02 Vectorization.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/rdemarqui/sentiment_analysis/blob/main/02%20Vectorization.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Sentiment Analysis\n",
        "## Vectorization"
      ],
      "metadata": {
        "id": "dZpm8cGCojAP"
      },
      "id": "dZpm8cGCojAP"
    },
    {
      "cell_type": "code",
      "source": [
        "seed_value = 42"
      ],
      "metadata": {
        "id": "iQnrTNjxLfl8"
      },
      "id": "iQnrTNjxLfl8",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "39805500",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-08-16T19:50:38.338557Z",
          "start_time": "2023-08-16T19:50:37.039133Z"
        },
        "id": "39805500"
      },
      "outputs": [],
      "source": [
        "import multiprocessing, os, time\n",
        "from sklearn.model_selection import GridSearchCV, KFold, cross_val_score\n",
        "from sklearn import metrics\n",
        "from lightgbm import LGBMClassifier\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set(style=\"whitegrid\")\n",
        "np.random.seed(seed_value)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "path = os.getcwd()\n",
        "path"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "yVJJ69iue9jW",
        "outputId": "18bb6f3b-bbd6-401e-cb61-004beaad69e8"
      },
      "id": "yVJJ69iue9jW",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'/content'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "20c692eb",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-08-16T19:50:38.358234Z",
          "start_time": "2023-08-16T19:50:38.341560Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "20c692eb",
        "outputId": "adbe23a0-c8c8-4a49-a94d-5616de2f6d38"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "cores = multiprocessing.cpu_count()\n",
        "cores"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def vector_method_score(X_train, y_train, X_test, y_test, n_splits=5, shuffle=True, scoring='roc_auc'):\n",
        "  # Check cross-validation score\n",
        "  kfold = KFold(n_splits=n_splits, shuffle=shuffle, random_state=seed_value)\n",
        "  model = LGBMClassifier(n_jobs=-1, verbose=-1, random_state=seed_value)\n",
        "  scores = cross_val_score(model, X_train, y_train, cv=kfold, scoring=scoring)\n",
        "  cv_score = np.round(scores, 5)\n",
        "\n",
        "  # Check test score\n",
        "  model.fit(X_train, y_train);\n",
        "  y_pred = model.predict_proba(X_test)[:,1]\n",
        "  test_score = np.round(metrics.roc_auc_score(y_test, y_pred), 5)\n",
        "\n",
        "  return cv_score, test_score"
      ],
      "metadata": {
        "id": "7MpOWiPwIQA1"
      },
      "id": "7MpOWiPwIQA1",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def pipeline(X_train, y_train, X_test, y_test, vect_function, **kwargs):\n",
        "  start_time = time.time()\n",
        "  dict_csv_score = {}\n",
        "  dict_test_score = {}\n",
        "\n",
        "  for preproc_method in X_train.columns:\n",
        "    print(f\"Running: {preproc_method}\")\n",
        "    X_train_vec, X_test_vec = vect_function(X_train[preproc_method], X_test[preproc_method])\n",
        "    cv_score, test_score = vector_method_score(X_train_vec, y_train, X_test_vec, y_test)\n",
        "\n",
        "    dict_csv_score[preproc_method] = cv_score\n",
        "    dict_test_score[preproc_method] = [test_score]\n",
        "\n",
        "  end_time = time.time()\n",
        "  execution_time = end_time - start_time\n",
        "  hours, remainder = divmod(int(execution_time), 3600)\n",
        "  minutes, seconds = divmod(remainder, 60)\n",
        "  print(f\"\\nExecution time: {hours:02d}:{minutes:02d}:{seconds:02d}\")\n",
        "\n",
        "  return dict_csv_score, dict_test_score"
      ],
      "metadata": {
        "id": "5LQT8Ke3JUx-"
      },
      "id": "5LQT8Ke3JUx-",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def create_score_table(cv_score, test_score):\n",
        "  df_score = pd.DataFrame(cv_score).T\n",
        "  df_score['train_mean'] = df_score.mean(axis=1)\n",
        "  df_score['test'] = pd.DataFrame(test_score).T\n",
        "\n",
        "  return df_score"
      ],
      "metadata": {
        "id": "qCuCuvGAFFdf"
      },
      "id": "qCuCuvGAFFdf",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "id": "17220a72",
      "metadata": {
        "id": "17220a72"
      },
      "source": [
        "## Dataset\n",
        "---"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "from sklearn.model_selection import train_test_split\n",
        "from itertools import chain"
      ],
      "metadata": {
        "id": "83M1FiUCF4LQ"
      },
      "id": "83M1FiUCF4LQ",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "try:\n",
        "  df = pd.read_csv(os.path.join(path, 'data\\preprocessed_corpus.zip'),\n",
        "                   sep=',', compression='zip')\n",
        "except:\n",
        "  !wget https://github.com/rdemarqui/sentiment_analysis/raw/main/data/preprocessed_corpus.zip -q\n",
        "  df = pd.read_csv(os.path.join(path, 'preprocessed_corpus.zip'),\n",
        "                   sep=',', compression='zip')"
      ],
      "metadata": {
        "id": "VSOorgW0Zh-4"
      },
      "id": "VSOorgW0Zh-4",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "f773490d",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-08-16T19:50:38.876610Z",
          "start_time": "2023-08-16T19:50:38.861358Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "f773490d",
        "outputId": "e294cee6-21a3-446d-c6e9-fcf0118eb417"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(113005, 7)\n"
          ]
        }
      ],
      "source": [
        "columns = ['review_text_clean', 'review_text_clean_stop', 'review_text_clean_stem',\n",
        "           'review_text_clean_stop_stem', 'review_text_clean_lemma', 'review_text_clean_stop_lemma',\n",
        "           'overall_rating']\n",
        "corpus = df[columns].copy()\n",
        "corpus['overall_rating'] = corpus['overall_rating'].apply(lambda x: 1 if x >= 4 else 0)\n",
        "corpus = corpus.dropna()\n",
        "print(corpus.shape)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(corpus.iloc[:, :-1], corpus.iloc[:, -1:],\n",
        "                                                    test_size=0.2, random_state=seed_value)"
      ],
      "metadata": {
        "id": "s8cXg8GG6B6W"
      },
      "id": "s8cXg8GG6B6W",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 417
        },
        "id": "qz0kkZvg9ka-",
        "outputId": "edf02a99-e8bd-40e0-e6a8-d9a7fa128f01"
      },
      "id": "qz0kkZvg9ka-",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                        review_text_clean  \\\n",
              "77662   adorei produto muito pratico serviu ao propros...   \n",
              "78184   meu produto veio danificado porem nao sera mot...   \n",
              "105690  otimo produto travesseiro maravilhoso quantida...   \n",
              "96751   otima aquisicao agua geladinha fresca designer...   \n",
              "49691   produto atende corresponde as especificacoes s...   \n",
              "\n",
              "                                   review_text_clean_stop  \\\n",
              "77662   adorei produto pratico serviu proprosito prome...   \n",
              "78184   produto veio danificado porem motivo devolucao...   \n",
              "105690  otimo produto travesseiro maravilhoso quantida...   \n",
              "96751   otima aquisicao agua geladinha fresca designer...   \n",
              "49691   produto atende corresponde especificacoes cont...   \n",
              "\n",
              "                                   review_text_clean_stem  \\\n",
              "77662       ador produt muit pra serv ao proprosit promet   \n",
              "78184   meu produt vei danific por nao ser motiv par d...   \n",
              "105690  otim produt travess maravilh quant muit boa de...   \n",
              "96751   otim aquisica agu gel fresc design bonit ador ...   \n",
              "49691   produt atend correspond as especificaco sem co...   \n",
              "\n",
              "                              review_text_clean_stop_stem  \\\n",
              "77662               ador produt pra serv proprosit promet   \n",
              "78184         produt vei danific por motiv devoluca obrig   \n",
              "105690  otim produt travess maravilh quant boa pen gan...   \n",
              "96751   otim aquisica agu gel fresc design bonit ador ...   \n",
              "49691   produt atend correspond especificaco cont pen ...   \n",
              "\n",
              "                                  review_text_clean_lemma  \\\n",
              "77662   adorar produto muito pratico servir a o propro...   \n",
              "78184   meu produto vir danificar porem nao sera motiv...   \n",
              "105690  otimo produto travesseiro maravilhoso quantida...   \n",
              "96751   otimo aquisicao aguo geladinho fresco designer...   \n",
              "49691   produto atender corresponder o especificacoe s...   \n",
              "\n",
              "                             review_text_clean_stop_lemma  \n",
              "77662   adorar produto pratico servir proprosito prometer  \n",
              "78184   produto vir danificar porem motivo devolucao o...  \n",
              "105690  otimo produto travesseiro maravilhoso quantida...  \n",
              "96751   otimo aquisicao aguo geladinho fresco designer...  \n",
              "49691   produto atender corresponder especificacoes co...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-4ac0f851-14c3-404b-948f-60587fb94d03\" 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>review_text_clean</th>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>77662</th>\n",
              "      <td>adorei produto muito pratico serviu ao propros...</td>\n",
              "      <td>adorei produto pratico serviu proprosito prome...</td>\n",
              "      <td>ador produt muit pra serv ao proprosit promet</td>\n",
              "      <td>ador produt pra serv proprosit promet</td>\n",
              "      <td>adorar produto muito pratico servir a o propro...</td>\n",
              "      <td>adorar produto pratico servir proprosito prometer</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>78184</th>\n",
