Natural language querying of databases: an information extraction approach in the conceptual query language

Natural language (NL) interfaces for database (DB) query formulation have always been recognized as a much-needed enhancement for DB end-users. NL systems, however, have shortcomings that have led some DB researchers to question their practicality. The drawbacks stem primarily from their weak interpretative power. This weakness is, to a large extent, due to their inability to deal with the nuances in human use of natural language. Some studies, however, show that NL database systems are practical and useful in narrow domains like DB querying. One way of addressing the difficulty with NL database query languages (DBQLs) is to combine concept-based DBQL paradigms with NL approaches to enhance the overall ease-of-use of the query interface. In this study, the conceptual query language-with-natural language (CQL/NL) is proposed. CQL/NL uses information extraction methods to filter NL query statements for search predicates that are derived from constructs on conceptual schemas. In this way, it avoids the computational difficulty with full-hedged NL parsing. In the main, we draw on certain concepts in natural language processing and computational linguistics to develop CQL/NL. (C) 2000 Academic Press.