Classificador_Fake_News.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Classificador Fake News.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "mount_file_id": "1yDkpl8KFo2U-ArevFILpom_fNcGbfOKb",
      "authorship_tag": "ABX9TyNdAPSbmmZXwhbzCyPo/zs5"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "<img alt=\"Colaboratory logo\" width=\"15%\" src=\"https://i.postimg.cc/8CdBQXmP/JPLavor.png\">\n",
        "\n",
        "#### **Data Science**\n",
        "*by [jplavorr](https://linktr.ee/jplavorr)*  \n",
        "\n",
        "---"
      ],
      "metadata": {
        "id": "0G6uam5-Y23V"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Classificador de Fake News\n",
        "\n",
        "\n",
        "Vivemos na era digital, onde a informação está mais acessível do que nunca, e continua ficando cada vez mais acessível, de forma mais ágil e ampla.\n",
        "\n",
        "Apesar dos benefícios que isso traz, também existem problemas e riscos com isso. Aqui, vamos nos preocupar com um dos maiores deles: ***Fake News***\n",
        "\n",
        "\n",
        "Com certeza você já se deparou com uma dessas. Em uma era de informação, a desinformação acaba sendo o nosso maior problema.\n",
        "\n",
        "Em suma, ***Fake News*** são, como a tradução literal sugere, Notícias Falsas. Elas podem ir desde um número diferente em uma estatística apresentada, até a notícia de uma acontecimento completamente falso.\n",
        "\n"
      ],
      "metadata": {
        "id": "jfZKIQGnHXiX"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "u11tGuInNo4d",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5054edd1-882a-43ea-e3fa-a310defb2db5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[?25l\r\u001b[K     |▊                               | 10 kB 22.7 MB/s eta 0:00:01\r\u001b[K     |█▍                              | 20 kB 14.1 MB/s eta 0:00:01\r\u001b[K     |██▏                             | 30 kB 10.3 MB/s eta 0:00:01\r\u001b[K     |██▉                             | 40 kB 8.3 MB/s eta 0:00:01\r\u001b[K     |███▌                            | 51 kB 4.3 MB/s eta 0:00:01\r\u001b[K     |████▎                           | 61 kB 5.1 MB/s eta 0:00:01\r\u001b[K     |█████                           | 71 kB 5.6 MB/s eta 0:00:01\r\u001b[K     |█████▊                          | 81 kB 5.8 MB/s eta 0:00:01\r\u001b[K     |██████▍                         | 92 kB 6.5 MB/s eta 0:00:01\r\u001b[K     |███████                         | 102 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████▉                        | 112 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████▌                       | 122 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████▏                      | 133 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████                      | 143 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████▋                     | 153 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████▍                    | 163 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████                    | 174 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████▊                   | 184 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████▌                  | 194 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████▏                 | 204 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████▉                 | 215 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████▋                | 225 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████▎               | 235 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████               | 245 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████▊              | 256 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████▍             | 266 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████▏            | 276 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████▉            | 286 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████▌           | 296 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████▎          | 307 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████          | 317 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████▊         | 327 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████▍        | 337 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████        | 348 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████▉       | 358 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▌      | 368 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▏     | 378 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████     | 389 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▋    | 399 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▍   | 409 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████   | 419 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▊  | 430 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▌ | 440 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▏| 450 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▉| 460 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 462 kB 5.2 MB/s \n",
            "\u001b[?25h"
          ]
        }
      ],
      "source": [
        "!pip install tensorflow -q"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# importando os pacotes necessarios\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import re\n",
        "import nltk\n",
        "from nltk.corpus import stopwords\n",
        "from nltk.stem.porter import PorterStemmer\n",
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "import missingno as msno\n",
        "\n",
        "# configurando o noteebook\n",
        "np.random.seed(42)\n",
        "sns.set_style()\n",
        "\n",
        "# filtrar mensagens de warning\n",
        "import warnings\n",
        "warnings.filterwarnings('ignore')\n"
      ],
      "metadata": {
        "id": "MmW63RSjI207"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "nltk.download(\"stopwords\")"
      ],
      "metadata": {
        "id": "lAyJ9UrvKJe2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f5bcbfc6-bf04-40de-a182-a08c57334e86"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/stopwords.zip.\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "stopwords = stopwords.words(\"english\")"
      ],
      "metadata": {
        "id": "rM0f3Rl6KPFh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Fake News Dataset\n",
        "\n",
        "O conjunto de dados ([que pode ser encontrado aqui](https://www.uvic.ca/ecs/ece/isot/datasets/fake-news/index.php)) contém dois tipos de artigos falsos e notícias reais. Esse conjunto de dados foi coletado de fontes do mundo real; os artigos verídicos foram obtidos rastreando artigos da Reuters.com (Noticiário local na Internet). Quanto aos artigos de notícias falsas, eles foram coletados de diferentes fontes. As notícias falsas artigos foram coletados de sites não confiáveis ​​que foram sinalizados pelo Politifact (um organização nos EUA) e Wikipedia. O conjunto de dados contém diferentes tipos de artigos em diferentes\n",
        "tópicos, no entanto, a maioria dos artigos se concentra em tópicos de notícias políticas e mundiais.\n",
        "\n",
        "\n",
        "O conjunto de dados consiste em dois arquivos CSV. O primeiro arquivo chamado “True.csv” contém mais de 12.600\n",
        "artigos de reuter.com. O segundo arquivo chamado “Fake.csv” contém mais de 12.600 artigos de\n",
        "diferentes recursos de agências de notícias falsas. Cada artigo inclui as seguintes informações:\n",
        "\n",
        "- Título do artigo.\n",
        "- Texto.\n",
        "- Tipo.  \n",
        "- Data em que o artigo foi publicado. \n",
        "\n",
        "Para corresponder aos dados de notícias falsas coletados para kaggle.com,\n",
        "focamos principalmente na coleta de artigos de 2016 a 2017. Os dados coletados foram limpos e processadas, porém, as pontuações e os erros nas fake news foram mantidos no texto."
      ],
      "metadata": {
        "id": "9qqpXfyPQDuU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fake = pd.read_csv(\"/content/drive/MyDrive/Fake.csv\")"
      ],
      "metadata": {
        "id": "p4dioup7JPW0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "true = pd.read_csv(\"/content/drive/MyDrive/True.csv\")"
      ],
      "metadata": {
        "id": "FRngpoCOn2oH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "fake.head()"
      ],
      "metadata": {
        "id": "3fK9Ch9Mn4_j",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "e3a54843-3bca-4cf2-fcfb-358aefeae9a9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                               title  \\\n",
              "0   Donald Trump Sends Out Embarrassing New Year’...   \n",
              "1   Drunk Bragging Trump Staffer Started Russian ...   \n",
              "2   Sheriff David Clarke Becomes An Internet Joke...   \n",
              "3   Trump Is So Obsessed He Even Has Obama’s Name...   \n",
              "4   Pope Francis Just Called Out Donald Trump Dur...   \n",
              "\n",
              "                                                text subject  \\\n",
              "0  Donald Trump just couldn t wish all Americans ...    News   \n",
              "1  House Intelligence Committee Chairman Devin Nu...    News   \n",
              "2  On Friday, it was revealed that former Milwauk...    News   \n",
              "3  On Christmas day, Donald Trump announced that ...    News   \n",
              "4  Pope Francis used his annual Christmas Day mes...    News   \n",
              "\n",
              "                date  \n",
              "0  December 31, 2017  \n",
              "1  December 31, 2017  \n",
              "2  December 30, 2017  \n",
              "3  December 29, 2017  \n",
              "4  December 25, 2017  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-48a77af5-cdc3-4262-9177-4cc15459cf55\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>title</th>\n",
              "      <th>text</th>\n",
              "      <th>subject</th>\n",
              "      <th>date</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Donald Trump Sends Out Embarrassing New Year’...</td>\n",
              "      <td>Donald Trump just couldn t wish all Americans ...</td>\n",
              "      <td>News</td>\n",
              "      <td>December 31, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Drunk Bragging Trump Staffer Started Russian ...</td>\n",
              "      <td>House Intelligence Committee Chairman Devin Nu...</td>\n",
              "      <td>News</td>\n",
              "      <td>December 31, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Sheriff David Clarke Becomes An Internet Joke...</td>\n",
              "      <td>On Friday, it was revealed that former Milwauk...</td>\n",
              "      <td>News</td>\n",
              "      <td>December 30, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Trump Is So Obsessed He Even Has Obama’s Name...</td>\n",
              "      <td>On Christmas day, Donald Trump announced that ...</td>\n",
              "      <td>News</td>\n",
              "      <td>December 29, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Pope Francis Just Called Out Donald Trump Dur...</td>\n",
              "      <td>Pope Francis used his annual Christmas Day mes...</td>\n",
              "      <td>News</td>\n",
              "      <td>December 25, 2017</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-48a77af5-cdc3-4262-9177-4cc15459cf55')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-48a77af5-cdc3-4262-9177-4cc15459cf55 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-48a77af5-cdc3-4262-9177-4cc15459cf55');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "true.head()"
      ],
      "metadata": {
        "id": "7lb11n2XyhCk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "6d5962e5-ff94-41f6-c55b-9b764dbad872"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                               title  \\\n",
              "0  As U.S. budget fight looms, Republicans flip t...   \n",
              "1  U.S. military to accept transgender recruits o...   \n",
              "2  Senior U.S. Republican senator: 'Let Mr. Muell...   \n",
              "3  FBI Russia probe helped by Australian diplomat...   \n",
              "4  Trump wants Postal Service to charge 'much mor...   \n",
              "\n",
              "                                                text       subject  \\\n",
              "0  WASHINGTON (Reuters) - The head of a conservat...  politicsNews   \n",
              "1  WASHINGTON (Reuters) - Transgender people will...  politicsNews   \n",
              "2  WASHINGTON (Reuters) - The special counsel inv...  politicsNews   \n",
              "3  WASHINGTON (Reuters) - Trump campaign adviser ...  politicsNews   \n",
              "4  SEATTLE/WASHINGTON (Reuters) - President Donal...  politicsNews   \n",
              "\n",
              "                 date  \n",
              "0  December 31, 2017   \n",
              "1  December 29, 2017   \n",
              "2  December 31, 2017   \n",
              "3  December 30, 2017   \n",
              "4  December 29, 2017   "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-7f8c66ed-f493-4921-9aa3-617d1995b189\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>title</th>\n",
              "      <th>text</th>\n",
              "      <th>subject</th>\n",
              "      <th>date</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>As U.S. budget fight looms, Republicans flip t...</td>\n",
              "      <td>WASHINGTON (Reuters) - The head of a conservat...</td>\n",
              "      <td>politicsNews</td>\n",
              "      <td>December 31, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>U.S. military to accept transgender recruits o...</td>\n",
              "      <td>WASHINGTON (Reuters) - Transgender people will...</td>\n",
              "      <td>politicsNews</td>\n",
              "      <td>December 29, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Senior U.S. Republican senator: 'Let Mr. Muell...</td>\n",
              "      <td>WASHINGTON (Reuters) - The special counsel inv...</td>\n",
              "      <td>politicsNews</td>\n",
              "      <td>December 31, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>FBI Russia probe helped by Australian diplomat...</td>\n",
              "      <td>WASHINGTON (Reuters) - Trump campaign adviser ...</td>\n",
              "      <td>politicsNews</td>\n",
              "      <td>December 30, 2017</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Trump wants Postal Service to charge 'much mor...</td>\n",
              "      <td>SEATTLE/WASHINGTON (Reuters) - President Donal...</td>\n",
              "      <td>politicsNews</td>\n",
              "      <td>December 29, 2017</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-7f8c66ed-f493-4921-9aa3-617d1995b189')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-7f8c66ed-f493-4921-9aa3-617d1995b189 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-7f8c66ed-f493-4921-9aa3-617d1995b189');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Shape of real news dataset\n",
        "print(\"Real news: \", true.shape)\n",
        "#Shape of fake news dataset\n",
        "print(\"Fake News: \", fake.shape)"
      ],
      "metadata": {
        "id": "3l0mw8CmD8Hy",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "151d55a5-bd9c-42e4-ab29-8fb406b0d46a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Real news:  (21417, 4)\n",
            "Fake News:  (23481, 4)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Análise Exploratória de Dados\n",
        "\n",
        "Primeiro, vamos realizar algumas análises para que possamos entender possíveis tendências em nossos dados. Como os datasets de notícias reais e notícias falsas são separados, precisaremos rotular a nossa variável e concatenar os dois datasets. Para o conjunto de dados de notícias reais, fazemos um loop no número de linhas e adicionamos um 0 ao dataframe de notícias reais. Para o conjunto de dados de notícias falsas, repetimos esse procedimento, mas adicionamos 1 (que é nossa variável alvo)."