              "      <td>meu produto veio danificado porem nao sera mot...</td>\n",
              "      <td>produto veio danificado porem motivo devolucao...</td>\n",
              "      <td>meu produt vei danific por nao ser motiv par d...</td>\n",
              "      <td>produt vei danific por motiv devoluca obrig</td>\n",
              "      <td>meu produto vir danificar porem nao sera motiv...</td>\n",
              "      <td>produto vir danificar porem motivo devolucao o...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>105690</th>\n",
              "      <td>otimo produto travesseiro maravilhoso quantida...</td>\n",
              "      <td>otimo produto travesseiro maravilhoso quantida...</td>\n",
              "      <td>otim produt travess maravilh quant muit boa de...</td>\n",
              "      <td>otim produt travess maravilh quant boa pen gan...</td>\n",
              "      <td>otimo produto travesseiro maravilhoso quantida...</td>\n",
              "      <td>otimo produto travesseiro maravilhoso quantida...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>96751</th>\n",
              "      <td>otima aquisicao agua geladinha fresca designer...</td>\n",
              "      <td>otima aquisicao agua geladinha fresca designer...</td>\n",
              "      <td>otim aquisica agu gel fresc design bonit ador ...</td>\n",
              "      <td>otim aquisica agu gel fresc design bonit ador ...</td>\n",
              "      <td>otimo aquisicao aguo geladinho fresco designer...</td>\n",
              "      <td>otimo aquisicao aguo geladinho fresco designer...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>49691</th>\n",
              "      <td>produto atende corresponde as especificacoes s...</td>\n",
              "      <td>produto atende corresponde especificacoes cont...</td>\n",
              "      <td>produt atend correspond as especificaco sem co...</td>\n",
              "      <td>produt atend correspond especificaco cont pen ...</td>\n",
              "      <td>produto atender corresponder o especificacoe s...</td>\n",
              "      <td>produto atender corresponder especificacoes co...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4ac0f851-14c3-404b-948f-60587fb94d03')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-4ac0f851-14c3-404b-948f-60587fb94d03 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-4ac0f851-14c3-404b-948f-60587fb94d03');\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",
              "\n",
              "\n",
              "<div id=\"df-fa9f6137-b7ce-4f02-8a27-917fcbaaf9f2\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-fa9f6137-b7ce-4f02-8a27-917fcbaaf9f2')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-fa9f6137-b7ce-4f02-8a27-917fcbaaf9f2 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Checking y proportion\n",
        "print(f'Train proportion\\n{y_train.value_counts()/y_train.shape[0]}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S579M1OV6r_r",
        "outputId": "0ef986c7-7dec-4e4a-87c1-f1587f8e960a"
      },
      "id": "S579M1OV6r_r",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train proportion\n",
            "overall_rating\n",
            "1                 0.700323\n",
            "0                 0.299677\n",
            "dtype: float64\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = X_train.astype(str)\n",
        "y_train = y_train.values.ravel()\n",
        "X_test = X_test.astype(str)\n",
        "y_test = y_test.values.ravel()"
      ],
      "metadata": {
        "id": "GuYn5oZEqW0t"
      },
      "id": "GuYn5oZEqW0t",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check Vocabulary size\n",
        "plt.figure(figsize=(4, 4))\n",
        "voc_size = {}\n",
        "for column in X_train.columns:\n",
        "  full_text = ' '.join(X_train[column].tolist())\n",
        "  words = full_text.split()\n",
        "  vocabulary = set(words)\n",
        "  voc_size[column] = len(vocabulary)\n",
        "\n",
        "df_vocabulary_size = pd.DataFrame(list(voc_size.items()), columns=['Corpus', 'Voc_size'])\n",
        "\n",
        "ax = sns.barplot(x='Voc_size', y='Corpus',\n",
        "                 data=df_vocabulary_size,\n",
        "                 orient='h', color='steelblue')\n",
        "plt.title('Vocabulary size')\n",
        "plt.plot();"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "id": "mp-HFOnJ9KhV",
        "outputId": "72dff834-03ca-405a-b553-999ca135c46b"
      },
      "id": "mp-HFOnJ9KhV",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 400x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "bdbe9969",
      "metadata": {
        "id": "bdbe9969"
      },
      "source": [
        "## Sparse\n",
        "---"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "e6861466",
      "metadata": {
        "id": "e6861466"
      },
      "source": [
        "### Bag of Word\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "80626732",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-07-25T17:42:46.472927Z",
          "start_time": "2023-07-25T17:42:46.429443Z"
        },
        "id": "80626732"
      },
      "outputs": [],
      "source": [
        "from sklearn.feature_extraction.text import CountVectorizer"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def apply_bag_of_word(X_train, X_test):\n",
        "  vectorizer = CountVectorizer(lowercase=False, dtype=np.float32)\n",
        "  X_train_vec = vectorizer.fit_transform(X_train)\n",
        "  X_test_vec = vectorizer.transform(X_test)\n",
        "\n",
        "  return X_train_vec, X_test_vec"
      ],
      "metadata": {
        "id": "vUucY72sIH8v"
      },
      "id": "vUucY72sIH8v",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_bow, test_score_bow = pipeline(X_train, y_train,\n",
        "                                        X_test, y_test,\n",
        "                                        apply_bag_of_word)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "O_c8ocdlMVT2",
        "outputId": "fb877afe-8d1d-4621-a5e9-af7de9b863b3"
      },
      "id": "O_c8ocdlMVT2",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:04:56\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_bow = create_score_table(cv_score_bow, test_score_bow)\n",
        "df_score_bow"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "9oGIlj7lHCpq",
        "outputId": "5b07d81c-64ad-4bd1-ba7c-aab34c1970f8"
      },
      "id": "9oGIlj7lHCpq",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.97844  0.97962  0.97932  0.97953  0.98056   \n",
              "review_text_clean_stop        0.97225  0.97122  0.97212  0.97027  0.97296   \n",
              "review_text_clean_stem        0.97942  0.98028  0.97947  0.98015  0.98120   \n",
              "review_text_clean_stop_stem   0.97264  0.97255  0.97281  0.97183  0.97446   \n",
              "review_text_clean_lemma       0.97825  0.97962  0.97927  0.97966  0.98104   \n",
              "review_text_clean_stop_lemma  0.97202  0.97071  0.97222  0.97110  0.97353   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.979494  0.97790  \n",
              "review_text_clean_stop          0.971764  0.97048  \n",
              "review_text_clean_stem          0.980104  0.97894  \n",
              "review_text_clean_stop_stem     0.972858  0.97228  \n",
              "review_text_clean_lemma         0.979568  0.97847  \n",
              "review_text_clean_stop_lemma    0.971916  0.97080  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-86077f1b-ff62-4d00-acf6-a30e863a2bec\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.97844</td>\n",
              "      <td>0.97962</td>\n",
              "      <td>0.97932</td>\n",
              "      <td>0.97953</td>\n",
              "      <td>0.98056</td>\n",
              "      <td>0.979494</td>\n",
              "      <td>0.97790</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.97225</td>\n",
              "      <td>0.97122</td>\n",
              "      <td>0.97212</td>\n",
              "      <td>0.97027</td>\n",
              "      <td>0.97296</td>\n",
              "      <td>0.971764</td>\n",
              "      <td>0.97048</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.97942</td>\n",
              "      <td>0.98028</td>\n",
              "      <td>0.97947</td>\n",
              "      <td>0.98015</td>\n",
              "      <td>0.98120</td>\n",
              "      <td>0.980104</td>\n",
              "      <td>0.97894</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.97264</td>\n",
              "      <td>0.97255</td>\n",
              "      <td>0.97281</td>\n",
              "      <td>0.97183</td>\n",
              "      <td>0.97446</td>\n",
              "      <td>0.972858</td>\n",
              "      <td>0.97228</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97825</td>\n",
              "      <td>0.97962</td>\n",
              "      <td>0.97927</td>\n",
              "      <td>0.97966</td>\n",
              "      <td>0.98104</td>\n",
              "      <td>0.979568</td>\n",
              "      <td>0.97847</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.97202</td>\n",
              "      <td>0.97071</td>\n",
              "      <td>0.97222</td>\n",
              "      <td>0.97110</td>\n",
              "      <td>0.97353</td>\n",
              "      <td>0.971916</td>\n",
              "      <td>0.97080</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-86077f1b-ff62-4d00-acf6-a30e863a2bec')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-86077f1b-ff62-4d00-acf6-a30e863a2bec 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-86077f1b-ff62-4d00-acf6-a30e863a2bec');\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",
              "\n",
              "\n",
              "<div id=\"df-6cf3005f-6d62-454c-8d2b-37a0fa26ba5e\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-6cf3005f-6d62-454c-8d2b-37a0fa26ba5e')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-6cf3005f-6d62-454c-8d2b-37a0fa26ba5e button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "a7567b78",
      "metadata": {
        "id": "a7567b78"
      },
      "source": [
        "### TF-IDF\n",
        "---"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "69fe16a6",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-07-25T17:42:46.842187Z",
          "start_time": "2023-07-25T17:42:46.799777Z"
        },
        "id": "69fe16a6"
