Predicting the Popularity of Topics based on User Sentiment in Microblogging Websites

摘要

Behavioral economics show us that emotions play an important role inindividual behavior and decision-making. Does this also affect collectivedecision making in a community? Here we investigate whether the communitysentiment energy of a topic is related to the spreading popularity of thetopic. To compute the community sentiment energy of a topic, we first analyzethe sentiment of a user on the key phrases of the topic based on the recenttweets of the user. Then we compute the total sentiment energy of all users inthe community on the topic based on the Markov Random Field (MRF) model andgraph entropy model. Experiments on two communities find the linear correlationbetween the community sentiment energy and the real spreading popularity oftopics. Based on the finding, we proposed two models to predict the popularityof topics. Experimental results show the effectiveness of the two models andthe helpful of sentiment in predicting the popularity of topics. Experimentsalso show that community sentiment affects collective decision making ofspreading a topic or not in the community.