Development of a Phishing Detection System Using Support Vector Machine

Abstract

Phishing represents a significant and escalating threat within the cyber domain, inflicting substantial financial losses on internet users annually. This illicit practice leverages both social engineering tactics and technological means to unlawfully obtain sensitive information from individuals online. Despite numerous studies and publications exploring various methodologies to combat phishing, the number of victims continues to surge due to the inefficiencies of current security measures. The inherently anonymous and unregulated nature of the internet further compounds its susceptibility to phishing attacks. While it's commonly believed that successful phishing endeavours involve the creation of replica messages or websites to deceive users, this notion has not undergone systematic examination to identify potential vulnerabilities. This paper endeavours to fill this gap by conducting a comprehensive evaluation of phishing, synthesizing diverse research perspectives and methodologies. It introduces an innovative classification method utilizing Support Vector Machine (SVM), achieving an impressive accuracy rate of 96.4% in detecting phishing attempts. By implementing this model to distinguish between phishing and legitimate URLs, the proposed solution offers a valuable tool for individuals and organizations to promptly identify and mitigate phishing threats. The findings of this study hold significant implications for bolstering internet security measures and enhancing user awareness in navigating potentially malicious online content.