NLP_twitter_sentiment_analysis.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "## Importing Essentials"
      ],
      "metadata": {
        "id": "AVk38gKQI183"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VxgQj38d_TtF"
      },
      "outputs": [],
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import tensorflow as tf"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CQqahMHXAgOl"
      },
      "outputs": [],
      "source": [
        "# Tensorflow NLP kit(only so much for our needs)\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from tensorflow.keras.layers import Embedding, LSTM, Dense, Dropout, Bidirectional\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "i8_69-z7KDJ1"
      },
      "outputs": [],
      "source": [
        "# Classification metric\n",
        "from sklearn.metrics import classification_report"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yyk1m57kBBrU",
        "outputId": "5bf226c1-5541-4179-951e-06c1d83562eb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/stopwords.zip.\n",
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n"
          ]
        }
      ],
      "source": [
        "# NLTK NLP kit(only so much for our needs)\n",
        "import nltk\n",
        "nltk.download('stopwords')\n",
        "nltk.download('wordnet')\n",
        "from nltk.corpus import stopwords\n",
        "from nltk.stem.wordnet import WordNetLemmatizer\n",
        "# Visualization\n",
        "from wordcloud import WordCloud\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UjvvGsiT6VD6",
        "outputId": "4c65ebf3-0ed8-499c-a414-9d46f4a660fc"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "udxsLwOY-hU8"
      },
      "outputs": [],
      "source": [
        "import zipfile\n",
        "with zipfile.ZipFile('/content/drive/MyDrive/Datasets/trainingandtestdata.zip', 'r') as zip_ref:\n",
        "  zip_ref.extractall('/content')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iDcqixyi_Dq8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "2224fbb6-bef9-496a-a839-14499b03ff0b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   target         ids                          date      flag  \\\n",
              "0       0  1467810369  Mon Apr 06 22:19:45 PDT 2009  NO_QUERY   \n",
              "1       0  1467810672  Mon Apr 06 22:19:49 PDT 2009  NO_QUERY   \n",
              "2       0  1467810917  Mon Apr 06 22:19:53 PDT 2009  NO_QUERY   \n",
              "3       0  1467811184  Mon Apr 06 22:19:57 PDT 2009  NO_QUERY   \n",
              "4       0  1467811193  Mon Apr 06 22:19:57 PDT 2009  NO_QUERY   \n",
              "\n",
              "              user                                               text  \n",
              "0  _TheSpecialOne_  @switchfoot http://twitpic.com/2y1zl - Awww, t...  \n",
              "1    scotthamilton  is upset that he can't update his Facebook by ...  \n",
              "2         mattycus  @Kenichan I dived many times for the ball. Man...  \n",
              "3          ElleCTF    my whole body feels itchy and like its on fire   \n",
              "4           Karoli  @nationwideclass no, it's not behaving at all....  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d96ce49d-a5e2-462c-864a-2875affd0d45\">\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>target</th>\n",
              "      <th>ids</th>\n",
              "      <th>date</th>\n",
              "      <th>flag</th>\n",
              "      <th>user</th>\n",
              "      <th>text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>1467810369</td>\n",
              "      <td>Mon Apr 06 22:19:45 PDT 2009</td>\n",
              "      <td>NO_QUERY</td>\n",
              "      <td>_TheSpecialOne_</td>\n",
              "      <td>@switchfoot http://twitpic.com/2y1zl - Awww, t...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0</td>\n",
              "      <td>1467810672</td>\n",
              "      <td>Mon Apr 06 22:19:49 PDT 2009</td>\n",
              "      <td>NO_QUERY</td>\n",
              "      <td>scotthamilton</td>\n",
              "      <td>is upset that he can't update his Facebook by ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0</td>\n",
              "      <td>1467810917</td>\n",
              "      <td>Mon Apr 06 22:19:53 PDT 2009</td>\n",
              "      <td>NO_QUERY</td>\n",
              "      <td>mattycus</td>\n",
              "      <td>@Kenichan I dived many times for the ball. Man...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0</td>\n",
              "      <td>1467811184</td>\n",
              "      <td>Mon Apr 06 22:19:57 PDT 2009</td>\n",
              "      <td>NO_QUERY</td>\n",
              "      <td>ElleCTF</td>\n",
              "      <td>my whole body feels itchy and like its on fire</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0</td>\n",
              "      <td>1467811193</td>\n",
              "      <td>Mon Apr 06 22:19:57 PDT 2009</td>\n",
              "      <td>NO_QUERY</td>\n",
              "      <td>Karoli</td>\n",
