IoT Anomaly Detection Using Picture Fuzzy Clustering Approach

Enhancing the efficacy of security systems without compromising performance remains one of the most challenging areas in cybersecurity research. Several approaches have shown promise in detecting anomalies in network data, among which fuzzy set-based methods are particularly notable. The Internet of Things (IoT) comprises a vast network of interconnected digital devices that continuously generate massive volumes of data and perform real-time computations. Due to their constant exposure to the Internet, these de-vices are highly vulnerable to threats from hackers and intruders. Such malicious activi-ties are categorized as anomalies, and detecting them within the IoT environment presents a compelling research challenge. Picture fuzzy sets (PFSs), which extend intuitionistic fuzzy sets (IFSs) by incorporating a neutrality parameter alongside membership and non-membership values, provide a robust framework for modelling the imprecision, vagueness, and uncertainty inherent in IoT datasets. In this article, we propose a Picture Fuzzy c-Means (PFCM) clustering-based method for detecting anomalies in IoT data. This algorithm represents an advanced variant of the classical Fuzzy c-Means (FCM) clustering technique. Given the complex and nuanced nature of uncertainty in IoT data, the pro-posed PFCM approach offers a more effective means of identifying anomalous records. Additionally, the computational complexity of the proposed method is analysed. Experimental evaluations using real-world datasets, along with comparative analyses against FCM and Intuitionistic Fuzzy c-Means (IFCM) algorithms, demonstrate the superior performance of the proposed approach