      ],
      "metadata": {
        "id": "lnpLJ5fSB0co"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Adicionar label 1 no fake\n",
        "fake[\"Label\"] = 1"
      ],
      "metadata": {
        "id": "-F1uWHiODcqt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Adicionar label 0 no true\n",
        "true[\"Label\"] = 0"
      ],
      "metadata": {
        "id": "O18fRdSKEK0c"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#concatenando os datasets\n",
        "news = pd.concat([true,fake])\n",
        "#Sample vai misturar nosso dataset\n",
        "news = news.sample(frac = 1)"
      ],
      "metadata": {
        "id": "xeD-ee3hEMoT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "news.head()"
      ],
      "metadata": {
        "id": "qo2dIXEREP71",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "19863e37-64ac-4833-e815-23c8658bffc0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                   title  \\\n",
              "799     BREAKING: GOP Chairman Grassley Has Had Enoug...   \n",
              "6500    Failed GOP Candidates Remembered In Hilarious...   \n",
              "3590    Mike Pence’s New DC Neighbors Are HILARIOUSLY...   \n",
              "1377   California AG pledges to defend birth control ...   \n",
              "11059  AZ RANCHERS Living On US-Mexico Border Destroy...   \n",
              "\n",
              "                                                    text       subject  Label  \\\n",
              "799    Donald Trump s White House is in chaos, and th...          News      1   \n",
              "6500   Now that Donald Trump is the presumptive GOP n...          News      1   \n",
              "3590   Mike Pence is a huge homophobe. He supports ex...          News      1   \n",
              "1377   SAN FRANCISCO (Reuters) - California Attorney ...  politicsNews      0   \n",
              "11059  Twisted reasoning is all that comes from Pelos...      politics      1   \n",
              "\n",
              "                                                     all  \n",
              "799     BREAKING: GOP Chairman Grassley Has Had Enoug...  \n",
              "6500    Failed GOP Candidates Remembered In Hilarious...  \n",
              "3590    Mike Pence’s New DC Neighbors Are HILARIOUSLY...  \n",
              "1377   California AG pledges to defend birth control ...  \n",
              "11059  AZ RANCHERS Living On US-Mexico Border Destroy...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-cef6fc04-ebd4-43de-905a-b781dfde9de3\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>title</th>\n",
              "      <th>text</th>\n",
              "      <th>subject</th>\n",
              "      <th>Label</th>\n",
              "      <th>all</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>799</th>\n",
              "      <td>BREAKING: GOP Chairman Grassley Has Had Enoug...</td>\n",
              "      <td>Donald Trump s White House is in chaos, and th...</td>\n",
              "      <td>News</td>\n",
              "      <td>1</td>\n",
              "      <td>BREAKING: GOP Chairman Grassley Has Had Enoug...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6500</th>\n",
              "      <td>Failed GOP Candidates Remembered In Hilarious...</td>\n",
              "      <td>Now that Donald Trump is the presumptive GOP n...</td>\n",
              "      <td>News</td>\n",
              "      <td>1</td>\n",
              "      <td>Failed GOP Candidates Remembered In Hilarious...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3590</th>\n",
              "      <td>Mike Pence’s New DC Neighbors Are HILARIOUSLY...</td>\n",
              "      <td>Mike Pence is a huge homophobe. He supports ex...</td>\n",
              "      <td>News</td>\n",
              "      <td>1</td>\n",
              "      <td>Mike Pence’s New DC Neighbors Are HILARIOUSLY...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1377</th>\n",
              "      <td>California AG pledges to defend birth control ...</td>\n",
              "      <td>SAN FRANCISCO (Reuters) - California Attorney ...</td>\n",
              "      <td>politicsNews</td>\n",
              "      <td>0</td>\n",
              "      <td>California AG pledges to defend birth control ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11059</th>\n",
              "      <td>AZ RANCHERS Living On US-Mexico Border Destroy...</td>\n",
              "      <td>Twisted reasoning is all that comes from Pelos...</td>\n",
              "      <td>politics</td>\n",
              "      <td>1</td>\n",
              "      <td>AZ RANCHERS Living On US-Mexico Border Destroy...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-cef6fc04-ebd4-43de-905a-b781dfde9de3')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-cef6fc04-ebd4-43de-905a-b781dfde9de3 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-cef6fc04-ebd4-43de-905a-b781dfde9de3');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Plotando dados ausentes\n",
        "msno.bar(news, figsize=(12,5));"
      ],
      "metadata": {
        "id": "0r1y4vugEczi",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 415
        },
        "outputId": "f8b1e33a-3aa7-4681-e330-87de1780de3e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x360 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "news.drop(columns = \"date\", axis = 1, inplace=True)"
      ],
      "metadata": {
        "id": "ELH_CFyOzE0X"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Agora temos exatamente o dataset que precisamos para comecar nosso modelo de machine learning."
      ],
      "metadata": {
        "id": "kaNFFZwJPVu8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "all_list = [words for words in news[\"title\"] + \" \" + news[\"text\"] + \" \" + news[\"subject\"]]"
      ],
      "metadata": {
        "id": "XjkeOXlaTF6W"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "news[\"all\"] = all_list"
      ],
      "metadata": {
        "id": "s89_rSg6TZzU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nTipo das Variáveis presentes no DataSet:\", news.dtypes)\n",
        "print('\\nValores únicos por variável:', news.nunique())\n",
        "print(\"\\nNome das variáveis:\", news.columns.values)"
      ],
      "metadata": {
        "id": "B3TFPBXvElC3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c6accb38-f74f-402b-9ee6-cefed62d5fe8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Tipo das Variáveis presentes no DataSet: title      object\n",
            "text       object\n",
            "subject    object\n",
            "Label       int64\n",
            "all        object\n",
            "dtype: object\n",
            "\n",
            "Valores únicos por variável: title      38729\n",
            "text       38646\n",
            "subject        8\n",
            "Label          2\n",
            "all        44685\n",
            "dtype: int64\n",
            "\n",
            "Nome das variáveis: ['title' 'text' 'subject' 'Label' 'all']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Como temos 21.417 amostras de notícias reais e 23.481 amostras de notícias falsas, há uma proporção de aproximadamente 48:52 notícias reais: notícias falsas. Isso significa que nosso conjunto de dados é relativamente equilibrado."
      ],
      "metadata": {
        "id": "vgfe5ZV9P_q2"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Tratamento de Texto"
      ],
      "metadata": {
        "id": "TI6-NOYr-VOE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "stemmer = PorterStemmer() "
      ],
      "metadata": {
        "id": "bF9PzJUqdCRI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df = news.copy()"
      ],
      "metadata": {
        "id": "elUnf7c2uKlG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.head()"
      ],
      "metadata": {
        "id": "LoJuOfHquNhk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "outputId": "b17abe63-ceae-4ec0-d000-68e3ea028ef9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                   title  \\\n",
              "799     BREAKING: GOP Chairman Grassley Has Had Enoug...   \n",
              "6500    Failed GOP Candidates Remembered In Hilarious...   \n",
              "3590    Mike Pence’s New DC Neighbors Are HILARIOUSLY...   \n",
              "1377   California AG pledges to defend birth control ...   \n",
              "11059  AZ RANCHERS Living On US-Mexico Border Destroy...   \n",
              "\n",
              "                                                    text       subject  Label  \\\n",
              "799    Donald Trump s White House is in chaos, and th...          News      1   \n",
              "6500   Now that Donald Trump is the presumptive GOP n...          News      1   \n",
              "3590   Mike Pence is a huge homophobe. He supports ex...          News      1   \n",
              "1377   SAN FRANCISCO (Reuters) - California Attorney ...  politicsNews      0   \n",
              "11059  Twisted reasoning is all that comes from Pelos...      politics      1   \n",
              "\n",
              "                                                     all  \\\n",
              "799     BREAKING: GOP Chairman Grassley Has Had Enoug...   \n",
              "6500    Failed GOP Candidates Remembered In Hilarious...   \n",
              "3590    Mike Pence’s New DC Neighbors Are HILARIOUSLY...   \n",
              "1377   California AG pledges to defend birth control ...   \n",
              "11059  AZ RANCHERS Living On US-Mexico Border Destroy...   \n",
              "\n",
              "                                               unstemmed  \\\n",
              "799    [breaking , gop, chairman, grassley, has, had,...   \n",
              "6500   [failed, gop, candidates, remembered, in, hila...   \n",
              "3590   [mike, pence s, new, dc, neighbors, are, hilar...   \n",
              "1377   [california, ag, pledges, to, defend, birth, c...   \n",
              "11059  [az, ranchers, living, on, us mexico, border, ...   \n",
              "\n",
              "                                                 stemmed  \n",
              "799    [breaking , gop, chairman, grassley, enough , ...  \n",
              "6500   [fail, gop, candid, rememb, hilari, mock, eulo...  \n",
              "3590   [mike, pence , new, dc, neighbor, hilari, trol...  \n",
              "1377   [california, ag, pledg, defend, birth, control...  \n",
              "11059  [az, rancher, live, us mexico, border, destroy...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-16f4294f-509a-471c-bba6-286c6c4f9fcd\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>title</th>\n",
              "      <th>text</th>\n",
              "      <th>subject</th>\n",
              "      <th>Label</th>\n",
