Explicit schematic information in knowledge representation and acquisition

Abstract

Schematic descriptions of a domain knowledge, called an ontology ( van Heijst et al., 1996), are very useful in facilitating and formalizing the knowledge acquisition process in knowledge-based systems development. This paper introduces the use of explicit ontologies in the knowledge representation and acquisition in KEshell ( Wu, 1991), an expert-system-building shell. We have designed a `rule schema+rule body' representation, which represents schematic descriptions of a domain knowledge as an integral part of the domain knowledge in its rule schemata. The rule schemata are acquired and in turn used to guide the acquisition of actual domain knowledge in a structured interactive knowledge transfer program, Sikt, with which subject experts/professionals can interact directly to build executable knowledge bases without the aid of computer programmers. Also, the user of KEshell can start with Prolog rules and explicit rule schemata will be automatically generated from these rules.