A Reestricted Domain Intelligent Question Answering System Based On Web

Since 90s of the 20th Century, Internet has developed rapidly all over the world. The increasing information on the web has promoted the development of nature language processing technology and also put forward high request for nature language processing technology: people want to acquire the useful information in the unsystematic web world fleetly and exactly. Automatic question answering system (QAS) has grown to satisfy the people's need. People can put common questions to system, and then the system searches relevant answers from corpus, returns answers to users. People can obtain useful information by QAS expediently.However, current QAS commonly matches the keywords to the questions in the question-database according to the users' inputs, but it would bring obvious shortcoming to express the semantics of questions only by keywords. To solve this limitation, in this paper, we discussed the correlative technique of question answering system and put forward how to establish an effective retrieval model, combining the method of distinguishing the answer type, expanding the keywords of question, which guarantees the system to provide the proper answers to user and achieve its intelligence ultimately.Our research integrates the knowledge and technology of information retrieval computational linguistics data mining etc. Experiment result indicates that our method is feasible to ameliorate the information retrieval in web circumstance and improves the intelligence of system markedly. Applying this system in distance education, it provides an advanced teaching environment based on WWW for wide users, while afford exact answers in time and avoid study obstacle.