.70      0.71      1029\n",
      "\n",
      "   micro avg       0.63      0.63      0.63      2348\n",
      "   macro avg       0.61      0.60      0.60      2348\n",
      "weighted avg       0.63      0.63      0.63      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[499 125 166]\n",
      " [172 252 105]\n",
      " [219  93 717]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 25.01459s\n",
      "Predict Time: 16.56266s\n",
      "Acurácia    : 46.59%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.50      0.15      0.23       790\n",
      "      NEUTRO       0.34      0.42      0.38       529\n",
      "    POSITIVO       0.51      0.73      0.60      1029\n",
      "\n",
      "   micro avg       0.47      0.47      0.47      2348\n",
      "   macro avg       0.45      0.43      0.40      2348\n",
      "weighted avg       0.47      0.47      0.43      2348\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[118 232 440]\n",
      " [ 36 224 269]\n",
      " [ 81 196 752]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#all_datasets\n",
    "r_words = {}\n",
    "for word in vec_count.get_feature_names():\n",
    "    idx = weights_count.index.get_loc(word)\n",
    "    w2c_val = .1\n",
    "    try:\n",
    "        w2c_val = model.wv[word]\n",
    "    except:\n",
    "        pass\n",
    "    r_words[word] = (weights_tfidf.iloc[idx].values + weights_count.iloc[idx].values) * w2c_val\n",
    "lwor = list(r_words.keys())\n",
    "X = np.asarray(list(r_words.values()))\n",
    "weights = pd.DataFrame(X, index=lwor)\n",
    "X = X.T\n",
    "\n",
    "normalizer = Normalizer(copy=False)\n",
    "X = normalizer.fit_transform(X)\n",
    "\n",
    "print(\"\\nall_datasets\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X, avalencias))\n",
    "        matriz_resultados['all']['tfidf+count+w2c'][classifier.__class__.__name__] = acc\n",
    "        train_time['all']['tfidf+count+w2c'][classifier.__class__.__name__] = tt\n",
    "        predict_time['all']['tfidf+count+w2c'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:52:27.074607Z",
     "start_time": "2018-11-29T00:51:49.047494Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "wPAWw7qtgY3g",
    "outputId": "decaf607-0583-4d80-dad1-e71abac53af4"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10651726, 13109500)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#tweets_mg\n",
    "# Count\n",
    "vec_count = CountVectorizer()\n",
    "X_count = vec_count.fit_transform(atweets_mg)\n",
    "weights_count = pd.DataFrame(np.round(X_count.toarray().T, 8), index=vec_count.get_feature_names())\n",
    "\n",
    "# TF-IDF\n",
    "vec_tfidf = TfidfVectorizer()\n",
    "X_tfidf = vec_tfidf.fit_transform(atweets_mg)\n",
    "weights_tfidf = pd.DataFrame(np.round(X_tfidf.toarray().T, 8), index=vec_tfidf.get_feature_names())\n",
    "\n",
    "# Word2Vec preprocessing\n",
    "frases_w2v = []\n",
    "for frase in atweets_mg:\n",
    "    bigram = []\n",
    "    p_frase = word_tokenize(frase)\n",
    "    for m, palavra in enumerate(p_frase):\n",
    "        next_p = None\n",
    "        try:\n",
    "            next_p = p_frase[m+1]\n",
    "        except:\n",
    "            pass\n",
    "        bigram += [f'{palavra}']\n",
    "#         if next_p:\n",
    "#             bigram += [f'{palavra} {next_p}']\n",
    "    frases_w2v += [bigram]\n",
    "\n",
    "# Word2Vec\n",
    "model = gensim.models.Word2Vec(\n",
    "    sentences=frases_w2v,\n",
    "    sg=1,\n",
    "    hs=1,\n",
    "    size=1,\n",
    "    window=5,\n",
    "    min_count=1,\n",
    "    seed=0,\n",
    "    workers=10)\n",
    "model.train(frases_w2v, total_examples=len(frases_w2v), epochs=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:53:46.687350Z",
     "start_time": "2018-11-29T00:52:27.076667Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 326
    },
    "colab_type": "code",
    "id": "HAytLOTZiRkF",
    "outputId": "79f1600e-0c4b-4bbd-f939-b616fd138470",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "tweets_mg\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 2.02521s\n",
      "Predict Time: 0.00892s\n",
      "Acurácia    : 62.49%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.27      0.05      0.08       201\n",
      "      NEUTRO       0.52      0.78      0.63       362\n",
      "    POSITIVO       0.79      0.76      0.78       426\n",
      "\n",
      "   micro avg       0.62      0.62      0.62       989\n",
      "   macro avg       0.53      0.53      0.50       989\n",
      "weighted avg       0.59      0.62      0.58       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 10 163  28]\n",
      " [ 22 284  56]\n",
      " [  5  97 324]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 1.11889s\n",
      "Predict Time: 0.03274s\n",
      "Acurácia    : 63.9%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.33      0.10      0.15       201\n",
      "      NEUTRO       0.53      0.82      0.64       362\n",
      "    POSITIVO       0.87      0.74      0.80       426\n",
      "\n",
      "   micro avg       0.64      0.64      0.64       989\n",
      "   macro avg       0.57      0.55      0.53       989\n",
      "weighted avg       0.63      0.64      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 20 165  16]\n",
      " [ 31 298  33]\n",
      " [ 10 102 314]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.19286s\n",
      "Predict Time: 12.9362s\n",
      "Acurácia    : 61.38%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.33      0.26      0.29       201\n",
      "      NEUTRO       0.53      0.64      0.58       362\n",
      "    POSITIVO       0.81      0.76      0.78       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.56      0.55      0.55       989\n",
      "weighted avg       0.61      0.61      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 52 127  22]\n",
      " [ 78 232  52]\n",
      " [ 26  77 323]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 10.24124s\n",
      "Predict Time: 0.01035s\n",
      "Acurácia    : 61.27%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.35      0.31      0.33       201\n",
      "      NEUTRO       0.57      0.54      0.55       362\n",
      "    POSITIVO       0.75      0.82      0.78       426\n",
      "\n",
      "   micro avg       0.61      0.61      0.61       989\n",
      "   macro avg       0.55      0.56      0.55       989\n",
      "weighted avg       0.60      0.61      0.61       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 63  98  40]\n",
      " [ 93 195  74]\n",
      " [ 26  52 348]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.02654s\n",
      "Predict Time: 0.00368s\n",
      "Acurácia    : 62.89%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.36      0.21      0.26       201\n",
      "      NEUTRO       0.54      0.69      0.61       362\n",
      "    POSITIVO       0.81      0.77      0.79       426\n",
      "\n",
      "   micro avg       0.63      0.63      0.63       989\n",
      "   macro avg       0.57      0.56      0.55       989\n",
      "weighted avg       0.62      0.63      0.62       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 42 134  25]\n",
      " [ 57 251  54]\n",
      " [ 19  78 329]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 5.61091s\n",
      "Predict Time: 3.16858s\n",
      "Acurácia    : 55.71%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.31      0.17      0.22       201\n",
      "      NEUTRO       0.50      0.48      0.49       362\n",
      "    POSITIVO       0.65      0.80      0.72       426\n",
      "\n",
      "   micro avg       0.56      0.56      0.56       989\n",
      "   macro avg       0.49      0.49      0.48       989\n",
      "weighted avg       0.53      0.56      0.53       989\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[ 35 108  58]\n",
