Spam_Detection_Model__Small_Data_Pipelines.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/sflyranger/Spam-Detection-Pipelines-/blob/main/Spam_Detection_Model__Small_Data_Pipelines.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color='Pickle'>Spam Detection Modeling on Imbalanced Data</font>**\n",
        "\n",
        "This notebook serves as a practical guide for building a classification model using pipelines. The process is broken down into two main parts:\n",
        "\n",
        "- **<font color='Pickle'>Pipeline Creation:</font>** In this first portion, I’ll be creating **three separate pipelines** using smaller samples of the dataset. Each pipeline will be designed to address the challenges of working with imbalanced data, particularly in the context of spam detection.\n",
        "  \n",
        "- **<font color='Pickle'>Model Selection & Final Pipeline:</font>** The second notebook will focus on evaluating the performance of these three pipelines. I’ll select the **best-performing model** from the three and use it to create a final, robust pipeline. This final pipeline will be trained on a **larger random sample** of the dataset to enhance the model’s generalization and effectiveness.\n",
        "\n",
        "This structured approach ensures that we carefully evaluate multiple strategies before committing to the final model, especially given the complexities of working with imbalanced data in spam detection."
      ],
      "metadata": {
        "id": "0sVkAt4zo6wY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Import necessary libraries\n",
        "import pandas as pd\n",
        "from pathlib import Path\n",
        "import sys\n",
        "import spacy\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Import the joblib library for saving and loading models\n",
        "import joblib\n",
        "\n",
        "# Import scikit-learn classes for building models\n",
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.base import TransformerMixin, BaseEstimator\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.compose import ColumnTransformer\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "from sklearn.model_selection import learning_curve\n",
        "\n",
        "\n",
        "# XGB classifier import\n",
        "from xgboost import XGBClassifier\n",
        "\n",
        "from imblearn.over_sampling import SMOTE\n",
        "\n",
        "# Import required stopwords\n",
        "import nltk\n",
        "from nltk.corpus import stopwords as nltk_stopwords  # Stopwords corpus\n",
        "nltk.download('stopwords')  # Download the stopwords corpus\n",
        "\n",
        "# Import the scipy library for working with sparse matrices\n",
        "from scipy.sparse import csr_matrix\n",
        "# importing counter for class distributions\n",
        "from collections import Counter"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IwTmVhCyCukQ",
        "outputId": "e2a4d01b-39f2-470f-fa35-4c73e9eff8e1"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Package stopwords is already up-to-date!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%load_ext autoreload\n",
        "%autoreload 2"
      ],
      "metadata": {
        "id": "aZBPh-HqClzf",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3ef07b01-4a8f-44fb-b139-a2473f093c24"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The autoreload extension is already loaded. To reload it, use:\n",
            "  %reload_ext autoreload\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "if 'google.colab' in str(get_ipython()):  # If the code is running in Colab\n",
        "    # mount google drive\n",
        "    from google.colab import drive\n",
        "    drive.mount('/content/drive')\n",
        "\n",
        "\n",
        "    !pip install -U nltk -qq # -U ensures the most updated version\n",
        "    !pip install -U spacy -qq # -qq limits the output commands to only the most important messages\n",
        "    !python -m spacy download en_core_web_sm -qq # Small model from spacy\n",
        "    !pip install -U pyspellchecker -qq. #pyspell checker for extendedmanualfeatures class\n",
        "    # set the base path to a Google Drive folder\n",
        "    base_path = '/content/drive/MyDrive/Colab Notebooks'\n",
        "\n",
        "    sys.path.append('/content/drive/MyDrive/Colab Notebooks/custom-functions')# enabling custom function usage"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eg9Tej6EB_8S",
        "outputId": "95647043-b8af-4f09-ad47-de9c24b3c667"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n",
            "\u001b[38;5;2m✔ Download and installation successful\u001b[0m\n",
            "You can now load the package via spacy.load('en_core_web_sm')\n",
            "\u001b[38;5;3m⚠ Restart to reload dependencies\u001b[0m\n",
            "If you are in a Jupyter or Colab notebook, you may need to restart Python in\n",
            "order to load all the package's dependencies. You can do this by selecting the\n",
            "'Restart kernel' or 'Restart runtime' option.\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m6.8/6.8 MB\u001b[0m \u001b[31m17.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Double checking path\n",
        "sys.path"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EwpehImPDhyU",
        "outputId": "b475f64e-0fa6-482d-90df-c0e0fee029ee"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content',\n",
              " '/env/python',\n",
              " '/usr/lib/python310.zip',\n",
              " '/usr/lib/python3.10',\n",
              " '/usr/lib/python3.10/lib-dynload',\n",
              " '',\n",
              " '/usr/local/lib/python3.10/dist-packages',\n",
              " '/usr/lib/python3/dist-packages',\n",
              " '/usr/local/lib/python3.10/dist-packages/IPython/extensions',\n",
              " '/usr/local/lib/python3.10/dist-packages/setuptools/_vendor',\n",
              " '/root/.ipython',\n",
              " '/content/drive/MyDrive/Colab Notebooks/custom-functions']"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Appending the data set path\n",
        "base_folder = Path(base_path)\n",
        "data_folder = base_folder/'datasets/spam_ds'\n",
        "model_folder = base_folder/'models/nlp_fall_2024/spam'"
      ],
      "metadata": {
        "id": "nVnpdpa3GcNs"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Making the directory for my models in google drive.\n",
        "model_folder.mkdir(parents=True, exist_ok=True)"
      ],
      "metadata": {
        "id": "g7kRLf6MlKf3"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# location of the spam data set\n",
        "spam_file = data_folder/'spam.csv'\n",
        "\n",
        "# Creating the dataframe object for the Spam data set\n",
        "spam_df = pd.read_csv(spam_file, encoding='ISO-8859-1')"
      ],
      "metadata": {
        "id": "-0xlX6b9I0Jx"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color='Pickle'> **We can see below that there are three additional columns within the data-set that have no meaning. I will start by renaming the v1 and v2 columns to label and message respectively then deleting the 3 unnamed columns.**</font>"
      ],
      "metadata": {
        "id": "HIuuABcuLdfw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spam_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "Gyd-wlYyLR4t",
        "outputId": "ecd97ad4-5af2-4959-ac25-ef8faff7215a"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     v1                                                 v2 Unnamed: 2  \\\n",
              "0   ham  Go until jurong point, crazy.. Available only ...        NaN   \n",
              "1   ham                      Ok lar... Joking wif u oni...        NaN   \n",
              "2  spam  Free entry in 2 a wkly comp to win FA Cup fina...        NaN   \n",
              "3   ham  U dun say so early hor... U c already then say...        NaN   \n",
              "4   ham  Nah I don't think he goes to usf, he lives aro...        NaN   \n",
              "\n",
              "  Unnamed: 3 Unnamed: 4  \n",
              "0        NaN        NaN  \n",
              "1        NaN        NaN  \n",
              "2        NaN        NaN  \n",
              "3        NaN        NaN  \n",
              "4        NaN        NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-5439bc4e-7c00-4ce8-9442-306018469259\" 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>v1</th>\n",
              "      <th>v2</th>\n",
              "      <th>Unnamed: 2</th>\n",
              "      <th>Unnamed: 3</th>\n",
              "      <th>Unnamed: 4</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ham</td>\n",
              "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ham</td>\n",
              "      <td>Ok lar... Joking wif u oni...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>spam</td>\n",
              "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>ham</td>\n",
              "      <td>U dun say so early hor... U c already then say...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>ham</td>\n",
              "      <td>Nah I don't think he goes to usf, he lives aro...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</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-5439bc4e-7c00-4ce8-9442-306018469259')\"\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-5439bc4e-7c00-4ce8-9442-306018469259 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-5439bc4e-7c00-4ce8-9442-306018469259');\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-9b8c7fff-f5dd-4444-b4da-c6aefaa5361a\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-9b8c7fff-f5dd-4444-b4da-c6aefaa5361a')\"\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",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-9b8c7fff-f5dd-4444-b4da-c6aefaa5361a button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "spam_df",
              "summary": "{\n  \"name\": \"spam_df\",\n  \"rows\": 5572,\n  \"fields\": [\n    {\n      \"column\": \"v1\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"spam\",\n          \"ham\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"v2\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 5169,\n        \"samples\": [\n          \"Did u download the fring app?\",\n          \"Pass dis to all ur contacts n see wat u get! Red;i'm in luv wid u. Blue;u put a smile on my face. Purple;u r realy hot. Pink;u r so swt. Orange;i thnk i lyk u. Green;i realy wana go out wid u. Yelow;i wnt u bck. Black;i'm jealous of u. Brown;i miss you Nw plz giv me one color\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Unnamed: 2\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 43,\n        \"samples\": [\n          \" GOD said\",\n          \" SHE SHUDVETOLD U. DID URGRAN KNOW?NEWAY\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Unnamed: 3\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 10,\n        \"samples\": [\n          \" \\\\\\\"OH No! COMPETITION\\\\\\\". Who knew\",\n          \" why to miss them\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"Unnamed: 4\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          \"GNT:-)\\\"\",\n          \" one day these two will become FREINDS FOREVER!\\\"\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Setting up label encoder for future use\n",
        "label_encoder = LabelEncoder()"
      ],
      "metadata": {
        "id": "qg96-XFzXhT5"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Renaming\n",
        "spam_df= spam_df.rename(columns = {'v1': 'label','v2': 'message'})\n",
        "\n",
        "# Deleting\n",
        "spam_df = spam_df.drop(columns = {'Unnamed: 2','Unnamed: 3','Unnamed: 4'})\n",
        "\n",
        "# Resetting index\n",
        "spam_df = spam_df.reset_index(drop = True)\n",
        "\n",
        "# New df\n",
        "spam_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "rtj3NZYNL4fQ",
        "outputId": "871944be-f690-46de-c94b-c674b3f50360"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  label                                            message\n",
              "0   ham  Go until jurong point, crazy.. Available only ...\n",
              "1   ham                      Ok lar... Joking wif u oni...\n",
              "2  spam  Free entry in 2 a wkly comp to win FA Cup fina...\n",
              "3   ham  U dun say so early hor... U c already then say...\n",
              "4   ham  Nah I don't think he goes to usf, he lives aro..."
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-67964329-bb49-407c-b9bb-8df897485536\" 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>message</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ham</td>\n",
              "      <td>Go until jurong point, crazy.. Available only ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>ham</td>\n",
              "      <td>Ok lar... Joking wif u oni...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>spam</td>\n",
              "      <td>Free entry in 2 a wkly comp to win FA Cup fina...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>ham</td>\n",
              "      <td>U dun say so early hor... U c already then say...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>ham</td>\n",
              "      <td>Nah I don't think he goes to usf, he lives aro...</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-67964329-bb49-407c-b9bb-8df897485536')\"\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-67964329-bb49-407c-b9bb-8df897485536 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-67964329-bb49-407c-b9bb-8df897485536');\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-e3ac633a-6a63-44a5-a02d-e3893e3fd267\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-e3ac633a-6a63-44a5-a02d-e3893e3fd267')\"\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",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-e3ac633a-6a63-44a5-a02d-e3893e3fd267 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "spam_df",
              "summary": "{\n  \"name\": \"spam_df\",\n  \"rows\": 5572,\n  \"fields\": [\n    {\n      \"column\": \"label\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 2,\n        \"samples\": [\n          \"spam\",\n          \"ham\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"message\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 5169,\n        \"samples\": [\n          \"Did u download the fring app?\",\n          \"Pass dis to all ur contacts n see wat u get! Red;i'm in luv wid u. Blue;u put a smile on my face. Purple;u r realy hot. Pink;u r so swt. Orange;i thnk i lyk u. Green;i realy wana go out wid u. Yelow;i wnt u bck. Black;i'm jealous of u. Brown;i miss you Nw plz giv me one color\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print('Spam Data Shape:', spam_df.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZgUGIm6zMrGo",
        "outputId": "9081a72a-0db9-45b3-84e5-5a9f0c124216"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Spam Data Shape: (5572, 2)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Finding percentages of ham and spam\n",
        "spam_df['label'].value_counts(normalize = True) # Normalize shows the proportion."
