A Novel Reposting Prediction Method Based on Quantified Microblog Hotness in Sina Weibo

Abstract

Reposting is a very important behavior in Sina Weibo. Predicting the reposting actions will be great benefit to outbreak events detecting, rumors spread preventing and market advertising. While many researches had been devoted to mine users’ behavior patterns or investigate the social network structure features, few attempts have been done to study how the hotness of microblog content influence users’ reposting actions. In this paper, we performed an interesting research on reposting prediction in Sina Weibo by quantified the microblog hotness. We evaluated the popularity of hot topics by their audience size and calculated microblog hotness based on the associated topic popularity. On the base of this work, we proposed a reposting prediction method based on microblog hotness and some other benchmark features. Experimental results on real-world Sina Weibo dataset showed the effectiveness of our approach.