A time-aware similarity-based trust computational model for social internet of things

The recent integration of social networking concepts with the emerging yet promising paradigm of Internet of Things (IoT) has facilitated the emergence of a new domain named Social Internet of Things (SIoT) to socialize smart objects by emulating the human behavior. The interaction between these smart objects depend on social relations amongst them, wherein each object establishes these relationships autonomously via rules set out by its owners for creating virtual communities and working environments with shared interests. Trust is a concept that quantifies the reliability of these autonomous relationships and hence enriches cooperation between the SIoT objects. This paper explores the concept of time-aware similarity-based trust to address the issue of trust while establishing a social relationship concerning virtual communities, work environment, and friendship similarity. The proposed trust model segregates the trust in two types, i.e., direct trust and indirect trust, wherein direct trust signifies the direct perception of a trustor towards a trustee, whereas, indirect trust depends on recommendations from the neighboring objects of a trustor. Moreover, a dynamic weighted sum method is employed to aggregate all the similarities so as to obtain a single trust score at any given time. We conducted a comprehensive experimental study based on a real-world data set and the simulation results demonstrate that our method can efficiently classify trustworthy and untrustworthy objects.