Analysis and Mining of Internet Public Opinion Based on LDA Subject Classification

Abstract

This paper uses Python, R language, Gephi and other software to crawl and classify the comment content of Weibo hot search events. Using word cloud, co-occurrence social network graphs, LDA topic classification visualization methods, this paper regularizes and integrates public opinions of hot events. Through this research, we can get the influence of public opinion mediators, public opinion objects, and government forces on the network public opinion and put forward corresponding improvement suggestions. We hope to contribute to the government’s governance and prevention of online public opinion during the spread of COVID-19 and other public hot events.