Performance Evaluation of Machine Learning Algorithms for Arrhythmia Classification in ECG Signals

Automated arrhythmia detection using electrocardiography (ECG) has the potential to improve diagnosis and treatment of cardiac conditions. In this study, we evaluated the performance of various machine learning (ML) algorithms for ECG categorization using single or multi-lead ECG data from the MIT-BIH database. We compared pre-trained models, ML models, and various data collection approaches for the diagnosis of normal beat (NOR), left bundle branch block beat (LBBB), right bundle branch block beat (RBBB), premature atrial contraction (PAC), and premature ventricular contraction (PVC). Our results show that an artificial neural network (ANN) using a radial basis function kernel and a dataset of ECG samples achieves an outstanding accuracy of 100%. This approach has the potential to be used in real clinical applications, especially in hospital settings where various ECG datasets are gathered through outpatient clinics.