SMS Spam Detection Using Machine Learning

An intelligent SMS spam detection system that uses machine learning to classify messages as spam or legitimate (ham). The system combines TF-IDF text vectorization, dimensionality reduction, handcrafted linguistic features, and XGBoost classification to achieve high accuracy.

🎯 Project Overview

This project addresses the growing problem of SMS spam by building an automated, data-driven detection system. Traditional rule-based filters struggle with evolving spam patterns, so we use machine learning to learn distinguishing patterns automatically.

Key Features

• High Accuracy: Achieves 98.5% accuracy on test data
• Low False Positives: Maintains precision of 97.8%
• Robust Performance: AUC score of 99.1%
• Real-time Classification: Lightweight model suitable for production deployment
• Comprehensive Feature Engineering: Combines text and statistical features

👥 Team Members

• Saif Mohamed
• Muhab Abdelraouf
• Adham Tamer
• Ahmed Sharif
• Ahmed Ramadan
• Amr Hamoda

📊 Results

Metric	Score
Accuracy	98.5%
Precision	97.8%
Recall	90.6%
AUC	99.1%

Confusion Matrix:

[[963   3]
 [ 14 135]]

🛠️ Technology Stack

• Python 3.x
• scikit-learn: Feature extraction and preprocessing
• XGBoost: Gradient boosting classifier
• NLTK: Text tokenization and stopword removal
• Pandas & NumPy: Data manipulation
• TQDM: Progress tracking

📁 Project Structure

sms-spam-detection/
├── data/
│   └── spam.csv                 # Dataset (not included, see below)
├── notebooks/
│   └── sms_spam_detection.ipynb # Full implementation notebook
├── src/
│   ├── __init__.py
│   ├── preprocessing.py         # Text cleaning and feature engineering
│   ├── model.py                 # Model training and prediction
│   └── utils.py                 # Helper functions
├── models/
│   └── trained_model.pkl        # Saved model (generated after training)
├── requirements.txt             # Project dependencies
├── README.md                    # This file
└── LICENSE                      # Project license

🚀 Getting Started

Prerequisites

• Python 3.7 or higher
• pip package manager

Installation

1. Clone the repository:

git clone https://github.com/yourusername/sms-spam-detection.git
cd sms-spam-detection

2. Install dependencies:

pip install -r requirements.txt

3. Download NLTK data:

python -c "import nltk; nltk.download('punkt'); nltk.download('stopwords')"

Dataset

This project uses the SMS Spam Collection Dataset.

Download the dataset:

1. Download spam.csv from the link above
2. Place it in the data/ directory

Usage

Training the Model

from src.model import SMSSpamDetector

# Initialize and train
detector = SMSSpamDetector()
detector.load_data('data/spam.csv')
detector.train()

# Evaluate
detector.evaluate()

Making Predictions

# Predict single message
message = "Congratulations! You've won a free iPhone. Click here to claim."
result = detector.predict(message)
print(f"Prediction: {result['label']}")
print(f"Spam Probability: {result['probability']:.4f}")

Using the Notebook

Open and run notebooks/sms_spam_detection.ipynb for the complete implementation with visualizations and detailed explanations.

🔬 Methodology

1. Data Preprocessing

• Text cleaning (URL removal, special characters, lowercase conversion)
• Tokenization and stopword removal

2. Feature Engineering

Text Features:

• TF-IDF vectorization (top 3,000 features)
• Dimensionality reduction using Truncated SVD (200 components)

Handcrafted Features:

• Message length (character count)
• Number of words
• Number of sentences
• Average word length
• Stopword ratio

3. Model Training

• Algorithm: XGBoost Classifier
• Hyperparameters:
  • n_estimators: 300
  • max_depth: 8
  • learning_rate: 0.15
• Custom threshold: 0.3 for spam classification (optimized for recall)

4. Evaluation

Multiple metrics ensure comprehensive performance assessment:

• Accuracy, Precision, Recall
• ROC-AUC score
• Confusion matrix analysis

🌐 Real-World Applications

This system can be integrated into:

• Mobile network operator messaging pipelines
• Smartphone messaging applications
• Email filtering systems
• Customer support platforms

The lightweight architecture enables real-time processing with minimal latency.

🔮 Future Enhancements

• Deep learning models (LSTM, BERT)
• Multilingual spam detection
• Active learning for continuous improvement
• Web API for easy integration
• Mobile app demonstration

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the project
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

📧 Contact

For questions or feedback, please open an issue in this repository.

🙏 Acknowledgments

• SMS Spam Collection Dataset from UCI Machine Learning Repository
• scikit-learn and XGBoost communities
• NLTK project for natural language processing tools