Research and application of public opinion retrieval based on user behavior modeling

This paper designs a system of network public opinion analysis based on the specific way of “matching Topic with Opinion” which can organize public opinion data, avoid the redundant data and retain the original information structure of the opinion. And this article proposes a user model founded on the user access behavior to scientifically classify and represent the relevant theory of the users’ retrieval behavior, then to analyze and manage the retrieval results which can to provide the accurate and relevant search results of public opinion information for users.