A Robust Machine Learning Framework for Phishing Attack Detection Using RF, DT, and XGBoost

Abstract

Phishing attacks have emerged as one of the most serious cyber threats, particularly in the domain of social engineering, affecting individuals and organizations in their daily activities and business operations. These attacks often result in severe financial losses and the compromise of sensitive information. Traditional phishing detection methods, such as blocklist-based approaches, are ineffective against newly emerging phishing websites due to delays in updating blocklist databases. In recent years, Machine Learning (ML) has demonstrated significant potential across various domains, including cybersecurity, by enabling the development of intelligent and adaptive threat detection systems.

This study proposes a phishing attack detection framework using three machine learning algorithms: Random Forest (RF), Decision Tree (DT), and Extreme Gradient Boosting (XGBoost). To enhance the performance of these models, various feature selection and classification techniques are incorporated. The primary objective of this research is to develop a robust and efficient phishing detection model capable of providing improved accuracy and reliability in identifying phishing attacks. The performance of the proposed models is evaluated using key metrics such as Accuracy, Precision, and Recall. The proposed approach offers notable improvements over traditional phishing detection methods and contributes to advancing cybersecurity solutions for safeguarding individuals and organizations against phishing threats.