Stability of classification performance on an adaptive neuro fuzzy inference system for disease complication prediction

It is crucial to detect disease complications caused by metabolic syndromes early. High cholesterol, high glucose, and blood pressure are indicators of syndrome. The aim this study use adaptive neuro fuzzy inference system (ANFIS) predict potential compare its performance other classifiers, namely random forest (RF), C4.5, naïve Bayesian classification (NBC) algorithms. Fuzzy subtractive clustering used construct membership functions rules throughout the process. This analyzed 148 different data sets. Cholesterol, systolic, diastolic all included in collection. learning process was conducted using a hybrid algorithm. consequent parameters adjusted forward leastsquare approach, while premise backward gradient-descent determined following indicators: accuracy, sensitivity, specification, precision, area under curve (AUC), root mean squared error (RMSE). results training prove that ANFIS an "excellent classification" classifier. has proven have very good stability across six parameters. properties implementation strongly support stability.