TECHNOLOGY OF EXPERT SYSTEMS.

Abstract

Expert systems are built in many ways, involving rules, networks, frames, and combinations thereof, with all sorts of variations within these categories with respect to knowledge representation and control. The purpose of this article is to introduce expert systems and to discuss methods for building them, including the advantages and disadvantages of each method. The computer resources needed to build and run expert systems are reviewed. The most cost-effective computer resource must still be determined. In most cases, approach selection at least narrows the choice for resources; in some cases, approach and resources can be selected together. However, this hardly reduces the complexity of the choice. To make matters worse, computer resources are changing as rapidly as the new system-building approaches are being developed.