              "      <td>@nationwideclass no, it's not behaving at all....</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d96ce49d-a5e2-462c-864a-2875affd0d45')\"\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-d96ce49d-a5e2-462c-864a-2875affd0d45 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-d96ce49d-a5e2-462c-864a-2875affd0d45');\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": 7
        }
      ],
      "source": [
        "columns = [\"target\", \"ids\", \"date\", \"flag\", \"user\", \"text\"]\n",
        "training_data = pd.read_csv('/content/training.1600000.processed.noemoticon.csv', encoding='latin_1', names=columns)\n",
        "testing_data = pd.read_csv('/content/testdata.manual.2009.06.14.csv', encoding='latin_1', names=columns)\n",
        "training_data.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L8DnOqZSV0EB"
      },
      "source": [
        "## Cleaning the data/ Preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LP-0PK1NWLKz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7649c60a-aef5-44bb-fbb7-1d02fe460044"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', \"you're\", \"you've\", \"you'll\", \"you'd\", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', \"she's\", 'her', 'hers', 'herself', 'it', \"it's\", 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', \"that'll\", 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', \"don't\", 'should', \"should've\", 'now', 'd', 'll', 'm', 'o', 're', 've', 'y', 'ain', 'aren', \"aren't\", 'couldn', \"couldn't\", 'didn', \"didn't\", 'doesn', \"doesn't\", 'hadn', \"hadn't\", 'hasn', \"hasn't\", 'haven', \"haven't\", 'isn', \"isn't\", 'ma', 'mightn', \"mightn't\", 'mustn', \"mustn't\", 'needn', \"needn't\", 'shan', \"shan't\", 'shouldn', \"shouldn't\", 'wasn', \"wasn't\", 'weren', \"weren't\", 'won', \"won't\", 'wouldn', \"wouldn't\"]\n"
          ]
        }
      ],
      "source": [
        "print(stopwords.words('english'))\n",
        "emoticons = [\":)\", ':-)', ': )', ':D', '=)', ':(', ':-(', ': (']\n",
        "stpwrd = nltk.corpus.stopwords.words('english')\n",
        "stpwrd.extend(emoticons)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gpTGgw1XGVs7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "227ffcd8-14b6-4e33-c631-76478f532110"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', \"you're\", \"you've\", \"you'll\", \"you'd\", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', \"she's\", 'her', 'hers', 'herself', 'it', \"it's\", 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', \"that'll\", 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', \"don't\", 'should', \"should've\", 'now', 'd', 'll', 'm', 'o', 're', 've', 'y', 'ain', 'aren', \"aren't\", 'couldn', \"couldn't\", 'didn', \"didn't\", 'doesn', \"doesn't\", 'hadn', \"hadn't\", 'hasn', \"hasn't\", 'haven', \"haven't\", 'isn', \"isn't\", 'ma', 'mightn', \"mightn't\", 'mustn', \"mustn't\", 'needn', \"needn't\", 'shan', \"shan't\", 'shouldn', \"shouldn't\", 'wasn', \"wasn't\", 'weren', \"weren't\", 'won', \"won't\", 'wouldn', \"wouldn't\", ':)', ':-)', ': )', ':D', '=)', ':(', ':-(', ': (']\n"
          ]
        }
      ],
      "source": [
        "print(stpwrd)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ROlAww5GnBos"
      },
      "outputs": [],
      "source": [
        "import re # import regex\n",
        "\n",
        "def preprocess_data(data):\n",
        "  data = data.drop_duplicates(subset=['text'], keep='first')\n",
        "\n",
        "  # Remove emojis from data\n",
        "  data = data.astype(str).apply(lambda x: x.str.encode('ascii', 'ignore').str.decode('ascii'))\n",
        "\n",
        "  # Remove emoticons as stopwords so it will happen when tokenizing our sentences hence not needed here\n",
        "\n",
        "  # Remove urls\n",
        "  def remove_username_link_nums(text):\n",
        "    text = re.sub('@[^\\s]+', '', text)\n",
        "    text = re.sub('http[^\\s]+', '', text)\n",
        "    text = re.sub(r'\\d+', '', text)\n",
        "    return text\n",
        "\n",
        "  data['text'] = data['text'].apply(remove_username_link_nums)\n",
        "\n",
        "  # # turn text to lowercase\n",
        "  data['text'] = data['text'].map(lambda x: x.lower())\n",
        "\n",
        "  # Remove stopwords and lemmatize the text\n",
        "  lem = WordNetLemmatizer()\n",
        "  def remove_stopwords(text): \n",
        "    tokens = []\n",
        "    for token in text.split():\n",
        "      if token not in stpwrd:\n",
        "        tokens.append(lem.lemmatize(token, \"v\"))\n",
        " \n",
        "    return  \" \".join(tokens)\n",
        "\n",
        "  data.text = data.text.apply(lambda x: remove_stopwords(x))\n",
        "\n",
        "\n",
        "  return data\n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Analysing our main data\n",
        "It is essential that we take a look at the data we're finally going to use for building our model.\n",
        "We'll also divide our data into testing and training data without wasting any more time."
