Optimized AI-Driven Semantic Web Approach for Enhancing Phishing Detection in E-Commerce Platforms
Akshat Gaurav, Kwok Tai Chui, Varsha Arya, Razaz Waheeb Attar, Shavi Bansal, Ahmed Alhomoud, Kostas E. Psannis
Abstract
For e-commerce systems, phishing attempts remain a major threat, so sophisticated detection techniques using Semantic Web and artificial intelligence are very necessary. An efficient AI-driven Semantic Web method for phishing detection enhancement is presented in this work. The approach uses the Chi-square feature selection approach along with the Adaptive Differential Evolution with Optional External Archive (JADE) algorithm to optimize the hyperparameters of a Convolutional Neural Network (CNN) model. Having grown up on a large collection of more than 11,000 webpages, the model attained 93% accuracy. Although alternative models sometimes exceeded it in accuracy, the suggested method always showed the lowest loss values throughout all epochs, therefore stressing its stability and efficiency. Comparative study using conventional models confirms its resilience against phishing attacks for protecting e-commerce systems.