| With the popularity of the Internet and the development of social democratization process, the enthusiasm of the citizen participation in public affairs discussion is growing. When citizens participate in the discussion, the form of the conversation content mostly is unstructured and semi-structured text information. How to mining the user’s knowledge from this information, has become a hot research topic in the field of data mining. In this paper, from the perspective of topic, we put forward a topic clustering method of text of public affairs discussion, and discover the issues which are concerned by public, and discover the public’s attitude in discussion through sentiment analysis.The main content of this paper include:(1) Studying the relevant literature of text clustering, topic clustering and sentiment analysis at home and abroad, then summarize the basic method of topic clustering and sentiment analysis.(2) Describing the relevant concepts of public affairs and topic extracting, and analysis the existing method of text processing, text feature extracting and semantic similarity calculating.(3)Designing the framework of the topic clustering and sentiment analysis of public affairs discussion, and put forward the method of text topic extracting and the method of topic clustering which is based on the "HowNet" semantic similarity computation.(4)Using the clustering method we raised, we cluster the text of "consumption endowment pattern discussion" and analysis of the sentiment, experiments show that the method of topic clustering which is based on the "HowNet" semantic similarity computation that we raised is effective. |
