Expert systems and uncertainty in medical diagnosis: a proposal for fuzzy-ANP hybridisation

Abstract

Major research efforts in decision support systems for medical diagnosis have focused on the hypothesis testing and management of uncertain and imprecise decision variables aimed at obtaining optimal inferences. In this study, a survey of such systems and methods is carried out. The study identifies medical diagnosis as involving imprecise information in a multi-criteria decision situation. We propose a hybrid fuzzy based analytic network process (ANP) system for managing the hierarchical structuring and imprecision associated with medical diagnosis. The justification being that fuzzy-ANP system is capable of accommodating inherent uncertainty and vagueness in multi-attribute/multi-criteria decision-making with hierarchical structuring and feedback dependence.