International Conference on Information and Knowledge Technology

فارسی

Home / شانزدهمین کنفرانس بین المللی فناوری اطلاعات و دانش

An Adaptive Mutation-Enhanced EHO-SVM Framework for Intrusion Detection in IoMT Environments

Authors :

Amirhossein Damia¹ Erfaneh Khanmohammadi²

1- دانشگاه صنعتی خواجه نصیرالدین طوسی 2- دانشگاه صنعتی امیرکبیر

Keywords :

Internet of Medical Things،Support Vector Machine،Feature Selection،Intrusion Detection System،Elk Herding Algorithm،Gaussian Mutation

Abstract :

With the rapid expansion of Internet of Things (IoT) applications in the healthcare domain, a novel concept known as the Internet of Medical Things (IoMT) has emerged, providing an intelligent platform for monitoring, diagnosing, and treating patients. However, the continuous connectivity of these systems to the internet exposes them to cyber-attacks, which can have serious and irreversible consequences for patients and healthcare centers. Therefore, the development of intelligent and accurate systems for detecting attacks in IoMT has become a major challenge in the fields of cybersecurity and machine learning. This paper suggests a new method that needs to be used to enhance the effectiveness and accuracy of attack detection systems. Elk Herding Optimization (EHO) algorithm was improved and another version optimized by the use of adaptive Gaussian mutation was developed to have a better trade off between exploration and exploitation in a search. Testing of this enhanced version of the program on different benchmark functions indicated that the enhanced version of the program has better performance in terms of final accuracy and stability of results as opposed to the base version. This algorithm was then applied to optimize the hyperparameters of a Support Vector Machine (SVM) and to select effective features for detecting cyber-attacks in IoMT data. Furthermore, the improved version was compared with Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), showing that the proposed IEHO outperformed them in all evaluation criteria, including classification accuracy, number of selected features, and result stability.