      ],
      "metadata": {
        "id": "XpjBypPQJHj9"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ASH7B9MT08hA"
      },
      "outputs": [],
      "source": [
        "train_neg = training_data[:10000]\n",
        "train_pos = training_data[1590000:]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "UPnl3KkDBJZn"
      },
      "outputs": [],
      "source": [
        "test_neg = training_data[10000:12000]\n",
        "test_pos = training_data[1588000:1590000]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "90EWCJwV1R2n"
      },
      "outputs": [],
      "source": [
        "test_neg = preprocess_data(test_neg.drop(['ids', 'date', 'flag', 'user'], axis=1))\n",
        "test_pos = preprocess_data(test_pos.drop(['ids', 'date', 'flag', 'user'], axis=1))\n",
        "train_neg = preprocess_data(train_neg.drop(['ids', 'date', 'flag', 'user'], axis=1))\n",
        "train_pos = preprocess_data(train_pos.drop(['ids', 'date', 'flag', 'user'], axis=1))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HwwTf3f_4aKm",
        "outputId": "9c35c9e2-9b79-4e4a-8212-1a905dfb970e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10000, 6)"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ],
      "source": [
        "train_neg.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PgFQko981GW3"
      },
      "outputs": [],
      "source": [
        "train_examples =pd.concat([train_neg, train_pos])\n",
        "test_examples =pd.concat([test_neg, test_pos])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7eU6G1Cj4khg",
        "outputId": "ad71b757-54cb-4c65-8726-7a0717aec314"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(19955, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ],
      "source": [
        "train_examples.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "h38VOPWl3RIK",
        "outputId": "8c786e07-20b6-471a-8155-72f20dab37ac"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(3996, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ],
      "source": [
        "test_examples.shape"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Fix : The following code would run without any error but with new update of nltk, wordprocessng not done without omw\n",
        "nltk.download('omw-1.4') "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "g89qR-qOOyfn",
        "outputId": "878b251a-1e27-44a3-aaec-36752f9098a5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[nltk_data] Downloading package omw-1.4 to /root/nltk_data...\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "pIihkI5sHYLp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "3c6c8d8f-be11-4024-f31f-8038b7dd4db1"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  target                                               text\n",
              "0      0  - awww, that's bummer. shoulda get david carr ...\n",
              "1      0  upset can't update facebook texting it... migh...\n",
              "2      0   dive many time ball. manage save % rest go bound\n",
              "3      0                    whole body feel itchy like fire\n",
              "4      0    no, behave all. i'm mad. here? can't see there."
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f05a4bab-6c40-40ec-a158-c374236fce08\">\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>target</th>\n",
              "      <th>text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>- awww, that's bummer. shoulda get david carr ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0</td>\n",
              "      <td>upset can't update facebook texting it... migh...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0</td>\n",
              "      <td>dive many time ball. manage save % rest go bound</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0</td>\n",
              "      <td>whole body feel itchy like fire</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0</td>\n",
              "      <td>no, behave all. i'm mad. here? can't see there.</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f05a4bab-6c40-40ec-a158-c374236fce08')\"\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-f05a4bab-6c40-40ec-a158-c374236fce08 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-f05a4bab-6c40-40ec-a158-c374236fce08');\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": 18
        }
      ],
      "source": [
        "train_data = preprocess_data(training_data.drop(['ids', 'date', 'flag', 'user'], axis=1))\n",
        "test_data = preprocess_data(testing_data.drop(['ids', 'date', 'flag', 'user'], axis=1))\n",
        "\n",
        "train_data.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DQq1uVp0RU0g",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "873d8406-695a-48a2-87ca-507a016958dc"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "target    0\n",
              "text      0\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "# taking care of missing data\n",
        "train_data.isna().sum().sort_values()\n",
        "# since no data is missing, let's move on with the workflow"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Y8Mbozl_2EY1",
        "outputId": "9f5e474d-cd2b-4133-90f3-11d647db69f7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "target\n",
            "0    9976\n",
            "4    9979\n",
            "dtype: int64\n"
          ]
        }
      ],
      "source": [
        "print(train_examples.groupby('target').size())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eFfS6ymDc-vM",
        "outputId": "bbb2b89c-07e0-4146-88b8-22503477ed5e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "target\n",
            "0    1996\n",
            "4    2000\n",
            "dtype: int64\n"
          ]
        }
      ],
      "source": [
        "print(test_examples.groupby('target').size())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_JG-G_9i3ET_",
        "outputId": "dddc1aef-d331-418c-ba79-9736d25d40a7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "target\n",
            "0    1996\n",
            "4    2000\n",
            "dtype: int64\n"
          ]
        }
      ],
      "source": [
        "print(test_examples.groupby('target').size())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "GR73W38ieg8J",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "23247092-07d3-49a3-eca4-23eda0b328eb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Shape of training data : (1581466, 2)\n"
          ]
        }
      ],
      "source": [
        "print(f\"Shape of training data : {train_data.shape}\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_examples.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Clsf-_kNP4wt",
        "outputId": "358493c1-c996-4d86-ef80-7949fbc61e36"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(19955, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "LpP_lBJeQXmC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Start tokenizing our tweets!\n",
        "We're done organizing our data.\n",
        "The next task is to tokenize our data to make to \"assign meaning to words\".\n",
        "Tokenization is where the actual nlp begins i.e. step no. 1."
