Don't count the number of friends when you are spreading information in social networks

Abstract

The problem of spreading information in social networks is a topic of considerable recent interest, but the conventional influence maximisation problem which selects a set of any arbitrary k nodes in a network as the initially activated nodes might be inadequate in a real-world social network -- cyber-stalkers try to initially spread a rumour through their neighbours only rather than arbitrary users selected from the entire network. To consider this more practical scenario, Kim and Eiko [16] introduced the optimisation problem to find influential neighbours to maximise information diffusion. We extend this model by introducing several important parameters such as user propagation rate on his (or her) neighbours to provide a more general and practical information diffusion model. We performed intensive simulations on several real-world network topologies (emails, blogs, Twitter and Facebook) to develop more effective information spreading schemes under this model. Unlike the results of previous research, our experimental results shows that information can be efficiently propagated in social networks using the propagation rate alone, even without consideration of the number of friends information. Moreover, we found that the naive random spreading would be used to efficiently spread information if k increases sufficiently (e.g. k = 4).