      " [ 61 175 126]\n",
      " [ 18  67 341]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#tweets_mg\n",
    "r_words = {}\n",
    "for word in vec_count.get_feature_names():\n",
    "    idx = weights_count.index.get_loc(word)\n",
    "    w2c_val = .1\n",
    "    try:\n",
    "        w2c_val = model.wv[word]\n",
    "    except:\n",
    "        pass\n",
    "    r_words[word] = (weights_tfidf.iloc[idx].values + weights_count.iloc[idx].values) * w2c_val\n",
    "lwor = list(r_words.keys())\n",
    "X = np.asarray(list(r_words.values()))\n",
    "weights = pd.DataFrame(X, index=lwor)\n",
    "X = X.T\n",
    "\n",
    "normalizer = Normalizer(copy=False)\n",
    "X = normalizer.fit_transform(X)\n",
    "\n",
    "print(\"\\ntweets_mg\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X, aval_tweets_mg))\n",
    "        matriz_resultados['tweets']['tfidf+count+w2c'][classifier.__class__.__name__] = acc\n",
    "        train_time['tweets']['tfidf+count+w2c'][classifier.__class__.__name__] = tt\n",
    "        predict_time['tweets']['tfidf+count+w2c'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:54:56.183028Z",
     "start_time": "2018-11-29T00:53:46.689538Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "LGu8X9bxgZGI",
    "outputId": "5779fd2b-7993-4c12-eef9-9de872ec32f1"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20972654, 22315000)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#titulo_noticias\n",
    "# Count\n",
    "vec_count = CountVectorizer()\n",
    "X_count = vec_count.fit_transform(atitulo_noticias)\n",
    "weights_count = pd.DataFrame(np.round(X_count.toarray().T, 8), index=vec_count.get_feature_names())\n",
    "\n",
    "# TF-IDF\n",
    "vec_tfidf = TfidfVectorizer()\n",
    "X_tfidf = vec_tfidf.fit_transform(atitulo_noticias)\n",
    "weights_tfidf = pd.DataFrame(np.round(X_tfidf.toarray().T, 8), index=vec_tfidf.get_feature_names())\n",
    "\n",
    "# Word2Vec preprocessing\n",
    "frases_w2v = []\n",
    "for frase in atitulo_noticias:\n",
    "    bigram = []\n",
    "    p_frase = word_tokenize(frase)\n",
    "    for m, palavra in enumerate(p_frase):\n",
    "        next_p = None\n",
    "        try:\n",
    "            next_p = p_frase[m+1]\n",
    "        except:\n",
    "            pass\n",
    "        bigram += [f'{palavra}']\n",
    "#         if next_p:\n",
    "#             bigram += [f'{palavra} {next_p}']\n",
    "    frases_w2v += [bigram]\n",
    "\n",
    "# Word2Vec\n",
    "model = gensim.models.Word2Vec(\n",
    "    sentences=frases_w2v,\n",
    "    sg=1,\n",
    "    hs=1,\n",
    "    size=1,\n",
    "    window=5,\n",
    "    min_count=1,\n",
    "    seed=0,\n",
    "    workers=10)\n",
    "model.train(frases_w2v, total_examples=len(frases_w2v), epochs=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:14.341281Z",
     "start_time": "2018-11-29T00:54:56.185605Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 377
    },
    "colab_type": "code",
    "id": "vvKIhKXQif9K",
    "outputId": "76418614-74e3-4e4c-fae7-d5bff5cedc28",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "titulo_noticias\n",
      "Modelo      : MultinomialNB\n",
      "Train Time  : 0.03827s\n",
      "Predict Time: 0.00794s\n",
      "Acurácia    : 66.23%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.67      0.79      0.73       635\n",
      "      NEUTRO       1.00      0.03      0.06       130\n",
      "    POSITIVO       0.65      0.66      0.65       594\n",
      "\n",
      "   micro avg       0.66      0.66      0.66      1359\n",
      "   macro avg       0.77      0.49      0.48      1359\n",
      "weighted avg       0.69      0.66      0.63      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[503   0 132]\n",
      " [ 42   4  84]\n",
      " [201   0 393]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : ComplementNB\n",
      "Train Time  : 0.03733s\n",
      "Predict Time: 0.00786s\n",
      "Acurácia    : 64.61%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.70      0.69      0.70       635\n",
      "      NEUTRO       0.41      0.31      0.35       130\n",
      "    POSITIVO       0.63      0.67      0.65       594\n",
      "\n",
      "   micro avg       0.65      0.65      0.65      1359\n",
      "   macro avg       0.58      0.56      0.57      1359\n",
      "weighted avg       0.64      0.65      0.64      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[440  27 168]\n",
      " [ 26  40  64]\n",
      " [165  31 398]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LogisticRegression\n",
      "Train Time  : 3.4431s\n",
      "Predict Time: 0.01708s\n",
      "Acurácia    : 66.52%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.68      0.77      0.72       635\n",
      "      NEUTRO       0.50      0.11      0.18       130\n",
      "    POSITIVO       0.65      0.68      0.67       594\n",
      "\n",
      "   micro avg       0.67      0.67      0.67      1359\n",
      "   macro avg       0.61      0.52      0.52      1359\n",
      "weighted avg       0.65      0.67      0.65      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[487   4 144]\n",
      " [ 47  14  69]\n",
      " [181  10 403]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : RandomForestClassifier\n",
      "Train Time  : 3.97777s\n",
      "Predict Time: 0.06379s\n",
      "Acurácia    : 64.68%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.65      0.78      0.71       635\n",
      "      NEUTRO       0.51      0.16      0.25       130\n",
      "    POSITIVO       0.65      0.61      0.63       594\n",
      "\n",
      "   micro avg       0.65      0.65      0.65      1359\n",
      "   macro avg       0.60      0.52      0.53      1359\n",
      "weighted avg       0.64      0.65      0.63      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[495   5 135]\n",
      " [ 51  21  58]\n",
      " [216  15 363]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : KNeighborsClassifier\n",
      "Train Time  : 0.38568s\n",
      "Predict Time: 30.83668s\n",
      "Acurácia    : 60.56%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.62      0.75      0.68       635\n",
      "      NEUTRO       0.39      0.22      0.28       130\n",
      "    POSITIVO       0.62      0.54      0.57       594\n",
      "\n",
      "   micro avg       0.61      0.61      0.61      1359\n",
      "   macro avg       0.54      0.50      0.51      1359\n",
      "weighted avg       0.60      0.61      0.59      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[476  18 141]\n",
      " [ 46  28  56]\n",
      " [249  26 319]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : MLPClassifier\n",
      "Train Time  : 20.13697s\n",
      "Predict Time: 0.01733s\n",
      "Acurácia    : 64.75%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.70      0.70      0.70       635\n",
      "      NEUTRO       0.42      0.38      0.40       130\n",
      "    POSITIVO       0.64      0.65      0.65       594\n",
      "\n",
      "   micro avg       0.65      0.65      0.65      1359\n",
      "   macro avg       0.58      0.58      0.58      1359\n",
      "weighted avg       0.65      0.65      0.65      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[443  28 164]\n",
      " [ 29  49  52]\n",
      " [165  41 388]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : LinearSVC\n",
      "Train Time  : 0.04004s\n",
      "Predict Time: 0.00716s\n",
      "Acurácia    : 66.08%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.69      0.72      0.70       635\n",
      "      NEUTRO       0.50      0.30      0.37       130\n",
      "    POSITIVO       0.65      0.67      0.66       594\n",
      "\n",
      "   micro avg       0.66      0.66      0.66      1359\n",
      "   macro avg       0.61      0.57      0.58      1359\n",