              "      <th>all</th>\n",
              "      <th>unstemmed</th>\n",
              "      <th>stemmed</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>799</th>\n",
              "      <td>BREAKING: GOP Chairman Grassley Has Had Enoug...</td>\n",
              "      <td>Donald Trump s White House is in chaos, and th...</td>\n",
              "      <td>News</td>\n",
              "      <td>1</td>\n",
              "      <td>BREAKING: GOP Chairman Grassley Has Had Enoug...</td>\n",
              "      <td>[breaking , gop, chairman, grassley, has, had,...</td>\n",
              "      <td>[breaking , gop, chairman, grassley, enough , ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6500</th>\n",
              "      <td>Failed GOP Candidates Remembered In Hilarious...</td>\n",
              "      <td>Now that Donald Trump is the presumptive GOP n...</td>\n",
              "      <td>News</td>\n",
              "      <td>1</td>\n",
              "      <td>Failed GOP Candidates Remembered In Hilarious...</td>\n",
              "      <td>[failed, gop, candidates, remembered, in, hila...</td>\n",
              "      <td>[fail, gop, candid, rememb, hilari, mock, eulo...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3590</th>\n",
              "      <td>Mike Pence’s New DC Neighbors Are HILARIOUSLY...</td>\n",
              "      <td>Mike Pence is a huge homophobe. He supports ex...</td>\n",
              "      <td>News</td>\n",
              "      <td>1</td>\n",
              "      <td>Mike Pence’s New DC Neighbors Are HILARIOUSLY...</td>\n",
              "      <td>[mike, pence s, new, dc, neighbors, are, hilar...</td>\n",
              "      <td>[mike, pence , new, dc, neighbor, hilari, trol...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1377</th>\n",
              "      <td>California AG pledges to defend birth control ...</td>\n",
              "      <td>SAN FRANCISCO (Reuters) - California Attorney ...</td>\n",
              "      <td>politicsNews</td>\n",
              "      <td>0</td>\n",
              "      <td>California AG pledges to defend birth control ...</td>\n",
              "      <td>[california, ag, pledges, to, defend, birth, c...</td>\n",
              "      <td>[california, ag, pledg, defend, birth, control...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11059</th>\n",
              "      <td>AZ RANCHERS Living On US-Mexico Border Destroy...</td>\n",
              "      <td>Twisted reasoning is all that comes from Pelos...</td>\n",
              "      <td>politics</td>\n",
              "      <td>1</td>\n",
              "      <td>AZ RANCHERS Living On US-Mexico Border Destroy...</td>\n",
              "      <td>[az, ranchers, living, on, us mexico, border, ...</td>\n",
              "      <td>[az, rancher, live, us mexico, border, destroy...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-16f4294f-509a-471c-bba6-286c6c4f9fcd')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-16f4294f-509a-471c-bba6-286c6c4f9fcd button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-16f4294f-509a-471c-bba6-286c6c4f9fcd');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Separa a frase em uma lista de sentencas.\n",
        "df['unstemmed'] = df['all'].str.split()"
      ],
      "metadata": {
        "id": "QwjLaPh3uIFB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Se tiver algo diferente de palavras, ele ira preencher com espaco em branco\n",
        "df['unstemmed'] = df['unstemmed'].apply(lambda x: [re.sub(r\"[^A-Za-z]\", \" \", y).lower() for y in x])"
      ],
      "metadata": {
        "id": "89Mm3SBZwghC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Aplica o porter stemmer nessas palavras\n",
        "df['stemmed'] = df['unstemmed'].apply(lambda x: [stemmer.stem(y) for y in x if not y in stopwords])"
      ],
      "metadata": {
        "id": "_VuIdk1Fuxhi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Exclui stopwords.\n",
        "df['stemmed_stopwords'] = df[\"stemmed\"].apply(lambda x: [item for item in x if item not in stopwords])"
      ],
      "metadata": {
        "id": "lbyYiEn1qDdw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df[\"features\"] = df['stemmed_stopwords'].apply(lambda x: ' '.join(x))"
      ],
      "metadata": {
        "id": "fTKqGYxW00UF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_prod = df[[\"features\", \"Label\"]]"
      ],
      "metadata": {
        "id": "rQoe6f9Y7rFM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_prod.head()"
      ],
      "metadata": {
        "id": "v74vmZvR74Ul",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "204e7a81-2f40-464e-b5d7-3151cb142cc7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                features  Label\n",
              "799    breaking  gop chairman grassley enough  demand...      1\n",
              "6500   fail gop candid rememb hilari mock eulog  vide...      1\n",
              "3590   mike pence  new dc neighbor hilari troll homop...      1\n",
              "1377   california ag pledg defend birth control insur...      0\n",
              "11059  az rancher live us mexico border destroy nanci...      1"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-26e4ade3-663d-4885-984b-e1abf331c819\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>features</th>\n",
              "      <th>Label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>799</th>\n",
              "      <td>breaking  gop chairman grassley enough  demand...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6500</th>\n",
              "      <td>fail gop candid rememb hilari mock eulog  vide...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3590</th>\n",
              "      <td>mike pence  new dc neighbor hilari troll homop...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1377</th>\n",
              "      <td>california ag pledg defend birth control insur...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11059</th>\n",
              "      <td>az rancher live us mexico border destroy nanci...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-26e4ade3-663d-4885-984b-e1abf331c819')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-26e4ade3-663d-4885-984b-e1abf331c819 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-26e4ade3-663d-4885-984b-e1abf331c819');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_prod.shape"
      ],
      "metadata": {
        "id": "EAZRea0K77em",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "54096582-2ff8-471a-ecb7-bfc1e373e062"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(44898, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Pré-Processamento de Strings - NLP\n",
        "\n",
        "Aqui neste projeto, vamos nos concentrar no texto e como você pode construir o classificador baseado em modelos de NLP. Começaremos examinando o sentimento no texto e aprenderemos a construir modelos que entendem o texto treinado em texto rotulado e, em seguida, podemos classificar o novo texto com base no que eles viram.\n",
        "\n",
        "Quando estávamos lidamos com imagens, é relativamente fácil para nós alimentá-las em uma rede neural, pois os valores dos pixels já eram números. E a rede poderia aprender parâmetros de funções que poderiam ser usados ​​para ajustar classes a rótulos. Mas o que acontece com o texto? Como podemos fazer isso com frases e palavras?\n",
        "\n",
        "\n",
        "## **Codificações baseadas em palavras**\n",
        "\n",
        "Poderíamos pegar codificações de caracteres para cada caractere em um conjunto. Por exemplo, os valores ASCII. Mas isso nos ajudará a entender o significado de uma palavra?\n",
        "\n",
        "Então, por exemplo, considere a palavra 'LISTEN' como mostrado aqui. Uma codificação de caracteres simples comum é ASCII, o American Standard Code for Information Interchange com os valores mostrados aqui:\n",
        "\n",
        "<img alt=\"Colaboratory logo\" width=\"50%\" src=\"https://i.postimg.cc/sXJ722t2/imagem0.png\">\n",
        "\n",
        "Podemos pensar que poderia ter uma palavra como LISTEN codificada usando esses valores. Mas o problema com isso, é claro, é que a semântica da palavra não está codificada nas letras. Isso pode ser demonstrado usando a palavra 'SILENT' que tem um significado muito diferente e quase oposto, mas com exatamente as mesmas letras.\n",
        "\n",
        "Portanto, parece que treinar uma rede neural apenas com as letras pode ser uma tarefa assustadora. Então, que tal se considerarmos as palavras? E se pudéssemos dar um valor às palavras e usar esses valores no treinamento de uma rede? Agora podemos estar chegando a algum lugar. \n",
        "\n",
        "Então, por exemplo, considere esta frase, \"I love my dog\". Que tal dar um valor a cada palavra? Qual é esse valor não importa. É que temos um valor por palavra, e o valor é sempre o mesmo para a mesma palavra.\n",
        "\n",
        "Assim, uma codificação simples para a frase seria, por exemplo, dar à palavra 'I' o valor um. A seguir, poderíamos dar às palavras 'Love', 'my' e 'dog' os valores 2, 3 e 4 respectivamente\n",
        "\n",
        "<img alt=\"Colaboratory logo\" width=\"70%\" src=\"https://i.postimg.cc/WbkCfLr4/imagem1.png\">\n",
        "\n",
        "Então a frase, \"I love my dog\" seria codificada como 1, 2, 3, 4. Então agora, e se eu tiver a frase, \"I love my cat\"? Bem, já codificamos as palavras 'I love my' como 1, 2, 3. Assim, podemos reutilizá-las e criar um novo token para cat, que não vimos antes. Então vamos fazer disso o número 5.\n",
        "\n",
        "<img alt=\"Colaboratory logo\" width=\"70%\" src=\"https://i.postimg.cc/DzrbCZfc/imagem2.png\">\n",
        "\n",
        "Agora, se olharmos apenas para os dois conjuntos de codificações, podemos começar a ver alguma semelhança entre as sentenças. \"I love my dog\" é 1, 2, 3, 4 e \"I love my cat\" é 1, 2, 3, 5. Então esse é pelo menos o começo e como podemos começar a treinar uma rede neural baseada em palavras. Felizmente, o TensorFlow e o Keras nos fornecem algumas APIs que tornam isso muito simples.\n",
        "\n",
        "Mas antes de começarmos a fazer os Tokens, iremos separar os nossos dados entre treino e teste."
      ],
      "metadata": {
        "id": "V_0RuxR5qUEc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# separando os datasets em treino, validação e teste\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "train, test = train_test_split(df_prod)\n",
        "train, validation = train_test_split(train, test_size = 0.2)\n",
        "print('Entradas por Dataset')\n",
        "print('Dataset de Treino: ',len(train), 'linhas.')\n",
        "print('Dataset de Validação: ', len(validation), 'linhas.')\n",
        "print('Dataset de Teste: ', len(test), 'linhas.')"
