Abstract
The most influential nodes are capable of generating maximum influence and widest information spread to their connected nodes in mobile social networks. Hence identifying high influential users plays an important role in monitoring network public opinion. However, influence maximization is an open and challenging issue, and existing works only focus on online social networks but fail to characterize the user social relations and social behavior. In this paper, firstly, we present a novel method based on location and preference relationships to calculate the social influence of users, and the initial nodes are identified by community partition. Secondly, according to the stochastic information spreading process in MSNs, we propose the Fluctuate Particle Swarm Optimization (FPSO)-algorithm to solve the problem of changing communication relationship between users at the optimization stage. Finally, we use extensive experiments on the real datasets to prove the effectiveness of the proposed FPSO-algorithm. The experimental evaluation shows that in the real networks, our proposal achieves better performance than other related methods.