HEART RATE VARIABILITY-BASED STRESS DETECTION USING DEEP NEURAL NETWORK MODELS

Abstract

This research makes use of Heart Rate Variability (HRV) to gauge people's levels of stress by analyzing physiological data through models trained on deep neural networks. Readings of heart rate variability (HRV) obtained from an electrocardiogram (ECG) are employed in the investigation due to their significance as indicators of autonomic nervous system activity that are associated with stress responses. In order to enhance the accuracy and utility of stress recognition, advanced deep neural network techniques, such as Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) models, are implemented in lieu of conventional machine learning methodologies. By means of feature extraction, classification, and data preparation, the proposed system is capable of distinguishing between stressful and non-stressful scenarios in real time. For instance, researchers have demonstrated that deep learning models are more adept at learning new features, making predictions, and managing intricate physiological trends. This investigation enhances healthcare monitoring systems that can identify early indicators of stress and mental illness, thereby facilitating the maintenance of one's own health in daily life.