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 178
        },
        "id": "yWKeXFkPNBa8",
        "outputId": "fd33fd34-2ea9-4382-938e-c270d10845a6"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "label\n",
              "ham     0.865937\n",
              "spam    0.134063\n",
              "Name: proportion, dtype: float64"
            ],
            "text/html": [
              "<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>proportion</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>label</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>ham</th>\n",
              "      <td>0.865937</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>spam</th>\n",
              "      <td>0.134063</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div><br><label><b>dtype:</b> float64</label>"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color='Pickle'>**Observations on Class Imbalance**</font>\n",
        "\n",
        "We can observe a <font color='Pickle'>**significant class imbalance**</font> in the dataset:\n",
        "\n",
        "- <font color='Pickle'>**Ham**:</font> Represents the proportion of email messages that are not spam, making up approximately **87%** of the dataset.\n",
        "- <font color='Pickle'>**Spam**:</font> Represents the spam emails, accounting for only about **13%** of the data.\n",
        "\n",
        "This imbalance highlights the need for careful consideration in model training to ensure that the <font color='Pickle'>**minority class (Spam)**</font> is adequately represented and that the model doesn't become biased towards the <font color='Pickle'>**majority class (Ham)**</font>.\n"
      ],
      "metadata": {
        "id": "-3CFrz11OWME"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color='Pickle'>Metric for Evaluating the Model:**\n",
        "\n",
        "## **<font color='Pickle'>F1 Score**\n",
        "\n",
        "Given the stark class imbalance between Spam and Ham, it's crucial to select the most appropriate metric for evaluating our model, as the traditional metric of accuracy will not suffice. Always predicting messages as ham would yield an accuracy of 87%, but this wouldn't truly reflect the model's performance in classifying spam.\n",
        "\n",
        "The optimal metric for this situation is the **<font color='Pickle'>F1 Score</font>**, a harmonic mean of two critical metrics: **<font color='Pickle'>Precision</font>** and **<font color='Pickle'>Recall</font>**.\n",
        "\n",
        "### **<font color='Pickle'>Precision</font>**\n",
        "Precision is calculated using the formula:\n",
        "\n",
        "<font color='Pickle'>**Precision = TP / (TP + FP)**</font>\n",
        "\n",
        "where:\n",
        "- **<font color='Pickle'>TP</font>** (True Positives): The number of correctly classified spam messages.\n",
        "- **<font color='Pickle'>FP</font>** (False Positives): The number of messages incorrectly classified as spam.\n",
        "\n",
        "A high **<font color='Pickle'>Precision</font>** value indicates that the model makes few mistakes when predicting positive instances (i.e., spam).\n",
        "\n",
        "### **<font color='Pickle'>Recall</font>**\n",
        "Recall (also known as sensitivity or true positive rate) is the ratio of true positive predictions to the total number of actual positive instances within the dataset. It is represented by the formula:\n",
        "\n",
        "<font color = 'Pickle'>**Recall = TP / (TP + FN)**\n",
        "\n",
        "where:\n",
        "- **<font color='Pickle'>FN</font>** (False Negatives): The number of messages that are actually spam but have been incorrectly classified as ham.\n",
        "\n",
        "**<font color='Pickle'>Recall</font>** answers the question: \"Out of all the positive instances in the dataset, how many did we correctly predict?\"\n",
        "\n",
        "### **<font color='Pickle'>F1 Score</font>**\n",
        "The **<font color='Pickle'>F1 Score</font>**, as mentioned above, is the harmonic mean of **<font color='Pickle'>Precision</font>** and **<font color='Pickle'>Recall</font>**, represented by the formula:\n",
        "\n",
        "**<font color='Pickle'> F1 Score = 2(Precision x Recall) / (Precision + Recall)**\n",
        "\n",
        "The harmonic mean is used here because it gives greater weight to lower values. This means that the **<font color='Pickle'>F1 Score</font>** will only be high if both **<font color='Pickle'>precision</font>** and **<font color='Pickle'>recall</font>** are high. If either metric is low, then the **<font color='Pickle'>F1 score</font>** will also be low. This ensures that a model with a precision of 0.8 and recall of 0.1 will have a lower **<font color='Pickle'>F1 score</font>** than a model with a precision and recall both at 0.6.\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "VpG5hjQqYMRF"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color='Pickle'>Classification Pipelines**\n",
        "Below I am creating three model Pipelines to test which is most effective at classification. Each Pipeline will be evaluated on 10% of the full data set. After I evaluate each model I will create a second notebook to use the best model weights to train a model on a randomly selected 40% of the data where it will then be evaluated on a test set of un-seen data.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "Bv9efsGmfvAX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### <font color ='pickle'>**Creating Pipeline data subsets**"
      ],
      "metadata": {
        "id": "VMQ1Vxb7lvO_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Sampling for Sparse Embedding Pipeline\n",
        "sparse_df = spam_df.sample(frac = 0.1,random_state = 42)\n",
        "\n",
        "# Feature Engineering Pipeline sampling\n",
        "fe_df = sparse_df\n",
        "\n",
        "# Feature engineering and Sparse Embedding Pipeline\n",
        "s_fe_df = sparse_df\n"
      ],
      "metadata": {
        "id": "nMJX6TIFgy5O"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color='Pickle'>Pipeline 1: Preprocessing + Sparse Embeddings Pipeline + ML model</font>**\n",
        "\n",
        "## **<font color='Pickle'>For this pipeline, I will:</font>**\n",
        "\n",
        "- Sample 10% of the entire dataset.\n",
        "- Split the subset into train and test sets.\n",
        "- Preprocess the data using the Spacy preprocessor\n",
        "- Create the data pipeline that uses a <font color = 'Pickle'>**TFIDF vectorizer**</font> and an <font color ='Pickle'>**XGBoost**</font> model.\n",
        "- Train the hyperparameters of the model using <font color = 'Pickle'>**RandomSearchCV**</font> and 5-fold cross-validation.\n",
        "- Fit the best parameters on the training data to evaluate the model.\n",
        "- Plot the learning curves of the model based on a model performance metric <font color = 'Pickle'>**(F1 score)**</font>\n",
        "- Evaluate the model on the test data set by printing the classification report.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "2S5cbbQ6tAew"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Splitting into training and temp data\n",
        "train_s_df,test_s_df = train_test_split(sparse_df, test_size = 0.4, random_state = 42)"
      ],
      "metadata": {
        "id": "cm6YlymHwQLz"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print('Train class distribution:',train_s_df['label'].value_counts())\n",
        "\n",
        "print('-'*30)\n",
        "print('Test class distribution:',test_s_df['label'].value_counts())\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4o6eoS2M2rvu",
        "outputId": "ecd618a7-c4da-4af4-d967-6d1303cbde1d"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train class distribution: label\n",
            "ham     303\n",
            "spam     31\n",
            "Name: count, dtype: int64\n",
            "------------------------------\n",
            "Test class distribution: label\n",
            "ham     187\n",
            "spam     36\n",
            "Name: count, dtype: int64\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Seperating independent and dependent columns and Label Encoding\n",
        "X_train_s = train_s_df['message'].values\n",
        "X_test_s = test_s_df['message'].values\n",
        "\n",
        "y_train_s = train_s_df['label'].values\n",
        "y_test_s = test_s_df['label'].values\n",
        "\n",
        "y_train_s_encoded = label_encoder.fit_transform(y_train_s)\n",
        "y_test_s_encoded = label_encoder.transform(y_test_s)\n",
        "\n",
        "# Checking class distributions\n",
        "train_distribution = Counter(y_train_s_encoded)\n",
        "print(\"Training set distribution:\", train_distribution)\n",
        "\n",
        "test_distribution = Counter(y_test_s_encoded)\n",
        "print(\"Test set distribution:\", test_distribution)"
      ],
      "metadata": {
        "id": "daBbeGPXj_mz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "867e3fcf-685d-47bc-8ad3-e9b8d7c1bb9f"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training set distribution: Counter({0: 303, 1: 31})\n",
            "Test set distribution: Counter({0: 187, 1: 36})\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# import custom-preprocessor from python file\n",
        "import CustomPreprocessorSpacy as cp"
      ],
      "metadata": {
        "id": "_L7xVvG8j1_3"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cp.SpacyPreprocessor?"
      ],
      "metadata": {
        "id": "lKRnxv_HIll8"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color='Pickle'>**The Spacy Preprocessor above is a full text preprocessing class allowing for multiple parameters.**</font>\n",
        "\n",
        "For full documentation refer to this link. https://github.com/explosion/spacy-models/releases/tag/en_core_web_sm-3.7.1\n",
        "\n",
        "#### **<font color='Pickle'>Attributes:**\n",
        "\n",
        "- **<font color='Pickle'>`model`</font>** *(str)*:  \n",
        "  The spaCy language model to be used for tokenization and other NLP tasks.\n",
        "\n",
        "- **<font color='Pickle'>`batch_size`</font>** *(int)*:  \n",
        "  The number of documents to process at once during spaCy's pipeline processing.\n",
        "\n",
        "- **<font color='Pickle'>`lemmatize`</font>** *(bool)*:  \n",
        "  If `True`, lemmatize tokens.\n",
        "\n",
        "- **<font color='Pickle'>`lower`</font>** *(bool)*:  \n",
        "  If `True`, convert all characters to lowercase.\n",
        "\n",
        "- **<font color='Pickle'>`remove_stop`</font>** *(bool)*:  \n",
        "  If `True`, remove stopwords.\n",
        "\n",
        "- **<font color='Pickle'>`remove_punct`</font>** *(bool)*:  \n",
        "  If `True`, remove punctuation.\n",
        "\n",
        "- **<font color='Pickle'>`remove_email`</font>** *(bool)*:  \n",
        "  If `True`, remove email addresses.\n",
        "\n",
        "- **<font color='Pickle'>`remove_url`</font>** *(bool)*:  \n",
        "  If `True`, remove URLs.\n",
        "\n",
        "- **<font color='Pickle'>`remove_num`</font>** *(bool)*:  \n",
        "  If `True`, remove numbers.\n",
        "\n",
        "- **<font color='Pickle'>`stemming`</font>** *(bool)*:  \n",
        "  If `True`, apply stemming to tokens *(mutually exclusive with lemmatization)*.\n",
        "\n",
        "- **<font color='Pickle'>`add_user_mention_prefix`</font>** *(bool)*:  \n",
        "  If `True`, add `'@'` as a separate token *(useful for user mentions in social media data)*.\n",
        "\n",
        "- **<font color='Pickle'>`remove_hashtag_prefix`</font>** *(bool)*:  \n",
        "  If `True`, do not separate `'#'` from the following text.\n",
        "\n",
        "- **<font color='Pickle'>`basic_clean_only`</font>** *(bool)*:  \n",
        "  If `True`, perform only basic cleaning *(HTML tags removal, line breaks, etc.)* and ignore other preprocessing steps.\n",
        "\n"
      ],
      "metadata": {
        "id": "qLofej9IEXMF"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color='Pickle'>**Preprocessing**\n"
      ],
      "metadata": {
        "id": "bDfXFYJbNRg8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Using the custom Spacy preprocessor to clean the train and test data,\n",
        "X_train_cleaned_s = cp.SpacyPreprocessor(model = 'en_core_web_sm',batch_size = 100).transform(X_train_s)\n",
        "X_test_cleaned_s = cp.SpacyPreprocessor(model = 'en_core_web_sm',batch_size = 100).transform(X_test_s)\n",
        "\n",
        "# Creating filepaths in drive\n",
        "file_X_train_cleaned_s = data_folder/'X_train_cleaned_sparse.pkl'\n",
        "file_X_test_cleaned_s = data_folder/'X_test_cleaned_sparse.pkl'\n",
        "\n",
        "# Uploading pkl files to drive\n",
        "joblib.dump(X_train_cleaned_s, file_X_train_cleaned_s)\n",
        "joblib.dump(X_test_cleaned_s, file_X_test_cleaned_s)"
      ],
      "metadata": {
        "id": "d5hHpP9eNcGd"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Creating filepaths in drive\n",
        "file_X_train_cleaned_s = data_folder/'X_train_cleaned_sparse.pkl'\n",
        "file_X_test_cleaned_s = data_folder/'X_test_cleaned_sparse.pkl'"
      ],
      "metadata": {
        "id": "227AviLLQOos"
      },
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### <font color='Pickle'>**Pipeline Set Up**</font>\n",
        "\n",
        "- Using the TF-IDF vectorizer with the following hyperparameters:\n",
        "  - **<font color='Pickle'>`min_df = 2`</font>**: This ensures that terms must appear in at least 2 documents to be considered. This helps in filtering out very rare words that might not contribute meaningfully to the model.\n",
        "  - **<font color='Pickle'>`max_df = 100`</font>**: Terms that appear in more than 100 documents are excluded. This helps in removing extremely common words or phrases that are too frequent to be useful in distinguishing between spam and non-spam.\n",
        "  - **<font color='Pickle'>`token_pattern = r\"\\b\\w[\\w@#.!?]*\\b\"`</font>**: This pattern is designed to capture a wide range of tokens, including words with special characters like `@`, `#`, `.` and `!`. It helps in recognizing abnormal spellings and patterns that are often indicative of spam.\n",
        "  - **<font color='Pickle'>`max_features = 1000`</font>**: Limits the number of features (terms) to the top 1000 based on term frequency. This reduces the dimensionality of the feature space while keeping the most informative terms.\n",
        "  - **<font color='Pickle'>`ngram_range = (1, 3)`</font>**: Extracts unigrams, bigrams, and trigrams. This helps capture common phrases and sequences like \"You won a\" or \"Congratulations you win,\" which are often used in spam emails.\n",
        "  - **<font color='Pickle'>`stop_words = None`</font>**: No stopwords are removed. Stopwords are retained because common phrases that include stopwords can be indicative of spam, such as \"click here\" or \"you have won.\"\n",
        "  - **<font color='Pickle'>`lowercase = False`</font>**: Text is not converted to lowercase, allowing the model to differentiate between capitalized and uncapitalized versions of words, which might be important in identifying spam (e.g., \"FREE\" vs. \"free\").\n",
        "\n"
      ],
      "metadata": {
        "id": "DzRVAiKGJs0Q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Importing the nltk stopwords library\n",
        "nltk_stop_words = nltk_stopwords.words('english')"
      ],
      "metadata": {
        "id": "N5U0qFugK9aJ"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "classifier_1 = Pipeline([\n",
        "    ('vectorizer',TfidfVectorizer(min_df = 2,\n",
        "                                   max_df = 100,\n",
        "                                   token_pattern =r\"\\b\\w[\\w@#.!?]*\\b\",\n",
        "                                   max_features = 1000,\n",
        "                                   ngram_range = (1,3), # Pulls in bigrams and trigrams for common phrasing such as 'You won a','Congratulations you win', etc.\n",
        "                                   stop_words = None, # Stop words can be helpful in classifying spam emails because of common phrases.\n",
        "                                   lowercase = False # Keeping both capitalized and uncapitalized versions of words.\n",
        "                                   )),\n",
        "    ('classifier',XGBClassifier(enable_categorical=True, eval_metric = 'logloss'))\n",
        "])"
      ],
      "metadata": {
        "id": "S4G9bzCS0fVh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### <font color='Pickle'>**Hyperparameter Tuning Setup**</font>\n",
        "\n",
        "- Using `RandomizedSearchCV` to tune the following hyperparameters:\n",
        "\n",
        "  - **<font color='Pickle'>`vectorizer__max_features`</font>**:\n",
        "    - **Values:** `[300, 500, 700, 1000]`\n",
        "    - **Explanation:** Controls the number of features (terms) considered by the TF-IDF vectorizer. This range helps explore different levels of feature dimensionality, allowing the model to capture the most relevant terms.\n",
        "\n",
        "  - **<font color='Pickle'>`classifier__n_estimators`</font>**:\n",
        "    - **Values:** `[20, 50, 100]`\n",
        "    - **Explanation:** Determines the number of boosting rounds in the XGBoost model. More boosting rounds can increase model complexity, allowing the model to learn more intricate patterns in the data.\n",
        "\n",
        "  - **<font color='Pickle'>`classifier__learning_rate`</font>**:\n",
        "    - **Values:** `[0.01, 0.1, 0.2]`\n",
        "    - **Explanation:** Sets the learning rate for each boosting round. Lower values make the model update more cautiously, which can help prevent overfitting, especially in smaller datasets.\n",
        "\n",
        "  - **<font color='Pickle'>`classifier__max_depth`</font>**:\n",
        "    - **Values:** `[3, 5, 7]`\n",
        "    - **Explanation:** Controls the maximum depth of the trees in the XGBoost model. Deeper trees allow the model to capture more complex patterns, but they also increase the risk of overfitting.\n",
        "\n",
        "  - **<font color='Pickle'>`classifier__scale_pos_weight`</font>**:\n",
        "    - **Values:** `[1, 10, 25, 50, 75, 99, 100, 1000]`\n",
        "    - **Explanation:** Adjusts the balance between positive and negative classes to handle class imbalance. The wide range of values allows exploration of how much emphasis should be placed on the minority class to achieve better classification results.\n"
      ],
      "metadata": {
        "id": "R-78VFkRYVhM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "param_grid_classifier_1 = {\n",
        "    'vectorizer__max_features':[300,500,700,1000], # Number of features to consider in the RandomSearch for TF-IDF\n",
        "    'classifier__n_estimators': [20,50,100], # Number of boosting rounds in XGB\n",
        "    'classifier__learning_rate': [0.01,0.1,0.2], # Learning rate for each boosting round\n",
