Predictive Modelling of Exam Outcomes Using Stress-Aware Learning from Wearable Biosignals
Abstract
1. Introduction
2. Material and Method
2.1. Data Collection
2.2. Preprocessing and Feature Engineering
2.3. Model Training and Evaluation
- TP (True Positive): Instances that are actually positive and correctly predicted as positive.
- TN (True Negative): Instances that are actually negative and correctly predicted as negative.
- FP (False Positive): Instances that are actually negative but incorrectly predicted as positive.
- FN (False Negative): Instances that are actually positive but incorrectly predicted as negative.
3. Results and Discussion
4. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Kang, W.; Kim, S.H.; Yoo, E.; Kim, S. Predicting Students’ Exam Scores Using Physiological Signals. arXiv 2023, arXiv:2301.12051. [Google Scholar] [CrossRef]
- Al-Alim, M.A.; Mubarak, R.; Salem, N.M.; Sadek, I. A machine-learning approach for stress detection using wearable sensors in free-living environments. Comput. Biol. Med. 2024, 179, 108918. [Google Scholar] [CrossRef]
- Alzukari, R.; Gallo, K. The Effect of Stress on Academic Achievement; Technical Report; Department of Counseling Educational Psychology and Foundations, Mississippi State University: Starkville, MS, USA, 2023. [Google Scholar]
- Gjoreski, M.; Gjoreski, H.; Lustrek, M.; Gams, M. Continuous stress detection using a wrist device: In laboratory and real life. In Proceedings of the ACM International Joint Conference on Pervasive and Ubiquitous Computing: Adjunct, Heidelberg, Germany, 12–16 September 2016; pp. 1185–1193. [Google Scholar]
- Rajendran, S.; Chamundeswari, S.; Sinha, A.A. Predicting the academic performance of middle- and high-school students using machine learning algorithms. Soc. Sci. Humanit. Open 2022, 6, 100357. [Google Scholar] [CrossRef]
- Neigel, P.; Vargo, A.; Tag, B.; Kise, K. Unobtrusive stress detection using wearables: Application and challenges in a university setting. Front. Comput. Sci. 2025, 7, 1575404. [Google Scholar] [CrossRef]
- Zavaleta, J.C.; Alva, R.Y.; Vargas, S.V.; Medina, E.F.; Somoza, Y.P.; Andrade-Arenas, L. Relationship between Stress and Academic Performance: An Analysis in Virtual Mode. Int. J. Adv. Comput. Sci. Appl. (IJACSA) 2021, 12, 823–833. [Google Scholar] [CrossRef]
- Sano, A.; Picard, R.W. Stress Recognition Using Wearable Sensors and Mobile Phones. In Proceedings of the 2013 Humaine Association Conference on Affective Computing and Intelligent Interaction, Geneva, Switzerland, 2–5 September 2013; pp. 671–676. [Google Scholar] [CrossRef]
- Setz, C.; Arnrich, B.; Schumm, J.; Marca, R.L.; Tröster, G.; Ehlert, U. Discriminating Stress From Cognitive Load Using a Wearable EDA Device. IEEE Trans. Inf. Technol. Biomed. 2010, 14, 410–417. [Google Scholar] [CrossRef] [PubMed]
- Amin, M.R.; Wickramasuriya, D.S.; Faghih, R.T. A Wearable Exam Stress Dataset for Predicting Grades using Physiological Signals. In Proceedings of the 2022 IEEE Healthcare Innovations and Point of Care Technologies (HI-POCT), Houston, TX, USA, 10–11 March 2022; pp. 30–36. [Google Scholar] [CrossRef]
- PhysioNet. A Wearable Exam Stress Dataset for Predicting Cognitive Performance in Real-World Settings. 2021. Available online: https://physionet.org/content/?topic=stress (accessed on 7 September 2025).