      },
      "outputs": [],
      "source": [
        "from sklearn.feature_extraction.text import TfidfVectorizer"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def apply_tf_idf(X_train, X_test):\n",
        "  vectorizer = TfidfVectorizer(lowercase=False, dtype=np.float32)\n",
        "  X_train_vec = vectorizer.fit_transform(X_train)\n",
        "  X_test_vec = vectorizer.transform(X_test)\n",
        "\n",
        "  return X_train_vec, X_test_vec"
      ],
      "metadata": {
        "id": "7Kzz9dRaIWm-"
      },
      "id": "7Kzz9dRaIWm-",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_tfidf, test_score_tfidf = pipeline(X_train, y_train,\n",
        "                                            X_test, y_test,\n",
        "                                            apply_tf_idf)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aia3UBD6I_S9",
        "outputId": "d68a2ca6-c327-4b0f-b9d1-2b673cceaf0e"
      },
      "id": "aia3UBD6I_S9",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:10:06\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_tfidf = create_score_table(cv_score_tfidf, test_score_tfidf)\n",
        "df_score_tfidf"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "qTHlPOF0G5yG",
        "outputId": "6ca5762f-327a-447a-fd14-71c8c4769648"
      },
      "id": "qTHlPOF0G5yG",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.97865  0.97960  0.97946  0.97966  0.98097   \n",
              "review_text_clean_stop        0.97287  0.97207  0.97290  0.97109  0.97396   \n",
              "review_text_clean_stem        0.97990  0.98048  0.97988  0.98017  0.98159   \n",
              "review_text_clean_stop_stem   0.97401  0.97337  0.97333  0.97262  0.97531   \n",
              "review_text_clean_lemma       0.97880  0.97958  0.97954  0.97997  0.98132   \n",
              "review_text_clean_stop_lemma  0.97272  0.97125  0.97255  0.97151  0.97390   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.979668  0.97811  \n",
              "review_text_clean_stop          0.972578  0.97143  \n",
              "review_text_clean_stem          0.980404  0.97907  \n",
              "review_text_clean_stop_stem     0.973728  0.97280  \n",
              "review_text_clean_lemma         0.979842  0.97882  \n",
              "review_text_clean_stop_lemma    0.972386  0.97166  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-188d0519-f916-4321-9f4d-c8ca45c7ded0\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.97865</td>\n",
              "      <td>0.97960</td>\n",
              "      <td>0.97946</td>\n",
              "      <td>0.97966</td>\n",
              "      <td>0.98097</td>\n",
              "      <td>0.979668</td>\n",
              "      <td>0.97811</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.97287</td>\n",
              "      <td>0.97207</td>\n",
              "      <td>0.97290</td>\n",
              "      <td>0.97109</td>\n",
              "      <td>0.97396</td>\n",
              "      <td>0.972578</td>\n",
              "      <td>0.97143</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.97990</td>\n",
              "      <td>0.98048</td>\n",
              "      <td>0.97988</td>\n",
              "      <td>0.98017</td>\n",
              "      <td>0.98159</td>\n",
              "      <td>0.980404</td>\n",
              "      <td>0.97907</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.97401</td>\n",
              "      <td>0.97337</td>\n",
              "      <td>0.97333</td>\n",
              "      <td>0.97262</td>\n",
              "      <td>0.97531</td>\n",
              "      <td>0.973728</td>\n",
              "      <td>0.97280</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97880</td>\n",
              "      <td>0.97958</td>\n",
              "      <td>0.97954</td>\n",
              "      <td>0.97997</td>\n",
              "      <td>0.98132</td>\n",
              "      <td>0.979842</td>\n",
              "      <td>0.97882</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.97272</td>\n",
              "      <td>0.97125</td>\n",
              "      <td>0.97255</td>\n",
              "      <td>0.97151</td>\n",
              "      <td>0.97390</td>\n",
              "      <td>0.972386</td>\n",
              "      <td>0.97166</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-188d0519-f916-4321-9f4d-c8ca45c7ded0')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-188d0519-f916-4321-9f4d-c8ca45c7ded0 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-188d0519-f916-4321-9f4d-c8ca45c7ded0');\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",
              "\n",
              "\n",
              "<div id=\"df-844d33ec-e1f1-4a85-8375-9ea42dfc3a9d\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-844d33ec-e1f1-4a85-8375-9ea42dfc3a9d')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-844d33ec-e1f1-4a85-8375-9ea42dfc3a9d button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "b27a3e58",
      "metadata": {
        "id": "b27a3e58"
      },
      "source": [
        "## Dense\n",
        "___"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "b6684fad",
      "metadata": {
        "id": "b6684fad"
      },
      "source": [
        "### Word2vec\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "dfc1b300",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-07-25T17:42:47.227462Z",
          "start_time": "2023-07-25T17:42:47.184701Z"
        },
        "id": "dfc1b300"
      },
      "outputs": [],
      "source": [
        "from gensim.models import Word2Vec"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def apply_word2vec(X_train, X_test, sg=0):\n",
        "  X_train_tokenized = [text.split() for text in X_train.values.tolist()]\n",
        "  X_test_tokenized = [text.split() for text in X_test.values.tolist()]\n",
        "\n",
        "  vectorizer = Word2Vec(vector_size=300, sg=sg, workers=cores)\n",
        "  vectorizer.build_vocab(X_train_tokenized)\n",
        "  vectorizer.train(X_train_tokenized, total_examples=vectorizer.corpus_count, epochs=vectorizer.epochs)\n",
        "\n",
        "  # A random vector for OOV it's better than zeros vector, because model can learn something\n",
        "  vector_shape = vectorizer.wv[list(vectorizer.wv.key_to_index)[0]].shape\n",
        "  np.random.seed(seed_value)\n",
        "  UNK_TOKEN = np.random.rand(*vector_shape)\n",
        "\n",
        "  def get_word_vector(word):\n",
        "    return vectorizer.wv[word] if word in vectorizer.wv else UNK_TOKEN\n",
        "\n",
        "  X_train_vec = [[get_word_vector(word) for word in text] for text in X_train_tokenized]\n",
        "  X_test_vec = [[get_word_vector(word) for word in text] for text in X_test_tokenized]\n",
        "\n",
        "  # Get mean of word vectors for each text\n",
        "  X_train_vec = [np.mean(np.array(vector).T, axis=1) for vector in X_train_vec]\n",
        "  X_test_vec = [np.mean(np.array(vector).T, axis=1) for vector in X_test_vec]\n",
        "\n",
        "  return X_train_vec, X_test_vec"
      ],
      "metadata": {
        "id": "RfOHj5nSqo9b"
      },
      "id": "RfOHj5nSqo9b",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "id": "0764cc79",
      "metadata": {
        "id": "0764cc79"
      },
      "source": [
        "#### CBOW - Continuous Bag of Words\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_cbow, test_score_cbow = pipeline(X_train, y_train,\n",
        "                                          X_test, y_test,\n",
        "                                          apply_word2vec, sg=0)"
      ],
      "metadata": {
        "id": "x1UkOpq3UShb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bf7279a0-729c-437a-ef12-33193471c35e"
      },
      "id": "x1UkOpq3UShb",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:14:14\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_cbow = create_score_table(cv_score_cbow, test_score_cbow)\n",
        "df_score_cbow"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "DtmeM9MiGoVo",
        "outputId": "25f25fa5-f7a3-4b02-b28a-b871f324659d"
      },
      "id": "DtmeM9MiGoVo",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.97590  0.97610  0.97668  0.97666  0.97793   \n",
              "review_text_clean_stop        0.97315  0.97329  0.97381  0.97312  0.97426   \n",
              "review_text_clean_stem        0.97575  0.97661  0.97678  0.97701  0.97840   \n",
              "review_text_clean_stop_stem   0.97270  0.97306  0.97298  0.97285  0.97440   \n",
              "review_text_clean_lemma       0.97711  0.97726  0.97754  0.97810  0.97958   \n",
              "review_text_clean_stop_lemma  0.97258  0.97293  0.97290  0.97218  0.97477   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.976654  0.97630  \n",
              "review_text_clean_stop          0.973526  0.97301  \n",
              "review_text_clean_stem          0.976910  0.97685  \n",
              "review_text_clean_stop_stem     0.973198  0.97242  \n",
              "review_text_clean_lemma         0.977918  0.97689  \n",
              "review_text_clean_stop_lemma    0.973072  0.97201  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-fbd83d7a-72a5-46e4-9bbe-92e7c5ddd25d\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.97590</td>\n",
              "      <td>0.97610</td>\n",
              "      <td>0.97668</td>\n",
              "      <td>0.97666</td>\n",
              "      <td>0.97793</td>\n",
              "      <td>0.976654</td>\n",
              "      <td>0.97630</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.97315</td>\n",
              "      <td>0.97329</td>\n",
              "      <td>0.97381</td>\n",
              "      <td>0.97312</td>\n",
              "      <td>0.97426</td>\n",
              "      <td>0.973526</td>\n",
              "      <td>0.97301</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.97575</td>\n",
              "      <td>0.97661</td>\n",
              "      <td>0.97678</td>\n",
              "      <td>0.97701</td>\n",
              "      <td>0.97840</td>\n",
              "      <td>0.976910</td>\n",
              "      <td>0.97685</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.97270</td>\n",
              "      <td>0.97306</td>\n",
              "      <td>0.97298</td>\n",
              "      <td>0.97285</td>\n",
              "      <td>0.97440</td>\n",
              "      <td>0.973198</td>\n",