        "    'classifier__max_depth': [3,5,7,9,11], # Maximum depth of trees (higher values make the model more complex)\n",
        "    'classifier__scale_pos_weight': [1,10,25,50,75,99,100,1000] # Class weight values to handle the class imbalance and some for exploration.\n",
        "}\n"
      ],
      "metadata": {
        "id": "4v78XmOeVnlX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### <font color='Pickle'>**Cross-Validation and Hyperparameter Tuning**</font>\n",
        "\n",
        "- **Cross-Validation Strategy**: Using **<font color='Pickle'>`RepeatedStratifiedKFold`</font>** with 5 splits and 1 repeat to ensure balanced class distribution across folds.\n",
        "  \n",
        "- **Hyperparameter Tuning with RandomizedSearchCV**:\n",
        "  - **Iterations**: Testing **<font color='Pickle'>`10`</font>** different hyperparameter combinations.\n",
        "  - **Scoring Metric**: Using **<font color='Pickle'>`f1_macro`</font>** to address class imbalance.\n",
        "  - **Parallel Processing**: Set **<font color='Pickle'>`n_jobs=-1`</font>** to utilize all available CPU cores.\n",
        "  - **Verbose Level**: Set to **<font color='Pickle'>`2`</font>** to provide detailed output for better understanding.\n"
      ],
      "metadata": {
        "id": "jraGPfV2bPAj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cv_strategy = RepeatedStratifiedKFold(n_splits = 5,n_repeats = 1, random_state = 42)\n",
        "\n",
        "grid_classifier_1 = RandomizedSearchCV(\n",
        "    estimator = classifier_1,\n",
        "    param_distributions = param_grid_classifier_1,\n",
        "    n_iter =300,# Number of different combinations to try\n",
        "    cv=cv_strategy,\n",
        "    scoring = 'f1_macro', #F1 scoring metric within the CV\n",
        "    verbose = 2, # Showing output for personal understanding\n",
        "    n_jobs = -1 # Ensureing the use of all available cores.\n",
        ")"
      ],
      "metadata": {
        "id": "PlmoPGSG_4hg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color='Pickle'>**Fitting model**"
      ],
      "metadata": {
        "id": "A7WtNaMTcDhp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train_cleaned_s = joblib.load(file_X_train_cleaned_s)\n",
        "grid_classifier_1.fit(X_train_cleaned_s, y_train_s_encoded)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        },
        "id": "pN2UA4CPcSam",
        "outputId": "b03cacf0-bdfe-439f-8f4b-d599299341ff"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 5 folds for each of 300 candidates, totalling 1500 fits\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[('vectorizer',\n",
              "                                              TfidfVectorizer(lowercase=False,\n",
              "                                                              max_df=100,\n",
              "                                                              max_features=1000,\n",
              "                                                              min_df=2,\n",
              "                                                              ngram_range=(1,\n",
              "                                                                           3),\n",
              "                                                              token_pattern='\\\\b\\\\w[\\\\w@#.!?]*\\\\b')),\n",
              "                                             ('classifier',\n",
              "                                              XGBClassifier(base_score=None,\n",
              "                                                            booster=None,\n",
              "                                                            callbacks=None,\n",
              "                                                            colsample_bylevel=None,\n",
              "                                                            co...\n",
              "                                                            n_jobs=None,\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={'classifier__learning_rate': [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        'classifier__max_depth': [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        'classifier__n_estimators': [20, 50,\n",
              "                                                                     100],\n",
              "                                        'classifier__scale_pos_weight': [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000],\n",
              "                                        'vectorizer__max_features': [300, 500,\n",
              "                                                                     700,\n",
              "                                                                     1000]},\n",
              "                   scoring='f1_macro', verbose=2)"
            ],
            "text/html": [
              "<style>#sk-container-id-7 {color: black;}#sk-container-id-7 pre{padding: 0;}#sk-container-id-7 div.sk-toggleable {background-color: white;}#sk-container-id-7 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-7 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-7 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-7 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-7 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-7 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-7 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-7 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-7 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-7 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-7 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-7 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-7 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-7 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-7 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-7 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-7 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-7 div.sk-item {position: relative;z-index: 1;}#sk-container-id-7 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-7 div.sk-item::before, #sk-container-id-7 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-7 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-7 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-7 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-7 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-7 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-7 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-7 div.sk-label-container {text-align: center;}#sk-container-id-7 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-7 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-7\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[(&#x27;vectorizer&#x27;,\n",
              "                                              TfidfVectorizer(lowercase=False,\n",
              "                                                              max_df=100,\n",
              "                                                              max_features=1000,\n",
              "                                                              min_df=2,\n",
              "                                                              ngram_range=(1,\n",
              "                                                                           3),\n",
              "                                                              token_pattern=&#x27;\\\\b\\\\w[\\\\w@#.!?]*\\\\b&#x27;)),\n",
              "                                             (&#x27;classifier&#x27;,\n",
              "                                              XGBClassifier(base_score=None,\n",
              "                                                            booster=None,\n",
              "                                                            callbacks=None,\n",
              "                                                            colsample_bylevel=None,\n",
              "                                                            co...\n",
              "                                                            n_jobs=None,\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={&#x27;classifier__learning_rate&#x27;: [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        &#x27;classifier__max_depth&#x27;: [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        &#x27;classifier__n_estimators&#x27;: [20, 50,\n",
              "                                                                     100],\n",
              "                                        &#x27;classifier__scale_pos_weight&#x27;: [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000],\n",
              "                                        &#x27;vectorizer__max_features&#x27;: [300, 500,\n",
              "                                                                     700,\n",
              "                                                                     1000]},\n",
              "                   scoring=&#x27;f1_macro&#x27;, verbose=2)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-25\" type=\"checkbox\" ><label for=\"sk-estimator-id-25\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomizedSearchCV</label><div class=\"sk-toggleable__content\"><pre>RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[(&#x27;vectorizer&#x27;,\n",
              "                                              TfidfVectorizer(lowercase=False,\n",
              "                                                              max_df=100,\n",
              "                                                              max_features=1000,\n",
              "                                                              min_df=2,\n",
              "                                                              ngram_range=(1,\n",
              "                                                                           3),\n",
              "                                                              token_pattern=&#x27;\\\\b\\\\w[\\\\w@#.!?]*\\\\b&#x27;)),\n",
              "                                             (&#x27;classifier&#x27;,\n",
              "                                              XGBClassifier(base_score=None,\n",
              "                                                            booster=None,\n",
              "                                                            callbacks=None,\n",
              "                                                            colsample_bylevel=None,\n",
              "                                                            co...\n",
              "                                                            n_jobs=None,\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={&#x27;classifier__learning_rate&#x27;: [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        &#x27;classifier__max_depth&#x27;: [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        &#x27;classifier__n_estimators&#x27;: [20, 50,\n",
              "                                                                     100],\n",
              "                                        &#x27;classifier__scale_pos_weight&#x27;: [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000],\n",
              "                                        &#x27;vectorizer__max_features&#x27;: [300, 500,\n",
              "                                                                     700,\n",
              "                                                                     1000]},\n",
              "                   scoring=&#x27;f1_macro&#x27;, verbose=2)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-26\" type=\"checkbox\" ><label for=\"sk-estimator-id-26\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;vectorizer&#x27;,\n",
              "                 TfidfVectorizer(lowercase=False, max_df=100, max_features=1000,\n",
              "                                 min_df=2, ngram_range=(1, 3),\n",
              "                                 token_pattern=&#x27;\\\\b\\\\w[\\\\w@#.!?]*\\\\b&#x27;)),\n",
              "                (&#x27;classifier&#x27;,\n",
              "                 XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
              "                               colsample_bylevel=None, colsample_bynode=None,\n",
              "                               colsample_bytree=None, device=None,\n",
              "                               early_stopping_rounds=None,\n",
              "                               enable_categori...\n",
              "                               feature_types=None, gamma=None, grow_policy=None,\n",
              "                               importance_type=None,\n",
              "                               interaction_constraints=None, learning_rate=None,\n",
              "                               max_bin=None, max_cat_threshold=None,\n",
              "                               max_cat_to_onehot=None, max_delta_step=None,\n",
              "                               max_depth=None, max_leaves=None,\n",
              "                               min_child_weight=None, missing=nan,\n",
              "                               monotone_constraints=None, multi_strategy=None,\n",
              "                               n_estimators=None, n_jobs=None,\n",
              "                               num_parallel_tree=None, random_state=None, ...))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-27\" type=\"checkbox\" ><label for=\"sk-estimator-id-27\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">TfidfVectorizer</label><div class=\"sk-toggleable__content\"><pre>TfidfVectorizer(lowercase=False, max_df=100, max_features=1000, min_df=2,\n",
              "                ngram_range=(1, 3), token_pattern=&#x27;\\\\b\\\\w[\\\\w@#.!?]*\\\\b&#x27;)</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-28\" type=\"checkbox\" ><label for=\"sk-estimator-id-28\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">XGBClassifier</label><div class=\"sk-toggleable__content\"><pre>XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
              "              colsample_bylevel=None, colsample_bynode=None,\n",
              "              colsample_bytree=None, device=None, early_stopping_rounds=None,\n",
              "              enable_categorical=True, eval_metric=&#x27;logloss&#x27;,\n",
              "              feature_types=None, gamma=None, grow_policy=None,\n",
              "              importance_type=None, interaction_constraints=None,\n",
              "              learning_rate=None, max_bin=None, max_cat_threshold=None,\n",
              "              max_cat_to_onehot=None, max_delta_step=None, max_depth=None,\n",
              "              max_leaves=None, min_child_weight=None, missing=nan,\n",
              "              monotone_constraints=None, multi_strategy=None, n_estimators=None,\n",
              "              n_jobs=None, num_parallel_tree=None, random_state=None, ...)</pre></div></div></div></div></div></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 143
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##<font color = 'Pickle'>**Understanding XGBoost's Log Loss and F1 Scoring**\n",
        "\n",
        "### <font color = 'Pickle'>**Why Doesn't XGBoost Include F1 Scoring?**\n",
        "\n",
        "When working with the **<font color='Pickle'>XGBoost</font>** model, it's important to understand that it primarily optimizes **<font color='Pickle'>log loss</font>** during training. **<font color='Pickle'>Log loss</font>**, or logistic loss, is a common objective function for binary classification tasks, and it helps the model learn to predict probabilities that are well-calibrated.\n",
        "\n",
        "### <font color = 'Pickle'>**But What About F1 Score?**\n",
        "\n",
        "Even though I'm using **<font color='Pickle'>f1_macro</font>** as the scoring metric in **<font color='Pickle'>RandomizedSearchCV</font>**, this doesn't mean **<font color='Pickle'>XGBoost</font>** is optimizing the **<font color='Pickle'>F1 score</font>** during training. Instead, **<font color='Pickle'>XGBoost</font>** sticks to optimizing **<font color='Pickle'>log loss</font>** because it's a differentiable function, which is essential for gradient-based optimization algorithms like **<font color='Pickle'>XGBoost</font>**.\n",
        "\n",
        "The **<font color='Pickle'>F1 score</font>**, on the other hand, is not differentiable, which makes it unsuitable as a direct loss function. However, that doesn't mean it's not useful—quite the opposite! By using **<font color='Pickle'>F1</font>** as the scoring metric in **<font color='Pickle'>RandomizedSearchCV</font>**, I'm making sure that the model configuration that performs best is one that balances precision and recall, which is crucial for my imbalanced dataset.\n",
        "\n",
        "### <font color = 'Pickle'>**How Does This Work in Practice?**\n",
        "\n",
        "- **During Training:** **<font color='Pickle'>XGBoost</font>** optimizes the **<font color='Pickle'>log loss</font>** function to learn the best model parameters.\n",
        "- **During Hyperparameter Tuning:** **<font color='Pickle'>RandomizedSearchCV</font>** evaluates each model configuration using the **<font color='Pickle'>F1 score</font>** (or whichever metric I specify). This way, the model that is selected is the one that performs best in terms of **<font color='Pickle'>F1 score</font>**, even though it was trained using **<font color='Pickle'>log loss</font>**.\n",
        "\n",
        "### <font color = 'Pickle'>**Why This Makes Sense**\n",
        "\n",
        "Using **<font color='Pickle'>log loss</font>** for training helps the model become confident in its probability predictions, while evaluating it with **<font color='Pickle'>F1 score</font>** ensures that I'm picking a model that effectively handles the class imbalance in my data. It's a balanced approach that leverages the strengths of both metrics to improve model performance.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "kaONcrLTSMXk"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "##<font color = 'Pickle'>**Getting the best Parameters**"
      ],
      "metadata": {
        "id": "yfTEhaevA2Rz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'Best cross-validation score: {grid_classifier_1.best_score_:.2f}')\n",
        "print(\"\\nBest Parameters:\",grid_classifier_1.best_params_)\n",
        "print(\"\\nBest Estimator:\", grid_classifier_1.best_estimator_)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fbtcEnVqgxK_",
        "outputId": "e18880f0-f6e3-477f-e462-2de9805660ad"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Best cross-validation score: 0.85\n",
            "\n",
            "Best Parameters: {'vectorizer__max_features': 700, 'classifier__scale_pos_weight': 100, 'classifier__n_estimators': 100, 'classifier__max_depth': 3, 'classifier__learning_rate': 0.2}\n",
            "\n",
            "Best Estimator: Pipeline(steps=[('vectorizer',\n",
