An expert system using an extended AND-OR graph

Abstract

The inference engine (1E) is a crucial component of expert systems (ES). As the management environment has become rapidly digitalized with the advent of the Internet, the traditional 1E is now facing a severe criticism - i.e., that it cannot effectively provide an agile, knowledge-based decision support suitable for a wide variety of problems. The objective of this paper is to propose a new type of ES called IMIXAO (Integer-Matrix-driven Inference based on an eXtended AND-OR graph), with an integer-matrix-driven 1E based on an extended AND-OR graph framework. Traditionally, the AND-OR graph is used for making inferences in the ES area. Theoretical backgrounds of IMIXAO are presented in this paper, focusing on what an extended AND-OR graph is in terms of inference. The experiment conducted in this study, with an illustrative example of dermatological diseases classification, revealed that the proposed IMIXAO can make precise and agile inferences in a complex situation. (c) 2006 Elsevier B.V. All rights reserved.