Phishing Detection and Prevention Using Django-Based Machine Learning Frameworks

Project Synopsis & Overview

This project develops an explainable adaptive ensemble learning framework for detecting phishing websites using URL, domain, HTML/JavaScript, and security-related features. Random Forest, XGBoost, and LightGBM are combined using an adaptive AUC-weighted mechanism to improve phishing classification. SHAP provides feature-level explanations for predictions, while a Django web application enables real-time URL scanning, probability scores, scan history, and transparent security analysis.

Complete Technical Specifications