부산대학교 도서관

When using natural language interfaces to databases (NLIDBs) for database queries that involve many tables, the resulting SQL query may include semantically implicit entities. These entities are related to the semantic meaning of a query, when upon referring to an entity (table), another entity (or entities) is (are) semantically implied with which the first entity is related, and the user might ignore the relationship between the two (or more) entities. In a previous work, this problem was addressed by using a semantic view. However, the configuration of the NLIDB is very complex for the database administrator for dealing with this kind of queries. This means that the semantic view can only be created and configured by the NLIDB developers. This article describes a method to configure a dictionary of semantic information to create a virtual semantic view in a NLIDB, thus facilitating the creation of semantic views in a NLIDB. [ABSTRACT FROM AUTHOR]