UCL logo

UCL Discovery

UCL home » Library Services » Electronic resources » UCL Discovery

Default databases: Extending the approach of deductive databases using default logic

Hunter, A; McBrien, P; (1998) Default databases: Extending the approach of deductive databases using default logic. DATA KNOWL ENG , 26 (2) 135 - 160.

Full text not available from this repository.

Abstract

Extending the relational data model using classical logic to give deductive databases has some significant benefits. In particular, classical logic rules offer an efficient representation: a universally quantified rule can represent many facts. However, classical logic does not support the representation of general rules or synonymously defaults.General rules are rules that are usually valid, but occasionally have exceptions. They are useful in a database since they can allow for the derivation of relations on the basis of incomplete information. The need for incorporating general rules into a database is reinforced when considering that participants in the development process may naturally describe rules for a deductive database in the form of general rules.In order to meet this need for using general rules in databases, we extend the notion of deductive databases. In particular, we use default logic, an extension of classical logic that has been developed for representing and reasoning with default knowledge, to formalize the use of general rules in deductive databases, to give what we call default databases. In this paper, we provide an overview of default logic, motivate its applicability to capturing general rules in databases, and then develop a framework for default databases. In particular, we propose a methodology for developing default databases that is based on entity-relationship modelling.

Type:Article
Title:Default databases: Extending the approach of deductive databases using default logic
Keywords:default logic, non-monotonic logic, incomplete information, default databases, inferential databases, deductive databases, entity-relationship modelling, data modelling, rule engineering, METHODOLOGY
UCL classification:UCL > School of BEAMS > Faculty of Engineering Science > Computer Science

Archive Staff Only: edit this record