exploration_preprocessing.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "mount_file_id": "1ujmXbi61rH3Hh4upS0r25aVw97dXikGE",
      "authorship_tag": "ABX9TyPCv5V9sHiVorKLKlpRpvqr"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# **Unzipping**"
      ],
      "metadata": {
        "id": "MpH5Pgsb_5tu"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "67_UCoiqAAam"
      },
      "outputs": [],
      "source": [
        "pip freeze"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pwd"
      ],
      "metadata": {
        "id": "j_tN7oMYKJcY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ls "
      ],
      "metadata": {
        "id": "e7XDim1lKnp1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!unzip drive/MyDrive/pubhealth/data/PUBHEALTH.zip -d drive/MyDrive/pubhealth/data/raw"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SeEBYed2KpDJ",
        "outputId": "e278cc13-741b-421d-9096-d47edd5429ac"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Archive:  drive/MyDrive/pubhealth/data/PUBHEALTH.zip\n",
            "replace drive/MyDrive/pubhealth/data/raw/PUBHEALTH/train.tsv? [y]es, [n]o, [A]ll, [N]one, [r]ename: N\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "ls drive/MyDrive/pubhealth/data/raw/PUBHEALTH/"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ClN7lozOKzpu",
        "outputId": "1818bb39-8169-4ed6-eda5-d6f34f5c5a9b"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "dev.tsv  test.tsv  train.tsv\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Exploratory Analysis and Preprocessing**"
      ],
      "metadata": {
        "id": "J0eDJukK_l4c"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd"
      ],
      "metadata": {
        "id": "-DWkrPKWMUtU"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train = pd.read_csv('drive/MyDrive/pubhealth/data/raw/PUBHEALTH/train.tsv', sep='\\t')\n",
        "dev = pd.read_csv('drive/MyDrive/pubhealth/data/raw/PUBHEALTH/dev.tsv', sep='\\t')\n",
        "test = pd.read_csv('drive/MyDrive/pubhealth/data/raw/PUBHEALTH/test.tsv', sep='\\t',index_col=0)"
      ],
      "metadata": {
        "id": "PpKk9nYINQOw"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "6BcwzkLa2uqv",
        "outputId": "22ed9b19-edce-4eab-ae76-0a5ff2433f11"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  claim_id                                              claim  \\\n",
              "0    15661  \"The money the Clinton Foundation took from fr...   \n",
              "1     9893    Annual Mammograms May Have More False-Positives   \n",
              "2    11358  SBRT Offers Prostate Cancer Patients High Canc...   \n",
              "3    10166  Study: Vaccine for Breast, Ovarian Cancer Has ...   \n",
              "4    11276  Some appendicitis cases may not require ’emerg...   \n",
              "\n",
              "       date_published                                        explanation  \\\n",
              "0      April 26, 2015  \"Gingrich said the Clinton Foundation \"\"took m...   \n",
              "1    October 18, 2011  This article reports on the results of a study...   \n",
              "2  September 28, 2016  This news release describes five-year outcomes...   \n",
              "3    November 8, 2011  While the story does many things well, the ove...   \n",
              "4  September 20, 2010  We really don’t understand why only a handful ...   \n",
              "\n",
              "                                       fact_checkers  \\\n",
              "0                                      Katie Sanders   \n",
              "1                                                      \n",
              "2  Mary Chris Jaklevic,Steven J. Atlas, MD, MPH,K...   \n",
              "3                                                      \n",
              "4                                                      \n",
              "\n",
              "                                           main_text  \\\n",
              "0  \"Hillary Clinton is in the political crosshair...   \n",
              "1  While the financial costs of screening mammogr...   \n",
              "2  The news release quotes lead researcher Robert...   \n",
              "3  The story does discuss costs, but the framing ...   \n",
              "4  \"Although the story didn’t cite the cost of ap...   \n",
              "\n",
              "                                             sources    label  \\\n",
              "0  https://www.wsj.com/articles/clinton-foundatio...    false   \n",
              "1                                                     mixture   \n",
              "2  https://www.healthnewsreview.org/wp-content/up...  mixture   \n",
              "3  http://clinicaltrials.gov/ct2/results?term=can...     true   \n",
              "4                                                        true   \n",
              "\n",
              "                                      subjects  \n",
              "0  Foreign Policy, PunditFact, Newt Gingrich,   \n",
              "1               Screening,WebMD,women's health  \n",
              "2      Association/Society news release,Cancer  \n",
              "3                  Cancer,WebMD,women's health  \n",