            "                 TfidfVectorizer(lowercase=False, max_df=100, max_features=700,\n",
            "                                 min_df=2, ngram_range=(1, 3),\n",
            "                                 token_pattern='\\\\b\\\\w[\\\\w@#.!?]*\\\\b')),\n",
            "                ('classifier',\n",
            "                 XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
            "                               colsample_bylevel=None, colsample_bynode=None,\n",
            "                               colsample_bytree=None, device=None,\n",
            "                               early_stopping_rounds=None,\n",
            "                               enable_categoric...\n",
            "                               feature_types=None, gamma=None, grow_policy=None,\n",
            "                               importance_type=None,\n",
            "                               interaction_constraints=None, learning_rate=0.2,\n",
            "                               max_bin=None, max_cat_threshold=None,\n",
            "                               max_cat_to_onehot=None, max_delta_step=None,\n",
            "                               max_depth=3, max_leaves=None,\n",
            "                               min_child_weight=None, missing=nan,\n",
            "                               monotone_constraints=None, multi_strategy=None,\n",
            "                               n_estimators=100, n_jobs=None,\n",
            "                               num_parallel_tree=None, random_state=None, ...))])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Saving the model**"
      ],
      "metadata": {
        "id": "H_0qrfHRsBYS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "file_best_estimator_pipeline1 = model_folder/'pipeline1_sparse_best_estimator.pkl'\n",
        "\n",
        "file_complete_grid_pipeline1 = model_folder/'pipeline1_sparse_complete_grid.pkl'\n",
        "\n",
        "joblib.dump(grid_classifier_1.best_estimator_, file_best_estimator_pipeline1)\n",
        "joblib.dump(grid_classifier_1, file_complete_grid_pipeline1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZA610u76qgrJ",
        "outputId": "64774da6-7836-4f15-8a94-548eadacee60"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/drive/MyDrive/Colab Notebooks/models/nlp_fall_2024/spam/pipeline1_sparse_complete_grid.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 145
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Reload the model**"
      ],
      "metadata": {
        "id": "ttUEK8als7Do"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "best_estimator_pipeline1 = joblib.load(file_best_estimator_pipeline1)\n",
        "\n",
        "complete_grid_pipeline1 = joblib.load(file_complete_grid_pipeline1)\n"
      ],
      "metadata": {
        "id": "T0feWuFbss5D"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Why Save and Reload the Model?**\n",
        "\n",
        "1. **<font color='Pickle'>Efficiency</font>:**\n",
        "   - After running **<font color='Pickle'>`RandomizedSearchCV`</font>**, I save the best model and grid search results using **<font color='Pickle'>`joblib`</font>**. Hyperparameter tuning is computationally expensive, and by saving the outputs, I can avoid repeating the time-consuming process. Reloading these saved objects later allows me to focus on analyzing the results, like plotting learning curves, without the need to rerun the entire tuning process.\n",
        "\n",
        "2. **<font color='Pickle'>Reproducibility</font>:**\n",
        "   - Saving and reloading the best model ensures consistency in my results. By working with the exact model configuration identified as the best, I avoid any variations that could arise from rerunning the grid search. This approach maintains the integrity of my analysis, especially when I need to reproduce or share my findings.\n",
        "\n",
        "3. **<font color='Pickle'>Experiment Management</font>:**\n",
        "   - In complex experiments involving multiple models or stages, saving intermediate results helps me keep everything organized. It’s easier to revisit specific parts of my experiment without having to rerun everything from scratch. This method allows me to separate the computationally intensive tasks from the analysis, making my workflow more modular and manageable.\n"
      ],
      "metadata": {
        "id": "KVbjyN4numoK"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Plotting the learning curve**"
      ],
      "metadata": {
        "id": "GtVKzeK5u39h"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from plot_learning_curve import plot_learning_curve"
      ],
      "metadata": {
        "id": "J3roocjPvsr2"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_learning_curve(\n",
        "    best_estimator_pipeline1,\n",
        "    'Learning Curves for XGBoost with F1 as scoring Metric',\n",
        "    X_train_cleaned_s,\n",
        "    y_train_s_encoded,\n",
        "    scoring = 'f1_macro',\n",
        "    n_jobs = -1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 397
        },
        "id": "2XezfBBHtg3C",
        "outputId": "002ecd8e-3fd3-422e-d40e-2e77c6a26452"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<module 'matplotlib.pyplot' from '/usr/local/lib/python3.10/dist-packages/matplotlib/pyplot.py'>"
            ]
          },
          "metadata": {},
          "execution_count": 148
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x500 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Observations**\n",
        "- **<font color='Pickle'>Learning Curve:</font>**\n",
        "   - The **<font color='Pickle'>training</font>** score starts at a low point and increases rapidly as more training examples are introduced into the model. This is an indicator that the model is learning well from the training data set. It rises to an almost perfect score near the end suggesting an **<font color='indianred'>overfit.</font>**\n",
        "\n",
        "   - The **<font color='Pickle'>CV</font>** curve follows a similar pattern improving with more examples but at a slightly lower rate, indicating a lower ability to generalize as more data is added. We can also see how the distance between the curves rises but stays fairly consistent.\n",
        "- **<font color='Pickle'>Scalability:</font>**\n",
        "   - This plot shows how scalable the model is with more data. The fit times here are increasing in a linear pattern as more data is added. The increase seems steady so the model will scale well with the dataset.\n",
        "- **<font color='Pickle'>Model Performance:</font>**\n",
        "   - The plot shows how the **<font color='Pickle'>model performance</font>** improves as the time for fitting increases. We can see a relatively consistent increase in the score as the training time increases but there is quite a bit of variation so we can't trust the fit times alone to capture the models performance variability.\n",
        "   "
      ],
      "metadata": {
        "id": "SZ7NSXNUFZOO"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = \"Pickle\">**Checking Cross Validation Scores and Train Scores**"
      ],
      "metadata": {
        "id": "lEBbhAR423BN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('Train scores:',best_estimator_pipeline1.score(X_train_cleaned_s,y_train_s_encoded))\n",
        "\n",
        "print('Cross Validation Scores:', complete_grid_pipeline1.best_score_)"
      ],
      "metadata": {
        "id": "bpJmJ1iSLRd4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4c57f7d9-2fdd-4a15-e38a-17c3043b056f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train scores: 0.9910179640718563\n",
            "Cross Validation Scores: 0.8549152241620905\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Evaluating on the Testing data set**"
      ],
      "metadata": {
        "id": "oyHfXYcV4BeO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Last Pipeline\n",
        "def final_pipeline(text):\n",
        "  cleaned_text = joblib.load(file_X_test_cleaned_s) # Pulling in the cleaned and preprocessed testing data\n",
        "  best_estimator_pipeline1 = joblib.load(file_best_estimator_pipeline1) # pulling in the best model parameters\n",
        "  predictions =best_estimator_pipeline1.predict(cleaned_text)\n",
        "  return predictions\n"
      ],
      "metadata": {
        "id": "YIDH6SB335Ic"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_test_pred_s = final_pipeline(X_test_s)"
      ],
      "metadata": {
        "id": "H0XQikOP6XHV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Classification Report**"
      ],
      "metadata": {
        "id": "bmE4OlVE6lB1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('Sparse embeddings model metrics through the tfidf vectorizer:\\n\\n', classification_report(y_test_s_encoded, y_test_pred_s))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LofKh-St6f9X",
        "outputId": "7c4a0500-5324-4c59-b1e4-5eec78242b4b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Sparse embeddings model metrics through the tfidf vectorizer:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.94      0.96      0.95       187\n",
            "           1       0.75      0.67      0.71        36\n",
            "\n",
            "    accuracy                           0.91       223\n",
            "   macro avg       0.84      0.81      0.83       223\n",
            "weighted avg       0.91      0.91      0.91       223\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Model Interpretation Summary**\n",
        "\n",
        "- **Class 0 (Ham - Majority Class):**\n",
        "  - **<font color='Pickle'>Precision (0.94):</font>** The model correctly identifies 94% of ham emails, making it highly reliable in avoiding false positives for non-spam emails.\n",
        "  - **<font color='Pickle'>Recall (0.96):</font>** With a recall of 96%, the model is doing an excellent job of capturing nearly all ham emails, missing only a small fraction.\n",
        "  - **<font color='Pickle'>F1-Score (0.95):</font>** The F1-score, which balances precision and recall, is very high, indicating strong performance on the majority class.\n",
        "\n",
        "- **Class 1 (Spam - Minority Class):**\n",
        "  - **<font color='Pickle'>Precision (0.75):</font>** When the model predicts an email as spam, it’s accurate 75% of the time, meaning that 25% of the emails classified as spam are actually not spam emails (false positives).\n",
        "  - **<font color='Pickle'>Recall (0.67):</font>** The model successfully identifies 67% of the actual spam emails, indicating that it’s missing about one-third of spam instances.\n",
        "  - **<font color='Pickle'>F1-Score (0.71):</font>** Since both precision and recall for spam are around 70%, the F1-score, which is the harmonic mean of these two metrics, is approximately 0.71. This reflects the model’s balanced but moderate performance on the minority class.\n",
        "\n",
        "- **Overall Takeaway:**\n",
        "  - The model excels in identifying **<font color='Pickle'>ham</font>** emails, as expected, but the **<font color='Pickle'>spam</font>** class shows decent precision with a need for improved recall. The F1-score for spam highlights the balance between precision and recall, but further enhancements could push this value higher, leading to better overall spam detection.\n",
        "\n"
      ],
      "metadata": {
        "id": "O-UYCcpcFvpl"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color='Pickle'>Pipeline 2: Preprocessing + Manual Feature Engineering + ML Model Pipeline</font>**\n",
        "\n",
        "## **<font color='Pickle'>For this pipeline, I will:</font>**\n",
        "\n",
        "- Use 10% of the entire dataset.\n",
        "- Split the subset into train and test sets.\n",
        "- Preprocess the data using the Spacy preprocessor.\n",
        "- **<font color='Pickle'>Generate manual features</font>** such as the count of words, characters, digits, exclamation marks, numbers, nouns, proper nouns, auxiliary verbs (AUX), verbs, adjectives, named entities, spelling mistakes, and others of my choosing.\n",
        "- Create the data pipeline that integrates these engineered features along with an <font color='Pickle'>**XGBoost**</font> model.\n",
        "- Train the hyperparameters of the model using <font color='Pickle'>**GridSearchCV**</font> with 5-fold cross-validation.\n",
        "- Fit the best parameters on the training data to evaluate the model.\n",
        "- Plot the learning curves of the model based on a model performance metric <font color='Pickle'>**(F1 score)**</font>.\n",
        "- Evaluate the model on the test dataset by printing the classification report.\n"
      ],
      "metadata": {
        "id": "y1eGVgV_M5rv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Resetting the training and testing data to use for the manual features\n",
        "train_fe_df,test_fe_df = train_s_df,test_s_df"
      ],
      "metadata": {
        "id": "KYEYA4XvNm9u"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print('Train class distribution:',train_fe_df['label'].value_counts())\n",
        "\n",
        "print('-'*30)\n",
        "print('Test class distribution:',test_fe_df['label'].value_counts())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PYGGt84XN1P7",
        "outputId": "9366e21d-73d3-468f-ca76-de980ed50bff"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train class distribution: label\n",
            "ham     303\n",
            "spam     31\n",
            "Name: count, dtype: int64\n",
            "------------------------------\n",
            "Test class distribution: label\n",
            "ham     187\n",
            "spam     36\n",
            "Name: count, dtype: int64\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Seperating indep and dependent columns and encoding\n",
        "X_train_fe = train_fe_df['message'].values\n",
        "X_test_fe = test_fe_df['message'].values\n",
        "\n",
        "y_train_fe = train_fe_df['label'].values\n",
        "y_test_fe = test_fe_df['label'].values\n",
        "\n",
        "y_train_fe_encoded = label_encoder.fit_transform(y_train_fe)\n",
        "y_test_fe_encoded = label_encoder.transform(y_test_fe)\n",
        "\n",
        "# Checking class distribution\n",
        "train_distribution = Counter(y_train_fe_encoded)\n",
        "print('Training distribution:',train_distribution)\n",
        "\n",
        "test_distribution = Counter(y_test_fe_encoded)\n",
        "print('Test distribution:',test_distribution)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JZZf_qynOCLY",
        "outputId": "abd232f0-90e3-43bf-8862-db6985a389cd"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training distribution: Counter({0: 303, 1: 31})\n",
            "Test distribution: Counter({0: 187, 1: 36})\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Generating Manual Features**\n",
        "- Note the usage of False for each of the parameters for the Spacy Preprocessor, this is because I will be adding my own manual features."
      ],
      "metadata": {
        "id": "ENfdo0xWPnSS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Cleaning the training set via SpacyPreprocessor\n",
        "X_train_cleaned_fe = cp.SpacyPreprocessor(model = 'en_core_web_sm',\n",
        "                                          lemmatize =False,\n",
        "                                          lower = False, remove_stop = False,\n",
        "                                          remove_punct = False,remove_email = False,\n",
        "                                          remove_url = False, add_user_mention_prefix = False,\n",
        "                                          basic_clean_only = True).transform(X_train_fe)\n",
        "# Cleaning the testing set\n",
        "X_test_cleaned_fe = cp.SpacyPreprocessor(model = 'en_core_web_sm',\n",
        "                                         lemmatize =False,\n",
        "                                          lower = False, remove_stop = False,\n",
        "                                          remove_punct = False,remove_email = False,\n",
        "                                          remove_url = False, add_user_mention_prefix = False,\n",
        "                                          basic_clean_only = True).transform(X_test_fe)\n",
        "# Saving our files for later use\n",
        "file_X_train_cleaned_fe = data_folder/'X_train_cleaned_fe.pkl'\n",
        "file_X_test_cleaned_fe = data_folder/'X_test_cleaned_fe.pkl'\n",
        "\n",
        "# Dumping files to drive\n",
        "joblib.dump(X_train_cleaned_fe,file_X_train_cleaned_fe)\n",
        "joblib.dump(X_test_cleaned_fe,file_X_test_cleaned_fe)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZS1KbeGlPt8V",
        "outputId": "26dc6365-24cf-439c-ed54-8bb0472ca03e"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/content/drive/MyDrive/Colab Notebooks/custom-functions/CustomPreprocessorSpacy.py:83: MarkupResemblesLocatorWarning: The input looks more like a filename than markup. You may want to open this file and pass the filehandle into Beautiful Soup.\n",
            "  soup = BeautifulSoup(text, \"html.parser\")\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/drive/MyDrive/Colab Notebooks/datasets/spam_ds/X_test_cleaned_fe.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install pyspellchecker"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sVnbABKUZOfw",
        "outputId": "3cd2a184-a617-4621-fc97-645f7957d115"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting pyspellchecker\n",
            "  Downloading pyspellchecker-0.8.1-py3-none-any.whl.metadata (9.4 kB)\n",
            "Downloading pyspellchecker-0.8.1-py3-none-any.whl (6.8 MB)\n",
            "\u001b[?25l   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/6.8 MB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K   \u001b[91m━━━━━━━━\u001b[0m\u001b[91m╸\u001b[0m\u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.5/6.8 MB\u001b[0m \u001b[31m44.3 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K   \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[90m╺\u001b[0m\u001b[90m━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.0/6.8 MB\u001b[0m \u001b[31m56.0 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K   \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[91m╸\u001b[0m \u001b[32m6.8/6.8 MB\u001b[0m \u001b[31m65.6 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K   \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[91m╸\u001b[0m \u001b[32m6.8/6.8 MB\u001b[0m \u001b[31m65.6 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m6.8/6.8 MB\u001b[0m \u001b[31m38.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: pyspellchecker\n",
            "Successfully installed pyspellchecker-0.8.1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from FeaturizerSpacy import ExtendedManualFeatures"
      ],
      "metadata": {
        "id": "L3_Tkmn_ZQ8u"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ExtendedManualFeatures??"
