Classification of Cardiac Arrhythmias using Biorthogonal Wavelets and Support Vector Machines

The classification of Electrocardiogram (ECG) is critical for diagnosis and treatment of patients with heart disorders. We present a technique for automatic the detection and classification of cardiac arrhythmias using biorthogonal wavelet functions and support vector machines (SVM). For feature extraction the biorthogonal wavelet transforms are applied to decompose the ECG signal into wavelet scales. Further, a soft thresholding technique is used to denoise and extract important cardiac events in the signal. Subsequently, we applied SVM classifier to discriminate the detected event features into normal or pathological ones. Numeric computations demonstrate that the efficient wavelet preprocessing provides an accurate estimation of important physiological features of ECG and its effect in the improvement of the SVM classification performance.