Fuzzy approach with a multi-expert approximate knowledge representation for medical diagnostic

Our work presents a new approach for medical diagnosis based on a fuzzy modeling and the use of a multi-expert approximate knowledge representation. The proposed approach attempts to make the final interpretations more objective. To do that we attempt to deal with conflicting expert opinions that can differ in their grading the importance of symptoms and can provide different medical knowledge. To derive a consensus, fuzzy preference relations are constructed and used to infer a diagnosis without averaging the grading assigned by the different diagnosticians. Further improvements would concern a methodological interpretation and exploitation of the final group's fuzzy preference relation in order to achieve the diagnostic process. We plan to apply the presented approach to the analysis and interpretation of biomedical images.<
>