Enhanced Phishing Detection : A Hybrid SVM-Genetic Algorithm Approach

The majority of public and financial institutions have recently upgraded and enhanced the direct online services they offer to their clients due to the rise in internet applications and users. However, the majority of web users are unaware of internet security measurements. Hence, attacks on various online platforms are gradually increasing. Attackers use various methods to steal users' sensitive information; one of the most common scams is phishing websites. Therefore, there is a need to fight these attacks and constantly improve detection technologies, including machine learning (ML) methods. ML methods classify whether a site is phishing or not based on a number of pieces of data obtained from other webpages. Therefore, this paper aims to present a model for detecting and classifying phishing websites using the Support Vector Machine (SVM) model optimized using the Genetic Algorithm (GA) to obtain the best classification accuracy. The collected dataset consists of 12,000 samples. The phishing URLs were collected from the PhishTank website, while the legitimate ones were from the Kaggle website. Furthermore, accuracy, precision, recall, and the F1-score were used to evaluate the performance of the presented method. The obtained results were compared to the results of previous research, which was conducted using SVM algorithms with Ant Colony Optimization (ACO). The attained results showed that the classification accuracy of the presented approach achieved 97.62%, which is higher than the traditional SVM model by 9.29% and almost equal to the SVM-ACO model.