      ],
      "metadata": {
        "id": "Q9Jr-GZXJ0fk"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NlFI4nS3Xo5A"
      },
      "outputs": [],
      "source": [
        "tokenizer = Tokenizer()\n",
        "tokenizer.fit_on_texts(train_examples.text)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8_Kzu81mcR97"
      },
      "outputs": [],
      "source": [
        "vocab_size = len(tokenizer.word_index) + 1 \n",
        "embedding_dim = 16 \n",
        "max_length = 50 \n",
        "trunc_type='post'  \n",
        "padding_type='post' \n",
        "oov_tok = \"<OOV>\" \n",
        "training_portion = .8"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uHIb423CYAYb"
      },
      "outputs": [],
      "source": [
        "train_sequences = tokenizer.texts_to_sequences(train_examples.text)\n",
        "test_sequences = tokenizer.texts_to_sequences(test_examples.text)\n",
        "\n",
        "X_train = pad_sequences(train_sequences, maxlen = max_length, padding = padding_type)\n",
        "\n",
        "X_test = pad_sequences(test_sequences, maxlen = max_length, padding = padding_type)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9zM8qvH3lKT-"
      },
      "outputs": [],
      "source": [
        "word_index = tokenizer.word_index"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MAT4ten57IJX",
        "outputId": "9beee999-5345-4126-bd47-e041e5b4e966"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "18991"
            ]
          },
          "metadata": {},
          "execution_count": 35
        }
      ],
      "source": [
        "len(word_index)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W4i2HsjniOc9",
        "outputId": "0f4f182d-8990-408b-b80f-15d975d2deff"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(19955, 50)"
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ],
      "source": [
        "X_train.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "-yeJgWFY051Z",
        "outputId": "30d5edd6-f669-4591-c68f-c9dfb6d96f34"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  target                                               text\n",
              "0      4  loooooooovvvvvveee kindle. dx cool, fantastic ...\n",
              "1      4    read kindle... love it... lee childs good read.\n",
              "2      4    ok, first assesment #kindle ...it fuck rocks!!!\n",
              "3      4  love kindle. i've mine months never look back....\n",
              "4      4                  fair enough. kindle think perfect"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-7ebd03da-41fe-469a-9d17-3ea784a41169\">\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>target</th>\n",
              "      <th>text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>4</td>\n",
              "      <td>loooooooovvvvvveee kindle. dx cool, fantastic ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>4</td>\n",
              "      <td>read kindle... love it... lee childs good read.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>4</td>\n",
              "      <td>ok, first assesment #kindle ...it fuck rocks!!!</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>love kindle. i've mine months never look back....</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>fair enough. kindle think perfect</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-7ebd03da-41fe-469a-9d17-3ea784a41169')\"\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-7ebd03da-41fe-469a-9d17-3ea784a41169 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-7ebd03da-41fe-469a-9d17-3ea784a41169');\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": 37
        }
      ],
      "source": [
        "test_data.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "MQFgH5etcG98"
      },
      "outputs": [],
      "source": [
        "y_train = train_examples.target.replace({'0': 0, '4' : 1}).values\n",
        "y_test = test_examples.target.replace({'0': 0, '4' : 1, '2':1}).values"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ITBWhQbOxKpI",
        "outputId": "da82abe2-9a54-4f67-92f2-c1a808012d65"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dtype('int64')"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ],
      "source": [
        "y_test.dtype"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0VV2M1nfcFOY"
      },
      "outputs": [],
      "source": [
        "import pickle\n",
        "with open('tokenizer.pickle', 'wb') as handle:\n",
        "    pickle.dump(tokenizer, handle, protocol=pickle.HIGHEST_PROTOCOL)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Word2Vec\n",
        "Word2vec is a technique for NLP. This algorithm uses a neural network model to learn word associations. This may also be called word embedding in various texts.\n",
        "For our purpose and to make the project easy, we will make use of Gensim, which will help us to develop word embeddings."
