Classification of Ventricular arrhythmia using a support vector machine based on morphological features

This paper proposes a method for the classification of ventricular arrhythmia using support vector machines (SVM). The features used in the SVMs were extracted automatically based on morphological information. Three different features were extracted: RR interval, QRS slope, and QRS shape similarity. Then, the SVM was used to classify five different electrocardiogram (ECG) heartbeat episodes. The Gaussian Radial Basis Function was utilized for the kernel function because the ECG beat episodes were treated as a non-linear pattern. The sensitivity of the classification used for the five beat episodes was 93.16%.