Design And Implementation Of Question Answering System In A Restricted Domain

With the fast spread speed of Internet and the explosion of information, it is a hot research that how to find out the information needed quickly and exactly from the vast of information source. The emergences of search engines satisfy the user's need of getting information needed in a certain degree, but usually, search engines return hundreds of thousands of documents in response to a query. If the users want to get the final answers, they have to search manually.Question Answering System (QAS) is the next generation of search engine. It allows the user to ask questions using natural language and returns precise answers. So compared with traditional search engines, QAS can retrieve the most precise answer to satisfy user's demands of searching. Question Analysis(QA), Information Retrieval(IR) and Answering Extraction(AE) are the three keys in QAS. We discuss all of them, design and implement a QAS based on web and computer domain feature.QA is the initial task of Chinese QAS, the result of which has a great effect on the following processing. We make the following research:in lexical analysis phrase, we insert computer word list based on general segmentation dictionary, exclude word ambiguity; to the query questioned in natural language, we classify the question using a combined method based on question word and question focus. The experiment shows it perform well.IR is a very important connecting part in QAS, whose performance effects the precise of AE module We address two strategy retrieval methods. It includes the local knowledge database retrieval and web retrieval. If retrievaled the answer in local knowledge, the answer will be given to the user directly; if failed, web retrieval will be done. When doing web retrieval, different retrieval strategy will be taken based on answer type obtained by QA module.In the phase of AE, answering choosing settle on user's undergoing of QAS directly. This part conferred oversimplified, and to some specific question types, completed using a flexible method.Finally, the implementation of QAS, the results and evaluations will be given.