Classification of Premature Ventricular Contraction based on ECG Signal using Multiorder Rényi Entropy

Electrocardiogram (ECG) signals are commonly used to analyze the heart abnormalities. Many researchers used this method due to the simplicity, inexpensive, and the non-invasiveness of the devices. The basic concept of ECG is by measuring the electrical activity of the heart using non-invasive electrodes placed in the body. Premature ventricular contraction (PVC) is one of the abnormalities of the human heart which produce extra heartbeat that started in the two lower ventricles. This `additional' heartbeat disrupts the regular heart rhythm. Early detection for PVC is essential, so in this research, we classify the PVC from ECG signal by using multi-order Rényi entropy as the feature extraction, and SVM as the classifier. We search for the optimum feature number needed for the detection system. Our proposed method showed a promising result, because we only use one feature extraction parameter, that is Rényi entropy, as the only feature which calculated in the different orders, and this made the computing complexity low. We got 95.8% of accuracy using six characteristics of entropy for PVC and normal ECG classification. Our proposed method was simple, low computation and need fewer number of features.