A Scalable URL-Based Phishing Websites Detection System Using Machine Learning Techniques

The Internet has become incredibly important to everyone in today's world. Everyone uses the Internet to conduct their online business, including buying, banking, booking, and other activities. As a result, phishing and other cyberattacks have become more prevalent. Phishing is a kind of attack that preys on unwary online users by tricking them into giving personal details together with passwords, usernames, credit card details, and social security numbers. By disguising a webpage as a reliable or reputable one, attackers trick Internet users into providing personal information. Although numerous anti-phishing techniques, including blacklists and white lists, rule-based detection, heuristics, and methods based on visual similarities have been developed to date, Internet users are still tricked into divulging important information on fake websites. ML is one of the best techniques for spotting these harmful actions. This is because the vast majority of phishing attacks have characteristics that machine learning can detect. We'll offer a phishing detection system in this research that uses machine learning, specifically supervised learning, to determine whether a website is authentic or fraudulent. In this work, the RF, DT, LR, and KNN supervised learning models are used. RF is regarded as the best model through comparisons because of its high accuracy and low false-negative value compared with other models.