Popular Microblogging Quality Prediction Model using Potential Fusion Characteristics

Abstract

In this paper, we put forward a potential fusion feature model to deal with the popularity prediction problem for Sina Microblogging. We analyze the characteristics of microblog forwarding number, number of views and comments. By combining the implicit relation between features, we fuse the n-gram model to establish a quality prediction model for microblogging. The characteristics of the model are combined with the regression model to predict the popularity of a microblog text. Our experiment results show that the model is effective.