      "weighted avg       0.65      0.66      0.65      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[460  18 157]\n",
      " [ 36  39  55]\n",
      " [174  21 399]]\n",
      "------------------------------\n",
      "\n",
      "Modelo      : SVC\n",
      "Train Time  : 10.48507s\n",
      "Predict Time: 6.32095s\n",
      "Acurácia    : 53.79%\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "    NEGATIVO       0.58      0.52      0.55       635\n",
      "      NEUTRO       0.28      0.12      0.16       130\n",
      "    POSITIVO       0.53      0.65      0.58       594\n",
      "\n",
      "   micro avg       0.54      0.54      0.54      1359\n",
      "   macro avg       0.46      0.43      0.43      1359\n",
      "weighted avg       0.53      0.54      0.53      1359\n",
      "\n",
      "Matrix de Confusão: \n",
      "[[329  21 285]\n",
      " [ 52  15  63]\n",
      " [189  18 387]]\n",
      "------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#titulo_noticias\n",
    "r_words = {}\n",
    "for word in vec_count.get_feature_names():\n",
    "    idx = weights_count.index.get_loc(word)\n",
    "    w2c_val = .1\n",
    "    try:\n",
    "        w2c_val = model.wjv[word]\n",
    "    except:\n",
    "        pass\n",
    "    r_words[word] = (weights_tfidf.iloc[idx].values + weights_count.iloc[idx].values) * w2c_val\n",
    "lwor = list(r_words.keys())\n",
    "X = np.asarray(list(r_words.values()))\n",
    "weights = pd.DataFrame(X, index=lwor)\n",
    "X = X.T\n",
    "\n",
    "normalizer = Normalizer(copy=False)\n",
    "X = normalizer.fit_transform(X)\n",
    "\n",
    "print(\"\\ntitulo_noticias\")\n",
    "for classifier in classifiers:\n",
    "    try:\n",
    "        acc, tt, pt = run_ml_model(classifier, **split_data(X, aval_titulo_noticias))\n",
    "        matriz_resultados['titulo_noticias']['tfidf+count+w2c'][classifier.__class__.__name__] = acc\n",
    "        train_time['titulo_noticias']['tfidf+count+w2c'][classifier.__class__.__name__] = tt\n",
    "        predict_time['titulo_noticias']['tfidf+count+w2c'][classifier.__class__.__name__] = pt\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Resultado dos Classificadores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Resultado dos Classificadores para todas as frases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:14.858496Z",
     "start_time": "2018-11-29T00:58:14.343811Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Acuraria (%):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >63.12%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >54.64%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >50.00%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >63.12%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >50.60%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >53.88%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.00%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >57.84%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >54.98%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >45.36%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >34.11%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >54.43%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >50.68%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >59.41%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >64.78%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >58.99%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >50.13%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >62.73%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >58.90%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >54.05%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >62.52%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >62.69%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >58.73%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >50.13%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >64.18%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >59.28%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >54.22%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >64.22%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >61.97%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >57.37%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >48.94%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >61.75%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >59.75%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >50.77%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >59.28%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >58.86%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >43.91%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >48.47%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >63.20%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >43.82%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >50.55%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.00%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >62.61%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >57.75%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >50.21%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >62.82%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >59.07%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >53.32%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >62.95%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >48.17%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >46.25%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >41.23%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >48.04%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >44.85%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >46.64%</td>\n",
       "                        <td id=\"T_82165db8_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >46.59%</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297c7aa518>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Acuraria (%):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(matriz_resultados['all'])\n",
    "df.style.apply(highlight_max).format({\n",
    "    'tfidf': '{:,.2f}%'.format,\n",
    "    'tfidf+lsa': '{:,.2f}%'.format,\n",
    "    'tfidf+lda': '{:,.2f}%'.format,\n",
    "    'count': '{:,.2f}%'.format,\n",
    "    'count+lsa': '{:,.2f}%'.format,\n",
    "    'count+lda': '{:,.2f}%'.format,\n",
    "    'tfidf+count+w2c': '{:,.2f}%'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:14.923632Z",
     "start_time": "2018-11-29T00:58:14.860983Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tempo para treinar o classificador (segundos):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >0.01132</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >0.02366</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >0.02311</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >0.03500</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >0.01517</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >0.01651</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >0.00243</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >0.04872</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >0.02413</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >0.00272</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >0.02668</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >0.01475</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >3.27147</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >0.22786</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >2.21711</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >0.79302</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >0.33628</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >0.77893</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >0.31192</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >0.08492</