Chen Liyong

Effective Prediction of Microblog User Retweet Behaviors Based on Markov Random Field

Prediction of user retweet behavior is of great significance to the topic detection and diffusion range control. Based on the data of Sina microblog, we propose an effective prediction framework based on Markov Random Field for finding the global optimal solution to improve the prediction accuracy. Under the proposed method, we systematically analyze the various factors affecting the retweet behavior and the retweet constraints between users, and then construct the prediction model to perform the prediction. Experimental analysis shows that the prediction accuracy can be improved significantly compared to traditional methods.