IDAKWO SAMUEL AMEH

Phishing has become a major cybercrime threat due to the growth of online transactions and digital communication. This project addresses the challenge by developing a web-based phishing detection system that uses machine learning and explainable AI (XAI) to analyze lexical, content-based, and host-based URL features for real-time classification. Built with Python, Flask, and an XGBoost model, the system achieved approximately 94.79% accuracy, demonstrating strong reliability and effectiveness. Its modular architecture includes a web interface for user input, a backend API for processing, a feature extraction module, and the XGBoost classifier. The XAI component provides clear explanations of predictions, improving user trust, awareness, and decision-making. The system was tested in both local and simulated cloud environments to verify scalability and stability. Overall, it offers a practical, transparent cybersecurity tool suitable for individuals and institutions, contributes to research on intelligent phishing detection, and provides a foundat