Machine Learning & Security
Academic Capstone
Cyber Scam & Phishing Detection System
An intelligent cybersecurity application that detects scam and phishing attacks across URLs, emails, SMS, and QR codes using Machine Learning models and explainable AI.
Project Overview & Background
An advanced cybersecurity solution engineered to protect users from multi-vector phishing and scam threats. Combining machine learning classifiers with threat intelligence and heuristic rules, the system inspects suspicious URLs, email headers/bodies, SMS messages, and QR codes. It integrates SHAP and LIME for explainable AI, allowing users and security analysts to understand exactly why a given message or link was flagged.
Key Features & Technical Contributions
- Multi-channel threat scanning across URLs, email content, SMS text, and QR code payloads.
- Trained and evaluated classification models using Scikit-learn with high accuracy on phishing benchmarks.
- Integrated explainable AI (XAI) using SHAP and LIME to generate interpretable feature importance graphs.
- Built a Django REST API backend to process on-demand text, link, and image scanning requests.
Ready to discuss this project?
Feel free to reach out if you have questions about the implementation or architecture.