HEARTBEAT DYNAMICS: A NOVEL EFFICIENT INTERPRETABLE FEATURE FOR ARRHYTHMIAS

Abstract

This project aims to advance arrhythmia detection by analyzing the intricate dynamics of heartbeat patterns. A novel and interpretable feature derived from heart rate variability (HRV) is introduced to capture subtle physiological variations associated with cardiac irregularities. Leveraging deep neural network (DNN) architectures, the proposed system performs multi-class stress detection, offering insights into the correlation between stress states and arrhythmic behavior. By integrating interpretability with deep learning, this approach enhances both the accuracy and transparency of arrhythmia diagnosis, contributing to more reliable and clinically meaningful cardiac monitoring systems.