Spam Detection Model

Introduction

Welcome to the Spam Detection Model project! This repository provides a robust machine learning solution for classifying email text as "Spam" or "Not Spam (Ham)" using advanced natural language processing (NLP) techniques and a Logistic Regression classifier. The project includes a user-friendly Streamlit web application for real-time predictions and a Jupyter Notebook for experimentation and model exploration.

Key Features

• Intuitive Web Interface: Easily interact with the model using a modern Streamlit app.

• Comprehensive Text Preprocessing: Utilizes NLTK for tokenization, stopword removal, and lemmatization to ensure high-quality input for the model.

• Accurate Machine Learning Model: Logistic Regression trained on vectorized email data for reliable spam detection.

• Jupyter Notebook Support: spam_detection.ipynb allows you to explore data, experiment with preprocessing, train models, and visualize results.

• Instant Results: Get immediate feedback on your email text classification.

Getting Started

Prerequisites

• Python 3.7 or higher

• Required Python packages (see requirement.txt)

Installation

1. Clone the repository or download the source code:

   git clone <repository_url>

2. Install dependencies:

   pip install -r requirement.txt

Running the Streamlit Application

1. Ensure spam_detection_model.pkl and vectorizer.pkl are present in the project directory.

2. Launch the Streamlit app:

   streamlit run spam_detection.py

3. Enter your email text in the provided text area and click Predict to view the classification result.

Using the Jupyter Notebook

1. Open spam_detection.ipynb in Jupyter Lab or Jupyter Notebook.

2. Explore data preprocessing, model training, evaluation, and predictions step by step.

3. Modify parameters or try different ML algorithms directly in the notebook.

How to Train Your Own Model

If you want to train the spam detection model from scratch or update it with new data, follow these steps:

1. Open spam_detection.ipynb.

2. Load your dataset containing emails and their labels (Spam or Ham).

3. Preprocess the text:

   • Tokenization

   • Stopword removal

   • Lemmatization

4. Vectorize the text using CountVectorizer or TfidfVectorizer.

5. Train a machine learning model, e.g., Logistic Regression, on the vectorized features.

6. Evaluate the model using metrics like accuracy, precision, recall, and F1-score.

7. Save the trained model and vectorizer using joblib:

   import joblib joblib.dump(model, 'spam_detection_model.pkl') joblib.dump(vectorizer, 'vectorizer.pkl')

8. Use the saved files in the Streamlit app for real-time predictions.

Example Usage

Subject: Congratulations! You have won a prize. Click here to claim your reward now.

Prediction: SPAM

Project Structure

• spam_detection.py : Streamlit application script

• spam_detection.ipynb : Jupyter Notebook for experimentation and model development

• spam_detection_model.pkl : Pre-trained machine learning model

• vectorizer.pkl : Text vectorizer for feature extraction

• requirement.txt : List of required Python packages

• Spam_detection_model/README.md : Project documentation

• Spam_detection_model/LICENSE : License information

License

This project is licensed under the MIT License. See the LICENSE for details.

Acknowledgements

• Developed by LovishTech

• Built with Streamlit, scikit-learn, NLTK, pandas, joblib, and Jupyter Notebook