              "      <td>0.97242</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97711</td>\n",
              "      <td>0.97726</td>\n",
              "      <td>0.97754</td>\n",
              "      <td>0.97810</td>\n",
              "      <td>0.97958</td>\n",
              "      <td>0.977918</td>\n",
              "      <td>0.97689</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.97258</td>\n",
              "      <td>0.97293</td>\n",
              "      <td>0.97290</td>\n",
              "      <td>0.97218</td>\n",
              "      <td>0.97477</td>\n",
              "      <td>0.973072</td>\n",
              "      <td>0.97201</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-fbd83d7a-72a5-46e4-9bbe-92e7c5ddd25d')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-fbd83d7a-72a5-46e4-9bbe-92e7c5ddd25d 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-fbd83d7a-72a5-46e4-9bbe-92e7c5ddd25d');\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",
              "\n",
              "\n",
              "<div id=\"df-d9fdbaf0-f4ad-4c6b-b8a5-5a562d793a79\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-d9fdbaf0-f4ad-4c6b-b8a5-5a562d793a79')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-d9fdbaf0-f4ad-4c6b-b8a5-5a562d793a79 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "7ba58812",
      "metadata": {
        "id": "7ba58812"
      },
      "source": [
        "#### SG - Skip-Gram\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_skgram, test_score_skgram = pipeline(X_train, y_train,\n",
        "                                              X_test, y_test,\n",
        "                                              apply_word2vec, sg=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J3l4qOaJVYir",
        "outputId": "684c7a09-0052-467b-ec1f-acbef11a42ae"
      },
      "id": "J3l4qOaJVYir",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:14:39\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_skgram = create_score_table(cv_score_skgram, test_score_skgram)\n",
        "df_score_skgram"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "emIMPJhbGbEY",
        "outputId": "28025db0-d3cf-47c8-ca4f-db8205de78a6"
      },
      "id": "emIMPJhbGbEY",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.97588  0.97616  0.97626  0.97715  0.97736   \n",
              "review_text_clean_stop        0.97288  0.97333  0.97393  0.97285  0.97420   \n",
              "review_text_clean_stem        0.97597  0.97716  0.97650  0.97777  0.97865   \n",
              "review_text_clean_stop_stem   0.97299  0.97276  0.97284  0.97266  0.97489   \n",
              "review_text_clean_lemma       0.97645  0.97719  0.97811  0.97824  0.97959   \n",
              "review_text_clean_stop_lemma  0.97256  0.97230  0.97233  0.97237  0.97415   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.976562  0.97596  \n",
              "review_text_clean_stop          0.973438  0.97302  \n",
              "review_text_clean_stem          0.977210  0.97648  \n",
              "review_text_clean_stop_stem     0.973228  0.97279  \n",
              "review_text_clean_lemma         0.977916  0.97692  \n",
              "review_text_clean_stop_lemma    0.972742  0.97221  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-92a5eece-4c10-48fb-971b-3307e2b4da82\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.97588</td>\n",
              "      <td>0.97616</td>\n",
              "      <td>0.97626</td>\n",
              "      <td>0.97715</td>\n",
              "      <td>0.97736</td>\n",
              "      <td>0.976562</td>\n",
              "      <td>0.97596</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.97288</td>\n",
              "      <td>0.97333</td>\n",
              "      <td>0.97393</td>\n",
              "      <td>0.97285</td>\n",
              "      <td>0.97420</td>\n",
              "      <td>0.973438</td>\n",
              "      <td>0.97302</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.97597</td>\n",
              "      <td>0.97716</td>\n",
              "      <td>0.97650</td>\n",
              "      <td>0.97777</td>\n",
              "      <td>0.97865</td>\n",
              "      <td>0.977210</td>\n",
              "      <td>0.97648</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.97299</td>\n",
              "      <td>0.97276</td>\n",
              "      <td>0.97284</td>\n",
              "      <td>0.97266</td>\n",
              "      <td>0.97489</td>\n",
              "      <td>0.973228</td>\n",
              "      <td>0.97279</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97645</td>\n",
              "      <td>0.97719</td>\n",
              "      <td>0.97811</td>\n",
              "      <td>0.97824</td>\n",
              "      <td>0.97959</td>\n",
              "      <td>0.977916</td>\n",
              "      <td>0.97692</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.97256</td>\n",
              "      <td>0.97230</td>\n",
              "      <td>0.97233</td>\n",
              "      <td>0.97237</td>\n",
              "      <td>0.97415</td>\n",
              "      <td>0.972742</td>\n",
              "      <td>0.97221</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-92a5eece-4c10-48fb-971b-3307e2b4da82')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-92a5eece-4c10-48fb-971b-3307e2b4da82 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-92a5eece-4c10-48fb-971b-3307e2b4da82');\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",
              "\n",
              "\n",
              "<div id=\"df-55210d3d-afd5-4da8-a5d4-9190ac45fea5\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-55210d3d-afd5-4da8-a5d4-9190ac45fea5')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-55210d3d-afd5-4da8-a5d4-9190ac45fea5 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "ccf5b5fa",
      "metadata": {
        "id": "ccf5b5fa"
      },
      "source": [
        "### FastText\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "d06acc07",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-07-25T17:42:55.467814Z",
          "start_time": "2023-07-25T17:42:55.223958Z"
        },
        "id": "d06acc07"
      },
      "outputs": [],
      "source": [
        "from gensim.models import FastText"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def apply_FastText(X_train, X_test):\n",
        "  X_train_tokenized = [text.split() for text in X_train.values.tolist()]\n",
        "  X_test_tokenized = [text.split() for text in X_test.values.tolist()]\n",
        "\n",
        "  vectorizer = FastText(vector_size=300, workers=cores)\n",
        "  vectorizer.build_vocab(X_train_tokenized)\n",
        "  vectorizer.train(X_train_tokenized, total_examples=vectorizer.corpus_count,epochs=vectorizer.epochs)\n",
        "\n",
        "  # A random vector for OOV it's better than zeros vector, because model can learn something\n",
        "  vector_shape = vectorizer.wv[list(vectorizer.wv.key_to_index)[0]].shape\n",
        "  np.random.seed(seed_value)\n",
        "  UNK_TOKEN = np.random.rand(*vector_shape)\n",
        "\n",
        "  def get_word_vector(word):\n",
        "    return vectorizer.wv[word] if word in vectorizer.wv else UNK_TOKEN\n",
        "\n",
        "  X_train_vec = [[get_word_vector(word) for word in text] for text in X_train_tokenized]\n",
        "  X_test_vec = [[get_word_vector(word) for word in text] for text in X_test_tokenized]\n",
        "\n",
        "  # Get mean of word vectors for each text\n",
        "  X_train_vec = [np.mean(np.array(vector).T, axis=1) for vector in X_train_vec]\n",
        "  X_test_vec = [np.mean(np.array(vector).T, axis=1) for vector in X_test_vec]\n",
        "\n",
        "  return X_train_vec, X_test_vec"
      ],
      "metadata": {
        "id": "CyfW7dlamBTJ"
      },
      "id": "CyfW7dlamBTJ",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_FastTest, test_score_FastTest = pipeline(X_train, y_train,\n",
        "                                                  X_test, y_test,\n",
        "                                                  apply_FastText)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3Wn6eExUwAK3",
        "outputId": "98c83e86-0514-45a5-df94-12db50ddc944"
      },
      "id": "3Wn6eExUwAK3",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:20:39\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_FastTest = create_score_table(cv_score_FastTest, test_score_FastTest)\n",
        "df_score_FastTest"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "5NmICQ-AGPPu",
        "outputId": "103758e6-1952-461e-db80-9308fcb0371f"
      },
      "id": "5NmICQ-AGPPu",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.96984  0.97060  0.97147  0.97154  0.97227   \n",
              "review_text_clean_stop        0.97005  0.97024  0.97133  0.96988  0.97192   \n",
              "review_text_clean_stem        0.97134  0.97138  0.97185  0.97318  0.97424   \n",
              "review_text_clean_stop_stem   0.96982  0.96904  0.97044  0.96946  0.97200   \n",
              "review_text_clean_lemma       0.97135  0.97191  0.97172  0.97258  0.97414   \n",
              "review_text_clean_stop_lemma  0.96863  0.96839  0.97037  0.96834  0.97116   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.971144  0.97011  \n",
              "review_text_clean_stop          0.970684  0.97032  \n",
              "review_text_clean_stem          0.972398  0.97148  \n",
              "review_text_clean_stop_stem     0.970152  0.96980  \n",
              "review_text_clean_lemma         0.972340  0.97167  \n",
              "review_text_clean_stop_lemma    0.969378  0.96919  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d1ec95b9-f7f0-4408-a561-e015e204387f\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.96984</td>\n",
              "      <td>0.97060</td>\n",
              "      <td>0.97147</td>\n",
              "      <td>0.97154</td>\n",
              "      <td>0.97227</td>\n",
              "      <td>0.971144</td>\n",
              "      <td>0.97011</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.97005</td>\n",
              "      <td>0.97024</td>\n",
              "      <td>0.97133</td>\n",
              "      <td>0.96988</td>\n",
              "      <td>0.97192</td>\n",
              "      <td>0.970684</td>\n",
              "      <td>0.97032</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.97134</td>\n",
              "      <td>0.97138</td>\n",
