With the development of big data technology, the number of Internet users in China has exceeded 1 billion. With the help of a mature Internet community, the commercial value of public opinion dissemination is gaining more and more attention. It has attracted many scholars in the fields of journalism, social psychology, law, computer, and programming. How to identify favorable or harmful information, how to guide the correct public opinion, and how to turn public opinion into commercial value have become research hot topics in recent years. This paper sets the research object as the hot topic data on the Internet community Zhihu and calculates the relative public opinion dissemination efficiency level in each DMU through the DEA model. Then, these topics would be grouped into eight categories so that we could analyze the relationship between the number of hot topics and the percentage of high-efficiency topics. Finally, based on the statistical results, we obtain three main modes of public opinion communication and we infer the principle of how the mode forms.