Research on Diagnosing Heart Disease Using Adaptive Network-based Fuzzy Interferences System

Abstract

The shape of ST segment of Electrocardiogram (ECG) is of great importance in diagnosing heart diseases. Based on feature points of ST segments which have been extracted from electrocardiogram (ECG) data with wavelet transform (WT), a five-input-and-single-output adaptive network-based fuzzy interferences system (ANFIS) is designed to classify the shapes of ST segments. In the system the if-then rule of Takagi-Sugeno is taken, and the combination of the gradient descent and the least-squares method is adopted to train the system. The effectiveness is demonstrated via the ECG data from the MIT-BIT and clinical ECG data.