Machine Learning Based Assessment of Mental Stress Using Wearable Sensor Data

Abstract

Mental stress is a major health issue that affects an individual\'s physical and psychological well-being. Conventional stress assessment techniques mainly depend on self-reporting methods and clinical evaluations, which are subjective and unsuitable for continuous monitoring. With the advancement of wearable sensor technology, physiological signals such as heart rate, electrodermal activity, skin temperature, and body movement can be collected in real time. This paper presents a machine learning-based approach for assessing mental stress using wearable sensor data. The proposed system involves data collection, preprocessing, feature extraction, and classification using machine learning algorithms. The developed framework can provide accurate and continuous stress monitoring, supporting applications in healthcare, workplace monitoring, and personal wellness management.

Introduction

The text presents a Machine Learning-Based Assessment of Mental Stress Using Wearable Sensor Data. The main objective is to develop a continuous, non-invasive system that can detect and classify mental stress using physiological signals collected from wearable devices.

Traditional stress assessment methods, such as questionnaires, interviews, and clinical evaluations, are subjective and cannot provide continuous real-time monitoring. Wearable sensors overcome this limitation by continuously measuring physiological parameters such as heart rate, Heart Rate Variability (HRV), Electrodermal Activity (EDA), skin temperature, and body movement.

The proposed system follows five major stages:

1. Sensor Data Acquisition – Wearable devices collect physiological signals associated with different stress conditions.
2. Data Preprocessing – Noise and unwanted variations are removed through filtering and normalization.
3. Feature Extraction – Relevant statistical, time-domain, and frequency-domain features are extracted from the sensor signals.
4. Machine Learning Classification – Extracted features are analysed using machine learning algorithms.
5. Stress Prediction – The trained model classifies an individual's stress level, such as low, moderate, or high.

The study considers several machine learning algorithms, including Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), Artificial Neural Network (ANN), Decision Tree (DT), and Gradient Boosting Machine (GBM). These models are trained using labelled wearable-sensor data and evaluated using accuracy, precision, recall, and F1-score. The best-performing model can then be selected for stress prediction.

The system also incorporates Internet of Things (IoT), artificial intelligence, signal processing, and wearable sensor technologies. IoT enables physiological data to be transmitted from wearable devices through Bluetooth, Wi-Fi, or cloud platforms, allowing continuous monitoring and potentially remote access to stress information.

The proposed application includes user registration and login, a data-viewing module for physiological measurements, an algorithm module, and a prediction module that determines the user's stress level from processed sensor data.

Key Findings / Benefits

• Enables continuous and real-time stress monitoring.
• Provides a non-invasive and user-friendly approach.
• Reduces dependence on subjective questionnaires and manual assessment.
• Uses multiple physiological signals to improve stress detection.
• Machine learning can identify complex patterns that may not be easily detected through conventional methods.
• Can potentially be applied in healthcare, workplaces, education, and personal wellness.

Future Scope

Future improvements could include deep learning models, personalized stress-detection systems that adapt to individual physiological patterns, and integration with IoT, cloud computing, smartphones, and mobile applications. These developments could make stress monitoring more accurate, personalized, remotely accessible, and capable of providing real-time intervention or alerts.

Conclusion

This paper presents a machine learning-based approach for mental stress assessment using wearable sensor data. The integration of wearable technology and artificial intelligence enables continuous and objective stress monitoring. By analyzing physiological signals, machine learning models can identify stress patterns and provide valuable support for healthcare and wellness applications.

References

[1] P. Schmidt et al., \"Introducing WESAD, a Multimodal Dataset for Wearable Stress and Affect Detection,\" ACM International Conference on Multimodal Interaction, 2018. [2] J. A. Healey and R. W. Picard, \"Detecting Stress During Real-World Driving Tasks Using Physiological Sensors,\" IEEE Transactions on Intelligent Transportation Systems, 2005. [3] A. Sano and R. W. Picard, \"Stress Recognition Using Wearable Sensors and Mobile Phones,\" IEEE Transactions on Affective Computing, 2013. [4] Y. S. Can, B. Arnrich, and C. Ersoy, \"Stress Detection in Daily Life Scenarios Using Smart Phones and Wearable Sensors,\" Journal of Biomedical Informatics, 2019.

Copyright

Copyright © 2026 Dr. M. Veerabhadra Rao, A. Anudurga. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.