Measurement-driven temporal analysis of information diffusion in online social networks

Abstract

The rapid development of online social networks (OSN) renders them a popular mechanism for information diffusion. Studying the temporal characteristics is critical in understanding the diffusion process. However, due to the lack of well-defined propagation data, hardly any study addresses the temporal feature of information diffusion in OSN. In this paper, we present a measurement study on information diffusion in the Renren social network. We investigate the latency of information propagation along social links and define the “activation time” for an OSN user, and find that the activation time follows the lognormal distribution. Based on this, we develop two new information diffusion models incorporating asynchronous activation times. Application of the models in the influence maximization problem shows that they capture the temporal diffusion behavior very well. This leads to fundamental ramifications to many related OSN applications.