Integrating Machine Learning and Biological Stress Indicators for Early Prediction of Occupational Burnout among Young Professionals
- 1University of Agriculture Faisalabad
- 2South East Technological University, Ireland
- 3Bahauddin Zakariya University
Abstract
Occupational burnout can develop slowly, making it difficult to detect in the early stages. The primary data for assessing the situation come periodically from the person experiencing it. The paper discusses the potential to detect occupational burnout by integrating biological, psychological, and work-related data through statistical analysis and machine learning. The research outlines biological sensors, physiological stress responses, and machine-learning-based prediction of mental pressure, focusing on heart rate variability, electrodermal activity, sleep, and physical activity. The study suggests using descriptive statistics and correlation analysis to assess relationships between major variables, and logistic regression as a simple statistical prediction method. Two machine learning models can be readily employed for pattern identification of early burnout risk: a Decision Tree and k-nearest neighbors (k-NN). The sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC) can be used to assess model performance. The main research question is whether biological information predicts outcomes beyond psychological and occupational information. The article addresses the issue through the MEBTF, which can overcome the problem by using multimodal measurement methods to detect changes in risk trajectories rather than diagnosing burnout. The research shows considerable potential; however, it has identified biological outcomes, a lack of external validation, and a small sample size. Future prospective research must address these issues.
This is a preprint / defensive disclosure to establish priority. This version is not peer-reviewed. An extended version will be submitted to a peer-reviewed journal
Related
- Nomogram models integrating multiregion radiomics, quantitative computed tomography, and clinical indicators for distinguishing connective tissue disease-associated interstitial lung disease from idiopathic pulmonary fibrosis and improving the prediction of gender-age-physiology stage
- Integrating Genotype With Clinical Indicators Improves Venous Thromboembolism Risk Prediction in Cancer Patients: A Retrospective Cohort Study From Xinjiang
- A nomogram prediction model for 30-day mortality risk in severe cerebral infarction patients integrating inflammation-related indicators and APACHE II score.
- Hearing thresholds among dental professionals and students: early identification of occupational hearing loss using conventional and advanced audiological assessments
- Quality of work life and professional ethics among occupational health professionals: a social exchange theory perspective.
BibTeX
@misc{kanwal2026integrating,
title = {{Integrating Machine Learning and Biological Stress Indicators for Early Prediction of Occupational Burnout among Young Professionals}},
author = {Nayab Kanwal and Tabassum Dr Mujahid and Hammad Muhammad},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23571582},
url = {https://zenodo.org/records/23571582},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Integrating Machine Learning and Biological Stress Indicators for Early Prediction of Occupational Burnout among Young Professionals
AU - Nayab Kanwal
AU - Tabassum Dr Mujahid
AU - Hammad Muhammad
PY - 2026
DA - 2026/10/22
PB - Zenodo
DO - 10.5281/zenodo.23571582
UR - https://zenodo.org/records/23571582
AB - Occupational burnout can develop slowly, making it difficult to detect in the early stages. The primary data for assessing the situation come periodically from the person experiencing it. The paper discusses the potential to detect occupational burnout by integrating biological, psychological, and work-related data through statistical analysis and machine learning. The research outlines biological sensors, physiological stress responses, and machine-learning-based prediction of mental pressure, focusing on heart rate variability, electrodermal activity, sleep, and physical activity. The study suggests using descriptive statistics and correlation analysis to assess relationships between major variables, and logistic regression as a simple statistical prediction method. Two machine learning models can be readily employed for pattern identification of early burnout risk: a Decision Tree and k-nearest neighbors (k-NN). The sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC) can be used to assess model performance. The main research question is whether biological information predicts outcomes beyond psychological and occupational information. The article addresses the issue through the MEBTF, which can overcome the problem by using multimodal measurement methods to detect changes in risk trajectories rather than diagnosing burnout. The research shows considerable potential; however, it has identified biological outcomes, a lack of external validation, and a small sample size. Future prospective research must address these issues. This is a preprint / defensive disclosure to establish priority. This version is not peer-reviewed. An extended version will be submitted to a peer-reviewed journal
ER -