Phishing Website Detection Using Machine Learning Algorithms

Abstract

Abstract— The complexity of phishing assaults has increased, making it exceedingly challenging for the typical person to discern the legitimacy of an email message link or website. Nowadays, phishing efforts are a prevalent and successful tactic used by cybercriminals. In this section we provide an intelligent solution that reduces security risks for people and businesses by giving users an interface to determine whether a URL is phishing or authentic. The foundation of the system is machine learning, namely supervised learning. We chose the Gradient Boosting method because of its excellent classification performance. Our goal is to develop a better classifier by evaluating the characteristics of phishing websites and selecting the optimal combination of them to train the classifier with. As a result, our paper has 97.4% correctness and a total of 32 characteristics.