              "      <td>0.97185</td>\n",
              "      <td>0.97318</td>\n",
              "      <td>0.97424</td>\n",
              "      <td>0.972398</td>\n",
              "      <td>0.97148</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.96982</td>\n",
              "      <td>0.96904</td>\n",
              "      <td>0.97044</td>\n",
              "      <td>0.96946</td>\n",
              "      <td>0.97200</td>\n",
              "      <td>0.970152</td>\n",
              "      <td>0.96980</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97135</td>\n",
              "      <td>0.97191</td>\n",
              "      <td>0.97172</td>\n",
              "      <td>0.97258</td>\n",
              "      <td>0.97414</td>\n",
              "      <td>0.972340</td>\n",
              "      <td>0.97167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.96863</td>\n",
              "      <td>0.96839</td>\n",
              "      <td>0.97037</td>\n",
              "      <td>0.96834</td>\n",
              "      <td>0.97116</td>\n",
              "      <td>0.969378</td>\n",
              "      <td>0.96919</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d1ec95b9-f7f0-4408-a561-e015e204387f')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-d1ec95b9-f7f0-4408-a561-e015e204387f 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-d1ec95b9-f7f0-4408-a561-e015e204387f');\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",
              "\n",
              "\n",
              "<div id=\"df-111c8aad-0dde-40bf-b138-240e5c56ef0a\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-111c8aad-0dde-40bf-b138-240e5c56ef0a')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-111c8aad-0dde-40bf-b138-240e5c56ef0a button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "02444b2c",
      "metadata": {
        "id": "02444b2c"
      },
      "source": [
        "### Doc2Vec\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "f4c71a6b",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-07-24T20:05:35.982840Z",
          "start_time": "2023-07-24T20:05:35.910763Z"
        },
        "id": "f4c71a6b"
      },
      "outputs": [],
      "source": [
        "from gensim.models import Doc2Vec\n",
        "from gensim.models.doc2vec import TaggedDocument"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def apply_doc2vec(X_train, X_test, dm=0):\n",
        "  X_train_tagged = [TaggedDocument(words=text.split(), tags=[str(i)]) for i, text in enumerate(X_train.values.tolist())]\n",
        "  X_test_tokenized = [text.split() for text in X_test.values.tolist()]\n",
        "\n",
        "  vectorizer = Doc2Vec(vector_size=300, workers=cores, dm=dm)\n",
        "  vectorizer.build_vocab(X_train_tagged)\n",
        "  vectorizer.train(X_train_tagged, total_examples=vectorizer.corpus_count, epochs=vectorizer.epochs)\n",
        "\n",
        "  X_train_vec = [vectorizer.dv[tag] for tag in vectorizer.dv.index_to_key]\n",
        "  X_test_vec = [vectorizer.infer_vector(text) for text in X_test_tokenized]\n",
        "\n",
        "  return X_train_vec, X_test_vec"
      ],
      "metadata": {
        "id": "HdgBXD3J3o_s"
      },
      "id": "HdgBXD3J3o_s",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "id": "18672e00",
      "metadata": {
        "id": "18672e00"
      },
      "source": [
        "#### DBOW - Distributed Bag of Words\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_dbow, test_score_dbow = pipeline(X_train, y_train,\n",
        "                                          X_test, y_test,\n",
        "                                          apply_doc2vec, dm=0)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5eb1a6de-2fc4-423b-9e8c-326a1a956bea",
        "id": "n3fJh_c4DLjR"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:18:00\n"
          ]
        }
      ],
      "id": "n3fJh_c4DLjR"
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_dbow = create_score_table(cv_score_dbow, test_score_dbow)\n",
        "df_score_dbow"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "PHfS7UPJF7to",
        "outputId": "7bd680fa-df7e-4e15-c719-63e230637d94"
      },
      "id": "PHfS7UPJF7to",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.96274  0.96404  0.96411  0.96350  0.96501   \n",
              "review_text_clean_stop        0.95132  0.94836  0.94986  0.94850  0.95263   \n",
              "review_text_clean_stem        0.96615  0.96652  0.96522  0.96619  0.96719   \n",
              "review_text_clean_stop_stem   0.95678  0.95522  0.95559  0.95527  0.95816   \n",
              "review_text_clean_lemma       0.96463  0.96574  0.96598  0.96527  0.96758   \n",
              "review_text_clean_stop_lemma  0.95473  0.95274  0.95161  0.95157  0.95432   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.963880  0.95209  \n",
              "review_text_clean_stop          0.950134  0.93693  \n",
              "review_text_clean_stem          0.966254  0.95312  \n",
              "review_text_clean_stop_stem     0.956204  0.94034  \n",
              "review_text_clean_lemma         0.965840  0.95175  \n",
              "review_text_clean_stop_lemma    0.952994  0.93994  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-cc1514f7-bc0a-49b5-8c4f-27ea2cff8cb8\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.96274</td>\n",
              "      <td>0.96404</td>\n",
              "      <td>0.96411</td>\n",
              "      <td>0.96350</td>\n",
              "      <td>0.96501</td>\n",
              "      <td>0.963880</td>\n",
              "      <td>0.95209</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.95132</td>\n",
              "      <td>0.94836</td>\n",
              "      <td>0.94986</td>\n",
              "      <td>0.94850</td>\n",
              "      <td>0.95263</td>\n",
              "      <td>0.950134</td>\n",
              "      <td>0.93693</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.96615</td>\n",
              "      <td>0.96652</td>\n",
              "      <td>0.96522</td>\n",
              "      <td>0.96619</td>\n",
              "      <td>0.96719</td>\n",
              "      <td>0.966254</td>\n",
              "      <td>0.95312</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.95678</td>\n",
              "      <td>0.95522</td>\n",
              "      <td>0.95559</td>\n",
              "      <td>0.95527</td>\n",
              "      <td>0.95816</td>\n",
              "      <td>0.956204</td>\n",
              "      <td>0.94034</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.96463</td>\n",
              "      <td>0.96574</td>\n",
              "      <td>0.96598</td>\n",
              "      <td>0.96527</td>\n",
              "      <td>0.96758</td>\n",
              "      <td>0.965840</td>\n",
              "      <td>0.95175</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.95473</td>\n",
              "      <td>0.95274</td>\n",
              "      <td>0.95161</td>\n",
              "      <td>0.95157</td>\n",
              "      <td>0.95432</td>\n",
              "      <td>0.952994</td>\n",
              "      <td>0.93994</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-cc1514f7-bc0a-49b5-8c4f-27ea2cff8cb8')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-cc1514f7-bc0a-49b5-8c4f-27ea2cff8cb8 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-cc1514f7-bc0a-49b5-8c4f-27ea2cff8cb8');\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",
              "\n",
              "\n",
              "<div id=\"df-aa4fc6bb-9fa4-4694-80e2-276de0afbc74\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-aa4fc6bb-9fa4-4694-80e2-276de0afbc74')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-aa4fc6bb-9fa4-4694-80e2-276de0afbc74 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "13d6ea4e",
      "metadata": {
        "id": "13d6ea4e"
      },
      "source": [
        "#### DM - Distributed Memory\n",
        "___"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_dm, test_score_dm = pipeline(X_train, y_train,\n",
        "                                      X_test, y_test,\n",
        "                                      apply_doc2vec, dm=1)"
      ],
      "metadata": {
        "id": "l49CmZ_jEIWF",
        "outputId": "3a1bebd4-535f-402d-b739-8b19c39062e3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "id": "l49CmZ_jEIWF",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:17:35\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_dm = create_score_table(cv_score_dm, test_score_dm)\n",
        "df_score_dm"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "qgMtuKWCFiOW",
        "outputId": "9d4000ca-6ab7-48d8-a028-2a38fb6d357d"
      },
      "id": "qgMtuKWCFiOW",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.96245  0.96439  0.96440  0.96277  0.96554   \n",
              "review_text_clean_stop        0.94938  0.94875  0.95040  0.94943  0.95184   \n",
              "review_text_clean_stem        0.96485  0.96597  0.96569  0.96404  0.96711   \n",
              "review_text_clean_stop_stem   0.95764  0.95584  0.95613  0.95546  0.95749   \n",
              "review_text_clean_lemma       0.96588  0.96580  0.96626  0.96633  0.96882   \n",
              "review_text_clean_stop_lemma  0.95360  0.95084  0.95261  0.95099  0.95408   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.963910  0.95050  \n",
              "review_text_clean_stop          0.949960  0.93719  \n",
              "review_text_clean_stem          0.965532  0.94671  \n",
              "review_text_clean_stop_stem     0.956512  0.94306  \n",
              "review_text_clean_lemma         0.966618  0.95374  \n",
              "review_text_clean_stop_lemma    0.952424  0.93533  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c2ae9e3c-4372-4ff9-8540-55e25de1e0e5\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.96245</td>\n",
              "      <td>0.96439</td>\n",
              "      <td>0.96440</td>\n",
              "      <td>0.96277</td>\n",
              "      <td>0.96554</td>\n",
              "      <td>0.963910</td>\n",
              "      <td>0.95050</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.94938</td>\n",
              "      <td>0.94875</td>\n",
              "      <td>0.95040</td>\n",
              "      <td>0.94943</td>\n",
              "      <td>0.95184</td>\n",
              "      <td>0.949960</td>\n",
