Hotspot Prediction And Analysis Of Propagating Of Topics Based On Microblog

Microblog, as a new media platform, develops rapidly since its speed of information dissemination, multi-source of information and high participation of user. Meanwhile, the character of information dissemination on Microblog leads the information to be vast and mussy, so the authenticity of the information and hot topics attract the attention of users on Microblog. Based on the popular application of analyzing Microblog, Hotspot Prediction and Analysis of Propagating of Topics are researched in this paper.In this paper, the topic source model assumption is used to predict the hot topics of Sina Microblog. By abstacting a group of users on Micro-blog as a topic source, and assuming that topic sources only generate two classes of topic:hot topic and none-hot topic, then the problem to predict hot topics is transformed to classifying the topics. A method, which is based on structured time series learning of attention of topics on Microblog, has been implemented and improved in this paper. The results of experiment shows that, the method can classify the hot topic with a90%true positive rate, and among them77.8%of hot topics are detected before Sina Micro-blog with a mean advantage of1.54hour.Besides, based on the relationship of forwarding and commenting of tweets and topics, the path of propagation of topics are reconstructed and visualized. Then the paper analyzes the transmission capacity of users who participated in the topics, and carry out a method to compute the key users of topics, which traces to the source of topics on Microblog effectively.