Phishing Detection Using Machine Learning

Phishing attacks remain one of the most prevalent cyber threats aimed at stealing sensitive information by impersonating legitimate entities. Traditional blacklist-based and rule-based security systems fail to detect newly generated phishing URLs due to their dynamic and evolving nature. Machine Learning (ML)-based models have emerged as an effective solution by learning patterns from URL features, webpage content, and host behavior. This paper reviews existing ML techniques for phishing detection, discusses important feature extraction methods, compares recent models, identifies gaps in current research, and proposes a hybrid approach combining feature-based and deep learning models for improved detection accuracy.