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >6.29359</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >1.06266</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >0.73606</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >10.64482</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >0.90832</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >0.56125</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >21.49860</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >500.15232</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >8.38732</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >32.79967</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >420.74058</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >5.18085</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >17.66538</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >66.92027</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >0.01453</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >0.02420</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >0.01860</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >0.02620</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >0.02481</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >0.00936</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >8.89272</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >6.65577</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >8.03179</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >21.97293</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >2.45774</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >1.15241</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >28.98565</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >0.85790</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >1.80184</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >1.81846</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >0.43936</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >0.61788</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >0.62617</td>\n",
       "                        <td id=\"T_82165db9_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >25.01459</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297cf0a198>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Tempo para treinar o classificador (segundos):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(train_time['all'])\n",
    "df.style.apply(highlight_min).format({\n",
    "    'tfidf': '{:,.5f}'.format,\n",
    "    'tfidf+lsa': '{:,.5f}'.format,\n",
    "    'tfidf+lda': '{:,.5f}'.format,\n",
    "    'count': '{:,.5f}'.format,\n",
    "    'count+lsa': '{:,.5f}'.format,\n",
    "    'count+lda': '{:,.5f}'.format,\n",
    "    'tfidf+count+w2c': '{:,.5f}'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:15.056630Z",
     "start_time": "2018-11-29T00:58:14.925288Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tempo para predizer novos itens (segundos):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >0.00073</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >0.00117</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >0.00117</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >0.00291</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >0.00157</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >0.00087</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >0.42257</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >7.79778</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >2.84385</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >0.62637</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >4.58188</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >2.04595</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >269.98029</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >0.00103</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >0.00967</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >0.00147</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >0.00056</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >0.00067</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >0.00079</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >0.02392</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >0.00242</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >0.00126</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >0.00124</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >0.00464</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >0.00079</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >0.00160</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >0.08318</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >0.02753</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >0.00986</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >0.01163</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >0.01651</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >0.00675</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >0.00676</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >0.04415</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >0.00138</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >0.00120</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >0.00125</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >0.00281</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >0.00102</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >0.00090</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >0.13253</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >0.07394</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >0.04380</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >0.15753</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >0.03132</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >0.03065</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >0.29381</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >0.50209</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >1.08235</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >1.06980</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >0.30641</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >0.27102</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >0.38366</td>\n",
       "                        <td id=\"T_82165dba_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >16.56266</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297ce16470>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Tempo para predizer novos itens (segundos):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(predict_time['all'])\n",
    "df.style.apply(highlight_min).format({\n",
    "    'tfidf': '{:,.5f}'.format,\n",
    "    'tfidf+lsa': '{:,.5f}'.format,\n",
    "    'tfidf+lda': '{:,.5f}'.format,\n",
    "    'count': '{:,.5f}'.format,\n",
    "    'count+lsa': '{:,.5f}'.format,\n",
    "    'count+lda': '{:,.5f}'.format,\n",
    "    'tfidf+count+w2c': '{:,.5f}'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Resultado dos Classificadores para os Tweets MG"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:15.192010Z",
