Automated diagnosis of Coronary Artery Disease using nonlinear features extracted from ECG signals

Coronary Artery Disease (CAD) is one of the hazardous heart disease which results in angina, Myocardial Infarction (MI) and Sudden Cardiac Death (SCD). CAD is a cardiac disorder in which a plague develops in the interior wall of the arteries resulting in blockage of blood reaching to the heart muscles. Electrocardiogram (ECG) is the cardiac signal which represents cardiac depolarisation and repolarisation regulated at the surface of the chest. The minute variations in amplitude and duration in the ECG wave specifies different pathological conditions which are tedious to interpret visually. Hence computer aided diagnostic systems are used to monitor ECG signals. In the present work, automated diagnosis of CAD is done using Discrete Wavelet Transform (DWT) and nonlinear feature extraction techniques like; Multivariate Multi-scale Entropy (MMSE), Tsallis entropy and renyi entropies. The extracted features after DWT are ranked based on t-value and fed to K Nearest Neighbour (KNN), Support Vector Machine (SVM), Probabilistic Neural Network (PNN) and Decision Tree (DT) classifiers for automated classification of normal and CAD classes. This technique provided the highest accuracy of 98.67% using KNN classifier. Hence, the proposed system can aid clinicians in faster and accurate diagnosis of CAD and thereby provide sufficient time for proper treatment.