Directed Belief Diffusion for Dynamic Social Networks

This paper presents a novel model of interaction patterns between individual agents in dynamic social networks. A simulation was developed implementing the directed diffusion protocol to propagate beliefs through a social network. Belief propagation through a social network is modeled as packet propagation through a wireless sensor network using the directed diffusion protocol. This data propagation model is compared against a network using basic directed diffusion; the analysis measures the level of stress the networks experiences during the experiment as well as the individual and general energy consumption used. The evaluation of both methods will provide a better understanding of belief propagation in social networks by studying the dynamics in a directed diffusion network. The results of the experiment can also lead to models of interaction patterns between individual agents on dynamic social networks that can be integrated with real data from telecommunication service providers and person-person contact data. Keywordssocial network models; wireless technologies; belief propagation; directed diffusion; gradients.