Provenance based Trust computation for Recommendation in Social Network

The term social network means to develop traditional communication among people. As it becomes online, the people in communication don't have face-to-face contact and also they may be known or unknown. Though known, there is no guarantee and confident that one who contact as a requester is a friend or foe. This paper aims towards developing a provenance-based trust model by utilizing the reputation of the mutual friend from the interaction he/she had with the requester. Provenance factors identified from the interaction information are used to compute the reputation. Because of the uncertainty in the behavior of a person in online social networking (OSN) sites, a stochastic model that evolves the randomness over time is applied and the trust computation is further refined. A statistical approach using Bayesian is used for recommending trustable request.