Classification of premature cardiac contractions based on RFECV and ensemble learning

Abstract

Premature cardiac contractions, including premature atrial contractions (PACs) and premature ventricular contractions (PVCs), are common arrhythmias that may increase the risk of cardiovascular complications when they occur frequently. Accurate classification of these events from electrocardiogram (ECG) signals remains challenging due to noise and signal variability. This study proposes a machine learning–based classification framework that combines recursive feature elimination with cross-validation for feature selection and an ensemble learning strategy to improve classification robustness. The approach was evaluated using the Massachusetts Institute of Technology – Beth Israel Hospital (MIT-BIH) Arrhythmia database and achieved high classification performance, with an accuracy of 95.34%, F1-score of 92.11%, and balanced precision and recall for PVC and PAC. In addition, SHapley Additive exPlanations (SHAP) were employed to identify the most influential features, enhancing model interpretability. The results demonstrate that the proposed framework provides a reliable and interpretable solution for distinguishing premature cardiac contractions, highlighting its potential application in clinical decision support systems.