A cost-effective, repeatable method for managing the development of multiple-expert expert systems

Abstract

A model and a methodology for managing the development of commercial expert systems involving multiple experts is presented. The model incorporates a view of the nature of human knowledge, the constraints of the technology, and pragmatic considerations of the organization in which the development takes place. The methodology engendered by this model introduces a new role, the Sherpa, and a new tool, the Knowledge Verification Template (KVT). This methodology has been successfully applied in multiple systems. Its specific application to the BioMark Program, a commercial preventive healthcare program, is presented.<>