Published Online:June 2026
Product Name:The IUP Journal of Information Technology
Product Type:Article
Product Code:IJIT020626
DOI:10.71329/IUPJIT/2026.22.2.32-58
Author Name:Priyanka Diwan, Swasti H S, Subhra Sulagana Debata, Kannagi A and Ananta Charan Ojha
Availability:YES
Subject/Domain:Engineering
Download Format:PDF
Pages:32-58
The paper proposes a hybrid framework that integrates machine learning and causal inference to enable both prediction and explanation of chronic stress. Using questionnaire responses collected from 300 participants, a Weighted Chronic Stress Index was developed to capture the differential contribution of six stress domains. Feature selection and SMOTE-based class balancing were applied prior to classification using Logistic Regression, Support Vector Machine (SVM), Random Forest, and XGBoost models. To move beyond correlational analysis, causal effects of domain-specific stressors were estimated using Inverse Probability Weighting, Double Machine Learning, and Causal Forests. The results demonstrate strong predictive performance across all models, with SVM achieving the highest classification accuracy and ROC-AUC. The proposed framework combines predictive accuracy with causal interpretability, providing a more comprehensive approach to chronic stress assessment and supporting the development of targeted and personalized stress intervention strategies.
Chronic stress is a persistent condition that disrupts both psychological and physiological functioning, often requiring extended periods for recovery. Unlike acute stress, it imposes sustained strain on the body, contributing to fatigue, impaired concentration, cardiovascular complications, and metabolic imbalance.