A method for optimization of fuzzy reasoning by genetic algorithms and its application to discrimination of myocardial heart disease

Abstract

A genetic algorithm (GA)-based method is proposed and implemented for determining the set of fuzzy membership functions that can provide an optimal classification of myocardial heart disease from ultrasonic images. Gaussian-distributed membership functions (GDMF's) constructed from the texture features inherent in the ultrasound images are used, and the coefficients acted as a set of parameters to adjust the magnitudes of the standard deviations of the GDMF's are employed. Optimal coefficients are determined through training process using the GA. The GA-based fuzzy classifier is used to discriminate two sets of echocardiographic images, namely, normal and abnormal cases, diagnosed by a highly trained physician. The results of the authors' experiments are very promising. The authors' achieve an average classification rate of 96%. The results indicate that the method has potential utility for computer-aided diagnosis of myocardial heart disease.