Understanding Of The Requirement In The Intelligent Query System

With the popularization of computers and the development of information technology, the information people need to deal with grows rapidly. So how to find information efficiently has become an urgent requirement. Nowadays, most of the search engines give so many results related to the key words the users input that the users can not find the needed information quickly and accurately. So people need more intelligent query system to find the information conveniently and accurately.Library intelligent query system implements an intelligent query about books in the library. It has a more nature interface, and the users can express their requirements with spoken language conveniently and accurately. And the system do semantic-based query to find exactly what the users want.This paper implements the requirement understanding part of the library intelligent query system, which analyzes query sentences that users input with spoken language, acquires book-related information and creates a semantic network for the content information if there is. Before analyzing the input sentence, we need to do some preprocessing, including segmentation, tagging, normalization and tagging domain-related semantic classes. According to the characteristics of Chinese spoken language, we do partial parsing to analyze the sentences. The grammar formalization based on context-free grammar with complex feature support is adopted in this system. The technology of pseudo unification of complex feature is used in the reduction of grammar rules. Semantic network is used to represent the semantic information of the content. In the semantic network, the vertex represents the concept and the edge represents the relation between concepts. This part we use complete parsing and grammars with condition checking and operations of semantic network to create the semantic network. The processing of complete parsing is implemented with Chart algorithm.A lot of samples of library query are selected to test the system thoroughly. The result illustrates that its accuracy is satisfying. Based on the test result, we give our future research emphases.