              "      <td>0.93719</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.96485</td>\n",
              "      <td>0.96597</td>\n",
              "      <td>0.96569</td>\n",
              "      <td>0.96404</td>\n",
              "      <td>0.96711</td>\n",
              "      <td>0.965532</td>\n",
              "      <td>0.94671</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.95764</td>\n",
              "      <td>0.95584</td>\n",
              "      <td>0.95613</td>\n",
              "      <td>0.95546</td>\n",
              "      <td>0.95749</td>\n",
              "      <td>0.956512</td>\n",
              "      <td>0.94306</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.96588</td>\n",
              "      <td>0.96580</td>\n",
              "      <td>0.96626</td>\n",
              "      <td>0.96633</td>\n",
              "      <td>0.96882</td>\n",
              "      <td>0.966618</td>\n",
              "      <td>0.95374</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.95360</td>\n",
              "      <td>0.95084</td>\n",
              "      <td>0.95261</td>\n",
              "      <td>0.95099</td>\n",
              "      <td>0.95408</td>\n",
              "      <td>0.952424</td>\n",
              "      <td>0.93533</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c2ae9e3c-4372-4ff9-8540-55e25de1e0e5')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-c2ae9e3c-4372-4ff9-8540-55e25de1e0e5 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-c2ae9e3c-4372-4ff9-8540-55e25de1e0e5');\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",
              "\n",
              "\n",
              "<div id=\"df-80ba2b64-aeee-47db-9823-8a6c4a93e785\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-80ba2b64-aeee-47db-9823-8a6c4a93e785')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-80ba2b64-aeee-47db-9823-8a6c4a93e785 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "aa5b01a0",
      "metadata": {
        "id": "aa5b01a0"
      },
      "source": [
        "### Transformer (Pre-trained model)\n",
        "___"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Bert"
      ],
      "metadata": {
        "id": "OBMmXo70rMgk"
      },
      "id": "OBMmXo70rMgk"
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Note 1:** In this stage, it's important that Colab uses `GPU`instead of `CPU`.\n",
        "\n",
        "**Note 2:** As already said in https://huggingface.co/blog/how-to-train-sentence-transformers, pre-trained transformers are slow to perform the task (even using GPU), so we decided to run in a small portion of data. It's results won't be used in the final conclusions. For text embedding, it's better to use **sentence transfomers**, available in Hugging Face (implemented in next topic).\n",
        "\n",
        "**Note 3:** Even with reduced data, we can see that Bert performed better with text without remove stop words, stematize and lemmatize (as expected)."
      ],
      "metadata": {
        "id": "QSpCOu9tDgG3"
      },
      "id": "QSpCOu9tDgG3"
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install transformers -q"
      ],
      "metadata": {
        "id": "6dZ4GkH-ly2I"
      },
      "id": "6dZ4GkH-ly2I",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "device = \"cuda:0\" if torch.cuda.is_available() else \"cpu\"\n",
        "device"
      ],
      "metadata": {
        "id": "D4DrVyKTdSgO",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "0878a33e-6345-4bbb-aa04-be1ca82b1d3b"
      },
      "id": "D4DrVyKTdSgO",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'cuda:0'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from transformers import AutoModel\n",
        "from transformers import AutoTokenizer"
      ],
      "metadata": {
        "id": "_rHxE2KsmzLi"
      },
      "id": "_rHxE2KsmzLi",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = AutoModel.from_pretrained('neuralmind/bert-base-portuguese-cased').to(device)\n",
        "tokenizer = AutoTokenizer.from_pretrained('neuralmind/bert-base-portuguese-cased',\n",
        "                                          do_lower_case=True, model_max_length=512)"
      ],
      "metadata": {
        "id": "a9EvPBXVquiO"
      },
      "id": "a9EvPBXVquiO",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def apply_tranformer(X_train, X_test, model=model, tokenizer=tokenizer):\n",
        "  model = model\n",
        "  tokenizer = tokenizer\n",
        "\n",
        "  X_train_list = X_train.values.tolist()\n",
        "  X_test_list = X_test.values.tolist()\n",
        "\n",
        "  # Get tokens form text\n",
        "  X_train_toc = [tokenizer.encode(doc, truncation=True,\n",
        "                                  return_tensors='pt').to(device) for doc in X_train_list]\n",
        "  X_test_toc = [tokenizer.encode(doc, truncation=True,\n",
        "                                 return_tensors='pt').to(device) for doc in X_test_list]\n",
        "\n",
        "  # Get mean vectors from converted tokens ([0, 1:-1] Last layer, ignore [CLS] and [SEP] special tokens)\n",
        "  with torch.no_grad():\n",
        "    X_train_vec = [torch.mean(model(tokens)[0][0, 1:-1].T, dim=1).cpu().data.numpy() for tokens in X_train_toc]\n",
        "    X_test_vec = [torch.mean(model(tokens)[0][0, 1:-1].T, dim=1).cpu().data.numpy() for tokens in X_test_toc]\n",
        "\n",
        "  return X_train_vec, X_test_vec"
      ],
      "metadata": {
        "id": "4Nijzzy0uAus"
      },
      "id": "4Nijzzy0uAus",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Testing with reduced data\n",
        "cv_score_bert, test_score_bert = pipeline(X_train.iloc[:1000, :], y_train[:1000],\n",
        "                                          X_test.iloc[:1000, :], y_test[:1000],\n",
        "                                          apply_tranformer,\n",
        "                                          model=model,\n",
        "                                          tokenizer=tokenizer)"
      ],
      "metadata": {
        "id": "vXu-LrUiwY_w",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8d9cf426-203c-4acc-e80e-808a23f1ac2c"
      },
      "id": "vXu-LrUiwY_w",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 00:07:09\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_bert = create_score_table(cv_score_bert, test_score_bert)\n",
        "df_score_bert"
      ],
      "metadata": {
        "id": "WQ_vWkX7FVY2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "outputId": "20b9b7ba-11ff-46ff-b09f-79de81b52aae"
      },
      "id": "WQ_vWkX7FVY2",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.98209  0.97775  0.94682  0.92932  0.96804   \n",
              "review_text_clean_stop        0.96932  0.94682  0.92691  0.93950  0.96791   \n",
              "review_text_clean_stem        0.92335  0.92307  0.92836  0.89634  0.94943   \n",
              "review_text_clean_stop_stem   0.90130  0.89515  0.87749  0.86294  0.91028   \n",
              "review_text_clean_lemma       0.97292  0.96814  0.94051  0.93419  0.95803   \n",
              "review_text_clean_stop_lemma  0.96353  0.94369  0.91234  0.91498  0.95764   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.960804  0.96099  \n",
              "review_text_clean_stop          0.950092  0.94229  \n",
              "review_text_clean_stem          0.924110  0.92251  \n",
              "review_text_clean_stop_stem     0.889432  0.89251  \n",
              "review_text_clean_lemma         0.954758  0.94173  \n",
              "review_text_clean_stop_lemma    0.938436  0.93649  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-87eb75e2-3829-4911-93c4-b7c15ecdff88\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.98209</td>\n",
              "      <td>0.97775</td>\n",
              "      <td>0.94682</td>\n",
              "      <td>0.92932</td>\n",
              "      <td>0.96804</td>\n",
              "      <td>0.960804</td>\n",
              "      <td>0.96099</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.96932</td>\n",
              "      <td>0.94682</td>\n",
              "      <td>0.92691</td>\n",
              "      <td>0.93950</td>\n",
              "      <td>0.96791</td>\n",
              "      <td>0.950092</td>\n",
              "      <td>0.94229</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.92335</td>\n",
              "      <td>0.92307</td>\n",
              "      <td>0.92836</td>\n",
              "      <td>0.89634</td>\n",
              "      <td>0.94943</td>\n",
              "      <td>0.924110</td>\n",
              "      <td>0.92251</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.90130</td>\n",
              "      <td>0.89515</td>\n",
              "      <td>0.87749</td>\n",
              "      <td>0.86294</td>\n",
              "      <td>0.91028</td>\n",
              "      <td>0.889432</td>\n",
              "      <td>0.89251</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97292</td>\n",
              "      <td>0.96814</td>\n",
              "      <td>0.94051</td>\n",
              "      <td>0.93419</td>\n",
              "      <td>0.95803</td>\n",
              "      <td>0.954758</td>\n",
              "      <td>0.94173</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.96353</td>\n",
              "      <td>0.94369</td>\n",
              "      <td>0.91234</td>\n",
              "      <td>0.91498</td>\n",
              "      <td>0.95764</td>\n",
              "      <td>0.938436</td>\n",
              "      <td>0.93649</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-87eb75e2-3829-4911-93c4-b7c15ecdff88')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-87eb75e2-3829-4911-93c4-b7c15ecdff88 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-87eb75e2-3829-4911-93c4-b7c15ecdff88');\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",
              "\n",
              "\n",
              "<div id=\"df-f751f94b-5fca-4785-b4d0-ea5be0176c3d\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-f751f94b-5fca-4785-b4d0-ea5be0176c3d')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-f751f94b-5fca-4785-b4d0-ea5be0176c3d button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Sentence Transformer"
      ],
      "metadata": {
        "id": "Y3vGM7sLDGCi"
      },
      "id": "Y3vGM7sLDGCi"
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Note:** In this stage, it's important that Colab uses `GPU`instead of `CPU`."