      ],
      "metadata": {
        "id": "8OCFk0XPKwmx"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_1fR8D57nbNY"
      },
      "outputs": [],
      "source": [
        "# Word2Vec\n",
        "import gensim\n",
        "\n",
        "# WORD2VEC \n",
        "W2V_SIZE = 228\n",
        "W2V_WINDOW = 7\n",
        "W2V_EPOCH = 10\n",
        "W2V_MIN_COUNT = 10\n",
        "\n",
        "documents = [_text.split() for _text in train_examples.text] \n",
        "w2v_model = gensim.models.word2vec.Word2Vec(size=W2V_SIZE, \n",
        "                                            window=W2V_WINDOW, \n",
        "                                            min_count=W2V_MIN_COUNT, \n",
        "                                            workers=8)\n",
        "w2v_model.build_vocab(documents)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6mlRvQPFnvRG",
        "outputId": "c0bf05ae-65cb-4ddf-9399-f47af1facde4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Vocab size 1868\n"
          ]
        }
      ],
      "source": [
        "words = w2v_model.wv.vocab.keys()\n",
        "vocab_size = len(words)\n",
        "print(\"Vocab size\", vocab_size)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Hc1IgyAmn4bG",
        "outputId": "c12be09f-4bcb-4720-fad4-9aa5cbd6ef28"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(870067, 1491990)"
            ]
          },
          "metadata": {},
          "execution_count": 44
        }
      ],
      "source": [
        "w2v_model.train(documents, total_examples=len(documents), epochs=W2V_EPOCH)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FBxBNkIYn9u1",
        "outputId": "cc73ad28-7fd4-4462-85f2-383d04006955"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-49-f1c7c3b5f77d>:1: DeprecationWarning: Call to deprecated `most_similar` (Method will be removed in 4.0.0, use self.wv.most_similar() instead).\n",
            "  w2v_model.most_similar(\"laugh\")\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('here,', 0.9996582269668579),\n",
              " ('people.', 0.9996448755264282),\n",
              " ('smile', 0.9996345639228821),\n",
              " ('is!', 0.9996265172958374),\n",
              " ('skin', 0.9996176958084106),\n",
              " ('&quot;', 0.9996099472045898),\n",
              " ('wow!', 0.9996069669723511),\n",
              " ('good,', 0.9995885491371155),\n",
              " ('is,', 0.9995819330215454),\n",
              " ('agree', 0.9995812773704529)]"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ],
      "source": [
        "w2v_model.most_similar(\"laugh\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vwoLnRZRoP-N",
        "outputId": "7900e0de-2d50-4002-e2f5-1480f36f1c4c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1869, 228)\n"
          ]
        }
      ],
      "source": [
        "# Embedding matrix for the embedding layer\n",
        "embedding_matrix = np.zeros((vocab_size+1, W2V_SIZE))\n",
        "for word, i in tokenizer.word_index.items():\n",
        "    if word in w2v_model.wv:\n",
        "        embedding_matrix[i] = w2v_model.wv[word]\n",
        "print(embedding_matrix.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Building our NLP model using LSTM\n",
        "LSTM (Long short-term memory) are special kinds of RNN, these are  designed to have a long-term memory making them capable of understanding the context better.\n",
        "Will these help us understand the context or sentiment of our tweets better, or not?\n",
        "There are previously many models built using different algorithms; let's see if using LSTMs make a staggering difference."
      ],
      "metadata": {
        "id": "37x0d7HzLmUr"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HUDxmUbMrMxo",
        "outputId": "888e50a2-e0e1-466f-89b3-9356ed1770da"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"Bidirectional_LSTM_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_1 (Embedding)     (None, 50, 228)           426132    \n",
            "                                                                 \n",
            " bidirectional_2 (Bidirectio  (None, 50, 256)          365568    \n",
            " nal)                                                            \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 50, 256)           0         \n",
            "                                                                 \n",
            " bidirectional_3 (Bidirectio  (None, 256)              394240    \n",
            " nal)                                                            \n",
            "                                                                 \n",
            " dropout_3 (Dropout)         (None, 256)               0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 64)                16448     \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 1)                 65        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 1,202,453\n",
            "Trainable params: 776,321\n",
            "Non-trainable params: 426,132\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model_1 = tf.keras.models.Sequential([\n",
        "                                      tf.keras.layers.Embedding(vocab_size+1, W2V_SIZE, weights=[embedding_matrix], input_length=max_length, trainable=False),\n",
