Enhancing Intrusion Detection in IoT Platforms Using a Novel Hybrid Gorilla Troops and Bird Swarm Optimization Algorithm

Abstract

In the era of advancing technology, the proliferation of the Internet of Things (IoT) has become pervasive, influencing various facets of  contemporary life. Intrusion Detection Systems (IDS) stand as a crucial guardian of these interconnected networks. Feature Selection  emerges as a pivotal element in the design of effective IDS, aiming to discern the optimal subset of features for accurate attack classification within an extensive feature set. This paper introduces an approach that enhances a Hybrid Gorilla Troops Optimizer (GTO)  algorithm and Bird Swarm Algorithm (BSA) with Step size parameters. The aim of adding the step size is to have controlled movement  and adaptively explore and exploit the search space, thereby enhancing the performance of the hybrid algorithm. The hybridization leverages the strengths of both algorithms in identifying the optimal feature subset. The resulting Improved Hybrid GTO with BSA  Algorithm (IGTO-BSA), utilizes metaheuristic techniques to boost feature selection, striking a balance between exploration and  exploitation, foster faster convergence and deliver superior solutions within reasonable time. The effectiveness of IGTO-BSA is assessed  using three diverse IoT datasets: NSL-KDD, CICIDS-2017, and UNSW-NB15. Performance evaluation is conducted using five metrics:  accuracy, sensitivity, specificity, computational time and the number of selected features. Comparative analysis with an existing  technique from the literature establishes the efficacy of the proposed approach.