Cardiac Arrhythmia Disorders Detection with Deep Learning Models

In this paper, the research of computer algorithms for automatic detection of heart rhythm disorders based on the analysis of electrocardiograms have been conducted. A new model of the electrocardiogram classifier is proposed, as an ensemble of a two-dimensional convolutional neural network and a long short-term memory model, which includes an attention layer. Computer experiments conducted on the MIT-BIH Physionet collection of ECG showed its high performance compared to other models of machine and deep learning. Proposed model successfully detected cardiac arrhythmia classes with an accuracy of 99.34%, AUC = 99% and recall = 99%. © 2022, Springer Nature Switzerland AG.