- Braithwaite, J.J.; Watson, D.G.; Jones, R.; Rowe, M. A Guide for Analysing Electrodermal Activity (EDA) and Skin Conductance Responses (SCRs) for Psychological Experiments; Technical Report; University of Birmingham: Birmingham, UK, 2015. [Google Scholar]
- Abd-Alrazaq, A.; Alajlani, M.; Ahmad, R.; AlSaad, R.; Aziz, S.; Ahmed, A.; Alsahli, M.; Damseh, R.; Sheikh, J. The Performance of Wearable AI in Detecting Stress Among Students: Systematic Review and Meta-Analysis. J. Med. Internet Res. 2024, 26, e52622. [Google Scholar] [CrossRef] [PubMed]
- Lisetti, C.L.; Nasoz, F. Using Noninvasive Wearable Computers to Recognize Human Emotions from Physiological Signals. EURASIP J. Adv. Signal Process. 2004, 2004, 929414. [Google Scholar] [CrossRef]
- Amin, O.B.; Mishra, V.; Tapera, T.M.; Volpe, R.; Sathyanarayana, A. Extending Stress Detection Reproducibility to Consumer Wearable Sensors. arXiv 2025, arXiv:2505.05694. [Google Scholar] [CrossRef]
- Li, R.; Liu, Z. Stress detection using deep neural networks. BMC Med Inform. Decis. Mak. 2020, 20, 285. [Google Scholar] [CrossRef] [PubMed]
- Hovsepian, K.; al’Absi, M.; Ertin, E.; Kamarck, T.; Nakajima, M.; Kumar, S. cStress: Towards a Gold Standard for Continuous Stress Assessment in the Mobile Environment. In Proceedings of the ACM International Conference on Ubiquitous Computing, Osaka, Japan, 7–11 September 2015; pp. 493–504. [Google Scholar] [CrossRef]
- Miranda-Correa, J.A.; Abadi, M.K.; Sebe, N.; Patras, I. AMIGOS: A Dataset for Affect, Personality and Mood Research on Individuals and Groups. IEEE Trans. Affect. Comput. 2021, 12, 479–493. [Google Scholar] [CrossRef]
- Amin, M.R.; Wickramasuriya, D.; Faghih, R.T. A Wearable Exam Stress Dataset for Predicting Cognitive Performance in Real-World Settings (Version 1.0.0), 26 May 2022. RRID:SCR-007345. Available online: https://physionet.org/content/wearable-exam-stress/1.0.0/ (accessed on 7 September 2025).
- Razavi, M.; McDonald, A.; Mehta, R.; Sasangohar, F. Evaluating Mental Stress Among College Students Using Heart Rate and Hand Acceleration Data Collected from Wearable Sensors. Hum. Comput. Interact. 2023, arXiv:2309.11097. [Google Scholar] [CrossRef]
Model | Exam | n_estimators | Max_Depth | Learning_Rate | C/reg | Kernel/Method |
---|---|---|---|---|---|---|
Random Forest | Midterm 1 | 100 | None | – | – | – |
Midterm 2 | 100 | None | – | – | – | |
Final | 100 | None | – | – | – | |
SVM | Midterm 1 | – | – | – | 10 | poly |
Midterm 2 | – | – | – | 0.1 | linear | |
Final | – | – | – | 10 | sigmoid | |
