Trust/distrust networks in social media are called weighted sign networks, in which edges are labeled with real numbers. An algorithm is proposed in this paper to improve trust prediction in WSNs by using local variables. Our algorithm, Strup, predicts the sign of edges through computing the stress of related nodes. Four new parameters are introduced to demonstrate the stress of nodes in the networks and to predict the sign of edges accurately. Considering these signs leads to more precise trust prediction. The experiment on four real-world WSNs showed that our proposed approach can easily combine with most of the existing weight prediction algorithms to improve them. Due to the tight relation between trust prediction and (opinion and emotion) prediction, we believe that our stress-based algorithm could be a promising solution in their challenging domains.

No abstract

No abstract

No abstract