Topic diffusion behavior tracking in online social network

Abstract

Information Diffusion in social network is of much importance in tracking public opinion, launching new products, and other applications. In this paper, we formally define the problem of Topic Diffusion Behavior Tracking and propose a novel model by investigating users' topic interest. Our algorithm is developed based on the combination of personal interest and friend influence. First, probability topic model is defined to model topic content efficiently by historical behavior. Second, to integrate topic content and friend influence, we develop a topic behavior tracking model based on random walk. Finally, we propose a novel measure called Topic-Interest-Rank (TIR), which ranks users according to how important they are in sociological phenomena, to predict the topic behavior in future. Comprehensive experimental studies on two different real world data sets show that our approach outperforms existing ones and well matches the practice. © 2012 Springer-Verlag Berlin Heidelberg.