Salp Swarm Optimization Search Based Feature Selection for Enhanced Phishing Websites Detection

Internet-connected devices are increasing rapidly. This facilitates transferring most of the real-world transactions to the cyber world. It follows that eCrime is growing continuously. Phishing is a cyber-attack carried out by intruders. They aim to deceive the users of the Internet to achieve their malicious goals. Therefore, experts have developed different approaches to protect financial transactions and personal login information of the users. Their primary concern is to detect the security breaches for online use of the Internet channels (e.g. emails, SMS, webpages, and social platforms). In this paper, we propose a new phishing detection system based on the Salp Swarm Algorithm (SSA). The main objective is to maximize the classification performance and minimize the number of features of the phishing system. Different transfer function (TF) families: S-TFs, V-TFs, X-TFs, U-TFs, and Z-TFs are used to convert the continuous SSA into binary. Sixteen different binary versions of the SSA algorithm are produced based on different TFs. A comparison analysis is performed to pick up the best binarization method. The phishing system is validated by comparing it with three state-of-the-art algorithms. The results show that BSSA with X-TFs achieved the best results in terms of the used evaluation measures.