      ],
      "metadata": {
        "id": "2zSUdXqoEql-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9bac6445-84e6-4626-e3d1-ca5f21be6da2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Entradas por Dataset\n",
            "Dataset de Treino:  26938 linhas.\n",
            "Dataset de Validação:  6735 linhas.\n",
            "Dataset de Teste:  11225 linhas.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train.head()"
      ],
      "metadata": {
        "id": "C6qDRi-gd_Bd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "05ad106d-a625-49e5-e16a-6107987b966a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                features  Label\n",
              "13627  russian media law pose threat free press  u s ...      0\n",
              "13478  russia accus u s  tri provok north korean lead...      0\n",
              "18485  father student releas n  korean prison slam ob...      1\n",
              "13689  import hate intolerance  u s ha import        ...      1\n",
              "7267   north korea care presid unit state unit nation...      0"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-debfb301-897c-448c-9b5a-7b7c6ff1a27b\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>features</th>\n",
              "      <th>Label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>13627</th>\n",
              "      <td>russian media law pose threat free press  u s ...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13478</th>\n",
              "      <td>russia accus u s  tri provok north korean lead...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18485</th>\n",
              "      <td>father student releas n  korean prison slam ob...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13689</th>\n",
              "      <td>import hate intolerance  u s ha import        ...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7267</th>\n",
              "      <td>north korea care presid unit state unit nation...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-debfb301-897c-448c-9b5a-7b7c6ff1a27b')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-debfb301-897c-448c-9b5a-7b7c6ff1a27b button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-debfb301-897c-448c-9b5a-7b7c6ff1a27b');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# preparando os dados para o modelo\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "\n",
        "vocab_size = 10000\n",
        "trunc_type = \"post\"\n",
        "pad_type = \"post\"\n",
        "oov_tok = \"<OOV>\""
      ],
      "metadata": {
        "id": "Czbv5j49xSE8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tokenizer = Tokenizer(num_words = vocab_size, oov_token=oov_tok)\n",
        "tokenizer.fit_on_texts(train.features)\n",
        "word_index = tokenizer.word_index"
      ],
      "metadata": {
        "id": "OoDYXTybeOBX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Vamos detalhar linha por linha. \n",
        "\n",
        "\n",
        "Tensorflow e Keras nos dão várias maneiras de codificar palavras, mas o que vou focar é o **tokenizer**. Isso fará o trabalho pesado para nós, gerando o dicionário de codificações de palavras e criando vetores a partir das frases.\n",
        "\n",
        "<p align=center>\n",
        "  <img src=\"https://www.kdnuggets.com/wp-content/uploads/text-tokens-tokenization-manning.jpg\" height=\"30%\">\n",
        "</p>\n",
        "\n",
        "Para ler mais sobre esse assunto e entender melhor os passos e parâmetros abaixo, deixarei links de referência [aqui](https://www.kdnuggets.com/2017/12/general-approach-preprocessing-text-data.html) e [aqui](https://www.kdnuggets.com/2020/03/tensorflow-keras-tokenization-text-data-prep.html).\n",
        "\n",
        "\n",
        "Em seguida, crio uma instância do tokenizer. O parâmetro `num_words`. Nesse caso, estou usando 10000, que é se refere ao número de palavras distintas nos dados. Se você estiver criando um conjunto de treinamento com base em muito texto, geralmente não saberá quantas palavras distintas exclusivas existem nesse texto. Então, ao definir esse hiperparâmetro, o que o tokenizer fará é pegar as 10000 principais palavras por volume e apenas codificá-las. É um atalho útil ao lidar com muitos dados e vale a pena experimentar quando você treinar com dados reais, como nosso caso.\n",
        "\n",
        "O método fit do tokenizer recebe os dados e os codifica. O tokenizer fornece uma propriedade de índice de palavras (`word_index`) que retorna um dicionário contendo pares chave-valor, em que a chave é a palavra e o valor é o token dessa palavra, que você pode inspecionar simplesmente imprimindo-a.\n",
        "\n",
        "### **Olhando mais para o Tokenizer**\n",
        "\n",
        "Em primeiro lugar, realmente precisamos de muitos dados de treinamento para obter um vocabulário amplo ou podemos acabar com frases pouco representativas. Em segundo lugar, em muitos casos, é uma boa ideia, em vez de simplesmente ignorar palavras não vistas, colocar um valor especial quando uma palavra não vista é encontrada. Você pode fazer isso com uma propriedade no tokenizer.\n",
        "\n",
        "Para isso, é preciso adicionar um token de propriedade oov ao construtor do tokenizer. Você pode ver agora que especifiquei que quero que o token `<00v>` para vocabulário externo seja usado para palavras que não estão no índice de palavras. Você pode usar o que quiser aqui, mas lembre-se de que deve ser algo único e distinto que não seja confundido com uma palavra real."
      ],
      "metadata": {
        "id": "dc8N7EMexosg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "word_index"
      ],
      "metadata": {
        "id": "ea_DoiqI25cH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a1ce486c-f5a4-4458-e656-09287b478d1e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'<OOV>': 1,\n",
              " 'trump': 2,\n",
              " 'said': 3,\n",
              " 'u': 4,\n",
              " 's': 5,\n",
              " 'would': 6,\n",
              " 'state': 7,\n",
              " 'presid': 8,\n",
              " 'republican': 9,\n",
              " 'new': 10,\n",
              " 'one': 11,\n",
              " 'say': 12,\n",
              " 'year': 13,\n",
              " 'peopl': 14,\n",
              " 'news': 15,\n",
              " 'also': 16,\n",
              " 'obama': 17,\n",
              " 'clinton': 18,\n",
              " 'like': 19,\n",
              " 'donald': 20,\n",
              " 'democrat': 21,\n",
              " 'govern': 22,\n",
              " 'time': 23,\n",
              " 'nation': 24,\n",
              " 'hous': 25,\n",
              " 'call': 26,\n",
              " 'polit': 27,\n",
              " 'support': 28,\n",
              " 'unit': 29,\n",
              " 'could': 30,\n",
              " 'campaign': 31,\n",
              " 'go': 32,\n",
              " 'report': 33,\n",
              " 'white': 34,\n",
              " 'reuters': 35,\n",
              " 'american': 36,\n",
              " 'make': 37,\n",
              " 'told': 38,\n",
              " 'right': 39,\n",
              " 'elect': 40,\n",
              " 'vote': 41,\n",
              " 'two': 42,\n",
              " 'get': 43,\n",
              " 'last': 44,\n",
              " 'it': 45,\n",
              " 'senat': 46,\n",
              " 'use': 47,\n",
              " 'first': 48,\n",
              " 'back': 49,\n",
              " 'includ': 50,\n",
              " 'work': 51,\n",
              " 'take': 52,\n",
              " 'want': 53,\n",
              " 'th': 54,\n",
              " 'offici': 55,\n",
              " 'washington': 56,\n",
              " 'law': 57,\n",
              " 'even': 58,\n",
              " 'us': 59,\n",
              " 'video': 60,\n",
              " 'group': 61,\n",
              " 'former': 62,\n",
              " 'parti': 63,\n",
              " 'hillari': 64,\n",
              " 'media': 65,\n",
              " 'court': 66,\n",
              " 'secur': 67,\n",
              " 'countri': 68,\n",
              " 'day': 69,\n",
              " 'show': 70,\n",
              " 'mani': 71,\n",
              " 'week': 72,\n",
              " 'may': 73,\n",
              " 'plan': 74,\n",
              " 'attack': 75,\n",
              " 'million': 76,\n",
              " 'accord': 77,\n",
              " 'think': 78,\n",
              " 'made': 79,\n",
              " 'russia': 80,\n",
              " 'come': 81,\n",
              " 'know': 82,\n",
              " 'bill': 83,\n",
              " 'twitter': 84,\n",
              " 'need': 85,\n",
              " 'polic': 86,\n",
              " 'offic': 87,\n",
              " 'ask': 88,\n",
              " 'america': 89,\n",
              " 'sinc': 90,\n",
              " 'way': 91,\n",
              " 'percent': 92,\n",
              " 'leader': 93,\n",
              " 't': 94,\n",
              " 'presidenti': 95,\n",
              " 'statement': 96,\n",
              " 'left': 97,\n",
              " 'tax': 98,\n",
              " 'public': 99,\n",
              " 'help': 100,\n",
              " 'north': 101,\n",
              " 'via': 102,\n",
              " 'imag': 103,\n",
              " 'tri': 104,\n",
              " 'see': 105,\n",
              " 'member': 106,\n",
              " 'world': 107,\n",
              " 'well': 108,\n",
              " 'feder': 109,\n",
              " 'month': 110,\n",
              " 'talk': 111,\n",
              " 'foreign': 112,\n",
              " 'russian': 113,\n",
              " 'forc': 114,\n",
              " 'part': 115,\n",
              " 'post': 116,\n",
              " 'deal': 117,\n",
              " 'administr': 118,\n",
              " 'investig': 119,\n",
              " 'rule': 120,\n",
              " 'order': 121,\n",
              " 'meet': 122,\n",
              " 'politicsnew': 123,\n",
              " 'look': 124,\n",
              " 'in': 125,\n",
              " 'china': 126,\n",
              " 'watch': 127,\n",
              " 'tuesday': 128,\n",
              " 'gener': 129,\n",
              " 'militari': 130,\n",
              " 'case': 131,\n",
              " 'power': 132,\n",
              " 'women': 133,\n",
              " 'issu': 134,\n",
              " 'end': 135,\n",
              " 'major': 136,\n",
              " 'war': 137,\n",
              " 'live': 138,\n",
              " 'depart': 139,\n",
              " 'still': 140,\n",
              " 'black': 141,\n",
              " 'congress': 142,\n",
              " 'much': 143,\n",
              " 'i': 144,\n",
              " 'thing': 145,\n",
              " 'that': 146,\n",
              " 'claim': 147,\n",
              " 'run': 148,\n",
              " 'comment': 149,\n",
              " 'a': 150,\n",
              " 'wednesday': 151,\n",
              " 'polici': 152,\n",
              " 'worldnew': 153,\n",
              " 'thursday': 154,\n",
              " 'anoth': 155,\n",
              " 'believ': 156,\n",
              " 'put': 157,\n",
              " 'critic': 158,\n",
              " 'move': 159,\n",
              " 'minist': 160,\n",
              " 'question': 161,\n",
              " 'friday': 162,\n",
              " 'muslim': 163,\n",
              " 'point': 164,\n",
              " 'immigr': 165,\n",
              " 'president': 166,\n",
              " 'candid': 167,\n",
              " 'featur': 168,\n",
              " 'three': 169,\n",
              " 'chang': 170,\n",
              " 'kill': 171,\n",
              " 'person': 172,\n",
              " 'monday': 173,\n",
              " 'conserv': 174,\n",
              " 'continu': 175,\n",
              " 'long': 176,\n",
              " 'give': 177,\n",
              " 'follow': 178,\n",
              " 'korea': 179,\n",
              " 'york': 180,\n",
              " 'allow': 181,\n",
              " 'citi': 182,\n",
              " 'fund': 183,\n",
              " 'man': 184,\n",
              " 'act': 185,\n",
              " 'protest': 186,\n",
              " 'next': 187,\n",
              " 'face': 188,\n",
              " 'recent': 189,\n",
              " 'repres': 190,\n",
              " 'whether': 191,\n",
              " 'interview': 192,\n",
              " 'com': 193,\n",
              " 'good': 194,\n",
