A Machine Learning Framework for Intrusion Detection System in IoT Networks Using an Ensemble Feature Selection Method

During the recent years, there has been an escalated growth of the Internet of Things (IoT) devices in our daily lives. Due to their inherent vulnerabilities, IoT networks are more susceptible to cyberattacks. Intrusion Detection Systems (IDSs) can provide a powerful defensive capability against security threats for IoT systems. Among various approaches for detecting malicious activities in IDSs, Machine Learning (ML) techniques have attracted significant attention in the research community. In this study, we propose an ML-based IDS framework based on a novel feature selection method for anomaly detection in IoT networks. The framework is evaluated by employing seven popular ML classification algorithms on TON_IoT network dataset with regard to accuracy, precision, recall, F1-score, and execution times. Finally, the results of the experiments illustrate the effectiveness of the proposed IDS and the best performing ML classifiers.