Cooperative Answering through Controlled Query Relaxation

This article presents methods to guide and control heuristically the relaxation of deductive and relational database queries. These methods enable a database system to compose responses that align with user needs. Query relaxation provides a user with an automatic means to identify new queries that are related to the user's original query. However, for large databases, many relaxations might be possible. These methods to control and restrict the relaxation process add focus and make relaxation more efficient. User restrictions over the database domain are expressed as user constraints, which restrict relaxed queries. The author presents a set of heuristics, based on semantic query-optimization techniques, for controlling relaxation. She then describes how the methods interact to relax queries, process user constraints, and apply the heuristic rules. Finally, she shows how to use electronic thesaurii as sources of semantic information and gives examples of two thesaurus-based applications of user-constrained relaxation.