In expert systems the heuristics used for combining the weight of evidence can be based on probabilistic, fuzzy set, or subjective confidence factors. Although the underlying assumptions for each of the methods differ, it can be shown that there are correspondences between them and that it is possible to develop a model of expert reasoning for medical consultation using any one of the methods. We have developed a system for representing expert knowledge, called ESMES, which is an outgrowth of the EXPERT scheme developed earlier at Rutgers. ESMES allows the use of alternative strategies in the solution of a consultation problem. We report on the performance of ESMES for a prototype glaucoma consultation model, using reasoning mechanisms similar to those of the EXPERT, MYCIN, INTERNIST I, and PROSPECTOR systems.
|Original language||English (US)|
|Number of pages||5|
|Journal||Proceedings of the Hawaii International Conference on System Science|
|State||Published - Dec 1 1983|
All Science Journal Classification (ASJC) codes