Two Flows of Approximativeness in Diagnostic Expert Systems
Several methods for approximative diagnostic problem solving under uncertainty have been adopted for expert systems (ES). Usually, however, no difference is made in application of those methods in controlling activation of diagnostic concepts for evaluation and in concept evaluation itself. Two different and independent control flows of approximative inference have been incorporated in our diagnostic expert system CODEX [l,3] Their principles are introduced briefly in our contribution.
KeywordsExpert System Priority Queue Concept Evaluation Certainty Factor Approximative Inference
Unable to display preview. Download preview PDF.
- 1.GYÁRFÁŠ, P., POPPER, M.: CODEX: Prototype driven backward and forward chaining computerized diagnostic expert system. In Trappl, R. (ed.): Cybernetics and system research, North-Holland, Amsterdam, 1984, (821–824).Google Scholar
- 2.HÁJEK, P. : Combining functions in consulting systems and dependence of premisses. In Plander, I. (ed.): Artificial intelligence and information-control systems of robots, North-Holland, Amsterdam, 1984Google Scholar
- 3.POPPER, M., GYÁRFÁŠ, F. : CODEX: A computer-based diagnostic expert system. In Plander, I. (ed.): Artificial intelligence and information-control systems of robots, North-Holland, 1984, (297–300).Google Scholar
- POPPER, M., KELEMEN, J. : An attempt to formalize the medical diagnostic problem solving. Computers and Artificial Intelligence, 5, 1984, (423–435).Google Scholar
- 6.SHORTLIFFE, E.H.: Computer-based medical consultations: MYCIN. Elsevier/North-Holland, New York, 1976.Google Scholar
- 7.SOWA, J.F.: Conceptual structures: Information processing in mind and machine. Addison-Jiesley, Reading, 1984.Google Scholar