URL-based Phishing Websites Detection via Machine Learning

Phishing is a cybersecurity attack that is used to trick victim users to provide sensitive information or deploy malicious software on their infrastructure. Depending on the target system and users, these attacks can inflict severe negative impacts on the system. Therefore, researchers have been working on developing phishing detection and prevention techniques to thwart these attacks. In this paper, we present an efficient phishing websites detection system that analyzes the phishing websites URL addresses to learn data patterns that can identify authentic and phishing websites. Our system employs machine learning techniques such as neural networks and decision trees to learn data patterns in websites URLs. We evaluate our system on a recent phishing websites dataset using classification accuracy as a performance indicator. Our best result shows that decision trees models provide 97.40% classification accuracy on the almost balanced-class dataset.