Abstract

Electrocardiogram (ECG) signals provide important information about heart function and structure. ECG signals are used to diagnose heart disease and classify cardiac arrhythmias. These signals are provided using the PQRSTU waveform. In this paper, extraction of the characteristics of P wave, PR interval, QRS complex, ST segment, QT interval, and T wave of ECG signal is performed using discrete wavelet transform method. After extracting the desired features, MLP neural network with 30 hidden layers and 20 epoch for training, were introduced for classification of cardiac arrhythmias. This article used the 70 samples of ECG signals MIT BIH database.