              " 'secretari': 195,\n",
              " 'intern': 196,\n",
              " 'never': 197,\n",
              " 'fact': 198,\n",
              " 'iran': 199,\n",
              " 'releas': 200,\n",
              " 'around': 201,\n",
              " 'place': 202,\n",
              " 'control': 203,\n",
              " 'old': 204,\n",
              " 'fbi': 205,\n",
              " 'compani': 206,\n",
              " 'appear': 207,\n",
              " 'busi': 208,\n",
              " 'win': 209,\n",
              " 'states': 210,\n",
              " 'accus': 211,\n",
              " 'top': 212,\n",
              " 'anti': 213,\n",
              " 'voter': 214,\n",
              " 'trade': 215,\n",
              " 'job': 216,\n",
              " 'becom': 217,\n",
              " 'program': 218,\n",
              " 'election': 219,\n",
              " 'sever': 220,\n",
              " 'fight': 221,\n",
              " 'open': 222,\n",
              " 'syria': 223,\n",
              " 'lead': 224,\n",
              " 'everi': 225,\n",
              " 'islam': 226,\n",
              " 'far': 227,\n",
              " 'the': 228,\n",
              " 'without': 229,\n",
              " 'expect': 230,\n",
              " 'inform': 231,\n",
              " 'respons': 232,\n",
              " 'start': 233,\n",
              " 'stop': 234,\n",
              " 'took': 235,\n",
              " 'famili': 236,\n",
              " 'head': 237,\n",
              " 'agenc': 238,\n",
              " 'effort': 239,\n",
              " 'money': 240,\n",
              " 'speech': 241,\n",
              " 'number': 242,\n",
              " 'concern': 243,\n",
              " 'nuclear': 244,\n",
              " 'provid': 245,\n",
              " 'action': 246,\n",
              " 'tell': 247,\n",
              " 'chief': 248,\n",
              " 'keep': 249,\n",
              " 'found': 250,\n",
              " 'close': 251,\n",
              " 'protect': 252,\n",
              " 'border': 253,\n",
              " 'name': 254,\n",
              " 'south': 255,\n",
              " 'hope': 256,\n",
              " 'fire': 257,\n",
              " 'email': 258,\n",
              " 'intellig': 259,\n",
              " 'justic': 260,\n",
              " 'commun': 261,\n",
              " 'region': 262,\n",
              " 'home': 263,\n",
              " 'turn': 264,\n",
              " 'press': 265,\n",
              " 'school': 266,\n",
              " 'decis': 267,\n",
              " 'director': 268,\n",
              " 'set': 269,\n",
              " 'n': 270,\n",
              " 'ad': 271,\n",
              " 'pay': 272,\n",
              " 'gun': 273,\n",
              " 'sourc': 274,\n",
              " 'high': 275,\n",
              " 'execut': 276,\n",
              " 'realli': 277,\n",
              " 'billion': 278,\n",
              " 'tweet': 279,\n",
              " 'author': 280,\n",
              " 'committe': 281,\n",
              " 'clear': 282,\n",
              " 'mean': 283,\n",
              " 'matter': 284,\n",
              " 'people': 285,\n",
              " 'poll': 286,\n",
              " 'system': 287,\n",
              " 'social': 288,\n",
              " 'al': 289,\n",
              " 'speak': 290,\n",
              " 'discuss': 291,\n",
              " 'legal': 292,\n",
              " 'him': 293,\n",
              " 'health': 294,\n",
              " 'country': 295,\n",
              " 'sign': 296,\n",
              " 'let': 297,\n",
              " 'ban': 298,\n",
              " 'announc': 299,\n",
              " 'fox': 300,\n",
              " 'possibl': 301,\n",
              " 'organ': 302,\n",
              " 'alreadi': 303,\n",
              " 'stori': 304,\n",
              " 'hold': 305,\n",
              " 'charg': 306,\n",
              " 'actual': 307,\n",
              " 'must': 308,\n",
              " 'today': 309,\n",
              " 'march': 310,\n",
              " 'stand': 311,\n",
              " 'and': 312,\n",
              " 'children': 313,\n",
              " 'pic': 314,\n",
              " 'care': 315,\n",
              " 'consid': 316,\n",
              " 'we': 317,\n",
              " 'happen': 318,\n",
              " 'john': 319,\n",
              " 'now': 320,\n",
              " 'current': 321,\n",
              " 'barack': 322,\n",
              " 'cnn': 323,\n",
              " 'illeg': 324,\n",
              " 'pass': 325,\n",
              " 'however': 326,\n",
              " 'got': 327,\n",
              " 'sunday': 328,\n",
              " 'years': 329,\n",
              " 'hand': 330,\n",
              " 'eu': 331,\n",
              " 'them': 332,\n",
              " 'great': 333,\n",
              " 'oper': 334,\n",
              " 'cruz': 335,\n",
              " 'four': 336,\n",
              " 'alleg': 337,\n",
              " 'governor': 338,\n",
              " 'yet': 339,\n",
              " 'servic': 340,\n",
              " 'co': 341,\n",
              " 'term': 342,\n",
              " 'judg': 343,\n",
              " 'least': 344,\n",
              " 'among': 345,\n",
              " 'free': 346,\n",
              " 'leav': 347,\n",
              " 'seen': 348,\n",
              " 'came': 349,\n",
              " 'remain': 350,\n",
              " 'might': 351,\n",
              " 'econom': 352,\n",
              " 'interest': 353,\n",
              " 'visit': 354,\n",
              " 'later': 355,\n",
              " 'refuge': 356,\n",
              " 'cut': 357,\n",
              " 'though': 358,\n",
              " 'attorney': 359,\n",
              " 'base': 360,\n",
              " 'find': 361,\n",
              " 'r': 362,\n",
              " 'someth': 363,\n",
              " 'so': 364,\n",
              " 'import': 365,\n",
              " 'read': 366,\n",
              " 'creat': 367,\n",
              " 'rais': 368,\n",
              " 'on': 369,\n",
              " 'seek': 370,\n",
              " 'human': 371,\n",
              " 'propos': 372,\n",
              " 'spokesman': 373,\n",
              " 'real': 374,\n",
              " 'went': 375,\n",
              " 'serv': 376,\n",
              " 'event': 377,\n",
              " 'team': 378,\n",
              " 'seem': 379,\n",
              " 'big': 380,\n",
              " 'posit': 381,\n",
              " 'known': 382,\n",
              " 'union': 383,\n",
              " 'receiv': 384,\n",
              " 'defens': 385,\n",
              " 'you': 386,\n",
              " 'earlier': 387,\n",
              " 'legisl': 388,\n",
              " 'aid': 389,\n",
              " 'w': 390,\n",
              " 'ever': 391,\n",
              " 'away': 392,\n",
              " 'held': 393,\n",
              " 'push': 394,\n",
              " 'senior': 395,\n",
              " 'past': 396,\n",
              " 'lawmak': 397,\n",
              " 'fail': 398,\n",
              " 'attempt': 399,\n",
              " 'wall': 400,\n",
              " 'local': 401,\n",
              " 'woman': 402,\n",
              " 'death': 403,\n",
              " 'return': 404,\n",
              " 'build': 405,\n",
              " 'student': 406,\n",
              " 'step': 407,\n",
              " 'gop': 408,\n",
              " 'government': 409,\n",
              " 'threat': 410,\n",
              " 'european': 411,\n",
              " 'involv': 412,\n",
              " 'record': 413,\n",
              " 'here': 414,\n",
              " 'second': 415,\n",
              " 'independ': 416,\n",
              " 'party': 417,\n",
              " 'littl': 418,\n",
              " 'bank': 419,\n",
              " 'relat': 420,\n",
              " 'play': 421,\n",
              " 'respond': 422,\n",
              " 'reason': 423,\n",
              " 'air': 424,\n",
              " 'decid': 425,\n",
              " 'confirm': 426,\n",
              " 'night': 427,\n",
              " 'process': 428,\n",
              " 'agre': 429,\n",
              " 'center': 430,\n",
              " 'up': 431,\n",
              " 'race': 432,\n",
              " 'arrest': 433,\n",
              " 'life': 434,\n",
              " 'requir': 435,\n",
              " 'financi': 436,\n",
              " 'better': 437,\n",
              " 'deni': 438,\n",
              " 'offer': 439,\n",
              " 'problem': 440,\n",
              " 'host': 441,\n",
              " 'result': 442,\n",
              " 'liber': 443,\n",
              " 'thi': 444,\n",
              " 'side': 445,\n",
              " 'reuter': 446,\n",
              " 'prime': 447,\n",
              " 'men': 448,\n",
              " 'account': 449,\n",
              " 'bring': 450,\n",
              " 'list': 451,\n",
              " 'target': 452,\n",
              " 'given': 453,\n",
              " 'area': 454,\n",
              " 'across': 455,\n",
              " 'suprem': 456,\n",
              " 'lot': 457,\n",
              " 'lie': 458,\n",
              " 'request': 459,\n",
              " 'privat': 460,\n",
              " 'arm': 461,\n",
              " 'opposit': 462,\n",
              " 'iraq': 463,\n",
              " 'special': 464,\n",
              " 'saudi': 465,\n",
              " 'hear': 466,\n",
              " 'suggest': 467,\n",
              " 'travel': 468,\n",
              " 'budget': 469,\n",
              " 'wrote': 470,\n",
              " 'line': 471,\n",
              " 'mexico': 472,\n",
              " 'facebook': 473,\n",
              " 'enough': 474,\n",
              " 'syrian': 475,\n",
              " 'defend': 476,\n",
              " 'noth': 477,\n",
              " 'out': 478,\n",
              " 'view': 479,\n",
              " 'done': 480,\n",
              " 'caus': 481,\n",
              " 'sander': 482,\n",
              " 'differ': 483,\n",
              " 'final': 484,\n",
              " 'agreement': 485,\n",
              " 'putin': 486,\n",
              " 'terrorist': 487,\n",
              " 'street': 488,\n",
              " 'full': 489,\n",
              " 'increas': 490,\n",
              " 'despit': 491,\n",
              " 'taken': 492,\n",
              " 'share': 493,\n",
              " 'hate': 494,\n",
              " 'house': 495,\n",
              " 'letter': 496,\n",
              " 'wire': 497,\n",
              " 'commit': 498,\n",
              " 'hard': 499,\n",
              " 'reform': 500,\n",
              " 'role': 501,\n",
              " 'mr': 502,\n",
              " 'comey': 503,\n",
              " 'her': 504,\n",
              " 'june': 505,\n",
              " 'five': 506,\n",
              " 'realdonaldtrump': 507,\n",
              " 'refus': 508,\n",
              " 'crime': 509,\n",
              " 'saturday': 510,\n",
              " 'manag': 511,\n",
              " 'univers': 512,\n",
              " 'staff': 513,\n",
              " 'there': 514,\n",
              " 'behind': 515,\n",
              " 'tie': 516,\n",
              " 'citizen': 517,\n",
              " 'east': 518,\n",
              " 'this': 519,\n",
              " 'sanction': 520,\n",
              " 'israel': 521,\n",
              " 'https': 522,\n",
              " 'sure': 523,\n",
              " 'ago': 524,\n",
              " 'ryan': 525,\n",
              " 'best': 526,\n",
              " 'rate': 527,\n",
              " 'feel': 528,\n",
              " 'instead': 529,\n",
              " 'along': 530,\n",
              " 'lawyer': 531,\n",
              " 'refer': 532,\n",
              " 'file': 533,\n",
              " 'less': 534,\n",
              " 'civil': 535,\n",
              " 'larg': 536,\n",
              " 'evid': 537,\n",
              " 'earli': 538,\n",
              " 'reach': 539,\n",
              " 'promis': 540,\n",
              " 'complet': 541,\n",
              " 'but': 542,\n",
              " 'nomine': 543,\n",
              " 'market': 544,\n",
              " 'warn': 545,\n",
              " 'join': 546,\n",
              " 'toward': 547,\n",
              " 'terror': 548,\n",
              " 'carri': 549,\n",
              " 'immedi': 550,\n",
              " 'britain': 551,\n",
              " 'word': 552,\n",
              " 'juli': 553,\n",
              " 'abl': 554,\n",
              " 'address': 555,\n",
              " 'st': 556,\n",
              " 'spend': 557,\n",
