Classifying Phishing Websites Using Multilayer Perceptron

Abstract

The prevalence of phishing as a cybercrime continues to escalate, posing significant threats to individuals' sensitive information. This paper addresses the urgent need for effective phishing detection methods, considering the limitations of existing approaches. The study employs Artificial Neural Networks, specifically Multilayer Perceptrons (MLP), trained using the backpropogation algorithm. The study also highlights MLP’s advantages in handling complex and noisy data. Through a comprehensive review of related works, the paper identifies gaps in current research and establishes the groundwork for an innovative phishing website classification framework. The proposed solution utilizes MLPs, offering a detailed explanation of the methodology, dataset, model architecture, and training processes. The research concludes by summarizing key findings, emphasizing the solution's contributions to cybersecurity, and outlining potential avenues for future research.