Analyzing and Predicting Forwarding Behaviors in Sina Weibo

Abstract

With the microblog becoming more and more popular, people spend much time on the microblogging site, such as Sina Weibo. Users can push a new microblog by themselves or forward a microblog which is interesting and valuable from their followers and followees, and forwarding behaviors have important influences in information spread. Studying forwarding behaviors enable researchers to understand which kind of information will be propagated in social network. In this paper, we experiment with a corpus of Sina Weibo messages and investigate five important features to find how these features affect forwarding behaviors. Then, we train a prediction model to forecast whether a microblog will be forwarded by a given user. After getting the parameters learned from the model, our model can achieve a precision and recall of 20.13 and 26.27 % on average.