              " 'racist': 558,\n",
              " 'council': 559,\n",
              " 'anyon': 560,\n",
              " 'document': 561,\n",
              " 'global': 562,\n",
              " 'debat': 563,\n",
              " 'threaten': 564,\n",
              " 'within': 565,\n",
              " 'approv': 566,\n",
              " 'demand': 567,\n",
              " 'cost': 568,\n",
              " 'due': 569,\n",
              " 'shoot': 570,\n",
              " 'januari': 571,\n",
              " 'crimin': 572,\n",
              " 'effect': 573,\n",
              " 'office': 574,\n",
              " 'level': 575,\n",
              " 'all': 576,\n",
              " 'note': 577,\n",
              " 'potenti': 578,\n",
              " 'review': 579,\n",
              " 'hour': 580,\n",
              " 'fear': 581,\n",
              " 'getti': 582,\n",
              " 'alli': 583,\n",
              " 'love': 584,\n",
              " 'someon': 585,\n",
              " 'messag': 586,\n",
              " 'describ': 587,\n",
              " 'turkey': 588,\n",
              " 'missil': 589,\n",
              " 'peac': 590,\n",
              " 'total': 591,\n",
              " 'd': 592,\n",
              " 'develop': 593,\n",
              " 'violenc': 594,\n",
              " 'water': 595,\n",
              " 'ralli': 596,\n",
              " 'thousand': 597,\n",
              " 'bush': 598,\n",
              " 'thought': 599,\n",
              " 'congression': 600,\n",
              " 'sexual': 601,\n",
              " 'challeng': 602,\n",
              " 'nomin': 603,\n",
              " 'late': 604,\n",
              " 'chairman': 605,\n",
              " 'advis': 606,\n",
              " 'hit': 607,\n",
              " 'sent': 608,\n",
              " 'is': 609,\n",
              " 'block': 610,\n",
              " 'direct': 611,\n",
              " 'diplomat': 612,\n",
              " 'to': 613,\n",
              " 'enforc': 614,\n",
              " 'young': 615,\n",
              " 'idea': 616,\n",
              " 'paul': 617,\n",
              " 'weapon': 618,\n",
              " 'entir': 619,\n",
              " 'cover': 620,\n",
              " 'six': 621,\n",
              " 'climat': 622,\n",
              " 'front': 623,\n",
              " 'access': 624,\n",
              " 'm': 625,\n",
              " 'again': 626,\n",
              " 'blame': 627,\n",
              " 'nearli': 628,\n",
              " 'fake': 629,\n",
              " 'daili': 630,\n",
              " 'oppos': 631,\n",
              " 'shot': 632,\n",
              " 'district': 633,\n",
              " 'negoti': 634,\n",
              " 'mark': 635,\n",
              " 'florida': 636,\n",
              " 'bad': 637,\n",
              " 'form': 638,\n",
              " 'friend': 639,\n",
              " 'remov': 640,\n",
              " 'individu': 641,\n",
              " 'pro': 642,\n",
              " 'latest': 643,\n",
              " 'gave': 644,\n",
              " 'april': 645,\n",
              " 'progress': 646,\n",
              " 'began': 647,\n",
              " 'outsid': 648,\n",
              " 'urg': 649,\n",
              " 'room': 650,\n",
              " 'detail': 651,\n",
              " 'accept': 652,\n",
              " 'parliament': 653,\n",
              " 'administration': 654,\n",
              " 'kind': 655,\n",
              " 'middl': 656,\n",
              " 'led': 657,\n",
              " 'whose': 658,\n",
              " 'movement': 659,\n",
              " 'august': 660,\n",
              " 'violat': 661,\n",
              " 'oil': 662,\n",
              " 'express': 663,\n",
              " 'key': 664,\n",
              " 'anyth': 665,\n",
              " 'tv': 666,\n",
              " 'lost': 667,\n",
              " 'central': 668,\n",
              " 'coalit': 669,\n",
              " 'firm': 670,\n",
              " 'stay': 671,\n",
              " 'son': 672,\n",
              " 'armi': 673,\n",
              " 'break': 674,\n",
              " 'conflict': 675,\n",
              " 'california': 676,\n",
              " 'no': 677,\n",
              " 'georg': 678,\n",
              " 'measur': 679,\n",
              " 'worker': 680,\n",
              " 'ahead': 681,\n",
              " 'moscow': 682,\n",
              " 'link': 683,\n",
              " 'near': 684,\n",
              " 'strong': 685,\n",
              " 'prevent': 686,\n",
              " 'associ': 687,\n",
              " 'launch': 688,\n",
              " 'futur': 689,\n",
              " 'explain': 690,\n",
              " 'lose': 691,\n",
              " 'added': 692,\n",
              " 'project': 693,\n",
              " 'non': 694,\n",
              " 'replac': 695,\n",
              " 'declar': 696,\n",
              " 'begin': 697,\n",
              " 'declin': 698,\n",
              " 'confer': 699,\n",
              " 'vice': 700,\n",
              " 'send': 701,\n",
              " 'victim': 702,\n",
              " 'network': 703,\n",
              " 'j': 704,\n",
              " 'understand': 705,\n",
              " 'constitut': 706,\n",
              " 'rather': 707,\n",
              " 'cite': 708,\n",
              " 'respect': 709,\n",
              " 'almost': 710,\n",
              " 'establish': 711,\n",
              " 'healthcar': 712,\n",
              " 'risk': 713,\n",
              " 'everyon': 714,\n",
              " 'cannot': 715,\n",
              " 'christian': 716,\n",
              " 'not': 717,\n",
              " 'he': 718,\n",
              " 'guy': 719,\n",
              " 'favor': 720,\n",
              " 'freedom': 721,\n",
              " 'board': 722,\n",
              " 'suspect': 723,\n",
              " 'western': 724,\n",
              " 'novemb': 725,\n",
              " 'answer': 726,\n",
              " 'often': 727,\n",
              " 'activ': 728,\n",
              " 'third': 729,\n",
              " 'session': 730,\n",
              " 'flynn': 731,\n",
              " 'met': 732,\n",
              " 'die': 733,\n",
              " 'myanmar': 734,\n",
              " 'pick': 735,\n",
              " 'photo': 736,\n",
              " 'texa': 737,\n",
              " 'test': 738,\n",
              " 'for': 739,\n",
              " 'british': 740,\n",
              " 'main': 741,\n",
              " 'korean': 742,\n",
              " 'prison': 743,\n",
              " 'tillerson': 744,\n",
              " 'insur': 745,\n",
              " 'decemb': 746,\n",
              " 'histori': 747,\n",
              " 'although': 748,\n",
              " 'small': 749,\n",
              " 'wrong': 750,\n",
              " 'appeal': 751,\n",
              " 'dollar': 752,\n",
              " 'conduct': 753,\n",
              " 'crowd': 754,\n",
              " 'enter': 755,\n",
              " 'secret': 756,\n",
              " 'chines': 757,\n",
              " 'milit': 758,\n",
              " 'capit': 759,\n",
              " 'limit': 760,\n",
              " 'h': 761,\n",
              " 'ministri': 762,\n",
              " 'industri': 763,\n",
              " 'truth': 764,\n",
              " 'low': 765,\n",
              " 'half': 766,\n",
              " 'town': 767,\n",
              " 'drop': 768,\n",
              " 'more': 769,\n",
              " 'attend': 770,\n",
              " 'wife': 771,\n",
              " 'danger': 772,\n",
              " 'thank': 773,\n",
              " 'jame': 774,\n",
              " 'true': 775,\n",
              " 'forward': 776,\n",
              " 'data': 777,\n",
              " 'pressur': 778,\n",
              " 'west': 779,\n",
              " 'alway': 780,\n",
              " 'father': 781,\n",
              " 'short': 782,\n",
              " 'probabl': 783,\n",
              " 'especi': 784,\n",
              " 'initi': 785,\n",
              " 'page': 786,\n",
              " 'me': 787,\n",
              " 'relationship': 788,\n",
              " 'resign': 789,\n",
              " 'reveal': 790,\n",
              " 'crisi': 791,\n",
              " 'either': 792,\n",
              " 'publish': 793,\n",
              " 'hack': 794,\n",
              " 'aim': 795,\n",
              " 'agent': 796,\n",
              " 'bomb': 797,\n",
              " 'murder': 798,\n",
              " 'activist': 799,\n",
              " 'emerg': 800,\n",
              " 'flag': 801,\n",
              " 'invest': 802,\n",
              " 'corrupt': 803,\n",
              " 'regard': 804,\n",
              " 'prosecutor': 805,\n",
              " 'octob': 806,\n",
              " 'destroy': 807,\n",
              " 'land': 808,\n",
              " 'popular': 809,\n",
              " 'phone': 810,\n",
              " 'german': 811,\n",
              " 'berni': 812,\n",
              " 'endors': 813,\n",
              " 'incom': 814,\n",
              " 'argu': 815,\n",
              " 'septemb': 816,\n",
              " 'paid': 817,\n",
              " 'committee': 818,\n",
              " 'mayor': 819,\n",
              " 'prepar': 820,\n",
              " 'if': 821,\n",
              " 'hundr': 822,\n",
              " 'at': 823,\n",
              " 'rubio': 824,\n",
              " 'v': 825,\n",
              " 'corpor': 826,\n",
              " 'reject': 827,\n",
              " 'nov': 828,\n",
              " 'cia': 829,\n",
              " 'expert': 830,\n",
              " 'biggest': 831,\n",
              " 'assault': 832,\n",
              " 'don': 833,\n",
              " 'class': 834,\n",
              " 'research': 835,\n",
              " 'centuri': 836,\n",
              " 'drug': 837,\n",
              " 'regul': 838,\n",
              " 'price': 839,\n",
              " 'resid': 840,\n",
              " 'spoke': 841,\n",
              " 'mike': 842,\n",
              " 'speaker': 843,\n",
              " 'david': 844,\n",
              " 'repeal': 845,\n",
              " 'seat': 846,\n",
              " 'african': 847,\n",
              " 'london': 848,\n",
              " 'februari': 849,\n",
              " 'ground': 850,\n",
              " 'soon': 851,\n",
              " 'michael': 852,\n",
              " 'deputi': 853,\n",
              " 'admit': 854,\n",
              " 'everyth': 855,\n",
              " 'ted': 856,\n",
              " 'car': 857,\n",
              " 'mass': 858,\n",
              " 'addit': 859,\n",
              " 'energi': 860,\n",
              " 'agenda': 861,\n",
              " 'robert': 862,\n",
              " 'opinion': 863,\n",
              " 'cooper': 864,\n",
              " 'largest': 865,\n",
              " 'do': 866,\n",
              " 'star': 867,\n",
              " 'leadership': 868,\n",
              " 'lower': 869,\n",
              " 'strike': 870,\n",
              " 'success': 871,\n",
              " 'moment': 872,\n",
              " 'remark': 873,\n",
              " 'grow': 874,\n",
              " 'l': 875,\n",
              " 'present': 876,\n",
              " 'brought': 877,\n",
              " 'bas': 878,\n",
              " 'benefit': 879,\n",
              " 'influenc': 880,\n",
              " 'ambassador': 881,\n",
              " 'food': 882,\n",
              " 'arabia': 883,\n",
              " 'merkel': 884,\n",
              " 'employe': 885,\n",
              " 'promot': 886,\n",
              " 'seriou': 887,\n",
              " 'book': 888,\n",
              " 'rise': 889,\n",
              " 'prove': 890,\n",
              " 'seven': 891,\n",
              " 'minor': 892,\n",
              " 'politician': 893,\n",
              " 'spent': 894,\n",
              " 'religi': 895,\n",
              " 'absolut': 896,\n",
              " 'stage': 897,\n",
              " 'becam': 898,\n",
              " 'appar': 899,\n",
              " 'course': 900,\n",
              " 'brief': 901,\n",
              " 'brexit': 902,\n",
              " 'insid': 903,\n",
              " 'design': 904,\n",
              " 'longer': 905,\n",
              " 'debt': 906,\n",
              " 'primari': 907,\n",
              " 'mccain': 908,\n",
              " 'germani': 909,\n",
              " 'obamacar': 910,\n",
              " 'pretti': 911,\n",
              " 'carolina': 912,\n",
              " 'down': 913,\n",
              " 'articl': 914,\n",
              " 'singl': 915,\n",
              " 'counti': 916,\n",