        "                                      tf.keras.layers.Bidirectional(LSTM(128, return_sequences=True)),\n",
        "                                      tf.keras.layers.Dropout(0.2),\n",
        "                                      tf.keras.layers.Bidirectional(LSTM(128)),\n",
        "                                      tf.keras.layers.Dropout(0.2),\n",
        "                                      tf.keras.layers.Dense(64, activation='relu'),\n",
        "                                      tf.keras.layers.Dense(1, activation='sigmoid')\n",
        "                                                           \n",
        "], name='Bidirectional_LSTM_1')\n",
        "\n",
        "model_1.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZFu7QD9fwx9Y",
        "outputId": "2389f50a-39bf-404f-fef2-806c76b5f316"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "target    object\n",
              "text      object\n",
              "dtype: object"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ],
      "source": [
        "test_data.dtypes"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4-NdYlZfufvZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8c1eede0-02bc-405f-bc89-4cbd5490ba59"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "624/624 [==============================] - 24s 20ms/step - loss: 0.6748 - accuracy: 0.5726 - val_loss: 0.6593 - val_accuracy: 0.5891\n",
            "Epoch 2/50\n",
            "624/624 [==============================] - 13s 20ms/step - loss: 0.6514 - accuracy: 0.6116 - val_loss: 0.6492 - val_accuracy: 0.6054\n",
            "Epoch 3/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6436 - accuracy: 0.6234 - val_loss: 0.6429 - val_accuracy: 0.6161\n",
            "Epoch 4/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6400 - accuracy: 0.6256 - val_loss: 0.6470 - val_accuracy: 0.6044\n",
            "Epoch 5/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6365 - accuracy: 0.6264 - val_loss: 0.6317 - val_accuracy: 0.6261\n",
            "Epoch 6/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6309 - accuracy: 0.6343 - val_loss: 0.6296 - val_accuracy: 0.6289\n",
            "Epoch 7/50\n",
            "624/624 [==============================] - 12s 20ms/step - loss: 0.6320 - accuracy: 0.6333 - val_loss: 0.6310 - val_accuracy: 0.6339\n",
            "Epoch 8/50\n",
            "624/624 [==============================] - 13s 20ms/step - loss: 0.6274 - accuracy: 0.6353 - val_loss: 0.6561 - val_accuracy: 0.6044\n",
            "Epoch 9/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6259 - accuracy: 0.6390 - val_loss: 0.6264 - val_accuracy: 0.6309\n",
            "Epoch 10/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6231 - accuracy: 0.6411 - val_loss: 0.6248 - val_accuracy: 0.6299\n",
            "Epoch 11/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6206 - accuracy: 0.6419 - val_loss: 0.6284 - val_accuracy: 0.6246\n",
            "Epoch 12/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6172 - accuracy: 0.6463 - val_loss: 0.6275 - val_accuracy: 0.6276\n",
            "Epoch 13/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6173 - accuracy: 0.6460 - val_loss: 0.6261 - val_accuracy: 0.6296\n",
            "Epoch 14/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6142 - accuracy: 0.6464 - val_loss: 0.6218 - val_accuracy: 0.6326\n",
            "Epoch 15/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6136 - accuracy: 0.6523 - val_loss: 0.6212 - val_accuracy: 0.6426\n",
            "Epoch 16/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6109 - accuracy: 0.6541 - val_loss: 0.6229 - val_accuracy: 0.6431\n",
            "Epoch 17/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6088 - accuracy: 0.6511 - val_loss: 0.6575 - val_accuracy: 0.6001\n",
            "Epoch 18/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.6077 - accuracy: 0.6558 - val_loss: 0.6213 - val_accuracy: 0.6419\n",
            "Epoch 19/50\n",
            "624/624 [==============================] - 13s 20ms/step - loss: 0.6036 - accuracy: 0.6558 - val_loss: 0.6238 - val_accuracy: 0.6341\n",
            "Epoch 20/50\n",
            "624/624 [==============================] - 15s 24ms/step - loss: 0.6025 - accuracy: 0.6567 - val_loss: 0.6226 - val_accuracy: 0.6391\n",
            "Epoch 21/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5975 - accuracy: 0.6621 - val_loss: 0.6220 - val_accuracy: 0.6391\n",
            "Epoch 22/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5939 - accuracy: 0.6656 - val_loss: 0.6302 - val_accuracy: 0.6431\n",
            "Epoch 23/50\n",
            "624/624 [==============================] - 12s 19ms/step - loss: 0.5865 - accuracy: 0.6720 - val_loss: 0.6378 - val_accuracy: 0.6244\n",
            "Epoch 24/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5853 - accuracy: 0.6731 - val_loss: 0.6271 - val_accuracy: 0.6476\n",
            "Epoch 25/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5787 - accuracy: 0.6755 - val_loss: 0.6329 - val_accuracy: 0.6419\n",
            "Epoch 26/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5710 - accuracy: 0.6852 - val_loss: 0.6422 - val_accuracy: 0.6319\n",
