Popularity Prediction of Burst Event in Microblogging

Every day, thousands of burst events are generated in microblogging first, and then affect the public opinion to a large degree. Thus, it is quite necessary to find out “how hot the burst event will be in the future”. In this paper, we propose a prediction model which combines the analysis of event content and users’ interest to predict the volume of the burst event in the implicit network. Particularly, it is assumed that different user has different influence power and different interest in the burst event. The popularity of an event depends on the volumes produced by the users infected in the past and its historical popularity. Experimental results show the superior performance of our approach.