     "start_time": "2018-11-29T00:58:15.059668Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Acuraria (%):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >59.25%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >57.84%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >52.07%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >59.56%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >57.13%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >50.35%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.00%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >61.17%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >60.47%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >53.59%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >50.76%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >58.04%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >48.33%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >61.38%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >63.90%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >64.31%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >55.71%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >61.48%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >62.49%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >54.80%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >62.89%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >63.40%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >63.40%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >55.31%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >63.30%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >61.98%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >54.50%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >62.49%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >61.68%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >60.26%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >53.69%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >61.07%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >64.21%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >53.19%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >61.27%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >61.98%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >50.56%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >55.71%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >60.57%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >53.08%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >53.69%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.00%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >63.40%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >61.27%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >53.79%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >62.59%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >60.97%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >53.19%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >63.90%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >53.69%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >46.01%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >46.21%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >48.63%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >44.49%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >44.29%</td>\n",
       "                        <td id=\"T_82165dbb_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >55.71%</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297ced3e10>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Acuraria (%):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(matriz_resultados['tweets'])\n",
    "df.style.apply(highlight_max).format({\n",
    "    'tfidf': '{:,.2f}%'.format,\n",
    "    'tfidf+lsa': '{:,.2f}%'.format,\n",
    "    'tfidf+lda': '{:,.2f}%'.format,\n",
    "    'count': '{:,.2f}%'.format,\n",
    "    'count+lsa': '{:,.2f}%'.format,\n",
    "    'count+lda': '{:,.2f}%'.format,\n",
    "    'tfidf+count+w2c': '{:,.2f}%'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:15.268070Z",
     "start_time": "2018-11-29T00:58:15.193809Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tempo para treinar o classificador (segundos):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >0.00645</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >0.02089</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >0.01123</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >0.00620</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >0.00700</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >0.00678</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >0.00186</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >0.02979</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >0.01879</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >0.00149</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >0.00686</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >0.00604</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >0.19286</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >0.11271</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >0.82166</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >0.31729</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >0.18380</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >0.47325</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >0.17695</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >0.02654</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >2.41728</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >0.74969</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >0.30723</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >2.64707</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >0.34512</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >0.17787</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >2.02521</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >41.46804</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >6.01657</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >12.12654</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >33.42182</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >2.52885</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >8.12206</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >10.24124</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >0.00563</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >0.01163</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >0.00858</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >0.00594</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >0.00694</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >0.00333</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >3.62224</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >2.34820</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >2.51137</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >1.71502</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >0.79983</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >0.41222</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >1.11889</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >0.60397</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >0.76614</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >0.76851</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >0.25913</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >0.34286</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >0.32485</td>\n",
       "                        <td id=\"T_82165dbc_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >5.61091</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297cee59b0>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Tempo para treinar o classificador (segundos):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(train_time['tweets'])\n",
    "df.style.apply(highlight_min).format({\n",
    "    'tfidf': '{:,.5f}'.format,\n",
    "    'tfidf+lsa': '{:,.5f}'.format,\n",
    "    'tfidf+lda': '{:,.5f}'.format,\n",
    "    'count': '{:,.5f}'.format,\n",
    "    'count+lsa': '{:,.5f}'.format,\n",
    "    'count+lda': '{:,.5f}'.format,\n",
    "    'tfidf+count+w2c': '{:,.5f}'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:15.397799Z",
