Multi-Class Stress Detection Through Heart Rate Variability: A Deep Neural Network Based Study

Abstract

This study presents a multi-class stress detection system based on heart rate variability (HRV) analysis using a deep neural network (DNN) framework. Stress is a significant physiological and psychological factor that affects human health, productivity, and overall well-being, making its timely detection essential in healthcare and human-centered applications. Conventional stress detection methods often rely on binary classification or handcrafted statistical models, which may fail to capture the complex nonlinear patterns present in physiological signals. To address these limitations, the proposed work utilizes HRV features extracted from electrocardiogram (ECG) signals to classify multiple stress levels, enabling a more detailed and realistic assessment of an individual’s stress state. A deep neural network is employed to automatically learn discriminative patterns from HRV data and perform robust multi-class classification of stress conditions such as relaxed, low stress, moderate stress, and high stress. The proposed framework includes signal preprocessing, HRV feature extraction, feature normalization, and DNN-based classification for accurate stress recognition. Experimental evaluation demonstrates that the deep learning-based approach outperforms conventional machine learning models in terms of classification accuracy, robustness, and generalization. The proposed system offers a reliable and intelligent solution for continuous stress monitoring and has significant potential in wearable healthcare, mental health assessment, and real-time physiological monitoring applications.