An efficient trust inference model in online social networks using fuzzy petri nets

Abstract

SummarySharing and exchanging of information among the group of users is the objective of social networks. With the growing popularity of the internet applications, many of the interactions are among the unknown users. Estimation of the user trustworthiness prior to conversation with him can increase the confidence level of a person in taking right decisions. Facebook, Twitter, LinkedIn, Instagram and Google+ are some of the most popular social networks, where the similar kinds of users share their messages and photos. In this article, the proposed method “Trust Inference Model in Online Social Networks Using Fuzzy Petri Nets (TMFPN)” computes the user trust value through two phases, that is, direct and indirect using twitter user data set. The direct trust of a user is evaluated based on his social activities like (follower count, listed count, mentions received, re tweets received, and posts) using fuzzy inference model. For the users those are not in direct interaction, indirect trust is computed. Proposed method‐TMFPN gathers network information through trust propagation and models the social network as a fuzzy petri net. Where users and interactions are considered as places and transitions, respectively. Here the concurrent reasoning algorithm (CRA) is applied over the converted FPN, for the selection of reliable trustworthy paths between two users at multiple hops distance. Performance of the proposed method is verified theoretically and practically. In the experiment results, performance of TMFPN is compared with existing trust computing methods in computation of indirect trust of multihop away users.