Detecting Stress Levels using Physiological Signals
Students & Supervisors
Student Authors
Supervisors
Abstract
Stress has become a critical issue in modern life and severely impacts mental, physical and emotional health. Chronic stress is associated with cardiovascular disorders, immune dysfunction and depression making early and accurate detection essential. Traditional self-report methods, such as questionnaires, lack subjective and continuous monitoring capabilities. With the advancement of wearable sensing and machine learning, it is now possible to objectively detect stress through physiological signals. This study investigates supervised machine learning algorithms for stress level classification using the WESAD dataset, which includes multimodal signals such as electrodermal activity, blood volume pulse, heart rate, temperature and accelerometer data collected under different mental states. The data were preprocessed by outlier removal, normalization and dimensionality reduction using principal component analysis. Multiple models were evaluated using precision, accuracy, recall, and F1-score metrics. Experimental results show that the stacking ensemble learning model achieved the highest accuracy of 91.45 percent, outperforming other classifiers such as random forest and logistic regression. These results demonstrate that multimodal physiological features are effective indicators of stress and that ensemble learning significantly increases the reliability of detection, paving the way for real-time, wearable-based stress monitoring applications.
Keywords
Publication Details
- Type of Publication:
- Conference Name: 2026 IEEE 2nd International Conference on Quantum Photonics, Artificial Intelligence & Networking (QPAIN 2026)
- Date of Conference: 16/04/2026 - 16/04/2026
- Venue: IT Business Incubator, Chittagong University of Engineering and Technology (CUET), Chattogram, Bangladesh
- Organizer: IEEE