Authors: Soumitra Dutta
Addresses: INSEAD, Fontainebleau, France
Abstract: While the results obtained with medical expert systems such as MYCIN, INTERNIST and CADIAG, have been very encouraging, they have some limitations which prevent them from fully simulating the behaviour of a medical expert. Some of these deficiencies are external (i.e. peculiar to the domain of application) while others are internal (i.e. caused by the technology used). This paper explains these limitations of medical expert systems and addresses the issue of temporal reasoning in medical expert systems. Temporal reasoning is specially important for medical expert systems as the final diagnosis is often strongly affected by the sequence in which the symptoms develop. Current research on time structures in computer science is reviewed and a temporal model based on fuzzy set theory is developed. The proposed model allows a simple and natural representation of symptoms and provides for efficient computation of temporal relationships between symptoms. The applicability of the proposed temporal model to reasoning in expert systems is demonstrated.
Keywords: medical diagnosis; medical expert systems; medical applications; symptoms; temporal reasoning; fuzzy set theory; fuzzy logic; intelligent diagnosis.
International Journal of Computer Applications in Technology, 1990 Vol.3 No.1, pp.8 - 17
Published online: 11 Jun 2014 *Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article