              "4                                               "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-b939d691-7e70-4817-ab48-fbecca1da0f5\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>claim_id</th>\n",
              "      <th>claim</th>\n",
              "      <th>date_published</th>\n",
              "      <th>explanation</th>\n",
              "      <th>fact_checkers</th>\n",
              "      <th>main_text</th>\n",
              "      <th>sources</th>\n",
              "      <th>label</th>\n",
              "      <th>subjects</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>15661</td>\n",
              "      <td>\"The money the Clinton Foundation took from fr...</td>\n",
              "      <td>April 26, 2015</td>\n",
              "      <td>\"Gingrich said the Clinton Foundation \"\"took m...</td>\n",
              "      <td>Katie Sanders</td>\n",
              "      <td>\"Hillary Clinton is in the political crosshair...</td>\n",
              "      <td>https://www.wsj.com/articles/clinton-foundatio...</td>\n",
              "      <td>false</td>\n",
              "      <td>Foreign Policy, PunditFact, Newt Gingrich,</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>9893</td>\n",
              "      <td>Annual Mammograms May Have More False-Positives</td>\n",
              "      <td>October 18, 2011</td>\n",
              "      <td>This article reports on the results of a study...</td>\n",
              "      <td></td>\n",
              "      <td>While the financial costs of screening mammogr...</td>\n",
              "      <td></td>\n",
              "      <td>mixture</td>\n",
              "      <td>Screening,WebMD,women's health</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>11358</td>\n",
              "      <td>SBRT Offers Prostate Cancer Patients High Canc...</td>\n",
              "      <td>September 28, 2016</td>\n",
              "      <td>This news release describes five-year outcomes...</td>\n",
              "      <td>Mary Chris Jaklevic,Steven J. Atlas, MD, MPH,K...</td>\n",
              "      <td>The news release quotes lead researcher Robert...</td>\n",
              "      <td>https://www.healthnewsreview.org/wp-content/up...</td>\n",
              "      <td>mixture</td>\n",
              "      <td>Association/Society news release,Cancer</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>10166</td>\n",
              "      <td>Study: Vaccine for Breast, Ovarian Cancer Has ...</td>\n",
              "      <td>November 8, 2011</td>\n",
              "      <td>While the story does many things well, the ove...</td>\n",
              "      <td></td>\n",
              "      <td>The story does discuss costs, but the framing ...</td>\n",
              "      <td>http://clinicaltrials.gov/ct2/results?term=can...</td>\n",
              "      <td>true</td>\n",
              "      <td>Cancer,WebMD,women's health</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>11276</td>\n",
              "      <td>Some appendicitis cases may not require ’emerg...</td>\n",
              "      <td>September 20, 2010</td>\n",
              "      <td>We really don’t understand why only a handful ...</td>\n",
              "      <td></td>\n",
              "      <td>\"Although the story didn’t cite the cost of ap...</td>\n",
              "      <td></td>\n",
              "      <td>true</td>\n",
              "      <td></td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b939d691-7e70-4817-ab48-fbecca1da0f5')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-b939d691-7e70-4817-ab48-fbecca1da0f5 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-b939d691-7e70-4817-ab48-fbecca1da0f5');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#verifying the columns coherence\n",
        "train.columns.to_list() == dev.columns.to_list()== test.columns.to_list()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9DGVZUyP33pg",
        "outputId": "5fe5296a-fcbe-44f5-942a-b44855af180d"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Intuitive column selection though PCA study can be done based on tf-idf can be conducted**"
      ],
      "metadata": {
        "id": "DqydOw3UxiNu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "train_df = train[['label', 'claim','main_text','explanation','subjects','sources']]\n",
        "dev_df = dev[['label', 'claim','main_text','explanation','subjects']]\n",
        "test_df = test[['label', 'claim','main_text','explanation','subjects']]"
      ],
      "metadata": {
        "id": "dCg1VA6R3m7s"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_df.describe()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "id": "1Gaj7sDS73i7",
        "outputId": "9a3a2234-7f43-4086-feba-3b4198aeefe1"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       label                                              claim  \\\n",
              "count   9805                                               9824   \n",
              "unique     5                                               9824   \n",
              "top     true  \"The money the Clinton Foundation took from fr...   \n",
              "freq    5078                                                  1   \n",
              "\n",
              "                                                main_text explanation  \\\n",
