Évidences positives, négatives et préjugés dans le cadre de la théorie des fonctions de croyance

The online symptom checker MedVir allows medical diagnosis and can facilitate triage in emergency medicine. Diagnostic hypotheses and severity are estimated following the analysis of the patient’s answers to a series of questions. Several scores (positive, negative, specific) are assigned to each symptom according to a disease. In this paper, we propose an interpretation of the mechanism implemented in MedVir to exhibit diagnoses using the belief functions theory and prejudiced information fusion.