A Review on Anomaly Detection in Pacemaker Signal Patterns using One-Class SVM for Real-Time Cardiac Monitoring

Abstract

The advancement of real-time cardiac monitoring systems has become vital for proactive healthcare, especially in patients relying on pacemakers. Detecting anomalies in pacemaker signal patterns can serve as an early indicator of arrhythmias, device malfunction, or abnormal physiological changes. This review investigates the role of One-Class Support Vector Machine (OC-SVM) in identifying such anomalies with high sensitivity and minimal false positives. Unlike traditional supervised models that require extensive labeled datasets, OC-SVM offers a robust solution by modeling only the normal behavior and flagging deviations as potential anomalies. This paper surveys recent literature, compares performance metrics, and highlights the integration of OC-SVM with signal preprocessing, feature extraction, and edge-based deployment. The review concludes with insights into challenges, such as real-time implementation and false alarm reduction, and proposes future research directions to enhance the reliability of AI-driven cardiac monitoring systems.