              "count                                                9806        9824   \n",
              "unique                                               9589        9675   \n",
              "top     Doctors in India have been successful in treat...       false   \n",
              "freq                                                   12           3   \n",
              "\n",
              "           subjects sources  \n",
              "count          9804    9804  \n",
              "unique         5811    5063  \n",
              "top     Health News          \n",
              "freq           1242    4550  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-0552ef30-f9b4-483b-9d96-3fb5ebcdd67c\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label</th>\n",
              "      <th>claim</th>\n",
              "      <th>main_text</th>\n",
              "      <th>explanation</th>\n",
              "      <th>subjects</th>\n",
              "      <th>sources</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>9805</td>\n",
              "      <td>9824</td>\n",
              "      <td>9806</td>\n",
              "      <td>9824</td>\n",
              "      <td>9804</td>\n",
              "      <td>9804</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>unique</th>\n",
              "      <td>5</td>\n",
              "      <td>9824</td>\n",
              "      <td>9589</td>\n",
              "      <td>9675</td>\n",
              "      <td>5811</td>\n",
              "      <td>5063</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>top</th>\n",
              "      <td>true</td>\n",
              "      <td>\"The money the Clinton Foundation took from fr...</td>\n",
              "      <td>Doctors in India have been successful in treat...</td>\n",
              "      <td>false</td>\n",
              "      <td>Health News</td>\n",
              "      <td></td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>freq</th>\n",
              "      <td>5078</td>\n",
              "      <td>1</td>\n",
              "      <td>12</td>\n",
              "      <td>3</td>\n",
              "      <td>1242</td>\n",
              "      <td>4550</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-0552ef30-f9b4-483b-9d96-3fb5ebcdd67c')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-0552ef30-f9b4-483b-9d96-3fb5ebcdd67c 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-0552ef30-f9b4-483b-9d96-3fb5ebcdd67c');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_df.isna().sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zUkvvKgQ784Y",
        "outputId": "2013e7c0-5f9e-407d-8fa4-d3573a27b1e2"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "label          27\n",
              "claim           8\n",
              "main_text      26\n",
              "explanation     8\n",
              "subjects       28\n",
              "sources        28\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dev_df.describe()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "id": "ozDdi8t6-dbV",
        "outputId": "f4a8031a-8fc5-4539-d427-5b031bbfafa9"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       label                                              claim  \\\n",
              "count   1215                                               1219   \n",
              "unique     5                                               1219   \n",
              "top     true   A baby died at an unnamed medical facility be...   \n",
              "freq     629                                                  1   \n",
              "\n",
              "                                                main_text  \\\n",
              "count                                                1217   \n",
              "unique                                               1212   \n",
              "top     Beta-Propiolactone is in vaccines and is known...   \n",
              "freq                                                    3   \n",
              "\n",
              "                                              explanation     subjects  \n",
              "count                                                1219         1214  \n",
              "unique                                               1218          795  \n",
              "top     The largest study of its kind found new eviden...  Health News  \n",
              "freq                                                    2          160  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-6771c8c4-5c4e-4b25-8b2f-0379ba72fe4c\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label</th>\n",
              "      <th>claim</th>\n",
              "      <th>main_text</th>\n",
              "      <th>explanation</th>\n",
              "      <th>subjects</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>1215</td>\n",
              "      <td>1219</td>\n",
              "      <td>1217</td>\n",
              "      <td>1219</td>\n",
              "      <td>1214</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>unique</th>\n",
              "      <td>5</td>\n",
              "      <td>1219</td>\n",
              "      <td>1212</td>\n",
              "      <td>1218</td>\n",
              "      <td>795</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>top</th>\n",
