earticle

영어

Information diffusion in online social network especially in microblog system, can largely affect the public opinion and even the development of events, so the prediction of the future dynamics of diffusion could be very valuable at many aspects. In this research, a novel graph-based cascades construct algorithm is proposed, with which we build a prediction model for future information diffusion dynamics. By learning user influence features that related with the network topology and users’ interactions from a large scale real Weibo dataset, we successfully predicted the short term topic related tweets population dynamics.