Modeling and Analysis of Information Dissemination in Online Social Networks

Abstract

Information diffusion in online social networks has attracted public attention, and epidemic models have been applied to describe the diffusion process. However, online social networks exhibit a different propagation mechanism. To better characterize online information diffusion, we put forward a new model in which agents spontaneously participate in the discussions regarding a topic or may be persuaded by neighbors. Agents may ignore the information once they contact it. The attraction of the topic decays with time, and spreaders may also lose their interest in diffusing information. A real social network and virtual scale-free network are used as interaction topology. Results show unlike traditional epidemic models, the threshold of the interpersonal spreading rate with spontaneous diffusion drops to zero. The spontaneous diffusion mechanism lowers the threshold of diffusion process, but cannot cause a large extent of infection. The effect of the average network degree is quite different from that of the network diameter. The average degree of underlying network greatly improves the density of spreaders for any value of the interpersonal spreading rate, but the network diameter has a clear effect on the dynamics only when the spreading rate is large. Our work can help to understand and make effective measures of promoting or prohibiting information diffusion in social networks.