              " 'goe': 917,\n",
              " 'violent': 918,\n",
              " 'japan': 919,\n",
              " 'simpli': 920,\n",
              " 'economi': 921,\n",
              " 'mention': 922,\n",
              " 'togeth': 923,\n",
              " 'train': 924,\n",
              " 'captur': 925,\n",
              " 'avoid': 926,\n",
              " 'clearli': 927,\n",
              " 'probe': 928,\n",
              " 'product': 929,\n",
              " 'decad': 930,\n",
              " 'island': 931,\n",
              " 'similar': 932,\n",
              " 'ensur': 933,\n",
              " 'exist': 934,\n",
              " 'check': 935,\n",
              " 'wait': 936,\n",
              " 'colleg': 937,\n",
              " 'sit': 938,\n",
              " 'saw': 939,\n",
              " 'child': 940,\n",
              " 'then': 941,\n",
              " 'pull': 942,\n",
              " 'mind': 943,\n",
              " 'johnson': 944,\n",
              " 'commiss': 945,\n",
              " 'quit': 946,\n",
              " 'green': 947,\n",
              " 'heard': 948,\n",
              " 'produc': 949,\n",
              " 'transit': 950,\n",
              " 'radic': 951,\n",
              " 'c': 952,\n",
              " 'appoint': 953,\n",
              " 'mother': 954,\n",
              " 'kurdish': 955,\n",
              " 'below': 956,\n",
              " 'what': 957,\n",
              " 'impos': 958,\n",
              " 'medic': 959,\n",
              " 'buy': 960,\n",
              " 'financ': 961,\n",
              " 'northern': 962,\n",
              " 'iraqi': 963,\n",
              " 'self': 964,\n",
              " 'educ': 965,\n",
              " 'lawsuit': 966,\n",
              " 'repeatedli': 967,\n",
              " 'incid': 968,\n",
              " 'walk': 969,\n",
              " 'situat': 970,\n",
              " 'breitbart': 971,\n",
              " 'panel': 972,\n",
              " 'gain': 973,\n",
              " 'mainstream': 974,\n",
              " 'whole': 975,\n",
              " 'contact': 976,\n",
              " 'journalist': 977,\n",
              " 'san': 978,\n",
              " 'connect': 979,\n",
              " 'policy': 980,\n",
              " 'previous': 981,\n",
              " 'worri': 982,\n",
              " 'assist': 983,\n",
              " 'controversi': 984,\n",
              " 'donat': 985,\n",
              " 'f': 986,\n",
              " 'safe': 987,\n",
              " 'elector': 988,\n",
              " 'site': 989,\n",
              " 'rep': 990,\n",
              " 'drive': 991,\n",
              " 'rohingya': 992,\n",
              " 'fall': 993,\n",
              " 'separ': 994,\n",
              " 'determin': 995,\n",
              " 'trip': 996,\n",
              " 'ignor': 997,\n",
              " 'nato': 998,\n",
              " 'sex': 999,\n",
              " 'girl': 1000,\n",
              " ...}"
            ]
          },
          "metadata": {},
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "O próximo passo será transformar suas sentenças em listas de valores baseadas nesses tokens. Depois de tê-los, você provavelmente também precisará manipular essas listas, principalmente para fazer com que todas as frases tenham o mesmo tamanho, caso contrário, pode ser difícil treinar uma rede neural com eles. Lembre-se quando trabalhamos imagens, definimos uma camada de entrada com o tamanho da imagem que estamos alimentando na rede neural. Nos casos em que as imagens tivessem tamanhos diferentes, nós as redimensionávamos para caber. Bem, você vai enfrentar a mesma coisa com o texto. Felizmente, o TensorFlow inclui APIs para lidar com esses problemas."
      ],
      "metadata": {
        "id": "vi1UYD4T4WKF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# aplicando técnicas tokenizing e padding\n",
        "training_sequences = tokenizer.texts_to_sequences(np.array(train.features))"
      ],
      "metadata": {
        "id": "1W55uv8yySTm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Vamos começar criando uma lista de sequências, as frases codificadas com os tokens que geramos. Observe que todas as frases anteriores continham quatro palavras. Então esse é um pouco mais longo. Usaremos isso para demonstrar o preenchimento em um momento.\n",
        "\n",
        "O próximo pedaço de código é este, onde eu simplesmente chamo o tokenizer para obter textos para sequências, e ele os transformará em um conjunto de sequências. \n",
        "\n",
        "Por exemplo, se aplicarmos o tokenizer na frase \"I love my dog\", teremos o seguinte word_index: `{'my': 1, 'love': 2, 'dog': 3, 'i': 4}` e o `text_to_sequence` retornará a seguinte sequência: `[4, 2, 1, 3]`.\n"
      ],
      "metadata": {
        "id": "UCG3E-hg4qsR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "training_sequences[0:1]"
      ],
      "metadata": {
        "id": "OTAGu4u16K7D",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8fd8e911-ad2e-4a55-c08a-022181e5fbe6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[[113,\n",
              "  65,\n",
              "  57,\n",
              "  1494,\n",
              "  410,\n",
              "  346,\n",
              "  265,\n",
              "  4,\n",
              "  5,\n",
              "  7,\n",
              "  139,\n",
              "  56,\n",
              "  35,\n",
              "  4,\n",
              "  5,\n",
              "  7,\n",
              "  139,\n",
              "  3,\n",
              "  128,\n",
              "  10,\n",
              "  57,\n",
              "  181,\n",
              "  80,\n",
              "  260,\n",
              "  762,\n",
              "  451,\n",
              "  112,\n",
              "  65,\n",
              "  1579,\n",
              "  112,\n",
              "  796,\n",
              "  1494,\n",
              "  410,\n",
              "  346,\n",
              "  265,\n",
              "  649,\n",
              "  682,\n",
              "  47,\n",
              "  679,\n",
              "  3278,\n",
              "  203,\n",
              "  65,\n",
              "  7,\n",
              "  139,\n",
              "  1032,\n",
              "  4512,\n",
              "  5653,\n",
              "  3,\n",
              "  96,\n",
              "  128,\n",
              "  721,\n",
              "  7555,\n",
              "  50,\n",
              "  241,\n",
              "  65,\n",
              "  22,\n",
              "  73,\n",
              "  361,\n",
              "  1,\n",
              "  512,\n",
              "  371,\n",
              "  39,\n",
              "  2319,\n",
              "  80,\n",
              "  1022,\n",
              "  3403,\n",
              "  3,\n",
              "  80,\n",
              "  112,\n",
              "  796,\n",
              "  57,\n",
              "  47,\n",
              "  2442,\n",
              "  3653,\n",
              "  3063,\n",
              "  1,\n",
              "  5286,\n",
              "  292,\n",
              "  1939,\n",
              "  573,\n",
              "  2373,\n",
              "  694,\n",
              "  409,\n",
              "  302,\n",
              "  51,\n",
              "  57,\n",
              "  296,\n",
              "  8,\n",
              "  1377,\n",
              "  486,\n",
              "  510,\n",
              "  181,\n",
              "  682,\n",
              "  114,\n",
              "  112,\n",
              "  65,\n",
              "  1676,\n",
              "  15,\n",
              "  245,\n",
              "  113,\n",
              "  51,\n",
              "  112,\n",
              "  796,\n",
              "  1763,\n",
              "  183,\n",
              "  2493,\n",
              "  388,\n",
              "  2223,\n",
              "  80,\n",
              "  653,\n",
              "  42,\n",
              "  72,\n",
              "  29,\n",
              "  7,\n",
              "  435,\n",
              "  113,\n",
              "  7,\n",
              "  1178,\n",
              "  2554,\n",
              "  1474,\n",
              "  4,\n",
              "  5,\n",
              "  878,\n",
              "  2283,\n",
              "  206,\n",
              "  112,\n",
              "  796,\n",
              "  5653,\n",
              "  827,\n",
              "  3488,\n",
              "  4,\n",
              "  5,\n",
              "  113,\n",
              "  1927,\n",
              "  12,\n",
              "  36,\n",
              "  57,\n",
              "  760,\n",
              "  99,\n",
              "  231,\n",
              "  1101,\n",
              "  302,\n",
              "  1181,\n",
              "  1,\n",
              "  159,\n",
              "  115,\n",
              "  4793,\n",
              "  337,\n",
              "  80,\n",
              "  3111,\n",
              "  44,\n",
              "  13,\n",
              "  4,\n",
              "  5,\n",
              "  95,\n",
              "  31,\n",
              "  104,\n",
              "  100,\n",
              "  20,\n",
              "  2,\n",
              "  209,\n",
              "  4,\n",
              "  5,\n",
              "  259,\n",
              "  55,\n",
              "  211,\n",
              "  1529,\n",
              "  47,\n",
              "  113,\n",
              "  65,\n",
              "  880,\n",
              "  4,\n",
              "  5,\n",
              "  1278,\n",
              "  682,\n",
              "  438,\n",
              "  2050,\n",
              "  4,\n",
              "  5,\n",
              "  219,\n",
              "  1067,\n",
              "  112,\n",
              "  796,\n",
              "  57,\n",
              "  50,\n",
              "  65,\n",
              "  1579,\n",
              "  222,\n",
              "  1257,\n",
              "  1,\n",
              "  435,\n",
              "  30,\n",
              "  6257,\n",
              "  721,\n",
              "  241,\n",
              "  1960,\n",
              "  416,\n",
              "  80,\n",
              "  5653,\n",
              "  3,\n",
              "  153]]"
            ]
          },
          "metadata": {},
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Padding\n",
        "\n",
        "Uma vez que o tokenizer tenha criado as sequências, essas sequências podem ser passadas para sequências de preenchimento para que sejam preenchidas assim:\n",
        "\n",
        "Tokenizer Word_index:\n",
        "- `{'do': 8, 'you': 6, 'love': 3, 'i': 5, 'amazing': 11, 'my': 2, 'is': 10, 'think': 9,'dog': 4, '<00V>': 1, 'cat': 7} `\n",
        "\n",
        "Sequence:\n",
        "- `[[5, 3, 2, 4], [5, 3, 2, 7], [6, 3, 2, 4], [8, 6, 9, 2, 4, 10, 11]]`\n",
        "\n",
        "Padding:\n",
        "\n",
        "- `[[ 0 0 0 5 3 2 4]\n",
        "[ 0 0 0 5 3 2 7]\n",
        "[ 0 0 0 6 3 2 4]\n",
        "[ 8 6 9 2 4 10 11]]`\n",
        "\n",
        "Agora você pode ver que a lista de sentenças foi preenchida em uma matriz e que cada linha na matriz tem o mesmo comprimento. Ele conseguiu isso colocando o número apropriado de zeros antes da sentença. Então, no caso da sentença `[5, 3, 2, 4]`, ela realmente fez qualquer coisa. No caso da frase mais longa `[8, 6, 9, 2, 4, 10, 11]` aqui não precisou fazer nada.\n",
        "\n",
        "Muitas vezes você verá exemplos em que o preenchimento de zeros está após a frase e não antes, como você acabou de ver. Se você, como eu, estiver mais confortável com isso, você pode alterar o código para isso, adicionando o parâmetro `padding = post` como fizemos acima.\n",
        "\n",
        "Você deve ter notado que o número de colunas da matriz é de mesmo tamhanho da frase mais longa. Mas você pode substituir isso com o parâmetro `maxlen`. Por exemplo, se você quiser que suas frases tenham no máximo cinco palavras. Você pode dizer que `maxlen` é igual a cinco. Isso, obviamente, levará à pergunta:\n",
        "\n",
        "Se eu tiver frases mais longas que o comprimento máximo, perderei informações, mas de onde ? \n",
        "\n",
        "Assim como no preenchimento, o padrão é pré, o que significa que você perderá desde o início da frase. Se você quiser substituir isso para perder no final, poderá fazê-lo com o parâmetro `truncating`."
