A Survey on Phishing Website Detection Using Deep Neural Networks

Abstract

Phishing is a social engineering attack, where an attacker poses as a legitimate individual or institution and convinces a victim to divulge their details through human interaction. There has been a steep rise in phishing cases across the globe. A report by Cisco [1] shows that phishing was the reason for 90% of data breaches in 2021. Various detection models have been proposed in the past to counter such attacks. Some proposed models work on improving the detection rate of phishing URLs while others focus on reducing their detection time. Authors have used machine learning, deep learning, and various other novel mechanisms in feature selections that result in high algorithm performance. This study is a systematic analysis of recent work utilizing deep learning for phishing detection, highlighting the research methods, algorithms, programming tools, and datasets used in such studies. This study further proposes some guidelines for future research, which include standardizing documentation and performance reporting. These guidelines may help researchers in their quest to replicate others’ work and compare newly proposed methods with previously developed systems.