A Novel Deep Learning-based Model for the Efficient Classification of Electrocardiogram Signals

To manage healthcare, an electrocardiogram, often known as an "EKG" or "ECG", is a measurement of the electrical activity of the organ "heart". Deep Learning (DL) or Deep Neural Networks have recently attracted the attention of researchers in many other sectors, including healthcare and medicine. There has been a frequent rise in the number of researchers developing the model to classify and detect several diseases, out of which cardiac complications are the keen focus due to the mortality associated with it. Therefore, the objective of the present research is to develop a classification model for efficient and accurate classification of signals received from ECG. The present study uses a "deep neural network" for the classification of the ECG signal into a total of five criteria including Normal ECG, QRS Widening, ST Elevation, ST Depression, and Sinus Rhythm. The developed classification method is tested and evaluated with the "MITBIH arrhythmia database" which revealed significant matrices for all parameters such as "precision", "accuracy", "recall", and "F-1 score". In addition to that, the proposed model demonstrated competent results which further highlights the practical applicability of the model to implementation and adoption in the healthcare sector.