A novel Smishing defense approach based on meta-heuristic optimization algorithms

Abstract Phishing attacks have recently emerged as one of the most severe dangers to businesses, Internet users, service providers, cloud computing, and many other aspects of daily life. Most SMS phishing messages are short text with simple messages, new writing characteristics, and oversampling techniques for unbalanced data, so they are easy to spot. This paper introduces an approach for detecting SMS phishing based on machine learning algorithms. The suggested system incorporates feature extraction, oversampling, and selection and classification optimization algorithms. The Support vector machine (SVM) is used for feature extraction and categorization. Furthermore, the Adaptive Synthetic Sampling Approach technique was formerly an oversampling technique. The retrieved features are then analyzed using the hybridization of Reptile search and Prairie Dogs algorithms called (R-PDO), which are then used to choose the features’ ideal order. Experimental findings show that the R-PDO approach enhances the accuracy of the SMS phishing detection system. With just an average of 87.4 characteristics, the suggested method in this article achieves the best accuracy at 99.25%. The fi