SIoTPredict: A Framework for Predicting Relationships in the Social Internet of Things

Abstract

The Social Internet of Things (SIoT) is a new paradigm that integrates social network concepts with the Internet of Things (IoT). It boosts the discovery, selection and composition of services and information provided by distributed objects. In SIoT, searching for services is based on the utilization of the social structure resulted from the formed relationships. However, current approaches lack modelling and effective analysis of SIoT. In this work, we address this problem and specifically focus on modelling the SIoT’s evolvement. As the growing number of IoT objects with heterogeneous attributes join the social network, there is an urgent need for identifying the mechanisms by which SIoT structures evolve. We model the SIoT over time and address the suitability of traditional analytical procedures to predict future relationships (links) in the dynamic and heterogeneous SIoT. Specifically, we propose a framework, namely SIoTPredict, which includes three stages: i) collection of raw movement data of IoT devices, ii) generating temporal sequence networks of the SIoT, and iii) predicting relationships among IoT devices which are likely to occur. We have conducted extensive experimental studies to evaluate the proposed framework using real SIoT datasets and the results show the better performance of our framework.