XGBoost | Midterm 1 | 100 | 3 | 0.01 | – | – |
Midterm 2 | 100 | 3 | 0.01 | – | – | |
Final | 100 | 3 | 0.01 | – | – | |
CatBoost | Midterm 1 | 100 | 3 | 0.1 | L2 = 1 | – |
Midterm 2 | 100 | 3 | 0.01 | L2 = 1 | – | |
Final | 300 | 3 | 0.2 | L2 = 5 | – | |
GBM | Midterm 1 | 100 | 3 | 0.01 | – | – |
Midterm 2 | 200 | 4 | 0.1 | – | – | |
Final | 200 | 3 | 0.1 | – | – |
Model | Exam | n_estimators | Max_Depth | Learning_Rate | C/reg | Kernel/Method |
---|---|---|---|---|---|---|
Random Forest | Midterm 1 | 100 | None | – | - | - |
Midterm 2 | 100 | None | – | - | - | |
Final | 100 | None | – | - | - | |
SVM | Midterm 1 | – | – | – | 10 | poly |
Midterm 2 | – | – | – | 0.1 | linear | |
Final | – | – | – | 10 | sigmoid | |
XGBoost | Midterm 1 | 100 | 3 | 0.01 | – | – |
Midterm 2 | 100 | 3 | 0.01 | – | – | |
Final | 100 | 3 | 0.01 | – | – | |
CatBoost | Midterm 1 | 100 | 7 | 0.2 | L2 = 3 | – |
Midterm 2 | 100 | 3 | 0.2 | L2 = 3 | – | |
Final | 300 | 5 | 0.2 | L2 = 3 | – | |
GBM | Midterm 1 | 100 | 3 | - | – | – |
Midterm 2 | 200 | 3 | - | – | – | |
Final | 200 | 3 | - | – | – |
Model | Exam | n_estimators | Max_Depth | Learning_Rate | C/reg | Kernel/Method |
---|---|---|---|---|---|---|
Random Forest | Midterm 1 | 100 | None | – | – | – |
Midterm 2 | 100 | None | – | – | – | |
Final | 100 | None | – | – | – | |
SVM | Midterm 1 | – | – | - | 0.01 | linear |
Midterm 2 | – | – | – | 0.01 | rbf | |
Final | – | – | – | 0.01 | poly | |
XGBoost | Midterm 1 | 100 | 3 | 0.01 | – | – |
Midterm 2 | 100 | 3 | 0.01 | – | – | |
Final | 100 | 3 | 0.01 | – | – | |
CatBoost | Midterm 1 | 100 | 3 | 0.2 | L2 = 3 | – |
Midterm 2 | 100 | 3 | 0.2 | L2 = 3 | – | |
Final | 300 | 5 | 0.2 | L2 = 3 | – | |
GBM | Midterm 1 | 100 | 3 | - | – | – |
Midterm 2 | 200 | 3 | - | – | – | |
Final | 200 | 3 | - | – | – |
Model | Technique | Metric | Midterm 1 | Midterm 2 | Final |
---|---|---|---|---|---|
Random Forest | Hyperparameters tuning | Accuracy | 0.65 | 0.76 | 0.59 |
F1 Score | 0.56 | 0.66 | 0.42 | ||
Precision | 0.56 | 0.58 | 0.36 | ||
Recall | 0.65 | 0.76 | 0.59 | ||
SMOTE | Accuracy | 0.47 | 0.62 | 0.51 | |
F1 Score | 0.46 | 0.51 | 0.38 | ||
Precision | 0.49 | 0.48 | 0.33 | ||
Recall | 0.47 | 0.62 | 0.51 | ||
ANOVA + Dimension reduction | Accuracy | 0.91 | 0.69 | 0.61 | |
F1 Score | 0.91 | 0.65 | 0.59 | ||
Precision | 0.94 | 0.62 | 0.58 | ||
Recall | 0.91 | 0.70 | 0.61 | ||
SVM | Hyperparameters tuning | Accuracy | 0.66 | 0.65 | 0.57 |
F1 Score | 0.51 | 0.51 | 0.40 | ||
Precision | 0.43 | 0.42 | 0.38 | ||
Recall | 0.66 | 0.65 | 0.60 | ||
