Prediction Model of Microblog Retweeting Based on Naive Bayesian

Abstract

In this paper, we take Sina microblog as the research object to explore the features that influence microblog retweeting, as well as predict retweeting behavior. On the basis of obtaining a large number of microblog retweeting records, three features including number of historical interactions, interest similarity between users and microblog and similarity of active time are taken into account. We explore their exact influence on users' retweeting behavior, and establish a prediction model based on Naive Bayesian. Experiment indicates that the prediction model could achieve higher prediction accuracy with fewer features, which enables us to predict microblog retweeting timely in the process of dynamic public opinion propagation.