An artificial intelligence methodology for enhancing Sugeno defuzzification method

Abstract

This research is focused on investigating principal components analysis (PCA) for defuzzification which estimates singleton fuzzy values in a subjective way for Sugeno defuzzification method. This work goes beyond the approach of Sugeno defuzzification method to defuzzification where singleton fuzzy values are considered to be objective. The new artificial intelligence methodology is proposed for improving Sugeno defuzzification method by directly integrating it with a principal component analyser, a fuzzy inference engine, a knowledge base, and a user interface. For the chosen datasets, the artificial intelligence methodology improved accuracy by considering the difference between the predicted value and actual value. The improvements were up to 98%, 98%, 71%, and 95% for the Card Payment, non-communicable disease, communicable disease and Manas Prakriti respectively, while Sugeno defuzzification method shows significantly low accuracy by considering the difference between predicted value and actual value.