Iteratively Tracking Hot Topics on Public Opinion Based on Parallel Intelligence

Abstract

Internet public opinion has increasingly become an important factor affecting social stability and the harmony of international relations. For tracking public opinion topics, observing the drift of public opinion hot topics, and predicting the development trend of public opinion, this paper builds a parallel public opinion hot topics tracking system based on the Sina microblog platform, which can iteratively mine hot topics. The LDA model is used to extract the topic of the comments on the Sina microblog platform and the public opinion cellular automata is applied to predict its evolution trend in this system. Based on the deduction results of the public opinion cellular automata, the keywords for the next round are constrained and guided. The keywords generated in the new iteration are used as the crawling keywords in the next time window and updated iteratively with the event evolution. This method provides technical support for the government and related departments to grasp the development of public opinion situations accurately and in real time.