A Comprehensive Survey on Ensemble Machine Learning Approaches for Detection of Intrusion in IoT Networks
J. Jasmine Shirley, M. Geetha Priya · 2023
Threats to network security have been increasing leading to severe network attacks such that a simple firewall will not be sufficient to deter challenging and complicated attacks. Therefore, the utilization of intrusion detection systems (IDS) with other security devices is imperative to safeguard networks. The intrusion detection systems monitor traffic, collect, and analyze information, and raise an alert when there is a breach of security. It also prevents unauthorized access, stops intrusions, records them, and alerts the network admin to guarantee enhanced security of the system. Attack detection employing a single machine learning (ML) algorithm is not effective. As a result, ensemble learning was employed to combine different algorithms. To be more specific, ensemble learning is a popular analytical method that trains different algorithms to address similar problems, then combines the results to make a strong prediction that outperforms a single algorithm. The paper aims to offer a complete review of the most important ensemble strategies utilized to improve the efficacy of detecting intrusion in Internet of Things networks. The findings demonstrate that the two ensemble learning techniques majority voting and weighted averaging may be implemented with the utmost simplicity while yet producing accurate results.