      ],
      "metadata": {
        "id": "JAyNtwkFITlt"
      },
      "id": "JAyNtwkFITlt"
    },
    {
      "cell_type": "code",
      "source": [
        "device = !nvidia-smi -L\n",
        "device = \"cuda\" if device[0][:3] == 'GPU' else \"cpu\"\n",
        "device"
      ],
      "metadata": {
        "id": "3TWm69iMlMl9",
        "outputId": "30d47569-5c57-43d6-8fb9-386cc1588231",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "id": "3TWm69iMlMl9",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'cuda'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install sentence_transformers -q"
      ],
      "metadata": {
        "id": "5BFus9SGDrox"
      },
      "id": "5BFus9SGDrox",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "33883135",
      "metadata": {
        "ExecuteTime": {
          "end_time": "2023-07-24T20:06:49.567212Z",
          "start_time": "2023-07-24T20:06:45.593705Z"
        },
        "id": "33883135"
      },
      "outputs": [],
      "source": [
        "from sentence_transformers import SentenceTransformer"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model=SentenceTransformer('rufimelo/bert-large-portuguese-cased-sts', device='cuda')"
      ],
      "metadata": {
        "id": "cihdPcoo8m2B"
      },
      "id": "cihdPcoo8m2B",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def apply_sentece_transformer(X_train, X_test, model=model):\n",
        "  model = model\n",
        "  X_train_list = X_train.values.tolist()\n",
        "  X_test_list = X_test.values.tolist()\n",
        "\n",
        "  X_train_vec = model.encode(X_train_list)\n",
        "  X_test_vec = model.encode(X_test_list)\n",
        "\n",
        "  return X_train_vec, X_test_vec"
      ],
      "metadata": {
        "id": "bvhCjLBCGS6I"
      },
      "id": "bvhCjLBCGS6I",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cv_score_st, test_score_st = pipeline(X_train, y_train,\n",
        "                                      X_test, y_test,\n",
        "                                      apply_sentece_transformer,\n",
        "                                      model=model)"
      ],
      "metadata": {
        "id": "nfasjsREjxKD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7daef656-b6aa-45ab-9a79-73ec5ac3b24c"
      },
      "id": "nfasjsREjxKD",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Running: review_text_clean\n",
            "Running: review_text_clean_stop\n",
            "Running: review_text_clean_stem\n",
            "Running: review_text_clean_stop_stem\n",
            "Running: review_text_clean_lemma\n",
            "Running: review_text_clean_stop_lemma\n",
            "\n",
            "Execution time: 02:05:43\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_score_st = create_score_table(cv_score_st, test_score_st)\n",
        "df_score_st"
      ],
      "metadata": {
        "id": "0QtwIkJoDrah",
        "outputId": "f501e535-8561-40a9-b7c8-466ba975ee6b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        }
      },
      "id": "0QtwIkJoDrah",
      "execution_count": null,
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-8fb6ec61-fd31-4138-a501-21b8d7ebfa58\" 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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>train_mean</th>\n",
              "      <th>test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.98304</td>\n",
              "      <td>0.98473</td>\n",
              "      <td>0.98513</td>\n",
              "      <td>0.98466</td>\n",
              "      <td>0.98595</td>\n",
              "      <td>0.984702</td>\n",
              "      <td>0.98493</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.96851</td>\n",
              "      <td>0.96898</td>\n",
              "      <td>0.96987</td>\n",
              "      <td>0.96798</td>\n",
              "      <td>0.97079</td>\n",
              "      <td>0.969226</td>\n",
              "      <td>0.96812</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.96901</td>\n",
              "      <td>0.96887</td>\n",
              "      <td>0.96951</td>\n",
              "      <td>0.97033</td>\n",
              "      <td>0.97038</td>\n",
              "      <td>0.969620</td>\n",
              "      <td>0.97024</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.94964</td>\n",
              "      <td>0.94811</td>\n",
              "      <td>0.94804</td>\n",
              "      <td>0.94925</td>\n",
              "      <td>0.94988</td>\n",
              "      <td>0.948984</td>\n",
              "      <td>0.94946</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97978</td>\n",
              "      <td>0.98189</td>\n",
              "      <td>0.98143</td>\n",
              "      <td>0.98120</td>\n",
              "      <td>0.98371</td>\n",
              "      <td>0.981602</td>\n",
              "      <td>0.98176</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.96496</td>\n",
              "      <td>0.96572</td>\n",
              "      <td>0.96740</td>\n",
              "      <td>0.96469</td>\n",
              "      <td>0.96809</td>\n",
              "      <td>0.966172</td>\n",
              "      <td>0.96597</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8fb6ec61-fd31-4138-a501-21b8d7ebfa58')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-8fb6ec61-fd31-4138-a501-21b8d7ebfa58 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-8fb6ec61-fd31-4138-a501-21b8d7ebfa58');\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",
              "\n",
              "\n",
              "<div id=\"df-4a4d61ff-dc5f-4c2e-b031-3e2ec36e9153\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-4a4d61ff-dc5f-4c2e-b031-3e2ec36e9153')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-4a4d61ff-dc5f-4c2e-b031-3e2ec36e9153 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "text/plain": [
              "                                    0        1        2        3        4  \\\n",
              "review_text_clean             0.98304  0.98473  0.98513  0.98466  0.98595   \n",
              "review_text_clean_stop        0.96851  0.96898  0.96987  0.96798  0.97079   \n",
              "review_text_clean_stem        0.96901  0.96887  0.96951  0.97033  0.97038   \n",
              "review_text_clean_stop_stem   0.94964  0.94811  0.94804  0.94925  0.94988   \n",
              "review_text_clean_lemma       0.97978  0.98189  0.98143  0.98120  0.98371   \n",
              "review_text_clean_stop_lemma  0.96496  0.96572  0.96740  0.96469  0.96809   \n",
              "\n",
              "                              train_mean     test  \n",
              "review_text_clean               0.984702  0.98493  \n",
              "review_text_clean_stop          0.969226  0.96812  \n",
              "review_text_clean_stem          0.969620  0.97024  \n",
              "review_text_clean_stop_stem     0.948984  0.94946  \n",
              "review_text_clean_lemma         0.981602  0.98176  \n",
              "review_text_clean_stop_lemma    0.966172  0.96597  "
            ]
          },
          "execution_count": 22,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Results Analysis"
      ],
      "metadata": {
        "id": "6Q7I6YjC_WPH"
      },
      "id": "6Q7I6YjC_WPH"
    },
    {
      "cell_type": "code",
      "source": [
        "df_scores = {'Bag of Words': df_score_bow, 'TF-IDF':df_score_tfidf, 'W2V CBOW':df_score_cbow,\n",
        "             'W2V Skip-Gram':df_score_skgram, 'FastTest':df_score_FastTest, 'D2V DBOW':df_score_dbow,\n",
        "             'D2V DM':df_score_dm, 'Sent. Transf.':df_score_st}"
      ],
      "metadata": {
        "id": "-VYV-m9k_jvn"
      },
      "id": "-VYV-m9k_jvn",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Grouping all test results\n",
        "df_scores_vect = pd.DataFrame()\n",
        "for key in df_scores.keys():\n",
        "  df_scores_vect[key] = df_scores[key]['test'].values\n",
        "text_type = list(df_score_st.index.values)\n",
        "df_scores_vect = df_scores_vect.rename(index=dict(zip(df_scores_vect.index, text_type)))\n",
        "df_scores_vect['MEAN'] = round(df_scores_vect.mean(axis=1), 5)\n",
        "\n",
        "df_scores_vect"
      ],
      "metadata": {
        "id": "8Suf2i15_jyz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "outputId": "f7badc4e-f36e-4708-e0ae-1e8c633258d2"
      },
      "id": "8Suf2i15_jyz",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                              Bag of Words   TF-IDF  W2V CBOW  W2V Skip-Gram  \\\n",
              "review_text_clean                  0.97790  0.97811   0.97630        0.97596   \n",
              "review_text_clean_stop             0.97048  0.97143   0.97301        0.97302   \n",
              "review_text_clean_stem             0.97894  0.97907   0.97685        0.97648   \n",
              "review_text_clean_stop_stem        0.97228  0.97280   0.97242        0.97279   \n",
              "review_text_clean_lemma            0.97847  0.97882   0.97689        0.97692   \n",
              "review_text_clean_stop_lemma       0.97080  0.97166   0.97201        0.97221   \n",
              "\n",
              "                              FastTest  D2V DBOW   D2V DM  Sent. Transf.  \\\n",
              "review_text_clean              0.97011   0.95209  0.95050        0.98493   \n",
              "review_text_clean_stop         0.97032   0.93693  0.93719        0.96812   \n",
              "review_text_clean_stem         0.97148   0.95312  0.94671        0.97024   \n",