SMOTE | Accuracy | 0.75 | 0.92 | 0.69 | |
F1 Score | 0.69 | 0.91 | 0.72 | ||
Precision | 0.69 | 0.94 | 0.83 | ||
Recall | 0.75 | 0.92 | 0.70 | ||
ANOVA + Dimension reduction | Accuracy | 1.00 | 0.58 | 0.55 | |
F1 Score | 1.00 | 0.52 | 0.42 | ||
Precision | 1.00 | 0.52 | 0.35 | ||
Recall | 1.00 | 0.58 | 0.58 | ||
XGBoost | Hyperparameters tuning | Accuracy | 0.63 | 0.62 | 0.59 |
F1 Score | 0.50 | 0.48 | 0.52 | ||
Precision | 0.41 | 0.41 | 0.46 | ||
Recall | 0.64 | 0.62 | 0.59 | ||
SMOTE | Accuracy | 0.92 | 0.50 | 0.52 | |
F1 Score | 0.91 | 0.53 | 0.55 | ||
Precision | 0.94 | 0.45 | 0.46 | ||
Recall | 0.92 | 0.55 | 0.59 | ||
ANOVA + Dimension reduction | Accuracy | 0.90 | 0.61 | 0.62 | |
F1 Score | 0.90 | 0.62 | 0.61 | ||
Precision | 0.94 | 0.59 | 0.60 | ||
Recall | 0.90 | 0.65 | 0.62 | ||
CatBoost | Hyperparameters tuning | Accuracy | 0.73 | 0.71 | 0.67 |
F1 Score | 0.68 | 0.59 | 0.54 | ||
Precision | 0.74 | 0.56 | 0.57 | ||
Recall | 0.73 | 0.64 | 0.62 | ||
SMOTE | Accuracy | 1.00 | 0.95 | 0.78 | |
F1 Score | 1.00 | 0.94 | 0.78 | ||
Precision | 1.00 | 0.98 | 0.89 | ||
Recall | 1.00 | 0.96 | 0.78 | ||
ANOVA + Dimension reduction | Accuracy | 0.94 | 0.79 | 0.72 | |
F1 Score | 0.94 | 0.72 | 0.71 | ||
Precision | 0.97 | 0.73 | 0.70 | ||
Recall | 0.93 | 0.79 | 0.72 | ||
GBM | Hyperparameters tuning | Accuracy | 0.65 | 0.75 | 0.70 |
F1 Score | 0.51 | 0.66 | 0.59 | ||
Precision | 0.42 | 0.59 | 0.50 | ||
Recall | 0.65 | 0.77 | 0.70 | ||
SMOTE | Accuracy | 0.54 | 0.70 | 0.62 | |
F1 Score | 0.52 | 0.64 | 0.54 | ||
Precision | 0.54 | 0.69 | 0.57 | ||
Recall | 0.54 | 0.69 | 0.62 | ||
ANOVA + Dimension reduction | Accuracy | 0.93 | 0.78 | 0.75 | |
F1 Score | 0.93 | 0.71 | 0.71 | ||
Precision | 0.97 | 0.72 | 0.72 | ||
Recall | 0.92 | 0.77 | 0.72 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Lalwani, S.; Ferdowsi, S. Predictive Modelling of Exam Outcomes Using Stress-Aware Learning from Wearable Biosignals. Sensors 2025, 25, 5628. https://doi.org/10.3390/s25185628
Lalwani S, Ferdowsi S. Predictive Modelling of Exam Outcomes Using Stress-Aware Learning from Wearable Biosignals. Sensors. 2025; 25(18):5628. https://doi.org/10.3390/s25185628
Chicago/Turabian StyleLalwani, Sham, and Saideh Ferdowsi. 2025. "Predictive Modelling of Exam Outcomes Using Stress-Aware Learning from Wearable Biosignals" Sensors 25, no. 18: 5628. https://doi.org/10.3390/s25185628
APA StyleLalwani, S., & Ferdowsi, S. (2025). Predictive Modelling of Exam Outcomes Using Stress-Aware Learning from Wearable Biosignals. Sensors, 25(18), 5628. https://doi.org/10.3390/s25185628