ECG Classification for Heart Arrhythmia Using Deep Machine Learning

Abstract

Healthcare professionals commonly use Electrocardiogram (ECG) as a low-cost diagnostic tool for monitoring heart electrical signals. Arrhythmia, which is an abnormal heart signal, can be dangerous and cause death. The arrhythmia can be categorized in various types including tachycardia, bradycardia, supraventricular arrhythmias, and ventricular. The automated monitoring of arrhythmia and classification with ECG is very helpful for doctors. In this research we use deep machine learning for automated arrhythmia classification with the focus on the recent trends in arrhythmia classification. Using St. Mary’s University Deep Learning Platform, we conducted heavy and complex simulations to measure the performance of the various arrhythmia classification and detection models. Finally, we present the accuracy of the proposed deep learning algorithms, which surpasses the performance of the existing algorithms in precision and sensitivity.