Performance Evaluation on Detection of Phishing Websites Using Machine Learning Techniques

The rapid growth of the internet has revolutionized the way people communicate and conduct various online activities. However, this increased connectivity has also increased cyber threats, notably phishing assaults being among the most prevalent and damaging. By pretending to be legitimate, phishing attacks seek to deceive visitors into disclosing confidential information. Machine Learning (ML) has emerged as a pivotal tool for both attackers and defenders to enhance the sophistication of phishing websites and improve detection and prevention methods. Attackers use ML to create more convincing attacks, defenders leverage it to improve their detection and prevention capabilities. As a result, there is a critical need for effective techniques to detect and combat phishing websites. We employ methods such as 10-fold cross-validation and 75-25 splits and assess six classifiers: Naive Bayes, SMO, PART, JRIP, J48, and Random Forest. Our study investigates various feature extraction methods, classification algorithms, and evaluation metrics to identify the most effective approach. The results offer valuable insights into the strengths and limitations of each technique, guiding the development of robust cybersecurity defenses against phishing attacks.