            "Epoch 27/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5630 - accuracy: 0.6928 - val_loss: 0.6473 - val_accuracy: 0.6366\n",
            "Epoch 28/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5574 - accuracy: 0.6957 - val_loss: 0.6604 - val_accuracy: 0.6344\n",
            "Epoch 29/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5480 - accuracy: 0.7014 - val_loss: 0.6608 - val_accuracy: 0.6331\n",
            "Epoch 30/50\n",
            "624/624 [==============================] - 14s 22ms/step - loss: 0.5410 - accuracy: 0.7076 - val_loss: 0.6589 - val_accuracy: 0.6331\n",
            "Epoch 31/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.5303 - accuracy: 0.7177 - val_loss: 0.6741 - val_accuracy: 0.6421\n",
            "Epoch 32/50\n",
            "624/624 [==============================] - 12s 20ms/step - loss: 0.5183 - accuracy: 0.7240 - val_loss: 0.6661 - val_accuracy: 0.6404\n",
            "Epoch 33/50\n",
            "624/624 [==============================] - 13s 21ms/step - loss: 0.5074 - accuracy: 0.7305 - val_loss: 0.6950 - val_accuracy: 0.6379\n",
            "Epoch 34/50\n",
            "624/624 [==============================] - 14s 22ms/step - loss: 0.4973 - accuracy: 0.7364 - val_loss: 0.7381 - val_accuracy: 0.6316\n",
            "Epoch 35/50\n",
            "624/624 [==============================] - 14s 23ms/step - loss: 0.4828 - accuracy: 0.7444 - val_loss: 0.7809 - val_accuracy: 0.6379\n",
            "Epoch 36/50\n",
            "624/624 [==============================] - 15s 24ms/step - loss: 0.4754 - accuracy: 0.7510 - val_loss: 0.7546 - val_accuracy: 0.6304\n",
            "Epoch 37/50\n",
            "624/624 [==============================] - 15s 24ms/step - loss: 0.4649 - accuracy: 0.7568 - val_loss: 0.7543 - val_accuracy: 0.6269\n",
            "Epoch 38/50\n",
            "624/624 [==============================] - 15s 23ms/step - loss: 0.4496 - accuracy: 0.7678 - val_loss: 0.7942 - val_accuracy: 0.6161\n",
            "Epoch 39/50\n",
            "624/624 [==============================] - 15s 23ms/step - loss: 0.4401 - accuracy: 0.7697 - val_loss: 0.8064 - val_accuracy: 0.6331\n",
            "Epoch 40/50\n",
            "624/624 [==============================] - 15s 23ms/step - loss: 0.4277 - accuracy: 0.7786 - val_loss: 0.8194 - val_accuracy: 0.6279\n",
            "Epoch 41/50\n",
            "624/624 [==============================] - 13s 21ms/step - loss: 0.4089 - accuracy: 0.7913 - val_loss: 0.9051 - val_accuracy: 0.6239\n",
            "Epoch 42/50\n",
            "624/624 [==============================] - 15s 24ms/step - loss: 0.4105 - accuracy: 0.7883 - val_loss: 0.8946 - val_accuracy: 0.6316\n",
            "Epoch 43/50\n",
            "624/624 [==============================] - 17s 27ms/step - loss: 0.3993 - accuracy: 0.7959 - val_loss: 0.8943 - val_accuracy: 0.6274\n",
            "Epoch 44/50\n",
            "624/624 [==============================] - 16s 26ms/step - loss: 0.3828 - accuracy: 0.8018 - val_loss: 0.9307 - val_accuracy: 0.6369\n",
            "Epoch 45/50\n",
            "624/624 [==============================] - 15s 24ms/step - loss: 0.3798 - accuracy: 0.8041 - val_loss: 0.9526 - val_accuracy: 0.6184\n",
            "Epoch 46/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.3695 - accuracy: 0.8127 - val_loss: 0.9532 - val_accuracy: 0.6314\n",
            "Epoch 47/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.3584 - accuracy: 0.8205 - val_loss: 0.9781 - val_accuracy: 0.6286\n",
            "Epoch 48/50\n",
            "624/624 [==============================] - 11s 18ms/step - loss: 0.3514 - accuracy: 0.8244 - val_loss: 1.0337 - val_accuracy: 0.6286\n",
            "Epoch 49/50\n",
            "624/624 [==============================] - 12s 19ms/step - loss: 0.3428 - accuracy: 0.8266 - val_loss: 1.0588 - val_accuracy: 0.6336\n",
            "Epoch 50/50\n",
            "624/624 [==============================] - 12s 19ms/step - loss: 0.3380 - accuracy: 0.8334 - val_loss: 1.1120 - val_accuracy: 0.6216\n"
          ]
        }
      ],
      "source": [
        "model_1.compile(loss='binary_crossentropy',\n",
        "              optimizer=\"adam\",\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "history = model_1.fit(X_train, y_train, batch_size = 32, epochs=50, validation_data=(X_test, y_test), verbose=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lpuXACr73lnw",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6aeac90c-b6f1-4d67-db86-7c967abb0bc3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "125/125 [==============================] - 2s 8ms/step\n"
          ]
        }
      ],
      "source": [
        "y_pred = model_1.predict(X_test)\n",
        "y_pred = np.where(y_pred>0.5, 1, 0)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Computational limitations\n",
        "This project was built with zero monetary investment. The model performed with over an accuracy of 83% which is not much but the sample size taken was very small and the complexity of the RNN was very simple to make this model work fast. Training time ~ 10 mins for word vector, and RNN combined. This could have performed better with a better GPU and memory investment."
