This paper presents a branch-and-bound algorithm to process free-text natural language database queries based on the metadata search approach. The approach uses a metadata reference dictionary represented in a semantic graph of enterprise databases. User words, query cases, information models, and database values are vertices of the graph. The paper concludes with possible extensions to offer a full capability of free-text natural language database queries.
Keywords: -
E-mail: cast@kmitl.ac.th
Boonjing, V. . (2018). A Branch-and-Bound Algorithm for Natural Language Database Query Using Metadata Search. CURRENT APPLIED SCIENCE AND TECHNOLOGY, 23-34.

https://cast.kmitl.ac.th/articles/144340