Detection of arrhythmia based on discrete wavelet transform using artificial neural network and support vector machine

Arrhythmia is an abnormal rhythm of heart. Correct detection or interpretation is much more important. In this paper, we aim at detecting 5 types of ECG beats: (1) Normal beat (2) Right Bundle Branch Block, (3) Left Bundle Branch Block, (4) Premature Ventricular Contraction (5) Paced beat. We compared the performances of two different classifier using DWT feature extraction technique. The DWT technique uses detail coefficient (D1, D2, D3, D4, D5) and approximation coefficient A5 as a features. This feature sets is independently given to Artificial Neural network and Support Vector Machine classifiers. The experimental result shows that we obtained the highest sensitivity of 98.80% using Support Vector Machine as a classifier.