      ],
      "metadata": {
        "id": "iNbOKoEHV3cy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **<font color='Pickle'>Summary of Edits to the `ManualFeatures` Class</font>**\n",
        "\n",
        "- Below are the enhancements made to the original `ManualFeatures` class to implement a spellchecker.\n",
        "- The new method processes the `cleaned_text` input, **<font color='Pickle'>tokenizes</font>** the text into individual words, and uses the **<font color='Pickle'>SpellChecker</font>** object to identify and count misspelled words in each document.\n",
        "- This misspelled word count is then integrated into the feature set through the `transform` method.\n",
        "- The **<font color='Pickle'>transform method</font>** first calls the parent class's transform method to generate the base features and feature names.\n",
        "- It then appends the misspelled word count as a new feature, resulting in an updated array of features and names, including the new feature called `misspelled_count`.\n",
        "\n",
        "\n",
        "```python\n",
        "class ExtendedManualFeatures(ManualFeatures):\n",
        "    def __init__(self, spacy_model, batch_size=64, pos_features=True, ner_features=True, text_descriptive_features=True, spellcheck_features=True):\n",
        "        super().__init__(spacy_model, batch_size, pos_features, ner_features, text_descriptive_features)\n",
        "        self.spellcheck_features = spellcheck_features\n",
        "        self.spell = SpellChecker()\n",
        "\n",
        "    def get_spellcheck_features(self, cleaned_text):\n",
        "        misspelled_count = []\n",
        "\n",
        "        for doc in cleaned_text:\n",
        "            tokens = doc.split()  # Simple tokenization\n",
        "            misspelled = self.spell.unknown(tokens)  # Identify misspelled words\n",
        "            misspelled_count.append(len(misspelled))\n",
        "\n",
        "        return np.array(misspelled_count).reshape(-1, 1)\n",
        "\n",
        "    def transform(self, X, y=None):\n",
        "        # Use the transform method from the parent class\n",
        "        features, feature_names = super().transform(X)\n",
        "\n",
        "        if self.spellcheck_features:\n",
        "            spellcheck_features = self.get_spellcheck_features(X)\n",
        "            feature_names.append('misspelled_count')\n",
        "            features = np.hstack((features, spellcheck_features))\n",
        "\n",
        "        return features, feature_names\n"
      ],
      "metadata": {
        "id": "0SYejk-IfQMp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "featurizer = ExtendedManualFeatures(spacy_model = 'en_core_web_sm',batch_size = 1000)"
      ],
      "metadata": {
        "id": "4W0JctN_ZAZF"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Reloading and fitting the featurizer on train**"
      ],
      "metadata": {
        "id": "daCDq86Dqwqz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_trained_cleaned_fe = joblib.load(file_X_train_cleaned_fe)\n",
        "X_train_fe_features, feature_names = featurizer.fit_transform(X_train_cleaned_fe)\n",
        "feature_names"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UTV0JotqhuoY",
        "outputId": "49fe4b86-28e7-435f-eef7-713b8668a3dd"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['count_words',\n",
              " 'count_characters',\n",
              " 'count_characters_no_space',\n",
              " 'avg_word_length',\n",
              " 'count_numbers',\n",
              " 'count_sentences',\n",
              " 'noun_count',\n",
              " 'aux_count',\n",
              " 'verb_count',\n",
              " 'adj_count',\n",
              " 'ner',\n",
              " 'misspelled_count']"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color='Pickle'>**Pipeline**"
      ],
      "metadata": {
        "id": "BmS9_BxljInt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "classifier_2 = Pipeline([\n",
        "    ('classifier',XGBClassifier(enable_categorical = True, eval_metric = 'logloss'))\n",
        "])"
      ],
      "metadata": {
        "id": "Cy91tOyLiJiU"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color='Pickle'>**Parameter Grid**"
      ],
      "metadata": {
        "id": "LQJvTcmZjPbY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "param_grid_classifier_2 = {\n",
        "    'classifier__n_estimators': [20,50,100], # Number of boosting rounds in XGB\n",
        "    'classifier__learning_rate': [0.01,0.1,0.2], # Learning rate for each boosting round\n",
        "    'classifier__max_depth': [3,5,7,9,11], # Maximum depth of trees (higher values make the model more complex)\n",
        "    'classifier__scale_pos_weight': [1,10,25,50,75,99,100,1000] # Class weight values to handle the class imbalance and some for exploration.\n",
        "}"
      ],
      "metadata": {
        "id": "MigEkwvojN44"
      },
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color='Pickle'>**Specify Random Search**"
      ],
      "metadata": {
        "id": "ctPwvNN4jk5W"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cv_strategy = RepeatedStratifiedKFold(n_splits = 5,n_repeats = 1, random_state = 42)\n",
        "\n",
        "grid_classifier_2 = RandomizedSearchCV(\n",
        "    estimator = classifier_2,\n",
        "    param_distributions = param_grid_classifier_2,\n",
        "    n_iter = 300,\n",
        "    cv = cv_strategy, #Random Stratified 5 fold\n",
        "    scoring = 'f1_macro',\n",
        "    verbose = 2,\n",
        "    n_jobs = -1\n",
        ")"
      ],
      "metadata": {
        "id": "WsLv3Me2jfJi"
      },
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Fitting the model**"
      ],
      "metadata": {
        "id": "PByIXT6kkXxl"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# fitting on train\n",
        "grid_classifier_2.fit(X_train_fe_features, y_train_fe_encoded)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 552
        },
        "id": "iE2NOv_ikFVG",
        "outputId": "f28e7bf7-abab-4aaa-9f5e-67eb968c42a5"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 5 folds for each of 300 candidates, totalling 1500 fits\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[('classifier',\n",
              "                                              XGBClassifier(base_score=None,\n",
              "                                                            booster=None,\n",
              "                                                            callbacks=None,\n",
              "                                                            colsample_bylevel=None,\n",
              "                                                            colsample_bynode=None,\n",
              "                                                            colsample_bytree=None,\n",
              "                                                            device=None,\n",
              "                                                            early_stopping_rounds=None,\n",
              "                                                            enable_categorical=True,\n",
              "                                                            eval_metric='logloss',\n",
              "                                                            feature_types=...\n",
              "                                                            monotone_constraints=None,\n",
              "                                                            multi_strategy=None,\n",
              "                                                            n_estimators=None,\n",
              "                                                            n_jobs=None,\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={'classifier__learning_rate': [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        'classifier__max_depth': [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        'classifier__n_estimators': [20, 50,\n",
              "                                                                     100],\n",
              "                                        'classifier__scale_pos_weight': [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000]},\n",
              "                   scoring='f1_macro', verbose=2)"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[(&#x27;classifier&#x27;,\n",
              "                                              XGBClassifier(base_score=None,\n",
              "                                                            booster=None,\n",
              "                                                            callbacks=None,\n",
              "                                                            colsample_bylevel=None,\n",
              "                                                            colsample_bynode=None,\n",
              "                                                            colsample_bytree=None,\n",
              "                                                            device=None,\n",
              "                                                            early_stopping_rounds=None,\n",
              "                                                            enable_categorical=True,\n",
              "                                                            eval_metric=&#x27;logloss&#x27;,\n",
              "                                                            feature_types=...\n",
              "                                                            monotone_constraints=None,\n",
              "                                                            multi_strategy=None,\n",
              "                                                            n_estimators=None,\n",
              "                                                            n_jobs=None,\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={&#x27;classifier__learning_rate&#x27;: [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        &#x27;classifier__max_depth&#x27;: [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        &#x27;classifier__n_estimators&#x27;: [20, 50,\n",
              "                                                                     100],\n",
              "                                        &#x27;classifier__scale_pos_weight&#x27;: [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000]},\n",
              "                   scoring=&#x27;f1_macro&#x27;, verbose=2)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomizedSearchCV</label><div class=\"sk-toggleable__content\"><pre>RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[(&#x27;classifier&#x27;,\n",
              "                                              XGBClassifier(base_score=None,\n",
              "                                                            booster=None,\n",
              "                                                            callbacks=None,\n",
              "                                                            colsample_bylevel=None,\n",
              "                                                            colsample_bynode=None,\n",
              "                                                            colsample_bytree=None,\n",
              "                                                            device=None,\n",
              "                                                            early_stopping_rounds=None,\n",
              "                                                            enable_categorical=True,\n",
              "                                                            eval_metric=&#x27;logloss&#x27;,\n",
              "                                                            feature_types=...\n",
              "                                                            monotone_constraints=None,\n",
              "                                                            multi_strategy=None,\n",
              "                                                            n_estimators=None,\n",
              "                                                            n_jobs=None,\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={&#x27;classifier__learning_rate&#x27;: [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        &#x27;classifier__max_depth&#x27;: [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        &#x27;classifier__n_estimators&#x27;: [20, 50,\n",
              "                                                                     100],\n",
              "                                        &#x27;classifier__scale_pos_weight&#x27;: [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000]},\n",
              "                   scoring=&#x27;f1_macro&#x27;, verbose=2)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;classifier&#x27;,\n",
              "                 XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
              "                               colsample_bylevel=None, colsample_bynode=None,\n",
              "                               colsample_bytree=None, device=None,\n",
              "                               early_stopping_rounds=None,\n",
              "                               enable_categorical=True, eval_metric=&#x27;logloss&#x27;,\n",
              "                               feature_types=None, gamma=None, grow_policy=None,\n",
              "                               importance_type=None,\n",
              "                               interaction_constraints=None, learning_rate=None,\n",
              "                               max_bin=None, max_cat_threshold=None,\n",
              "                               max_cat_to_onehot=None, max_delta_step=None,\n",
              "                               max_depth=None, max_leaves=None,\n",
              "                               min_child_weight=None, missing=nan,\n",
              "                               monotone_constraints=None, multi_strategy=None,\n",
              "                               n_estimators=None, n_jobs=None,\n",
              "                               num_parallel_tree=None, random_state=None, ...))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">XGBClassifier</label><div class=\"sk-toggleable__content\"><pre>XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
              "              colsample_bylevel=None, colsample_bynode=None,\n",
              "              colsample_bytree=None, device=None, early_stopping_rounds=None,\n",
              "              enable_categorical=True, eval_metric=&#x27;logloss&#x27;,\n",
              "              feature_types=None, gamma=None, grow_policy=None,\n",
              "              importance_type=None, interaction_constraints=None,\n",
              "              learning_rate=None, max_bin=None, max_cat_threshold=None,\n",
              "              max_cat_to_onehot=None, max_delta_step=None, max_depth=None,\n",
              "              max_leaves=None, min_child_weight=None, missing=nan,\n",
              "              monotone_constraints=None, multi_strategy=None, n_estimators=None,\n",
              "              n_jobs=None, num_parallel_tree=None, random_state=None, ...)</pre></div></div></div></div></div></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Getting best parameters**"
      ],
      "metadata": {
        "id": "HUQV6-Y6lCps"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('Best Cross validation:',{grid_classifier_2.best_score_}) # calling the best score attribute\n",
        "print('\\nBest Parameters:',grid_classifier_2.best_params_) #best param attribute\n",
        "print('\\nBest estimator:', grid_classifier_2.best_estimator_) # best estimator attribute"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CKUdMM3hkqG5",
        "outputId": "47d20ef3-2b06-4d93-e755-1edbfcace6b5"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Best Cross validation: {0.9229371398596437}\n",
            "\n",
            "Best Parameters: {'classifier__scale_pos_weight': 10, 'classifier__n_estimators': 100, 'classifier__max_depth': 3, 'classifier__learning_rate': 0.2}\n",
            "\n",
            "Best estimator: Pipeline(steps=[('classifier',\n",
            "                 XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
            "                               colsample_bylevel=None, colsample_bynode=None,\n",
            "                               colsample_bytree=None, device=None,\n",
            "                               early_stopping_rounds=None,\n",
            "                               enable_categorical=True, eval_metric='logloss',\n",
            "                               feature_types=None, gamma=None, grow_policy=None,\n",
            "                               importance_type=None,\n",
            "                               interaction_constraints=None, learning_rate=0.2,\n",
            "                               max_bin=None, max_cat_threshold=None,\n",
            "                               max_cat_to_onehot=None, max_delta_step=None,\n",
            "                               max_depth=3, max_leaves=None,\n",
            "                               min_child_weight=None, missing=nan,\n",
            "                               monotone_constraints=None, multi_strategy=None,\n",
            "                               n_estimators=100, n_jobs=None,\n",
            "                               num_parallel_tree=None, random_state=None, ...))])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Saving the Model**"
      ],
      "metadata": {
        "id": "UMTr11fGlvFL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "file_best_estimator_pipeline_2 = model_folder/'pipeline2_fe_best_estimator.pkl'\n",
        "\n",
        "file_complete_grid_pipeline2 = model_folder/'pipeline2_fe_complete_grid.pkl'\n",
        "\n",
        "joblib.dump(grid_classifier_2.best_estimator_, file_best_estimator_pipeline_2)\n",
        "\n",
        "joblib.dump(grid_classifier_2,file_complete_grid_pipeline2)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Y5R5gBguloK-",
        "outputId": "15a9e2f1-5dfb-47b1-f8e4-fc326d35f209"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/drive/MyDrive/Colab Notebooks/models/nlp_fall_2024/spam/pipeline2_fe_complete_grid.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Reload the saved model**"
      ],
      "metadata": {
        "id": "E1repPUrmiR-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "best_estimator_pipeline2 = joblib.load(file_best_estimator_pipeline_2)\n",
        "\n",
        "complete_grid_pipeline2 = joblib.load(file_complete_grid_pipeline2)"
      ],
      "metadata": {
        "id": "kE0KlpM2mGXj"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Plot the learning curve for the model**"
      ],
      "metadata": {
        "id": "zXOfeY3TnFnr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from plot_learning_curve import plot_learning_curve"
      ],
      "metadata": {
        "id": "nyXAXShwnTeE"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_learning_curve(\n",
        "    best_estimator_pipeline2,\n",
        "    'Learning Curves for XGBoost with F1: Manual Features',\n",
        "    X_train_fe_features,\n",
        "    y_train_fe_encoded,\n",
        "    scoring = 'f1_macro',\n",
        "    n_jobs = -1)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        },
        "id": "ZvZSobiym_tA",
        "outputId": "fe0d7d26-0251-4c0d-a740-dd7f3d34f26c"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<module 'matplotlib.pyplot' from '/usr/local/lib/python3.10/dist-packages/matplotlib/pyplot.py'>"
            ]
          },
          "metadata": {},
          "execution_count": 35
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x500 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **<font color='Pickle'>Observations</font>**\n",
        "\n",
        "- **<font color='Pickle'>Learning Curve:</font>**  \n",
        "  The model is clearly **<font color='indianred'>overfitting.</font>**\n",
        " The training score remains locked at 1.0. The CV score starts off improving quickly, but we can see a noticeable dip in the middle before it levels off again. This lowers the gap between training and CV, increases it and lowers it again. This suggests the model's generalization ability fluctuates as it sees more data.\n",
        "\n",
        "- **<font color='Pickle'>Scalability of the Model:</font>**  \n",
        "  Fit times increase as the training size grows, which is expected. However, the increase is jagged, indicating spikes in computational demand. This could suggest some inconsistency in how the model handles larger datasets or the increasing complexity of the data. We can also see how the fit time has a higher variability as more data is added suggesting more fluctuation in scalability.\n",
