Selection of the Diagnostic Hypotheses through Fuzzy Decision in a Medical Diagnosis System

The paper presents an evaluation of the diagnostic hypotheses’ selection step within the medical diagnosis hybrid system DiaMed. Medical diagnoses are described in DiaMed through fuzzy decision functions that are used as discrimination functions for the selection. This type of selection is compared theoretically to a generic fuzzy inference model for medical diagnosis, and is evaluated on a differential diagnosis example, using the C discrimination index of Harrell.