Development of an Intrusion Detection System Using Mayfly Feature Selection and Support Vector Machine Algorithms

Abstract

Numerous security strategies have been used to control the threats associated with computer and network security. Methods such as access control, software and hardware firewall restrictions, and the encryption of private information. Nevertheless, these methods are insufficient because they all have serious drawbacks. As a result, using additional defense mechanisms, such as intrusion detection systems (IDS), becomes crucial. This research developed an effective intrusion detection system using mayfly feature selection and support vector machine algorithm. The SVM classifier achieved an accuracy of approximately 99.9% the precision is 99.75% sensitivity 100%, Fscore 99.87% While the training time display an average time of 1.4630sec. The results of this study suggest that security professionals and researchers should consider adopting ensemble methods like AdaBoost, especially when combined with robust base learners such as SVM, in the development of intrusion detection systems for IoT networks