        "\n",
        "- **<font color='Pickle'>Performance of the Model:</font>**  \n",
        "  There’s a dip in model performance as fit times increase, but it bounces back to a similar level by the end. While the model eventually recovers some performance, the initial dip points to inefficiencies, possibly linked to overfitting. We can see also how the variance within the line gets narrower as fit times increase, suggesting that as more data and times increase we may have more stable results.\n",
        "\n",
        "### **<font color='Pickle'>Next Steps</font>**\n",
        "\n",
        "- Regularization or simplifying the model should help tackle overfitting and smooth out performance.\n",
        "- Hyperparameter tuning (e.g., adjusting tree depth, learning rate, or regularization) could stabilize generalization and reduce fluctuations.\n",
        "- Adding more data or using augmentation techniques might improve CV performance and help smooth out the jagged learning curves.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "FBwIj_HdrJZC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Checking the train scores\n",
        "print(best_estimator_pipeline2.score(X_train_fe_features,y_train_fe_encoded))\n",
        "\n",
        "#Checking the CV scores\n",
        "print(complete_grid_pipeline2.best_score_)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kA-dSopYn3W-",
        "outputId": "c4736017-bf81-4f08-c4e7-1b669c1a99c6"
      },
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1.0\n",
            "0.9229371398596437\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color ='Pickle'>**Evaluating on the test dataset**\n"
      ],
      "metadata": {
        "id": "-MkO5QYFySk2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Final Pipeline\n",
        "def final_pipeline(text):\n",
        "  text_cleaned = joblib.load(file_X_test_cleaned_fe)\n",
        "  features,feature_names = featurizer.fit_transform(text_cleaned)\n",
        "  best_estimator_pipeline2 = joblib.load(file_best_estimator_pipeline_2)\n",
        "  predictions = best_estimator_pipeline2.predict(features)\n",
        "  return predictions"
      ],
      "metadata": {
        "id": "xyY6vhGRyMI1"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Making final predictions on the testing data set\n",
        "y_test_pred_fe = final_pipeline(X_test_fe)"
      ],
      "metadata": {
        "id": "sCnCIWb2zPE1"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Classification report**"
      ],
      "metadata": {
        "id": "6ZTaJklkzfxN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('\\nTest Set classification report:\\n\\n',\n",
        "      classification_report(y_test_fe_encoded,y_test_pred_fe))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4CHy-HFSzdbf",
        "outputId": "8e7e0681-a373-491e-98cc-5c3f5327f2ac"
      },
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Test Set classification report:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.98      0.96      0.97       187\n",
            "           1       0.82      0.89      0.85        36\n",
            "\n",
            "    accuracy                           0.95       223\n",
            "   macro avg       0.90      0.93      0.91       223\n",
            "weighted avg       0.95      0.95      0.95       223\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **<font color='Pickle'>Model Interpretation Summary</font>**\n",
        "\n",
        "- **<font color='Pickle'>Class 0 (Ham - Majority Class):</font>**\n",
        "  - **<font color='Pickle'>Precision (0.98):</font>** The model correctly identifies 98% of ham emails, making it highly accurate in avoiding false positives for non-spam emails.\n",
        "  - **<font color='Pickle'>Recall (0.96):</font>** With a recall of 96%, the model is doing an excellent job of capturing nearly all ham emails, missing only a small fraction.\n",
        "  - **<font color='Pickle'>F1-Score (0.97):</font>** The F1-score, which balances precision and recall, is very high, indicating strong performance on the majority class.\n",
        "\n",
        "- **<font color='Pickle'>Class 1 (Spam - Minority Class):</font>**\n",
        "  - **<font color='Pickle'>Precision (0.82):</font>** When the model predicts an email as spam, it’s accurate 82% of the time, meaning that 18% of the emails classified as spam are actually not spam emails (false positives). This is a solid improvement, but still something to keep an eye on.\n",
        "  - **<font color='Pickle'>Recall (0.89):</font>** The model successfully identifies 89% of the actual spam emails. There's room for improvement here, but it’s catching most of the spam.\n",
        "  - **<font color='Pickle'>F1-Score (0.85):</font>** The balance between precision and recall gives us an F1-score of 0.85, which shows decent handling of the minority class. While better than before, some overfitting could still be a concern.\n",
        "\n",
        "### **<font color='Pickle'>Overall Takeaway:</font>**\n",
        "\n",
        "- **<font color='Pickle'>Majority Class (Ham) Performance:</font>** The model nails it when it comes to identifying ham emails, with precision and recall both hitting near-perfect scores. This ensures legit emails don’t mistakenly get flagged as spam, which is a huge win.\n",
        "\n",
        "- **<font color='Pickle'>Spam Detection (Minority Class):</font>** The model does well in catching most spam emails, boosting the recall, but there’s a slight dip in precision, meaning some ham emails are misclassified as spam. The solid F1-score shows we’ve struck a good balance between catching spam and minimizing false positives.\n",
        "\n",
        "- **<font color='Pickle'>Watch Out for Overfitting:</font>** The high performance overall hints that we could be overfitting a bit, especially given the minority class performance. This could hurt generalization on unseen data, so we’ll need to monitor this closely.\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "p2BsUE_31C1W"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#<font color = \"Pickle\">**Third Pipeline: Preprocessing + Manual and TDIDF features + ML Modeling**"
      ],
      "metadata": {
        "id": "SjWjZJboTyue"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## **<font color='Pickle'>For this pipeline, I will:</font>**\n",
        "\n",
        "- Concatenate the features lists from both the Manual and TDIDF vectorizer pipelines into one dataset.\n",
        "- Ensure the Test data set remains constant\n",
        "- Create the data pipeline that integrates these engineered features witin a sparse transformer along with an <font color='Pickle'>**XGBoost**</font> model.\n",
        "- Train the hyperparameters of the model using <font color='Pickle'>**RandomSearchCV**</font> with 5-fold cross-validation.\n",
        "- Fit the best parameters on the training data to evaluate the model.\n",
        "- Plot the learning curves of the model based on a model performance metric <font color='Pickle'>**(F1 score)**</font>.\n",
        "- Evaluate the model on the test dataset by printing the classification report.\n"
      ],
      "metadata": {
        "id": "alYO9Z07URzY"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = \"Pickle\">**Concatenating the training datasets.**\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "se_sw4-uhOEq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Load in TFIDF cleaned training set\n",
        "X_train_cleaned_s = joblib.load(file_X_train_cleaned_s)"
      ],
      "metadata": {
        "id": "c46wNRfq0yrL"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Concatenating the cleaned text data with the features dataframe\n",
        "X_train_combo = pd.concat((pd.DataFrame(X_train_cleaned_s, columns = ['cleaned_text']),\n",
        "                           pd.DataFrame(X_train_fe_features, columns=feature_names)), axis = 1)\n",
        "\n",
        "\n",
        "y_train_combo = y_train_s_encoded\n",
        "y_test_combo = y_test_s_encoded\n",
        "\n",
        "X_test_combo = X_test_fe"
      ],
      "metadata": {
        "id": "r9SVPzAoiYGW"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color = 'Pickle'>**Below we will see the columns of the new data set that has all of the manual features created in the second pipeline and the cleaned text in one column.**\n",
        "- Next, I will create two pipelines, one to take the manual feature columns and transform them into a sparse matrix, and another to run tfidf vectorization on the cleaned text column of the data.\n",
        "- Then I will combine both of these pipelines into one and implement it within the classification pipeine.\n"
      ],
      "metadata": {
        "id": "1Y2I9scdk3sv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train_combo.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 348
        },
        "id": "eq3Ih136jwoT",
        "outputId": "6041d67a-0cd1-42c1-9af6-a271bb967589"
      },
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                        cleaned_text  count_words  \\\n",
              "0                                     fredericksburg          4.0   \n",
              "1                                          like jerk          8.0   \n",
              "2  u ask darren n pick u lor oso sian tmr haf 2 m...         18.0   \n",
              "3                          hey morning come ask:- pa          8.0   \n",
              "4          promise get soon text morning let know ok         29.0   \n",
              "\n",
              "   count_characters  count_characters_no_space  avg_word_length  \\\n",
              "0              32.0                       29.0         5.800000   \n",
              "1              35.0                       30.0         3.333333   \n",
              "2              74.0                       57.0         3.000000   \n",
              "3              41.0                       34.0         3.777778   \n",
              "4             124.0                       97.0         3.233333   \n",
              "\n",
              "   count_numbers  count_sentences  noun_count  aux_count  verb_count  \\\n",
              "0            0.0              1.0         1.0        1.0         0.0   \n",
              "1            0.0              2.0         1.0        2.0         1.0   \n",
              "2            1.0              1.0         7.0        0.0         5.0   \n",
              "3            0.0              1.0         2.0        0.0         1.0   \n",
              "4            0.0              2.0         2.0        2.0         6.0   \n",
              "\n",
              "   adj_count  ner  misspelled_count  \n",
              "0        0.0  1.0               1.0  \n",
              "1        0.0  1.0               2.0  \n",
              "2        1.0  2.0               7.0  \n",
              "3        0.0  0.0               2.0  \n",
              "4        0.0  0.0               2.0  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c4efd4c0-1b38-4fc4-b02c-4cf94070e9da\" 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>cleaned_text</th>\n",
              "      <th>count_words</th>\n",
              "      <th>count_characters</th>\n",
              "      <th>count_characters_no_space</th>\n",
              "      <th>avg_word_length</th>\n",
              "      <th>count_numbers</th>\n",
              "      <th>count_sentences</th>\n",
              "      <th>noun_count</th>\n",
              "      <th>aux_count</th>\n",
              "      <th>verb_count</th>\n",
              "      <th>adj_count</th>\n",
              "      <th>ner</th>\n",
              "      <th>misspelled_count</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>fredericksburg</td>\n",
              "      <td>4.0</td>\n",
              "      <td>32.0</td>\n",
              "      <td>29.0</td>\n",
              "      <td>5.800000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>like jerk</td>\n",
              "      <td>8.0</td>\n",
              "      <td>35.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>3.333333</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>u ask darren n pick u lor oso sian tmr haf 2 m...</td>\n",
              "      <td>18.0</td>\n",
              "      <td>74.0</td>\n",
              "      <td>57.0</td>\n",
              "      <td>3.000000</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>7.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>hey morning come ask:- pa</td>\n",
              "      <td>8.0</td>\n",
              "      <td>41.0</td>\n",
              "      <td>34.0</td>\n",
              "      <td>3.777778</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>promise get soon text morning let know ok</td>\n",
              "      <td>29.0</td>\n",
              "      <td>124.0</td>\n",
              "      <td>97.0</td>\n",
              "      <td>3.233333</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</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-c4efd4c0-1b38-4fc4-b02c-4cf94070e9da')\"\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-c4efd4c0-1b38-4fc4-b02c-4cf94070e9da 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-c4efd4c0-1b38-4fc4-b02c-4cf94070e9da');\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-687e1165-571c-4d74-8512-010b0e9e429e\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-687e1165-571c-4d74-8512-010b0e9e429e')\"\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",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-687e1165-571c-4d74-8512-010b0e9e429e button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "X_train_combo",
              "summary": "{\n  \"name\": \"X_train_combo\",\n  \"rows\": 334,\n  \"fields\": [\n    {\n      \"column\": \"cleaned_text\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 328,\n        \"samples\": [\n          \"problem\",\n          \"sms auction brand new nokia 7250 4 auction today auction free 2 join txt nokia 86021\",\n          \"tonight mate catch sleep new number way\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"count_words\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 9.910253896544257,\n        \"min\": 1.0,\n        \"max\": 59.0,\n        \"num_unique_values\": 44,\n        \"samples\": [\n          34.0,\n          5.0,\n          20.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"count_characters\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 49.91408498562112,\n        \"min\": 2.0,\n        \"max\": 276.0,\n        \"num_unique_values\": 136,\n        \"samples\": [\n          88.0,\n          79.0,\n          201.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"count_characters_no_space\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 40.44518551811166,\n        \"min\": 2.0,\n        \"max\": 220.0,\n        \"num_unique_values\": 121,\n        \"samples\": [\n          112.0,\n          7.0,\n          97.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"avg_word_length\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.7401222015299257,\n        \"min\": 1.0,\n        \"max\": 6.666666666666667,\n        \"num_unique_values\": 202,\n        \"samples\": [\n          4.4,\n          5.666666666666667,\n          3.8823529411764706\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"count_numbers\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.7777463085468055,\n        \"min\": 0.0,\n        \"max\": 5.0,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.0,\n          1.0,\n          5.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"count_sentences\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.2145866497338704,\n        \"min\": 1.0,\n        \"max\": 10.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          2.0,\n          5.0,\n          1.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"noun_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 3.6856238745024408,\n        \"min\": 0.0,\n        \"max\": 21.0,\n        \"num_unique_values\": 20,\n        \"samples\": [\n          1.0,\n          17.0,\n          16.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"aux_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.1610988527060633,\n        \"min\": 0.0,\n        \"max\": 7.0,\n        \"num_unique_values\": 7,\n        \"samples\": [\n          1.0,\n          2.0,\n          7.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"verb_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.9941044356814908,\n        \"min\": 0.0,\n        \"max\": 16.0,\n        \"num_unique_values\": 11,\n        \"samples\": [\n          2.0,\n          0.0,\n          16.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"adj_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.1494561292653498,\n        \"min\": 0.0,\n        \"max\": 8.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          1.0,\n          8.0,\n          0.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"ner\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1.3344926717859233,\n        \"min\": 0.0,\n        \"max\": 7.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          2.0,\n          5.0,\n          1.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"misspelled_count\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 2.7974757873770066,\n        \"min\": 0.0,\n        \"max\": 17.0,\n        \"num_unique_values\": 16,\n        \"samples\": [\n          1.0,\n          2.0,\n          5.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train_combo.info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Eu6JryZBmeEX",
        "outputId": "aa41747c-72b6-42f6-b26e-21224b17f1b7"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 334 entries, 0 to 333\n",
            "Data columns (total 13 columns):\n",
            " #   Column                     Non-Null Count  Dtype  \n",
            "---  ------                     --------------  -----  \n",
            " 0   cleaned_text               334 non-null    object \n",
            " 1   count_words                334 non-null    float64\n",
            " 2   count_characters           334 non-null    float64\n",
            " 3   count_characters_no_space  334 non-null    float64\n",
            " 4   avg_word_length            334 non-null    float64\n",
            " 5   count_numbers              334 non-null    float64\n",
            " 6   count_sentences            334 non-null    float64\n",
            " 7   noun_count                 334 non-null    float64\n",
            " 8   aux_count                  334 non-null    float64\n",
            " 9   verb_count                 334 non-null    float64\n",
            " 10  adj_count                  334 non-null    float64\n",
            " 11  ner                        334 non-null    float64\n",
            " 12  misspelled_count           334 non-null    float64\n",
            "dtypes: float64(12), object(1)\n",
            "memory usage: 34.0+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color ='Pickle'>**The cleaned text column will have tfidf vectorization ran on it and the rest will be in sparse matrix format.**"
      ],
      "metadata": {
        "id": "OZ7gdPZ5pY48"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Defining the sparse transformer class\n",
        "class SparseTransformer(BaseEstimator, TransformerMixin):\n",
        "    def __init__(self): #initializes the function\n",
        "      pass\n",
        "\n",
        "    def fit(self,X,y = None): #fits and returns the input\n",
        "      return self\n",
        "\n",
        "    def transform(self, X, y = None): #transforms the input to sparse format\n",
        "      return csr_matrix(X) # .toarray for Dense"
      ],
      "metadata": {
        "id": "4ox7bVTjpWW8"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# creating the sparse pipeline and the tfidf pipeline\n",
        "sparse_features = Pipeline([('sparse',SparseTransformer()),])\n",
        "\n",
        "vectorizer = Pipeline([('tfidf',TfidfVectorizer(min_df = 2,\n",
        "                                   max_df = 100,\n",
        "                                   token_pattern =r\"\\b\\w[\\w@#.!?]*\\b\",\n",
        "                                   max_features = 1000,\n",
        "                                   ngram_range = (1,3), # Pulls in bigrams and trigrams for common phrasing such as 'You won a','Congratulations you win', etc.\n",
        "                                   stop_words = None, # Stop words can be helpful in classifying spam emails because of common phrases.\n",
        "                                   lowercase = False # Keeping both capitalized and uncapitalized versions of words.\n",
        "                                   )),])\n"
      ],
      "metadata": {
        "id": "8Eg50icbqgTk"
      },
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ColumnTransformer??"