     "start_time": "2018-11-29T00:58:15.270852Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tempo para predizer novos itens (segundos):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >0.00053</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >0.00066</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >0.00069</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >0.00041</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >0.00043</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >0.00047</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >0.14588</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >1.00287</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >1.03105</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >0.10617</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >0.56981</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >0.26528</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >12.93620</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >0.00067</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >0.00076</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >0.00181</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >0.00035</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >0.00059</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >0.01041</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >0.00368</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >0.00165</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >0.01470</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >0.00077</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >0.00103</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >0.00186</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >0.00045</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >0.00892</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >0.01048</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >0.00564</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >0.00570</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >0.02620</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >0.00231</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >0.00311</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >0.01035</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >0.00054</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >0.01165</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >0.00087</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >0.00087</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >0.00047</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >0.00041</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.00000</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >0.05584</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >0.03586</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >0.04095</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >0.03709</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >0.01153</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >0.01528</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >0.03274</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >0.34286</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >0.38243</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >0.43126</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >0.18555</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >0.16097</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >0.15401</td>\n",
       "                        <td id=\"T_82165dbd_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >3.16858</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297ced3240>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Tempo para predizer novos itens (segundos):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(predict_time['tweets'])\n",
    "df.style.apply(highlight_min).format({\n",
    "    'tfidf': '{:,.5f}'.format,\n",
    "    'tfidf+lsa': '{:,.5f}'.format,\n",
    "    'tfidf+lda': '{:,.5f}'.format,\n",
    "    'count': '{:,.5f}'.format,\n",
    "    'count+lsa': '{:,.5f}'.format,\n",
    "    'count+lda': '{:,.5f}'.format,\n",
    "    'tfidf+count+w2c': '{:,.5f}'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Resultados dos Classificadores para os Títulos de Notícias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:15.601390Z",
     "start_time": "2018-11-29T00:58:15.400774Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Acuraria (%):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >65.19%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >57.17%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >48.64%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >60.49%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >48.64%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >39.74%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >64.61%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >60.49%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >54.97%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >50.40%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >47.76%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >54.89%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >47.83%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >60.56%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >68.65%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >62.99%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >53.42%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >65.86%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >61.59%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >47.53%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >66.08%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >65.86%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >63.72%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >53.79%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >66.96%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >60.71%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >47.31%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >66.52%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >66.00%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >61.52%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >48.86%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >67.18%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >58.65%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >46.43%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >64.75%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >65.34%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >46.73%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >53.57%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >65.56%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >46.73%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >47.46%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >66.23%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >62.62%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >60.41%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >52.24%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >63.80%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >60.71%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >48.93%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >64.68%</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >51.29%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >45.33%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >45.18%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >45.18%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >46.14%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >43.93%</td>\n",
       "                        <td id=\"T_82165dbe_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >53.79%</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f29758acf28>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Acuraria (%):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(matriz_resultados['titulo_noticias'])\n",
