In this paper a fuzzy relation-based framework is shown to be suitable to describe not only knowledge-based medical systems, explicitly using fuzzy approaches, but other ways of knowledge representation and processing. A particular example, the practically tested medical expert system Disco, is investigated from this point of view. The system is described in the fuzzy relation-based framework and compared with CADIAG-II-like systems that are a ``pattern" for computer-assisted diagnosis systems based on a fuzzy technology. Similarities and discrepancies in -- representation of knowledge, patient's information, inference mechanism and interpretation of results (diagnoses) -- of the systems are established.
This work can be considered as another step towards a general framework for computer-assisted medical diagnosis.
Keywords: fuzzy relations; medical diagnoses;
AMS: 03B52; 03E72; 62F15;
BACK to VOLUME 42 NO.6