      ],
      "metadata": {
        "id": "AyF53PdskyUe"
      },
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# creating the combined features transformer pipeline to be implemented in the last classifier\n",
        "combined_features = ColumnTransformer(\n",
        "      transformers = [\n",
        "          ('tfidf', vectorizer, 'cleaned_text'),],      # Calls the tfidf pipeline using vectorizer on the 'cleaned_text' column\n",
        "            remainder =sparse_features                 # Calls the sparse features pipeline on the remaining columns\n",
        ")"
      ],
      "metadata": {
        "id": "coCUtmy7rpVG"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Last classifier pipelines:**"
      ],
      "metadata": {
        "id": "qIYWUDDCstU2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "classifier_3 = Pipeline([('combined_features', combined_features),\n",
        "                         ('classifier', XGBClassifier(enable_categorical = True, eval_metric = 'logloss')),])"
      ],
      "metadata": {
        "id": "gSkzWtDNsqsR"
      },
      "execution_count": 49,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "classifier_3.get_params().keys()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "G40knyRCteq8",
        "outputId": "9edb5f15-f7b9-4768-c14c-823df9397eb4"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['memory', 'steps', 'verbose', 'combined_features', 'classifier', 'combined_features__n_jobs', 'combined_features__remainder__memory', 'combined_features__remainder__steps', 'combined_features__remainder__verbose', 'combined_features__remainder__sparse', 'combined_features__remainder', 'combined_features__sparse_threshold', 'combined_features__transformer_weights', 'combined_features__transformers', 'combined_features__verbose', 'combined_features__verbose_feature_names_out', 'combined_features__tfidf', 'combined_features__tfidf__memory', 'combined_features__tfidf__steps', 'combined_features__tfidf__verbose', 'combined_features__tfidf__tfidf', 'combined_features__tfidf__tfidf__analyzer', 'combined_features__tfidf__tfidf__binary', 'combined_features__tfidf__tfidf__decode_error', 'combined_features__tfidf__tfidf__dtype', 'combined_features__tfidf__tfidf__encoding', 'combined_features__tfidf__tfidf__input', 'combined_features__tfidf__tfidf__lowercase', 'combined_features__tfidf__tfidf__max_df', 'combined_features__tfidf__tfidf__max_features', 'combined_features__tfidf__tfidf__min_df', 'combined_features__tfidf__tfidf__ngram_range', 'combined_features__tfidf__tfidf__norm', 'combined_features__tfidf__tfidf__preprocessor', 'combined_features__tfidf__tfidf__smooth_idf', 'combined_features__tfidf__tfidf__stop_words', 'combined_features__tfidf__tfidf__strip_accents', 'combined_features__tfidf__tfidf__sublinear_tf', 'combined_features__tfidf__tfidf__token_pattern', 'combined_features__tfidf__tfidf__tokenizer', 'combined_features__tfidf__tfidf__use_idf', 'combined_features__tfidf__tfidf__vocabulary', 'classifier__objective', 'classifier__base_score', 'classifier__booster', 'classifier__callbacks', 'classifier__colsample_bylevel', 'classifier__colsample_bynode', 'classifier__colsample_bytree', 'classifier__device', 'classifier__early_stopping_rounds', 'classifier__enable_categorical', 'classifier__eval_metric', 'classifier__feature_types', 'classifier__gamma', 'classifier__grow_policy', 'classifier__importance_type', 'classifier__interaction_constraints', 'classifier__learning_rate', 'classifier__max_bin', 'classifier__max_cat_threshold', 'classifier__max_cat_to_onehot', 'classifier__max_delta_step', 'classifier__max_depth', 'classifier__max_leaves', 'classifier__min_child_weight', 'classifier__missing', 'classifier__monotone_constraints', 'classifier__multi_strategy', 'classifier__n_estimators', 'classifier__n_jobs', 'classifier__num_parallel_tree', 'classifier__random_state', 'classifier__reg_alpha', 'classifier__reg_lambda', 'classifier__sampling_method', 'classifier__scale_pos_weight', 'classifier__subsample', 'classifier__tree_method', 'classifier__validate_parameters', 'classifier__verbosity'])"
            ]
          },
          "metadata": {},
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Hyperparameter Pipeline:**"
      ],
      "metadata": {
        "id": "jeicuslVttqY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "param_grid_classifier_3 = {\n",
        "    'combined_features__tfidf__tfidf__max_features': [300,500,700,1000], # Max features to test through\n",
        "    'classifier__n_estimators': [20,50,100], # Number of boosting rounds in XGB\n",
        "    'classifier__learning_rate': [0.01,0.1,0.2], # Learning rate for each boosting round\n",
        "    'classifier__max_depth': [3,5,7,9,11], # Maximum depth of trees (higher values make the model more complex)\n",
        "    'classifier__scale_pos_weight': [1,10,25,50,75,99,100,1000] # Class weight values to handle the class imbalance and some for exploration.\n",
        "}"
      ],
      "metadata": {
        "id": "B8qc-G22tjL8"
      },
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Specifying RandomSearch:**"
      ],
      "metadata": {
        "id": "1Vo0edK-unr_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cv_strategy = RepeatedStratifiedKFold(n_splits = 5,n_repeats = 1, random_state = 42)\n",
        "\n",
        "grid_classifier_3 = RandomizedSearchCV(\n",
        "    estimator = classifier_3,\n",
        "    param_distributions = param_grid_classifier_3,\n",
        "    cv = cv_strategy,\n",
        "    n_iter = 300,\n",
        "    scoring = 'f1_macro',\n",
        "    verbose = 2,\n",
        "    n_jobs = -1\n",
        ")\n",
        ""
      ],
      "metadata": {
        "id": "M1lHSQY2uiuF"
      },
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Fitting the Model:**"
      ],
      "metadata": {
        "id": "hqSB7Pkivgk6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "grid_classifier_3.fit(X_train_combo,y_train_combo)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 236
        },
        "id": "kHdS6LhGvJ9a",
        "outputId": "4cad2bd3-e490-4435-d27a-c8763baea14d"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 5 folds for each of 300 candidates, totalling 1500 fits\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[('combined_features',\n",
              "                                              ColumnTransformer(remainder=Pipeline(steps=[('sparse',\n",
              "                                                                                           SparseTransformer())]),\n",
              "                                                                transformers=[('tfidf',\n",
              "                                                                               Pipeline(steps=[('tfidf',\n",
              "                                                                                                TfidfVectorizer(lowercase=False,\n",
              "                                                                                                                max_df=100,\n",
              "                                                                                                                max_features=1000,\n",
              "                                                                                                                min_df=2,\n",
              "                                                                                                                ngram_range=(1,\n",
              "                                                                                                                             3),\n",
              "                                                                                                                toke...\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={'classifier__learning_rate': [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        'classifier__max_depth': [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        'classifier__n_estimators': [20, 50,\n",
              "                                                                     100],\n",
              "                                        'classifier__scale_pos_weight': [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000],\n",
              "                                        'combined_features__tfidf__tfidf__max_features': [300,\n",
              "                                                                                          500,\n",
              "                                                                                          700,\n",
              "                                                                                          1000]},\n",
              "                   scoring='f1_macro', verbose=2)"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[(&#x27;combined_features&#x27;,\n",
              "                                              ColumnTransformer(remainder=Pipeline(steps=[(&#x27;sparse&#x27;,\n",
              "                                                                                           SparseTransformer())]),\n",
              "                                                                transformers=[(&#x27;tfidf&#x27;,\n",
              "                                                                               Pipeline(steps=[(&#x27;tfidf&#x27;,\n",
              "                                                                                                TfidfVectorizer(lowercase=False,\n",
              "                                                                                                                max_df=100,\n",
              "                                                                                                                max_features=1000,\n",
              "                                                                                                                min_df=2,\n",
              "                                                                                                                ngram_range=(1,\n",
              "                                                                                                                             3),\n",
              "                                                                                                                toke...\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={&#x27;classifier__learning_rate&#x27;: [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        &#x27;classifier__max_depth&#x27;: [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        &#x27;classifier__n_estimators&#x27;: [20, 50,\n",
              "                                                                     100],\n",
              "                                        &#x27;classifier__scale_pos_weight&#x27;: [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000],\n",
              "                                        &#x27;combined_features__tfidf__tfidf__max_features&#x27;: [300,\n",
              "                                                                                          500,\n",
              "                                                                                          700,\n",
              "                                                                                          1000]},\n",
              "                   scoring=&#x27;f1_macro&#x27;, verbose=2)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">RandomizedSearchCV</label><div class=\"sk-toggleable__content\"><pre>RandomizedSearchCV(cv=RepeatedStratifiedKFold(n_repeats=1, n_splits=5, random_state=42),\n",
              "                   estimator=Pipeline(steps=[(&#x27;combined_features&#x27;,\n",
              "                                              ColumnTransformer(remainder=Pipeline(steps=[(&#x27;sparse&#x27;,\n",
              "                                                                                           SparseTransformer())]),\n",
              "                                                                transformers=[(&#x27;tfidf&#x27;,\n",
              "                                                                               Pipeline(steps=[(&#x27;tfidf&#x27;,\n",
              "                                                                                                TfidfVectorizer(lowercase=False,\n",
              "                                                                                                                max_df=100,\n",
              "                                                                                                                max_features=1000,\n",
              "                                                                                                                min_df=2,\n",
              "                                                                                                                ngram_range=(1,\n",
              "                                                                                                                             3),\n",
              "                                                                                                                toke...\n",
              "                                                            num_parallel_tree=None,\n",
              "                                                            random_state=None, ...))]),\n",
              "                   n_iter=300, n_jobs=-1,\n",
              "                   param_distributions={&#x27;classifier__learning_rate&#x27;: [0.01, 0.1,\n",
              "                                                                      0.2],\n",
              "                                        &#x27;classifier__max_depth&#x27;: [3, 5, 7, 9,\n",
              "                                                                  11],\n",
              "                                        &#x27;classifier__n_estimators&#x27;: [20, 50,\n",
              "                                                                     100],\n",
              "                                        &#x27;classifier__scale_pos_weight&#x27;: [1, 10,\n",
              "                                                                         25, 50,\n",
              "                                                                         75, 99,\n",
              "                                                                         100,\n",
              "                                                                         1000],\n",
              "                                        &#x27;combined_features__tfidf__tfidf__max_features&#x27;: [300,\n",
              "                                                                                          500,\n",
              "                                                                                          700,\n",
              "                                                                                          1000]},\n",
              "                   scoring=&#x27;f1_macro&#x27;, verbose=2)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;combined_features&#x27;,\n",
              "                 ColumnTransformer(remainder=Pipeline(steps=[(&#x27;sparse&#x27;,\n",
              "                                                              SparseTransformer())]),\n",
              "                                   transformers=[(&#x27;tfidf&#x27;,\n",
              "                                                  Pipeline(steps=[(&#x27;tfidf&#x27;,\n",
              "                                                                   TfidfVectorizer(lowercase=False,\n",
              "                                                                                   max_df=100,\n",
              "                                                                                   max_features=1000,\n",
              "                                                                                   min_df=2,\n",
              "                                                                                   ngram_range=(1,\n",
              "                                                                                                3),\n",
              "                                                                                   token_pattern=&#x27;\\\\b\\\\w[\\\\w@#.!?]*\\\\b&#x27;))]),\n",
              "                                                  &#x27;cleaned_text&#x27;)])),\n",
              "                (&#x27;classifier&#x27;,\n",
              "                 XGBClassifier(base_score=N...\n",
              "                               feature_types=None, gamma=None, grow_policy=None,\n",
              "                               importance_type=None,\n",
              "                               interaction_constraints=None, learning_rate=None,\n",
              "                               max_bin=None, max_cat_threshold=None,\n",
              "                               max_cat_to_onehot=None, max_delta_step=None,\n",
              "                               max_depth=None, max_leaves=None,\n",
              "                               min_child_weight=None, missing=nan,\n",
              "                               monotone_constraints=None, multi_strategy=None,\n",