    "df.style.apply(highlight_max).format({\n",
    "    'tfidf': '{:,.2f}%'.format,\n",
    "    'tfidf+lsa': '{:,.2f}%'.format,\n",
    "    'tfidf+lda': '{:,.2f}%'.format,\n",
    "    'count': '{:,.2f}%'.format,\n",
    "    'count+lsa': '{:,.2f}%'.format,\n",
    "    'count+lda': '{:,.2f}%'.format,\n",
    "    'tfidf+count+w2c': '{:,.2f}%'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:15.663556Z",
     "start_time": "2018-11-29T00:58:15.608556Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tempo para treinar o classificador (segundos):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >0.02250</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >0.01448</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >0.01521</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >0.01115</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >0.00869</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >0.00805</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.03733</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >0.00397</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >0.02049</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >0.02605</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >0.00167</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >0.01027</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >0.00959</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >0.38568</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >0.17626</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >1.18968</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >0.42384</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >0.25415</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >0.55289</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >0.19812</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >0.04004</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >6.15197</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >0.65222</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >0.39617</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >6.46886</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >0.41299</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >0.22560</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >3.44310</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >149.83010</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >9.28450</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >17.96285</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >79.28158</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >4.45020</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >13.74030</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >20.13697</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >0.01788</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >0.01477</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >0.01080</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >0.00879</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >0.00921</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >0.00855</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.03827</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >11.44937</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >3.48411</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >4.32926</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >5.72836</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >1.08461</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >0.56448</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >3.97777</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >0.83014</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >1.04720</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >1.04952</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >0.33907</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >0.40392</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >0.36003</td>\n",
       "                        <td id=\"T_82b5531e_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >10.48507</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297ce4c748>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Tempo para treinar o classificador (segundos):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(train_time['titulo_noticias'])\n",
    "df.style.apply(highlight_min).format({\n",
    "    'tfidf': '{:,.5f}'.format,\n",
    "    'tfidf+lsa': '{:,.5f}'.format,\n",
    "    'tfidf+lda': '{:,.5f}'.format,\n",
    "    'count': '{:,.5f}'.format,\n",
    "    'count+lsa': '{:,.5f}'.format,\n",
    "    'count+lda': '{:,.5f}'.format,\n",
    "    'tfidf+count+w2c': '{:,.5f}'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-11-29T00:58:15.835303Z",
     "start_time": "2018-11-29T00:58:15.665856Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tempo para predizer novos itens (segundos):\n",
      "--------------------\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style  type=\"text/css\" >\n",
       "    #T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col1 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col2 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col0 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col3 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col4 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col5 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }    #T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col6 {\n",
       "            background-color:  green;\n",
       "             color:  white;\n",
       "             font-weight:  bold;\n",
       "        }</style><table id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7\" ><thead>    <tr>        <th class=\"blank level0\" ></th>        <th class=\"col_heading level0 col0\" >tfidf</th>        <th class=\"col_heading level0 col1\" >tfidf+lsa</th>        <th class=\"col_heading level0 col2\" >tfidf+lda</th>        <th class=\"col_heading level0 col3\" >count</th>        <th class=\"col_heading level0 col4\" >count+lsa</th>        <th class=\"col_heading level0 col5\" >count+lda</th>        <th class=\"col_heading level0 col6\" >tfidf+count+w2c</th>    </tr></thead><tbody>\n",
       "                <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row0\" class=\"row_heading level0 row0\" >ComplementNB</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col0\" class=\"data row0 col0\" >0.00187</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col1\" class=\"data row0 col1\" >0.00082</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col2\" class=\"data row0 col2\" >0.00080</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col3\" class=\"data row0 col3\" >0.00092</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col4\" class=\"data row0 col4\" >0.00051</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col5\" class=\"data row0 col5\" >0.00056</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row0_col6\" class=\"data row0 col6\" >0.00786</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row1\" class=\"row_heading level0 row1\" >KNeighborsClassifier</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row1_col0\" class=\"data row1 col0\" >0.44656</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row1_col1\" class=\"data row1 col1\" >1.30266</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row1_col2\" class=\"data row1 col2\" >1.53888</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row1_col3\" class=\"data row1 col3\" >0.15371</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row1_col4\" class=\"data row1 col4\" >0.94733</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row1_col5\" class=\"data row1 col5\" >0.43912</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row1_col6\" class=\"data row1 col6\" >30.83668</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row2\" class=\"row_heading