Title: An ANN-based HRV classifier for cardiac health prognosis

Authors: Ramesh Kumar Sunkaria; Vinod Kumar; Suresh Chandra Saxena; Achala M. Singhal

Addresses: Electrical Engineering Department, Indian Institute of Technology Roorkee, Roorkee (Uttaranchal), 247667, India ' Electrical Engineering Department, Indian Institute of Technology Roorkee, Roorkee (Uttaranchal), 247667, India ' Electrical Engineering Department, Indian Institute of Technology Roorkee, Roorkee (Uttaranchal), 247667, India ' Department of Cardiology, Himalayan Institute & Hospital Trust, Dehradun (Uttaranchal), 248140, India

Abstract: A multi-layer artificial neural network (ANN)-based heart rate variability (HRV) classifier has been proposed, which gives the cardiac health status as the output based on HRV of the patients independently of the cardiologists' view. The electrocardiogram (ECG) data of 46 patients were recorded in the out-patient department (OPD) of a hospital and HRV was evaluated using self-designed autoregressive-model-based technique. These patients suspected to be suffering from cardiac abnormalities were thoroughly examined by experienced cardiologists. On the basis of symptoms and other investigations, the attending cardiologists advised them to be classified into four categories as per the severity of cardiac health. Out of 46, the HRV data of 28 patients were used for training and data of 18 patients were used for testing of the proposed classifier. The cardiac health classification of each tested patient with the proposed classifier matches with the medical opinion of the cardiologists.

Keywords: ANNs; artificial neural networks; electrocardiograms; ECG; autonomic activity; HRV classifier; heart rate variability; autoregressive modelling; electronic healthcare; e-healthcare; cardiac health prognosis; outpatients.

DOI: 10.1504/IJEH.2014.064332

International Journal of Electronic Healthcare, 2014 Vol.7 No.4, pp.315 - 330

Accepted: 30 Apr 2013
Published online: 18 Aug 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article