This mini project explores email spam classification using a Multilayer Perceptron (MLP) neural network built with TensorFlow. The goal is to learn practical applications of neural networks by building a working text classification model that detects spam based on keywords in emails.
This is designed for hands-on learning and skill development in machine learning and natural language processing (NLP). It helps build understanding of:
- Processing and cleaning real-world email data.
- Feature extraction and text vectorization for ML models.
- Designing and training simple neural networks with TensorFlow.
- Evaluating model performance on classification tasks.
- Generating useful visual insights from data.
- Email data parsing from mbox files using Python's
mailboxlibrary. - Data preprocessing including text cleaning and time zone normalization.
- Visualizations showing email sending patterns and sender analysis.
- Spam/ham classification with a simple MLP with one hidden layer of 3 neurons.
- Learning focused on fundamentals of neural networks and TensorFlow workflows.
| Technology | Purpose |
|---|---|
 |
Programming language |
 |
Neural network framework |
 |
Data manipulation |
 |
Visualizations |
 |
Email parsing |
 |
Word cloud generation |
 |
Timezones handling |
Make sure you have Python 3.7+ installed. Install the required packages using pip:
pip install numpy pandas matplotlib mailbox wordcloud pytz tensorflow
- Download your email mbox files (
Inbox.mboxandSent.mbox). - Update the file paths in the script to point to the local mbox files.
- Run the script or Jupyter notebook: python mlp_email_spam_classifier.py
- View the training output and final evaluation accuracy.
├── Inbox.mbox # Inbox emails in mbox format
├── Sent.mbox # Sent emails in mbox format
├── mlp_email_spam_classifier.py # Main script with processing, modeling, and evaluation
├── README.md # This file
└── requirements.txt # Python dependencies (optional)
The model achieves 100% accuracy on this curated dataset of first 50 emails from inbox and sent folders merged, demonstrating successful keyword-based spam classification.
Below are some key plots generated by the model:
- This project serves as a learning tool focusing on the basics of neural networks and NLP.
- For production use, consider larger datasets and more advanced text processing techniques.
- Do not upload your personal email mbox files if publishing publicly; use sanitized sample data instead.
This project is licensed under the MIT License - see the LICENSE file for details.
Created by Purvi as part of personal learning and skill development.