level0 row2\" >LinearSVC</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col0\" class=\"data row2 col0\" >0.00096</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col1\" class=\"data row2 col1\" >0.00096</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col2\" class=\"data row2 col2\" >0.00084</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col3\" class=\"data row2 col3\" >0.00052</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col4\" class=\"data row2 col4\" >0.00048</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col5\" class=\"data row2 col5\" >0.00047</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row2_col6\" class=\"data row2 col6\" >0.00716</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row3\" class=\"row_heading level0 row3\" >LogisticRegression</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row3_col0\" class=\"data row3 col0\" >0.00667</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row3_col1\" class=\"data row3 col1\" >0.00164</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row3_col2\" class=\"data row3 col2\" >0.00103</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row3_col3\" class=\"data row3 col3\" >0.00307</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row3_col4\" class=\"data row3 col4\" >0.00128</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row3_col5\" class=\"data row3 col5\" >0.00056</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row3_col6\" class=\"data row3 col6\" >0.01708</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row4\" class=\"row_heading level0 row4\" >MLPClassifier</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row4_col0\" class=\"data row4 col0\" >0.01744</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row4_col1\" class=\"data row4 col1\" >0.00498</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row4_col2\" class=\"data row4 col2\" >0.00700</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row4_col3\" class=\"data row4 col3\" >0.02133</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row4_col4\" class=\"data row4 col4\" >0.00337</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row4_col5\" class=\"data row4 col5\" >0.00434</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row4_col6\" class=\"data row4 col6\" >0.01733</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row5\" class=\"row_heading level0 row5\" >MultinomialNB</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row5_col0\" class=\"data row5 col0\" >0.00189</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row5_col1\" class=\"data row5 col1\" >0.00198</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row5_col2\" class=\"data row5 col2\" >0.00099</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row5_col3\" class=\"data row5 col3\" >0.00106</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row5_col4\" class=\"data row5 col4\" >0.00057</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row5_col5\" class=\"data row5 col5\" >0.00631</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row5_col6\" class=\"data row5 col6\" >0.00794</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row6\" class=\"row_heading level0 row6\" >RandomForestClassifier</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row6_col0\" class=\"data row6 col0\" >0.19044</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row6_col1\" class=\"data row6 col1\" >0.03911</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row6_col2\" class=\"data row6 col2\" >0.05360</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row6_col3\" class=\"data row6 col3\" >0.06809</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row6_col4\" class=\"data row6 col4\" >0.01543</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row6_col5\" class=\"data row6 col5\" >0.02110</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row6_col6\" class=\"data row6 col6\" >0.06379</td>\n",
       "            </tr>\n",
       "            <tr>\n",
       "                        <th id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7level0_row7\" class=\"row_heading level0 row7\" >SVC</th>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row7_col0\" class=\"data row7 col0\" >0.51707</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row7_col1\" class=\"data row7 col1\" >0.55322</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row7_col2\" class=\"data row7 col2\" >0.58414</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row7_col3\" class=\"data row7 col3\" >0.18923</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row7_col4\" class=\"data row7 col4\" >0.19878</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row7_col5\" class=\"data row7 col5\" >0.18362</td>\n",
       "                        <td id=\"T_82b5531f_2753_11e9_80b4_d89c67dd8af7row7_col6\" class=\"data row7 col6\" >6.32095</td>\n",
       "            </tr>\n",
       "    </tbody></table>"
      ],
      "text/plain": [
       "<pandas.io.formats.style.Styler at 0x7f297cf94eb8>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('Tempo para predizer novos itens (segundos):')\n",
    "print('-' * 20)\n",
    "df = pd.DataFrame.from_dict(predict_time['titulo_noticias'])\n",
    "df.style.apply(highlight_min).format({\n",
    "    'tfidf': '{:,.5f}'.format,\n",
    "    'tfidf+lsa': '{:,.5f}'.format,\n",
    "    'tfidf+lda': '{:,.5f}'.format,\n",
    "    'count': '{:,.5f}'.format,\n",
    "    'count+lsa': '{:,.5f}'.format,\n",
    "    'count+lda': '{:,.5f}'.format,\n",
    "    'tfidf+count+w2c': '{:,.5f}'.format,\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Modelo escolhido e salvo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:57.879386Z",
     "start_time": "2018-12-21T16:55:55.978981Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo   : CalibratedClassifierCV\n",
      "Acurácia : 62.03%\n"
     ]
    }
   ],
   "source": [
    "vec_tfidf = TfidfVectorizer(ngram_range=(1, 2), max_features=1500)\n",
    "X_tfidf_svd = vec_tfidf.fit_transform(afrases)\n",
    "\n",
    "svd = TruncatedSVD(100, random_state=10)\n",
    "normalizer = MinMaxScaler(copy=False)\n",
    "lsa = make_pipeline(svd, normalizer)\n",
    "X_svd = lsa.fit_transform(X_tfidf_svd)\n",
    "\n",
    "svm = LinearSVC(max_iter=150, random_state=10)\n",
    "model = CalibratedClassifierCV(svm) \n",
    "model.fit(X_svd, avalencias)\n",
    "\n",
    "accuracy = np.round(model.score(X_svd, avalencias) * 100, 2)\n",
    "print(f'Modelo   : {model.__class__.__name__}')\n",
    "print(f'Acurácia : {accuracy}%')\n",
    "\n",
    "# filename = 'tfidf_valence.sav'\n",
    "# joblib.dump(vec_tfidf, filename)\n",
    "\n",
    "# filename = 'lsa_valence.sav'\n",
    "# joblib.dump(lsa, filename)\n",
    "\n",
    "# filename = 'model_valence.sav'\n",
    "# joblib.dump(model, filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:57.886265Z",
     "start_time": "2018-12-21T16:55:57.881202Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['NEGATIVO', 'NEUTRO', 'POSITIVO'], dtype='<U8')"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.classes_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-21T16:55:58.035817Z",
     "start_time": "2018-12-21T16:55:57.889329Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[37.31, 41.47, 21.22]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = model.predict_proba(X_svd)\n",
    "list(np.round(y[0] * 100, 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "include_colab_link": true,
   "name": "testes-valencia-projeto-final-ia369y-2sem-2018.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "hide_input": false,
  "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": "3.7.2"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {
    "height": "211px",
    "width": "252px"
   },
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}