              "      <td>true</td>\n",
              "      <td>A baby died at an unnamed medical facility be...</td>\n",
              "      <td>Beta-Propiolactone is in vaccines and is known...</td>\n",
              "      <td>The largest study of its kind found new eviden...</td>\n",
              "      <td>Health News</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>freq</th>\n",
              "      <td>629</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>160</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6771c8c4-5c4e-4b25-8b2f-0379ba72fe4c')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-6771c8c4-5c4e-4b25-8b2f-0379ba72fe4c 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-6771c8c4-5c4e-4b25-8b2f-0379ba72fe4c');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dev_df.isna().sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LSRoH7JgBN9j",
        "outputId": "2fc5153e-9ab7-47d4-cbb4-80f1cea83869"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "label          6\n",
              "claim          2\n",
              "main_text      4\n",
              "explanation    2\n",
              "subjects       7\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# dev_df.iloc[10].to_list() "
      ],
      "metadata": {
        "id": "zFQJC55kBYBe"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_df.describe()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "id": "aaAiH9LJBkls",
        "outputId": "1b20cb99-114a-4c07-ef22-b66b69237490"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       label                                              claim  \\\n",
              "count   1233                                               1235   \n",
              "unique     4                                               1235   \n",
              "top     true  A mother revealed to her child in a letter aft...   \n",
              "freq     599                                                  1   \n",
              "\n",
              "                                                main_text  \\\n",
              "count                                                1235   \n",
              "unique                                               1231   \n",
              "top     If the new coronavirus gets into your mouth, d...   \n",
              "freq                                                    2   \n",
              "\n",
              "                                              explanation     subjects  \n",
              "count                                                1235         1233  \n",
              "unique                                               1234          840  \n",
              "top     Plant-based food startup JUST, shortly to laun...  Health News  \n",
              "freq                                                    2          142  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-fa4eab4d-07ff-4d3a-ae69-f95bdac2a117\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label</th>\n",
              "      <th>claim</th>\n",
              "      <th>main_text</th>\n",
              "      <th>explanation</th>\n",
              "      <th>subjects</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>1233</td>\n",
              "      <td>1235</td>\n",
              "      <td>1235</td>\n",
              "      <td>1235</td>\n",
              "      <td>1233</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>unique</th>\n",
              "      <td>4</td>\n",
              "      <td>1235</td>\n",
              "      <td>1231</td>\n",
              "      <td>1234</td>\n",
              "      <td>840</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>top</th>\n",
              "      <td>true</td>\n",
              "      <td>A mother revealed to her child in a letter aft...</td>\n",
              "      <td>If the new coronavirus gets into your mouth, d...</td>\n",
              "      <td>Plant-based food startup JUST, shortly to laun...</td>\n",
              "      <td>Health News</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>freq</th>\n",
              "      <td>599</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>142</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-fa4eab4d-07ff-4d3a-ae69-f95bdac2a117')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-fa4eab4d-07ff-4d3a-ae69-f95bdac2a117 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-fa4eab4d-07ff-4d3a-ae69-f95bdac2a117');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_df.isna().sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cFE79RcUD7Qi",
        "outputId": "e708ac35-baf8-40f4-ded7-2339f4af5e1c"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "label          2\n",
              "claim          0\n",
              "main_text      0\n",
              "explanation    0\n",
              "subjects       2\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**There is seemingly something wrong with the labels of train and dev dataframes therefore this should be fixed + droping the na values**"
      ],
      "metadata": {
        "id": "cPk_m5ZlTGSY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(train_df.label.unique())\n",
        "print(dev_df.label.unique())\n",
