Smart Cities Hybridized to Prevent Phishing URL Attacks

Abstract

For intelligent phishing site recognition, this proposal introduces particle swarm optimization-based feature weights in order to improve phishing site detection. Particle Swarm Optimization (PSO) is used to identify phishing sites more accurately by checking multiple website properties. PSO-based recommended site feature weighting is used to rank web elements according to their importance in distinguishing real websites from phishing sites. Based on the test results, the PSO-based feature weighting significantly improved the classification accuracy, the true positive and negative rates, and the false negative and false positive rates. Phishing is the collection of personal information through fake websites, including passwords, account numbers, and credit card details. Attackers lure fake visitors by using attractive URLs. Recently, the Unified Resource Locator phishing was successfully detected using machine learning-based detection. K-nearest neighbors, decision trees, and random forests are just some of the machine learning classifiers used to determine if a site is real or not. This classification may make it easier to identify fake sites. A genetic algorithm, however, has been shown to improve the accuracy of feature selection and thus increase the detection efficiency.