              "review_text_clean_stop_stem    0.96980   0.94034  0.94306        0.94946   \n",
              "review_text_clean_lemma        0.97167   0.95175  0.95374        0.98176   \n",
              "review_text_clean_stop_lemma   0.96919   0.93994  0.93533        0.96597   \n",
              "\n",
              "                                 MEAN  \n",
              "review_text_clean             0.97074  \n",
              "review_text_clean_stop        0.96256  \n",
              "review_text_clean_stem        0.96911  \n",
              "review_text_clean_stop_stem   0.96162  \n",
              "review_text_clean_lemma       0.97125  \n",
              "review_text_clean_stop_lemma  0.96214  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-7d56db76-6221-4103-a992-e1c2fedcb61e\" 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>Bag of Words</th>\n",
              "      <th>TF-IDF</th>\n",
              "      <th>W2V CBOW</th>\n",
              "      <th>W2V Skip-Gram</th>\n",
              "      <th>FastTest</th>\n",
              "      <th>D2V DBOW</th>\n",
              "      <th>D2V DM</th>\n",
              "      <th>Sent. Transf.</th>\n",
              "      <th>MEAN</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>review_text_clean</th>\n",
              "      <td>0.97790</td>\n",
              "      <td>0.97811</td>\n",
              "      <td>0.97630</td>\n",
              "      <td>0.97596</td>\n",
              "      <td>0.97011</td>\n",
              "      <td>0.95209</td>\n",
              "      <td>0.95050</td>\n",
              "      <td>0.98493</td>\n",
              "      <td>0.97074</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop</th>\n",
              "      <td>0.97048</td>\n",
              "      <td>0.97143</td>\n",
              "      <td>0.97301</td>\n",
              "      <td>0.97302</td>\n",
              "      <td>0.97032</td>\n",
              "      <td>0.93693</td>\n",
              "      <td>0.93719</td>\n",
              "      <td>0.96812</td>\n",
              "      <td>0.96256</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stem</th>\n",
              "      <td>0.97894</td>\n",
              "      <td>0.97907</td>\n",
              "      <td>0.97685</td>\n",
              "      <td>0.97648</td>\n",
              "      <td>0.97148</td>\n",
              "      <td>0.95312</td>\n",
              "      <td>0.94671</td>\n",
              "      <td>0.97024</td>\n",
              "      <td>0.96911</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_stem</th>\n",
              "      <td>0.97228</td>\n",
              "      <td>0.97280</td>\n",
              "      <td>0.97242</td>\n",
              "      <td>0.97279</td>\n",
              "      <td>0.96980</td>\n",
              "      <td>0.94034</td>\n",
              "      <td>0.94306</td>\n",
              "      <td>0.94946</td>\n",
              "      <td>0.96162</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_lemma</th>\n",
              "      <td>0.97847</td>\n",
              "      <td>0.97882</td>\n",
              "      <td>0.97689</td>\n",
              "      <td>0.97692</td>\n",
              "      <td>0.97167</td>\n",
              "      <td>0.95175</td>\n",
              "      <td>0.95374</td>\n",
              "      <td>0.98176</td>\n",
              "      <td>0.97125</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>review_text_clean_stop_lemma</th>\n",
              "      <td>0.97080</td>\n",
              "      <td>0.97166</td>\n",
              "      <td>0.97201</td>\n",
              "      <td>0.97221</td>\n",
              "      <td>0.96919</td>\n",
              "      <td>0.93994</td>\n",
              "      <td>0.93533</td>\n",
              "      <td>0.96597</td>\n",
              "      <td>0.96214</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-7d56db76-6221-4103-a992-e1c2fedcb61e')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-7d56db76-6221-4103-a992-e1c2fedcb61e 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-7d56db76-6221-4103-a992-e1c2fedcb61e');\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",
              "\n",
              "\n",
              "<div id=\"df-ec462c19-1146-45e8-b6d3-cae5770b3c65\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-ec462c19-1146-45e8-b6d3-cae5770b3c65')\"\n",
              "            title=\"Suggest charts.\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\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-quickchart: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-quickchart {\n",
              "    background-color: #3B4455;\n",
              "    fill: #D2E3FC;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart: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",
              "    async function quickchart(key) {\n",
              "      const charts = await google.colab.kernel.invokeFunction(\n",
              "          'suggestCharts', [key], {});\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-ec462c19-1146-45e8-b6d3-cae5770b3c65 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 72
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(8, 4))\n",
        "columns_name = {'level_0':'Text Type', 'level_1': 'Vect Method', 0: 'Score(ROC)'}\n",
        "df_scores_vect_stack = df_scores_vect.stack().reset_index().rename(columns=columns_name)\n",
        "df_scores_vect_stack['Vect Method & Text'] =  df_scores_vect_stack['Vect Method'] + ' ' + df_scores_vect_stack['Text Type']\n",
        "\n",
        "ax = sns.boxplot(x='Vect Method',\n",
        "                 y='Score(ROC)',\n",
        "                 data=df_scores_vect_stack)\n",
        "ax.set_xticklabels(ax.get_xticklabels(), rotation=90)\n",
        "plt.title('Score by Vectorization Method')\n",
        "plt.plot()"
      ],
      "metadata": {
        "id": "I2f36BWsqWRn",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 542
        },
        "outputId": "b7b267a8-56e2-4d2f-b3bf-dd17cbf30fa2"
      },
      "id": "I2f36BWsqWRn",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[]"
            ]
          },
          "metadata": {},
          "execution_count": 42
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "c9ae3067",
      "metadata": {
        "id": "c9ae3067",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 436
        },
        "outputId": "1e3d03fc-be48-4526-ceb6-302e6a96fd3b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[]"
            ]
          },
          "metadata": {},
          "execution_count": 37
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 700x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.figure(figsize=(7, 4))\n",
        "df_top10 = df_scores_vect_stack.sort_values(by='Score(ROC)', ascending=False).head(10)\n",
        "ax = sns.barplot(x='Score(ROC)', y='Vect Method & Text',\n",
        "                 data=df_top10,\n",
        "                 orient='h', color='steelblue')\n",
        "ax.set_yticklabels(ax.get_yticklabels(), fontsize=9)\n",
        "plt.xlim(0.8, 1)\n",
        "for i, v in enumerate(df_top10['Score(ROC)']):\n",
        "    ax.text(v + 0.002, i, str(v), color='black', va='center', fontsize=6)\n",
        "plt.title('Top 10 Best Aproach')\n",
        "plt.plot()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10, 6))\n",
        "ax = sns.barplot(x='Vect Method', y='Score(ROC)',\n",
        "                 hue='Text Type',\n",
        "                 data=df_scores_vect_stack,\n",
        "                 )\n",
        "ax.set_xticklabels(ax.get_xticklabels(), rotation=90)\n",
        "plt.ylim(0.8, 1)\n",
        "ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))\n",
        "plt.title('Score Comparison')\n",
        "plt.plot()"
      ],
      "metadata": {
        "id": "sp2tP8hRr1kP",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 686
        },
        "outputId": "07cf6860-f5a6-481f-b767-3aa7c28404e6"
      },
      "id": "sp2tP8hRr1kP",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[]"
            ]
          },
          "metadata": {},
          "execution_count": 27
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.10.9"
    },
    "toc": {
      "base_numbering": 1,
      "nav_menu": {},
      "number_sections": false,
      "sideBar": true,
      "skip_h1_title": false,
      "title_cell": "Table of Contents",
      "title_sidebar": "Contents",
      "toc_cell": false,
      "toc_position": {
        "height": "calc(100% - 180px)",
        "left": "10px",
        "top": "150px",
        "width": "280.775px"
      },
      "toc_section_display": true,
      "toc_window_display": true
    },
    "varInspector": {
      "cols": {
        "lenName": 16,
        "lenType": 16,
        "lenVar": 40
      },
      "kernels_config": {
        "python": {
          "delete_cmd_postfix": "",
          "delete_cmd_prefix": "del ",
          "library": "var_list.py",
          "varRefreshCmd": "print(var_dic_list())"
        },
        "r": {
          "delete_cmd_postfix": ") ",
          "delete_cmd_prefix": "rm(",
          "library": "var_list.r",
          "varRefreshCmd": "cat(var_dic_list()) "
        }
      },
      "types_to_exclude": [
        "module",
        "function",
        "builtin_function_or_method",
        "instance",
        "_Feature"
      ],
      "window_display": false
    },
    "colab": {
      "provenance": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}