An ECG Beat Classification Using Adaptive Neuro-Fuzzy Inference System

Electrocardiography (ECG or EKG) uses electrodes to measure an electrical activity of heart. These heart signal allows comprehensive analysis of the heart condition. Acquisition of ECG signal from the human body is a non-invasive process that opens the door to new possibilities for the application of advanced signal processing and data analysis techniques in the diagnosis of heart diseases. With the availability and help of large database of ECG signal, a computationally intelligent system can be built and can take place of a cardiologist. The various abnormalities in the patient heart can be detected to identify various cardiac disease by use of Adaptive Neuro-Fuzzy Inference System (ANFIS) preprocessed by subtractive clustering. In this paper, six types of heartbeats are classified: Normal sinus rhythm, premature ventricular contraction (PVC), Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB), Atrial Premature Contraction (APC) and paced beats. The research aimed at detecting important characteristics (features of an ECG signal to determine whether the patient heartbeat is normal or irregular. The results from three different trials indicate an average accuracy of 98.43 %, average sensitivity of 95.3% and average specificity of 98.6%. These results are comparable to two artificial neural network (ANN) algorithms: gradient descent and Lavenberg Marquardt, as well as ANFIS preprocessed by grid partitioning.