Natural Language Interface by Optimization Model for Fuzzy Data Base Retrieval

Abstract

This paper copes with fuzzy database retrieval by natural language query with fuzzy expressions. The authors propose some simple methods to extract effective informations for database retrieval from the natural language (Japanese language) query. Those are mainly composed of Hirakana-retrieval dictionary, optimization model by 0-1 integer programming, and some additional dictionaries. The Hirakana-retrieval dictionary generates several interpretations of the query, and then the optimaization model determines an unique interpretation of the query.