Research on comprehensive information based Chinese information extraction system

Research and implementation for a testing Chinese information extraction (IE) system in special domain is introduced in this paper. The most important feature of this study is the application of comprehensive information concept into Chinese IE. According to comprehensive information theory, pragmatic information should be the key elements for better natural language processing and understanding, and should be used together with the other two commonly used kinds of information, which are syntactic and semantic information. Since IE is born for special application domain, which has clear pragmatic information, we believe that the integration of comprehensive information and IE is excellent. A testing system is designed and implemented according to this idea using the application domain of position changes in financial news, with a comprehensive information knowledge base as the backup. Testing results have shown that this method is promising.