STUDENT STRESS CLASSIFICATION BASED ON MACHINE LEARNING ALGORITHMS
https://doi.org/10.53360/2788-7995-2026-3(23)-6
Abstract
Survey instruments remain the most practical approach for large-scale monitoring of student stress in educational settings, and the Perceived Stress Scale (PSS) continues to be validated across contexts. This study evaluates how modern, off-the-shelf machine-learning methods perform on a tabular, multi-class student survey with pronounced class imbalance. We conduct a two-stage evaluation. First, we benchmark Decision Tree (DT), Random Forest (RF), and k-Nearest Neighbors (KNN) across five train-test regimes (90/10, 80/20, 70/30, 60/40, 50/50) to probe stability under different sample sizes. Second, we add an imbalance-aware experiment on an 80/20 split using SMOTE-style oversampling, and compare RF, Gradient Boosting (GB), and XGBoost (XGB). We report macro-averaged metrics suited for skewed multi-class targets. Across the five-split benchmark, RF and KNN are consistently stronger than a single DT. These outcomes align with recent evidence that tree-based ensembles dominate on typical tabular data and with reports that ensembles are reliable for stress-related classification. On the 80/20 split, RF and XGBoost tie at macro-F1 ≈ 0.73, while an additional SVM (RBF) reaches the best macro-F1 of 0.86. Clustering and projections confirm that stress responses form a continuum rather than discrete groups.
About the Authors
O. UmarovKazakhstan
Olzhas Umarov – third-year student
010000, Republic of Kazakhstan, Astana, 55/11 Mangilik El avenue, C1 block
Zh. Oralbekova
Kazakhstan
Zhanar Oralbekova – PhD, associate professor, acting Professor of the School of AI and Data Science
010000, Republic of Kazakhstan, Astana, 55/11 Mangilik El avenue, C1 block
Sh. Tutkyshbayeva
Kazakhstan
Shyryn Tutkyshbayeva – PhD, assistant professor of the School of Software Engineering
010000, Republic of Kazakhstan, Astana, 55/11 Mangilik El avenue, C1 block
D. Oralbekova
Kazakhstan
Dina Oralbekova – PhD, associate professor, Senior Researcher
050010, Republic of Kazakhstan, Almaty, 28 Shevchenko street
G. Tyulepberdinova
Kazakhstan
Gulnur Tyulepberdinova – candidate of physical and mathematical sciences, associate professor of the Department of Artificial Intelligence and Big Data
050040, Republic of Kazakhstan, Almaty, 71 Al-Farabi avenue
References
1. Brown T.A., Sellbom M. Further validation of the Personality Disorder Severity for ICD-11 (PDS-ICD-11) scale in a community mental health sample // Psychological Assessment. – 2023. – Vol. 35(8). – P. 706–714. https://doi.org/10.1037/pas0001253
2. Breiman L. Random forests // Machine Learning. – 2001. – Vol. 45. – P. 5–32. https://doi.org/10.1023/A:1010933404324
3. Chen T., Guestrin C. XGBoost: A scalable tree boosting system // KDD '16: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. – 2016. – P. 785–794. https://doi.org/10.1145/2939672.2939785
4. Grinsztajn L., Oyallon E., Varoquaux G. Why do tree-based models still outperform deep learning on typical tabular data? // 36th Conference on Neural Information Processing Systems (NeurIPS 2022) Track on Datasets and Benchmarks [Electron. resource]. – 2022. – Vol. 35. – P. 507-520. – URL: https://proceedings.neurips.cc/paper_files/paper/2022/file/0378c7692da36807bdec87ab043cdadc-Paper-Datasets_and_Benchmarks.pdf (date of access: 3.01.2026).
5. Shwartz-Ziv R., Armon A. Tabular data: Deep learning is not all you need // Information Fusion. – 2022. – Vol. 81. – P. 84–90. https://doi.org/10.1016/j.inffus.2021.11.011.
6. Anand R.V., Md A.Q., Urooj S., Mohan S., Alawad M.A., Adittya C. Enhancing diagnostic decision making: Ensemble learning techniques for reliable stress level classification // Diagnostics. – 2023. – Vol. 13(22): 3455. https://doi.org/10.3390/diagnostics13223455.
7. Grandini M., Bagli E., Visani G. Metrics for multi-class classification: An overview [Electron. resource] // arXiv:2008.05756. – 2020. – URL: https://arxiv.org/abs/2008.05756 (date of access: 3.01.2026).
8. Saito T., Rehmsmeier M. The precision–recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets // PLOS ONE. – 2015. – Vol. 10(3): e0118432. https://doi.org/10.1371/journal.pone.0118432
9. Raschka S. Model evaluation, model selection, and algorithm selection in machine learning [Electron. resource] // arXiv:1811.12808. – 2018. – URL: https://arxiv.org/abs/1811.12808 (date of access: 3.01.2026).
10. Lemaître G., Nogueira F., Aridas C.K. Imbalanced-learn: A Python toolbox to tackle the curse of imbalanced datasets in machine learning [Electron. resource] // Journal of Machine Learning Research. – 2017. – Vol. 18. – P. 1–5. – URL: https://www.jmlr.org/papers/volume18/16-365/16-365.pdf (date of access: 3.01.2026).
11. Kaggle. Student Stress Monitoring Datasets [Electron. resource]. – 2025. – URL: https://www.kaggle.com/datasets/mdsultanulislamovi/student-stress-monitoring-datasets (date of access: 15.09.2025).
Review
For citations:
Umarov O., Oralbekova Zh., Tutkyshbayeva Sh., Oralbekova D., Tyulepberdinova G. STUDENT STRESS CLASSIFICATION BASED ON MACHINE LEARNING ALGORITHMS. Bulletin of Shakarim University. Technical Sciences. 2026;(3(23)):54-62. https://doi.org/10.53360/2788-7995-2026-3(23)-6
JATS XML