        "print(test_df.label.unique())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eziW0ROTTF-Y",
        "outputId": "23938950-be17-4253-fbb3-8eb24f89037e"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['false' 'mixture' 'true' 'unproven' nan 'snopes']\n",
            "['unproven' 'true' 'false' 'mixture' nan\n",
            " 'National, Candidate Biography, Donald Trump, ']\n",
            "['false' 'true' 'unproven' 'mixture' nan]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**It shows clearly that there are , additionally to the nan values other wrong/outlier labels.**"
      ],
      "metadata": {
        "id": "wQRCxKbzafSN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_df = train_df.dropna()\n",
        "dev_df = dev_df.dropna()\n",
        "test_df = test_df.dropna()"
      ],
      "metadata": {
        "id": "Ct1n1qsobBje"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_df[train_df.label == 'snopes'].size == 0 "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Qf5p7VKBb4Dg",
        "outputId": "8e9f947c-9bcc-4f79-d195-59ef83d1e596"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(train_df.label.unique())\n",
        "print(dev_df.label.unique())\n",
        "print(test_df.label.unique())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JL9kgouLdQeO",
        "outputId": "5e617da1-66d7-4778-d470-018d79c44bef"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['false' 'mixture' 'true' 'unproven']\n",
            "['unproven' 'true' 'false' 'mixture']\n",
            "['false' 'true' 'unproven' 'mixture']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**To have an idea about sentences lengths for eac feature.**\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "MSE7Q3WrvzKH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#helper function to get average lengths of the texts in each columns.\n",
        "def tolist(col):\n",
        "  '''column of string to list\n",
        "  '''\n",
        "  return col.to_list()\n",
        "\n",
        "def avglist(strlist:list):\n",
        "  '''get average of a list of strings\n",
        "  '''\n",
        "  return sum([len(s) for s in strlist])/len(strlist)\n",
        "\n",
        "def average_columns_len(df,columns:list):\n",
        "  ''' given a list of columns in data frame , the function should return s dict having as values an average of the lenght of sentence present in the columns which are keys\n",
        "  '''\n",
        "  res = dict()\n",
        "  for col in columns:\n",
        "    res[col]=round(avglist(tolist(df[col])),2)\n",
        "  return res\n"
      ],
      "metadata": {
        "id": "3b15azAzb6q-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "average_columns_len(pub_health_train,['claim','main_text','subjects','explanation','sources'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4-tkmbET6c1L",
        "outputId": "7e0fa3e2-c05b-48f8-c0c9-cc3b548e65a4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'claim': 85.69,\n",
              " 'main_text': 4376.84,\n",
              " 'subjects': 40.08,\n",
              " 'explanation': 482.97,\n",
              " 'sources': 390.47}"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#"
      ],
      "metadata": {
        "id": "M6E6SrIpwkbH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Visualize label distribution over the dataset**"
      ],
      "metadata": {
        "id": "0CTOjz7Adcjo"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "labels = ['false', 'mixture', 'true', 'unproven']\n",
        "sns.set_theme(style=\"whitegrid\")\n",
        "fig, ax = plt.subplots(1,3,constrained_layout=True, figsize=(16, 5))\n",
        "axesSub = sns.countplot(x=\"label\", data=train_df, ax=ax[0], order=labels)\n",
        "axesSub.set_title('train')\n",
        "axesSub = sns.countplot(x=\"label\", data=dev_df, ax=ax[1], order=labels)\n",
        "axesSub.set_title('dev')\n",
        "axesSub = sns.countplot(x=\"label\", data=test_df, ax=ax[2], order=labels)\n",
        "axesSub.set_title('test')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 402
        },
        "id": "Rbul9CFMbke-",
        "outputId": "636a9e5a-ba94-444f-8623-6470b213bec8"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'test')"
            ]
          },
          "metadata": {},
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1152x360 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "highly imbalanced data, thinking either cuting down the 'true' labeld data"
      ],
      "metadata": {
        "id": "bE6US1fBozc8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# thinking about conducting a PCA on separated columns/features to see if any pattern appears for "
      ],
      "metadata": {
        "id": "LnUMOVTWILcN"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# save the clean data\n",
        "train_df.to_csv('drive/MyDrive/pubhealth/data/clean/exp1/clean_train.tsv', sep='\\t', index=False)\n",
        "dev_df.to_csv('drive/MyDrive/pubhealth/data/clean/exp1/clean_dev.tsv', sep='\\t', index=False)\n",
        "test_df.to_csv('drive/MyDrive/pubhealth/data/clean/exp1/clean_test.tsv', sep='\\t', index=False)"
      ],
      "metadata": {
        "id": "63lwvKh3wEWO"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "5ABKN1_mygAW"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}