      ],
      "metadata": {
        "id": "3TwcQ0Qi7LT5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "training_padded = pad_sequences(training_sequences,truncating=trunc_type, padding=pad_type)"
      ],
      "metadata": {
        "id": "c0CO8gVs7KiN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "max_length = len(training_padded[0])"
      ],
      "metadata": {
        "id": "QR-oslMJCB0D"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "max_length"
      ],
      "metadata": {
        "id": "gqlynb3OCDI4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5cbb872a-fa1e-4985-d149-30de2d671a6a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5132"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "validation_sequences = tokenizer.texts_to_sequences(np.array(validation.features))\n",
        "validation_padded = pad_sequences(validation_sequences, padding=pad_type, truncating=trunc_type, maxlen = max_length)"
      ],
      "metadata": {
        "id": "gIRzXjqp5O6H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# criando arrays com inputs\n",
        "x_train = np.copy(training_padded)\n",
        "x_val = np.copy(validation_padded)\n",
        "y_train = train['Label'].values\n",
        "y_val = validation['Label'].values"
      ],
      "metadata": {
        "id": "dwDwpJNSTr86"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# veriificando o tamanho dos arrays\n",
        "print('Tamanho dos Arrays')\n",
        "print('Treino: ', len(x_train), 'linhas.')\n",
        "print('Resposta: ', len(y_train), 'linhas.')"
      ],
      "metadata": {
        "id": "l9Jz87OnTinT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1f785717-13b2-4d4d-f074-661d53a89139"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Tamanho dos Arrays\n",
            "Treino:  26938 linhas.\n",
            "Resposta:  26938 linhas.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Word Embeddings\n",
        "\n",
        "Até então, estavamos transformando texto em sequências de números aonde cada número da sequência está associado com uma palavra. Por exemplo, você pode representar a palavra TensorFlow com o valor nove e, em seguida, substituir cada instância da palavra por nove em uma sequência. Usando ferramentas e TensorFlow, você pode processar strings para obter índices de todas as palavras em um corpus de strings e depois converter as strings em matrizes de números. Este é o começo para tirar o sentimento de suas frases. Mas agora, ainda é apenas uma sequência de números representando palavras. Então, a partir daí, como alguém realmente obteria sentimento?\n",
        "\n",
        "Bem, isso é algo que pode ser aprendido a partir de um corpus de palavras da mesma forma que características foram extraídas de imagens. Esse processo é chamado de incorporação, com a ideia de que palavras e palavras associadas são agrupadas como vetores em um espaço multidimensional.\n",
        "\n",
        "\n",
        "Até agora, criamos uma instância de tokenizer, dando-lhe o tamanho do nosso vocabulário. Agora ajustaremos o tokenizer em nosso conjunto de dados de treinamento. Assim que tivermos nosso índice de palavras, podemos substituir as strings que contêm as palavras pelo valor do token que criamos para elas. As frases terão comprimentos variantes, dessa forma preenchemos e truncamos as sentenças sequenciadas até que elas tenham o mesmo tamanho, determinado pelo parâmetro de comprimento máximo. Então fizemos o mesmo para as sequências de teste e validação. Observe que o índice de palavras são palavras derivadas do conjunto de treinamento, portanto, você deve esperar ver muito mais tokens de vocabulário no exame de teste. \n",
        "\n",
        "Agora é hora de definir nossa rede neural. \n",
        "\n",
        "# Rede Neural"
      ],
      "metadata": {
        "id": "mIPsT0pqT5uZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "embedding_dim = 16\n",
        "\n",
        "model = tf.keras.Sequential([\n",
        "    tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
        "    tf.keras.layers.GlobalAveragePooling1D(),\n",
        "    tf.keras.layers.Dense(24, activation='relu'),\n",
        "    tf.keras.layers.Dense(1, activation='sigmoid')\n",
        "])\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "awe2IlchlgWD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(loss='binary_crossentropy',\n",
        "                      optimizer=tf.keras.optimizers.Adam(learning_rate = 0.001), \n",
        "                      metrics=['accuracy'])\n",
        "history = model.fit(x_train, y_train, verbose = 1, epochs = 5, validation_data = (x_val, y_val),\n",
        "                   callbacks=[tf.keras.callbacks.EarlyStopping('val_loss', patience=6)])"
      ],
      "metadata": {
        "id": "E16aT-8pDyNy",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "988caee4-45b5-49c8-a00f-7d1e43cbd9cf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "842/842 [==============================] - 11s 13ms/step - loss: 0.0228 - accuracy: 0.9958 - val_loss: 0.0312 - val_accuracy: 0.9929\n",
            "Epoch 2/5\n",
            "842/842 [==============================] - 10s 11ms/step - loss: 0.0158 - accuracy: 0.9974 - val_loss: 0.0259 - val_accuracy: 0.9944\n",
            "Epoch 3/5\n",
            "842/842 [==============================] - 10s 12ms/step - loss: 0.0114 - accuracy: 0.9980 - val_loss: 0.0261 - val_accuracy: 0.9947\n",
            "Epoch 4/5\n",
            "842/842 [==============================] - 10s 12ms/step - loss: 0.0090 - accuracy: 0.9987 - val_loss: 0.0228 - val_accuracy: 0.9950\n",
            "Epoch 5/5\n",
            "842/842 [==============================] - 10s 12ms/step - loss: 0.0066 - accuracy: 0.9990 - val_loss: 0.0194 - val_accuracy: 0.9961\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Isso deve parecer muito familiar agora, exceto talvez por esta linha embedding. Essa é a chave para a análise de sentimentos de texto no TensorFlow, e é aí que a mágica realmente acontece."
      ],
      "metadata": {
        "id": "P5dIkbZQL5h9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "def plot_graphs(history, string):\n",
        "  plt.plot(history.history[string])\n",
        "  plt.plot(history.history['val_'+string])\n",
        "  plt.xlabel(\"Epochs\")\n",
        "  plt.ylabel(string)\n",
        "  plt.legend([string, 'val_'+string])\n",
        "  plt.show()\n",
        "  \n",
        "plot_graphs(history, \"accuracy\")\n",
        "plot_graphs(history, \"loss\")"
      ],
      "metadata": {
        "id": "p2R3d-ngMBbn",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 541
        },
        "outputId": "833cee7b-18f6-49d5-c991-e9f5ef9f09bf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Fazendo Previsões no Conjunto de Teste\n"
      ],
      "metadata": {
        "id": "PRk4L_Q5VgfR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# tratando os dados de teste\n",
        "test_sequences = tokenizer.texts_to_sequences(np.array(test.features))\n",
        "test_padded = pad_sequences(test_sequences, padding=pad_type, truncating=trunc_type, maxlen = max_length)"
      ],
      "metadata": {
        "id": "gIywJwySVgMS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# fazendo previsões\n",
        "preds = np.round(model.predict(test_padded))"
      ],
      "metadata": {
        "id": "comWIdrPVmvF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# verificando a acurácia\n",
        "acc = np.sum(1 if i==j else 0 for i,j in zip(preds, test[\"Label\"].values)) / len(test)\n",
        "print(\"Accuracy: \", acc )"
      ],
      "metadata": {
        "id": "90m5ij9xVpaD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4903f87b-0779-421c-b70f-50179cc246af"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy:  0.9960801781737194\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "e = model.layers[0]\n",
        "weights = e.get_weights()[0]\n",
        "print(weights. shape) # shape: (vocab_size, embedding dim) shape: (vocab_size, embedding_dim)"
      ],
      "metadata": {
        "id": "0EsitVH7GaoF",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "aad122ed-be1b-48e2-9d86-2837a4319d13"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(10000, 16)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "reverse_word_index = dict([(value, key) for (key, value) in word_index.items()])"
      ],
      "metadata": {
        "id": "9Ehw0bN6GMUN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import io\n",
        "out_v = io.open('vecs.tsv', 'w', encoding='utf-8')\n",
        "out_m = io.open('meta.tsv', 'w', encoding='utf-8')\n",
        "for word_num in range(1, vocab_size):\n",
        "  word = reverse_word_index[word_num]\n",
        "  embeddings = weights[word_num]\n",
        "  out_m.write(word + \"\\n\")\n",
        "  out_v.write('\\t'.join([str (x) for x in embeddings]) + \"\\n\")\n",
        "out_v.close()\n",
        "out_m.close()"
      ],
      "metadata": {
        "id": "2rlEpsxzGDOb"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "try:\n",
        "  from google.colab import files\n",
        "except ImportError:\n",
        "  pass\n",
        "else:\n",
        "  files.download('vecs.tsv')\n",
        "  files.download('meta.tsv')"
      ],
      "metadata": {
        "id": "WQGnkyztFNsJ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        },
        "outputId": "d5fa9aa0-3c08-4050-bc8f-1b3a94141181"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "download(\"download_b9583c2d-1150-45ba-88e4-4fcc2f1f2877\", \"vecs.tsv\", 1819358)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "download(\"download_44422976-c356-416c-bb71-34b86b3cabcf\", \"meta.tsv\", 72325)"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Apendice"
      ],
      "metadata": {
        "id": "zD569r2KVNK_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## TF IDF\n",
        "\n",
        "Uma contagem de palavras e verifica as palavras que mais acontecem para fake news e palavras para noticias reais. "
      ],
      "metadata": {
        "id": "gOR7Y026T4Qd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X = news[\"features\"].values\n",
        "Y = news[\"Label\"].values\n",
        "\n",
        "#Converter de texto para numero\n",
        "vectorizer = TfidfVectorizer()\n",
        "vectorizer.fit(X)\n",
        "\n",
        "X = vectorizer.transform(X)"
      ],
      "metadata": {
        "id": "7MtmrbvGU4fP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size = 0.2, stratify = Y, random_state = 42)"
      ],
      "metadata": {
        "id": "7w-qnB03U-Cf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Treinando o Modelo"
      ],
      "metadata": {
        "id": "dHlccLk7VUly"
      }
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "b0-jgdTYVJ4T"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}