Fuzzy query interface for a business database

Managers, in today's corporations, rely increasingly on the use of databases to obtain insights and updated information to make their decisions. This paper describes a flexible query interface based on fuzzy logic. Hence, queries in natural language with pre-defined syntactical structures are performed, and the system uses a fuzzy natural language process to provide answers. This process. uses the fuzzy translation rules of the meaning representation language PRUF, proposed by Zadeh (Intern. J. Man-Machine Studies 10 (1978) 395). The interface was built for a relational database of the 500 biggest non-financial Portuguese companies. The attributes considered are the economic and financial indicators. Examples of pseudo-natural language queries, such as "is company X very profitable?" or "are most private companies productive?", are presented to show the capabilities of this human-oriented interface. (C) 2003 Elsevier Science Ltd. All rights reserved.