Trust Analysis to Identify Malicious Nodes in the Social Internet of Things

In Social Internet of Things (SIoT), devices collaborate to request and provide services. As trust-related attacks can undermine the potential benefits offered by the SIoT, this study examines on-off and ballot-stuffing attacks that can impact the Internet of Things (IoT) and investigates a range of trust metrics that consider the characteristics of the attack types and SIoT environments. The relevance and significance of each trust feature is verified through descriptive analysis and multiple classifiers are employed to assess the effectiveness of the proposed model, with results being compared with existing state of the art solutions using the evaluation metrics such as Precision, Recall, and F-Measure. The results presented indicate a significant improvement in classifying malicious, benign and neutral nodes when compared to previous studies.