An Efficiency Clustering Approach based on Hedge Algebra in the IDS-based Machine Learning Models

The rapid proliferation of Internet of Things (IoT) devices raised security vulnerabilities within traditional network security solutions. Machine learning-based intrusion detection systems (IDS) demonstrate a significant paradigm for modern theoretical and practical applications. Intrusion Detection Systems are becoming more widely deployed closer to the network edge to anticipate massive attacks at an early stage. In addition to resource restrictions, studies consistently aim for explainability in attack detection techniques. This study proposes a novel clustering approach that integrates hedge algebra with traditional machine learning techniques to enhance IDS performance in edge computing environments. Experimental findings assessed on the CICIoT2023 dataset indicate that the efficacy of the proposed method surpasses that of the same approach studies in both cluster evaluation metrics and the attack detection rate of the overall IDS model.