      ],
      "metadata": {
        "id": "15mCJDOdS_TB"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7UVFDxb63rZQ",
        "outputId": "f6cacb3e-4b86-43d3-c637-554369064c50"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.64      0.55      0.59      1996\n",
            "           1       0.61      0.69      0.65      2000\n",
            "\n",
            "    accuracy                           0.62      3996\n",
            "   macro avg       0.62      0.62      0.62      3996\n",
            "weighted avg       0.62      0.62      0.62      3996\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(classification_report(y_test, y_pred))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kuCzKhdE4GqI"
      },
      "outputs": [],
      "source": [
        "model_1.save('Sentiment_LSTM_model_better.h5')\n",
        "with open('trainHistoryDict', 'wb') as file_pi:\n",
        "    pickle.dump(history.history, file_pi)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oqqai8J84u64"
      },
      "outputs": [],
      "source": [
        "from keras.models import load_model\n",
        "model_1 = load_model('Sentiment_LSTM_model_better.h5')\n",
        "# loading tokenizer\n",
        "with open('trainHistoryDict', 'rb') as file_pi:\n",
        "    history = pickle.load(file_pi)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iV5gaV8y-Zog",
        "outputId": "94049b20-1bf3-4072-f00f-1b78a55cc36b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])"
            ]
          },
          "metadata": {},
          "execution_count": 64
        }
      ],
      "source": [
        "history.keys()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 703
        },
        "id": "ur0KNggM9eYx",
        "outputId": "8e9c8238-a98c-43a4-f2be-b8abfdfa630b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.figure(figsize=(10,5))\n",
        "plt.plot(history['accuracy'])\n",
        "plt.plot(history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['Train accuracy', 'Test accuracy'], loc='lower right')\n",
        "plt.show()\n",
        "# History for loss\n",
        "plt.figure(figsize=(10,5))\n",
        "plt.plot(history['loss'])\n",
        "plt.plot(history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['Train loss', 'Test loss'], loc='upper right')\n",
        "plt.suptitle('Accuracy and loss for second model')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0fhHdevU-uiq"
      },
      "outputs": [],
      "source": [
        "def predict(text, include_neutral=True):\n",
        "    # Tokenize text\n",
        "    x_test = pad_sequences(tokenizer.texts_to_sequences([text]), maxlen=max_length)\n",
        "    # Predict\n",
        "    score = model_1.predict([x_test])[0]\n",
        "    if(score >=0.4 and score<=0.6):\n",
        "        label = \"Neutral\"\n",
        "    if(score >=0.4):\n",
        "        label = \"Negative\"\n",
        "    if(score <=0.6):\n",
        "        label = \"Positive\"\n",
        "\n",
        "    return {\"label\" : label,\n",
        "        \"score\": float(score)} "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D1BAPH_x--ap",
        "outputId": "03df42ef-499d-4079-b4cf-085eabd397ff"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 1s 1s/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'label': 'Positive', 'score': 0.5246134996414185}"
            ]
          },
          "metadata": {},
          "execution_count": 67
        }
      ],
      "source": [
        "predict(\" i love everyone\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predict(\" i hate twitter\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kndl3QydUOgg",
        "outputId": "9c48480a-3a65-4c0e-fdc2-7cae64e3604f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 41ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'label': 'Negative', 'score': 0.6320455074310303}"
            ]
          },
          "metadata": {},
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cq1097C9HpEC",
        "outputId": "7727a804-d398-4359-d72e-976958f44afc"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 26ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'label': 'Positive', 'score': 0.542919933795929}"
            ]
          },
          "metadata": {},
          "execution_count": 71
        }
      ],
      "source": [
        "predict(\"it is good to show love\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predict(\"I love how idiots are never killed in movies\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cZqIKVWkLAV6",
        "outputId": "1ea065be-715f-47ae-d569-d5b06f06140b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 23ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'label': 'Negative', 'score': 0.6007909774780273}"
            ]
          },
          "metadata": {},
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "sQkfOqNh9KhM"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}