              "                               n_estimators=None, n_jobs=None,\n",
              "                               num_parallel_tree=None, random_state=None, ...))])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">combined_features: ColumnTransformer</label><div class=\"sk-toggleable__content\"><pre>ColumnTransformer(remainder=Pipeline(steps=[(&#x27;sparse&#x27;, SparseTransformer())]),\n",
              "                  transformers=[(&#x27;tfidf&#x27;,\n",
              "                                 Pipeline(steps=[(&#x27;tfidf&#x27;,\n",
              "                                                  TfidfVectorizer(lowercase=False,\n",
              "                                                                  max_df=100,\n",
              "                                                                  max_features=1000,\n",
              "                                                                  min_df=2,\n",
              "                                                                  ngram_range=(1,\n",
              "                                                                               3),\n",
              "                                                                  token_pattern=&#x27;\\\\b\\\\w[\\\\w@#.!?]*\\\\b&#x27;))]),\n",
              "                                 &#x27;cleaned_text&#x27;)])</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">tfidf</label><div class=\"sk-toggleable__content\"><pre>cleaned_text</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-5\" type=\"checkbox\" ><label for=\"sk-estimator-id-5\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">TfidfVectorizer</label><div class=\"sk-toggleable__content\"><pre>TfidfVectorizer(lowercase=False, max_df=100, max_features=1000, min_df=2,\n",
              "                ngram_range=(1, 3), token_pattern=&#x27;\\\\b\\\\w[\\\\w@#.!?]*\\\\b&#x27;)</pre></div></div></div></div></div></div></div></div><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-6\" type=\"checkbox\" ><label for=\"sk-estimator-id-6\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">remainder</label><div class=\"sk-toggleable__content\"><pre></pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-7\" type=\"checkbox\" ><label for=\"sk-estimator-id-7\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">SparseTransformer</label><div class=\"sk-toggleable__content\"><pre>SparseTransformer()</pre></div></div></div></div></div></div></div></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-8\" type=\"checkbox\" ><label for=\"sk-estimator-id-8\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">XGBClassifier</label><div class=\"sk-toggleable__content\"><pre>XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
              "              colsample_bylevel=None, colsample_bynode=None,\n",
              "              colsample_bytree=None, device=None, early_stopping_rounds=None,\n",
              "              enable_categorical=True, eval_metric=&#x27;logloss&#x27;,\n",
              "              feature_types=None, gamma=None, grow_policy=None,\n",
              "              importance_type=None, interaction_constraints=None,\n",
              "              learning_rate=None, max_bin=None, max_cat_threshold=None,\n",
              "              max_cat_to_onehot=None, max_delta_step=None, max_depth=None,\n",
              "              max_leaves=None, min_child_weight=None, missing=nan,\n",
              "              monotone_constraints=None, multi_strategy=None, n_estimators=None,\n",
              "              n_jobs=None, num_parallel_tree=None, random_state=None, ...)</pre></div></div></div></div></div></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 53
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Getting the best parameters:**"
      ],
      "metadata": {
        "id": "zvbn9FO-xnr3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('Best Cross validation:',{grid_classifier_3.best_score_}) # calling the best score attribute\n",
        "print('\\nBest Parameters:',grid_classifier_3.best_params_) #best param attribute\n",
        "print('\\nBest estimator:', grid_classifier_3.best_estimator_) # best estimator attribute"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RzfDIRuPwC94",
        "outputId": "c4790f1e-f7d9-4548-e3ae-73c52ba32d2c"
      },
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Best Cross validation: {0.9364192305701156}\n",
            "\n",
            "Best Parameters: {'combined_features__tfidf__tfidf__max_features': 500, 'classifier__scale_pos_weight': 25, 'classifier__n_estimators': 50, 'classifier__max_depth': 3, 'classifier__learning_rate': 0.1}\n",
            "\n",
            "Best estimator: Pipeline(steps=[('combined_features',\n",
            "                 ColumnTransformer(remainder=Pipeline(steps=[('sparse',\n",
            "                                                              SparseTransformer())]),\n",
            "                                   transformers=[('tfidf',\n",
            "                                                  Pipeline(steps=[('tfidf',\n",
            "                                                                   TfidfVectorizer(lowercase=False,\n",
            "                                                                                   max_df=100,\n",
            "                                                                                   max_features=500,\n",
            "                                                                                   min_df=2,\n",
            "                                                                                   ngram_range=(1,\n",
            "                                                                                                3),\n",
            "                                                                                   token_pattern='\\\\b\\\\w[\\\\w@#.!?]*\\\\b'))]),\n",
            "                                                  'cleaned_text')])),\n",
            "                ('classifier',\n",
            "                 XGBClassifier(base_score=No...\n",
            "                               feature_types=None, gamma=None, grow_policy=None,\n",
            "                               importance_type=None,\n",
            "                               interaction_constraints=None, learning_rate=0.1,\n",
            "                               max_bin=None, max_cat_threshold=None,\n",
            "                               max_cat_to_onehot=None, max_delta_step=None,\n",
            "                               max_depth=3, max_leaves=None,\n",
            "                               min_child_weight=None, missing=nan,\n",
            "                               monotone_constraints=None, multi_strategy=None,\n",
            "                               n_estimators=50, n_jobs=None,\n",
            "                               num_parallel_tree=None, random_state=None, ...))])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Saving the model:**"
      ],
      "metadata": {
        "id": "saF-bmYPyNbP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "file_best_estimator_pipeline3 = model_folder/'pipeline3_best_estimator.pkl'\n",
        "\n",
        "file_complete_grid_pipeline3 = model_folder/'pipeline3_complete_grid.pkl'\n",
        "\n",
        "# dumping in drive\n",
        "joblib.dump(grid_classifier_3.best_estimator_, file_best_estimator_pipeline3)\n",
        "\n",
        "joblib.dump(grid_classifier_3,file_complete_grid_pipeline3)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2J56MTO9x2ft",
        "outputId": "53709ed5-113e-4a79-fcff-e7f7b8221fd7"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/drive/MyDrive/Colab Notebooks/models/nlp_fall_2024/spam/pipeline3_complete_grid.pkl']"
            ]
          },
          "metadata": {},
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Reloading the Model:**"
      ],
      "metadata": {
        "id": "anlMfFllycC5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "best_estimator_pipeline3 = joblib.load(file_best_estimator_pipeline3)\n",
        "\n",
        "complete_grid_pipeline3 = joblib.load(file_complete_grid_pipeline3)"
      ],
      "metadata": {
        "id": "mFlh_rmFybZ_"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Plotting the learning curve:**"
      ],
      "metadata": {
        "id": "Ay4XwKHYzTEp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plot_learning_curve(\n",
        "    best_estimator_pipeline3,\n",
        "    'Learning Curves for XGBoost with F1: Manual and TDIDF Features',\n",
        "    X_train_combo,\n",
        "    y_train_combo,\n",
        "    scoring = 'f1_macro',\n",
        "    n_jobs = -1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 267
        },
        "id": "x11fiy6SywAo",
        "outputId": "5427c048-522f-462a-ede4-430549cd23b8"
      },
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<module 'matplotlib.pyplot' from '/usr/local/lib/python3.10/dist-packages/matplotlib/pyplot.py'>"
            ]
          },
          "metadata": {},
          "execution_count": 59
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x500 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **<font color='Pickle'>Observations</font>**\n",
        "\n",
        "- **<font color='Pickle'>Learning Curve:</font>**  \n",
        "  This model still shows indication of **<font color='indianred'>overfitting.</font>**\n",
        " This is because of the model nearly perfectly fitting the training data. However we can see a consistent increase in the **<font color='Pickle'>CV</font>** scores of the validation data and a slight decrease in the **<font color='Pickle'>training score</font>**, indicating that this model may have the best generalization of the three despite being more complex.\n",
        "\n",
        "- **<font color='Pickle'>Scalability of the Model:</font>**  \n",
        "  As with the first model we can see a general upward and steady trend of fit times as the training examples increase. The fit times are slightly larger and have higher varaibility as the training examples increase than the other two models which can cause a bit of concern. However, in a limited compute environment the ability to generalize will hold a bit more weight since the fit times are not significantly higher.\n",
        "\n",
        "- **<font color='Pickle'>Performance of the Model:</font>**  \n",
        "We can see a steady improvement in model performance overall as the fit times increase. We can also see a slight taper in the shaded area of the graph as we move to the left, suggesting that as the fit times and data increase the model becomes more stable."
      ],
      "metadata": {
        "id": "dNUC3oyn0hWT"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**CV and Train Score:**"
      ],
      "metadata": {
        "id": "-KINjyT50uIP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Checking train\n",
        "print(best_estimator_pipeline3.score(X_train_combo,y_train_combo))\n",
        "\n",
        "#checking CV\n",
        "print(complete_grid_pipeline3.best_score_)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Lfztwn2DzjkG",
        "outputId": "0c2c8b1f-d903-4914-c19a-8e01a9c910ef"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.9940119760479041\n",
            "0.9364192305701156\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Evaluating on the test set:**"
      ],
      "metadata": {
        "id": "UzEr5KAV1MWg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def final_pipeline(text):\n",
        "  cleaned_text = joblib.load(file_X_test_cleaned_s) #pulling in the cleaned tfidf data from pipeline1\n",
        "  X_features, feature_names = featurizer.fit_transform(text)\n",
        "  X_combo = pd.concat((pd.DataFrame(cleaned_text,columns = ['cleaned_text']),\n",
        "                       pd.DataFrame(X_features, columns = feature_names)),axis = 1)\n",
        "  best_estimator_pipeline3 = joblib.load(file_best_estimator_pipeline3)\n",
        "  predictions = best_estimator_pipeline3.predict(X_combo)\n",
        "  return predictions"
      ],
      "metadata": {
        "id": "O7IrNyVF1G-X"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# predicting the final values using the testing set\n",
        "y_test_pred_combo = final_pipeline(X_test_combo)"
      ],
      "metadata": {
        "id": "JvHudM3w4M-A"
      },
      "execution_count": 62,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###<font color = 'Pickle'>**Classification Report:**"
      ],
      "metadata": {
        "id": "1EnSDM-14hgQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print('\\nTest set classification report:\\n\\n',\n",
        "      classification_report(y_test_combo, y_test_pred_combo))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9MTsXkDl4e2m",
        "outputId": "24e2af74-d2de-4ddc-eae3-71e15efb5d52"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Test set classification report:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.98      0.95      0.97       187\n",
            "           1       0.79      0.92      0.85        36\n",
            "\n",
            "    accuracy                           0.95       223\n",
            "   macro avg       0.88      0.93      0.91       223\n",
            "weighted avg       0.95      0.95      0.95       223\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **<font color='Pickle'>Model Interpretation Summary</font>**\n",
        "\n",
        "- **<font color='Pickle'>Class 0 (Ham - Majority Class):</font>**\n",
        "  - **<font color='Pickle'>Precision: 0.98</font>** – The model correctly identifies 98% of ham emails, making it highly reliable in avoiding false positives for legitimate, non-spam emails.\n",
        "  - **<font color='Pickle'>Recall: 0.95</font>** – It captures the vast majority of ham emails, missing only 5%, which is excellent coverage.\n",
        "  - **<font color='Pickle'>F1-Score: 0.97</font>** – The balance between precision and recall is strong, giving an overall F1-score of 0.97. This shows excellent performance on the majority class.\n",
        "\n",
        "- **<font color='Pickle'>Class 1 (Spam - Minority Class):</font>**\n",
        "  - **<font color='Pickle'>Precision: 0.79</font>** – When the model predicts an email as spam, it’s accurate 79% of the time. However, there are still some false positives, meaning 21% of emails flagged as spam are actually ham.\n",
        "  - **<font color='Pickle'>Recall: 0.92</font>** – It successfully identifies 92% of the actual spam emails, showing strong spam-catching ability, though there’s still a little room for improvement.\n",
        "  - **<font color='Pickle'>F1-Score: 0.85</font>** – The balance between precision and recall here gives us an F1-score of 0.85. This reflects solid handling of spam, but there’s still a chance for better performance.\n",
        "\n",
        "### **<font color='Pickle'>Overall Takeaway:</font>**\n",
        "\n",
        "- **<font color='Pickle'>Majority Class (Ham) Performance:</font>** The model is spot-on when it comes to identifying ham emails, with both precision and recall near perfect. Legitimate emails are rarely misclassified, making this a big win for accuracy and user experience.\n",
        "  \n",
        "- **<font color='Pickle'>Spam Detection (Minority Class):</font>** The model does a great job catching most spam emails, but with slightly lower precision, meaning some legitimate emails still get flagged as spam. The strong F1-score of 0.85 shows we’ve struck a good balance between catching spam and minimizing false positives.\n",
        "\n",
        "- **<font color='Pickle'>Watch Out for Overfitting:</font>** Despite the high overall performance, the slightly uneven handling of the minority class suggests some potential overfitting. This could impact generalization on unseen data, so it’s worth keeping an eye on as more data is tested.\n",
        "\n",
        "### **<font color='Pickle'>Performance Metrics Summary:</font>**\n",
        "- **<font color='Pickle'>Accuracy:</font>** 95%\n",
        "- **<font color='Pickle'>Macro Average F1-Score:</font>** 0.91\n",
        "- **<font color='Pickle'>Weighted Average F1-Score:</font>** 0.95\n",
        "\n"
      ],
      "metadata": {
        "id": "nligURfasFOr"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#<font color = \"Pickle\">**Final Model Choice**\n",
        "- <font color = 'Pickle'>**TFIDF + Manual Features (combo) Pipeline:**</font>\n",
        "  - The primary reason for this choice is the larger potential for variablity in hyperparameter tuning and the increase in model performance over time. I was caught between using the Manual features only or this pipeline because they both have <font color = \"pickle\">**F1-Scores**</font> of 0.85 for the <font color = \"pickle\">**spam**</font>  class and 0.97 for the <font color = \"pickle\">**ham**</font> class. Both models suffer from overfitting but the we can see a trend of the gap between the CV scores and training scores decrease more as more samples are added for the <font color = \"pickle\">**combination**</font>  pipeline."
      ],
      "metadata": {
        "id": "a8OMaNkDtLqO"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "hbz856vL4wjm"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}