Advancement of Phishing Attack Detection Using Machine Learning

Abstract

Phishing attacks continue to pose significant threats to individuals and organizations, necessitating the development of advanced detection mechanisms. This study, propose a novel approach to phishing attack detection leveraging Machine Learning (ML) techniques, focusing on the analysis of features extracted from phishing emails or URLs. The given labeled dataset is datasets containing both malicious and non-malicious instances, supervised ML models on which we demonstrate the effectiveness of this approach in accurately predicting the nature of incoming emails or URLs. Key advanced attributes such as alias symbol, URL length, and presence of advanced characters in URLs are extracted and utilized as features for classification. Various ML Algorithms, including Support Vector Machine, Logistic Regression and Artificial Neural Network are evaluated to identify the most effective classifier for this task. Experimental results on real-world datasets show the proposed approach's high accuracy, precision, and recall rates in detecting phishing attacks.