Phishdect: An Optimised Deep Neural Network Algorithm for Detecting Phishing Attacks in Online Platform

Abstract

In today’s world, phishing attacks are gradually increasing, resulting in individuals losing valuables, assets, personal information, etc., to unauthorized parties. In phishing, attackers craft malicious websites disguised as well-known, legitimate sites and send them to individuals to steal personal information and other related private details. The existing phishing attack detection approach suffers from overfitting, underfitting, vanishing gradients, and local minima, as it tries to optimize a highly non- convex and high-dimensional function resulting in a good fit of the model on the training data while failing to generalize well on new, unseen test data. However, from the literature, population-based WOA can avoid local optima and get a globally optimal solution. These advantages cause WOA to be an appropriate algorithm for solving different constrained or unconstrained optimization problems for practical applications without structural reformation to deep learning algorithms algorithm. Therefore, an efficient and accurate deep learning method is proposed in this study to determine whether a website is malicious using phishing attack datasets on MATLAB 2021a. The experimental results show that the proposed model attains the highest testing accuracy of 98% as against the classical MLP algorithms which achieved the highest testing accuracy of 93%. that, the proposed system achieved the highest precision score of 97%, recall of 98. % and F-score of 97% as against the other classical approaches.