A Dynamic Model on News Popularity Prediction in Online Social Networks

With the fast development of smartphone and wireless network, the real-time news spreads faster because people can use mobile client to broese images, video and audio contents. Many people share and comment to express their mends. redicting the popularity of the online content is a hot research poiont and many people concentate on finding out the law of information dissemination. In this paper, taking the Tencent News as a case, we observed that there exists a popularity migration, people somtimes leave comments on irrelevant topics, especially when a hot topic happens. The main contribution of this article is to solve the problem that few or no work for popularity prediction based on topic migration effect. We propose a model based on the reinfoced Poisson process model with the weak tie theory and competitive matrix. Our goal is to accurately estimate the popularity of a given viral topic at final based on the observation of its historical characteristics. Also, this method provides a better performance in the popularity prediction according to an empirical study.