A Hybrid Machine Learning and Blockchain Framework for Wireless Sensor Networks

This study proposes an enhanced method for detecting malicious nodes in wireless sensor networks (WSNs) by extending security measures based on least square-support vector machine (LS-SVM). Specifically, a Pearson correlation coefficient-based enhanced LS-SVM (ELS-SVM) is employed for identifying malicious nodes. Detection of such nodes through machine learning (ML) algorithms establishes a robust defense mechanism, enabling the monitoring of node behavior and the mitigation of potential threats. In this context, effective anomaly detection in WSNs utilizing ELS-SVM fosters a secure environment, further strengthened by transparent logging and secure data storage (SDS) via blockchain technology. SDS in WSNs is critical, and this research achieves it through a hybrid approach combining ELS-SVM and blockchain technology for malicious node detection (MND). By ensuring network reliability and data immutability, this hybrid mechanism offers enhanced protection against potential attacks.