Hybrid Optimization Algorithm to Mitigate Phishing URL Attacks In Smart Cities

Abstract

To improve the detection of phishing sites, this proposal introduces feature weights for intelligent phishing site detection based on hybrid bio-inspired algorithms. The proposed approach uses Gray Wolf Optimization (GWO) and Firefly Algorithm (FF), which examines a wide range of website attributes, to more precisely identify phishing sites. Then employs Artificial Neural Network (ANN) to classify different website elements according to the importance of each component in differentiating between legitimate and phishing websites using bioinspired-based recommended site feature weights. According to experimental findings, the suggested hybrid bioinspired-based feature weighting greatly improved classification accuracy, true positive (TPR), and negative rates (TNR), as well as precision and F1 score. Phishing is an online crime that entails the gathering of private information like passwords, account numbers, and credit card numbers. Attackers use alluring URLs to entice phony website visitors. Recently, Artificial Intelligence-based phishing detection has seen some success, and in this study, ANN was used to detect phishing. This ANN classifier may make phishing websites simpler to spot. However, it has been shown that the effectiveness of detection can be increased by applying a genetic algorithm to improve feature selection.