Fuzzy decision-making in electrocardiography

Abstract

This paper investigates the reasons for the use of fuzzy sets in the computerized diagnosis of electrocardiograms (ECG). The main points which suggest the use of fuzzy heuristics in this area are the imprecise class (i.e., disease) boundaries, the imprecise perception of measurements, and the imprecise relation between measurements and diseases. The reason to prefer fuzzy decision rules to the conventional yes/no approach is based upon the improved stability of a soft system in comparison with a hard but unrealistic decision-maker. The principal characteristics of the CARDS system developed in Padova are described.