Design and Implementation of Korean Query System using Frame-based Query Decomposition Method

A natural language query interface is an efficient retrieval tool for users who don't have enough knowledge about computers and databases. In this paper, we describe a design and implementation of a Korea query system. The Korean query processor transforms Korean queries into OQL (Object Query Language) statements and transfers them to OQL processor. The OQL processor retrieves data from the storage manager by referring the schema manager. The Korean query processor employs a frame-based query decomposition method to transform Korean queries into OQL statements. A Korean query is transformed into a predicate argument structure as its parsing result, which is decomposed into several query phrases. The decomposed query phrases are transformed into query frames, from which the target OQL statements are generated. We collect a large number of sample queries for the performance evaluation and about 87% of the queries are correctly transformed.