Research Of Database Query System Based On Natural Language Interface

Natural language processing is a very important category in AI, which mainlyfocuses on how to understand, process and generate natural language. In naturallanguage research, it has become one of the areas with promising prospect to applyadvanced natural language processing technology into developing database queryinterface system based on natural language.Due to its increasing development, database technology has been applied in allwalks of life; but only professionals have access to it. As a primary way ofinformation expression, natural language, combined with the database querytechnology, offers access to needed information for the non-professionals, which isof great theoretical and practical value. ER model in database is a fully-developedinformation model which can greatly enhance the realizability of query interface bycombining database ER model with natural Chinese language understanding.The thesis studies the basic principles and theories of natural languageprocessing, combines natural language understanding and database technology,designs and realizes a database Intelligent Retrieval Interface System-admissiongrades query-with visual c development tools. SQL language is used in data query.Access is employed in databases, and data access is realized with ActiveX DataObjects.The design of the knowledge base is indispensable in the construction of thesystem. The system builds different categories of knowledge base, which are thebasis of lexical, syntactic analysis-the essential steps in processing Chinese. Thethesis has adopted the bidirectional maximum matching method for wordprocessing, and eliminated ambiguity in a certain degree. To identify the unknownwords, the system has also made the corresponding processing. In terms of thecharacteristics of the database query, the thesis has discussed on the identificationprocess of query targets and query conditions in details. The ordinary staff queryrequirements is taken into consideration in studying target phrases of differentcategories and conditional expressions with multiple types.