Research Of Chinese Natural Language Understanding In Auto Question-Answer System

With the rapid development of Internet and the enlargement of network application,it will be more and more difficult for people to find information from Internet accurately and quickly. To resolve this, A new kind of information search system(Auto Question Answer Systems) has been proposed in the thesis. This kind of system is based on restricted domain knowledge and try to resolve the old search engine's defects. Auto Question Answer Systems is current a hot direction in natural language process research field, and it will undoubtedly become the next generation of search engine technology. It uses knowledge representation, information search, and other natural language process technology. Auto Question Answer System uses a different way for user to search information by input natural language question, not combination of keyword. And the result return to user is not some web pages but a concise and accurate answer. Compare to current search engine, Auto Question Answer System is more convenient and more efficient.The main achievements of the studies in the thesis list as follow: The thesis research the present situation of natural language process technology, especially the technology used in Chinese Auto Question Answer System field. Then based on domain knowledge database, a new improved resolution has been proposed for restricted domain Auto Question Answer System. This system is composed of four modules: Question Answer Database, Question Analysis module, Information Searching module and Information Extraction module. Question Answer Database store the domain knowledge, frequent ask question and origin text database. Question Analysis module is aim to analyze the input question from the syntax and accidence aspect so that to catch the real meaning of the user's question. Information Searching module uses the keywords to search the database, and try to find a list of answers with high correlativity. Information Extraction module's task is to process this list of answers, and extract the correct sentence or paragraph to user.