Design of Automated Website Phishing Detection using Sequential Mechanism of RCL Algorithm

Abstract

The phishing outbreaks in internet has become a major problem in web safety in recent years. The phishers will be stealing crucial economic data regarding the web user to perform economic break-in. In order to predict phishing websites, many blacklist-based phishing website recognition methods are used in this study. Traditional methods of detecting phishing websites rely on static features and rule-based schemes, which can be evaded by attackers. Recently, Deep Learning (DL) and Machine Learning (ML) models are employed for automated website phishing detection. With this motivation, this study develops an automated website phishing detection using the sequential mechanism of RCL algorithm. The proposed model employs Long-Short-Term Memory (LSTM), Convolutional Neural Network (CNN), and Random Forest (RF) models for the detection of attacks in the URLs and webpages by the similarity measurement of the decoy contents. The proposed model involves three major components namely, RF for URL phishing detection, CNN based phishing webpage detection, and LSTM based website classification (i.e., legitimate and phishing). The experimental result analysis of the RCL technique is tested on the benchmark dataset of Alexa and PhishTank. A comprehensive comparison study highlighted that the RCL algorithm accomplishes enhanced phishing detection performance over other existing techniques in terms of distinct evaluation metrics.