Study on question-answering system of restricted domain based on knowledge base

Large scale texts as the knowledge base of extracting answer can ensure the accuracy of information resources and also make information convenient for management. This paper presents the development of question answer system of restricted domain based on knowledge base and proposes an answer extraction method based on the sentence similarity and organization names, and which implements the study of question answering system of restricted domain based on knowledge base. The experiments prove that the introduction of organization names not only greatly improves the segmentation accuracy in the module of words segmentation but also improves the retrieval efficiency of system and the accuracy of answer extraction.