Predicting Depression Among University Students Using Machine Learning: A Comparative Study with Deployment Framework

Abstract

This study presents a machine learning-based framework for predicting depression among university students using academic, behavioral, and psychological variables. The proposed approach integrates comprehensive data preprocessing, exploratory data analysis, and the evaluation of multiple classification algorithms, including Logistic Regression, Random Forest, Support Vector Machine, and several ensemble methods. Among the evaluated models, AdaBoost achieved the best overall performance, with an accuracy of 84.3% and a ROC-AUC score of 0.91, demonstrating its effectiveness in capturing complex relationships within the dataset. The study also identifies key predictors of depression, such as academic pressure, satisfaction level, and suicidal thoughts, which are consistent with existing psychological research. In addition to model development, a web-based application was implemented using a Flask backend and React frontend to enable real-time depression risk prediction. This integration highlights the practical applicability of the proposed system as a decision-support tool for early detection and intervention. The findings indicate that ensemble learning methods, combined with user-friendly deployment, can significantly contribute to improving mental health monitoring systems in educational environments.

Keywords

• depression prediction
• machine learning
• adaboost
• student mental health
• classification

Ethical Statement

This study was conducted using a publicly available dataset and does not involve direct interaction with human or animal subjects. The data utilized in this research were obtained from an open-access source and do not contain personally identifiable information. Therefore, in accordance with ethical research guidelines and data protection regulations, this study does not require ethics committee approval. “This article does not require ethics committee approval.” All procedures performed in this study comply with general ethical standards in scientific research. The study does not include any clinical trials, experimental interventions, or data collection processes involving human participants, such as surveys, interviews, or observations conducted by the researchers. Regarding conflicts of interest, the authors declare that there are no financial, institutional, or personal relationships that could influence the results or interpretation of the study. “This article has no conflicts of interest with any individual or institution.”

Üniversite Öğrencilerinde Depresyonun Makine Öğrenmesi Kullanılarak Tahmin Edilmesi: Dağıtım Çerçevesi ile Karşılaştırmalı Bir Çalışma

Abstract

Bu çalışma, üniversite öğrencileri arasında depresyonu akademik, davranışsal ve psikolojik değişkenleri kullanarak tahmin etmek için makine öğrenmesi tabanlı bir çerçeve sunmaktadır. Önerilen yaklaşım; kapsamlı veri ön işleme, keşifsel veri analizi ve Lojistik Regresyon, Rastgele Orman, Destek Vektör Makineleri ve çeşitli topluluk (ensemble) yöntemleri dahil olmak üzere birden fazla sınıflandırma algoritmasının değerlendirilmesini içermektedir. Değerlendirilen modeller arasında AdaBoost, %84,3 doğruluk oranı ve 0,91 ROC-AUC skoru ile en iyi genel performansı göstermiş, veri seti içindeki karmaşık ilişkileri yakalamadaki etkinliğini ortaya koymuştur. Çalışma ayrıca akademik baskı, memnuniyet düzeyi ve intihar düşünceleri gibi depresyonun temel belirleyicilerini tanımlamış olup, bu bulgular mevcut psikolojik araştırmalarla tutarlıdır. Model geliştirmeye ek olarak, gerçek zamanlı depresyon riski tahmini yapabilmek için Flask tabanlı bir arka uç (backend) ve React tabanlı bir ön uç (frontend) kullanılarak web tabanlı bir uygulama geliştirilmiştir. Bu entegrasyon, önerilen sistemin erken tespit ve müdahale için bir karar destek aracı olarak pratik uygulanabilirliğini göstermektedir. Bulgular, kullanıcı dostu dağıtım ile birleştirilen topluluk öğrenme yöntemlerinin, eğitim ortamlarında ruh sağlığı izleme sistemlerinin geliştirilmesine önemli katkılar sağlayabileceğini göstermektedir.

Keywords

• depresyon tahmini
• makine öğrenmesi
• uyarlamalı arttırma algoritması
• öğrenci ruh sağlığı
• sınıflandırma

Ethical Statement

Bu çalışma, kamuya açık bir veri seti kullanılarak gerçekleştirilmiş olup insan veya hayvan deneklerle doğrudan etkileşim içermemektedir. Bu araştırmada kullanılan veriler açık erişimli bir kaynaktan elde edilmiş ve kişisel olarak tanımlanabilir herhangi bir bilgi içermemektedir. Bu nedenle, etik araştırma ilkeleri ve veri koruma düzenlemeleri doğrultusunda, bu çalışma için etik kurul onayı gerekmemektedir. “Bu makale etik kurul onayı gerektirmemektedir.” Bu çalışmada gerçekleştirilen tüm işlemler, bilimsel araştırmalarda geçerli genel etik standartlara uygundur. Çalışma; klinik deneyler, deneysel müdahaleler veya araştırmacılar tarafından yürütülen anket, görüşme ya da gözlem gibi insan katılımcıları içeren veri toplama süreçlerini kapsamamaktadır. Çıkar çatışmaları açısından, yazarlar çalışmanın sonuçlarını veya yorumlanmasını etkileyebilecek herhangi bir finansal, kurumsal ya da kişisel ilişkinin bulunmadığını beyan etmektedir. “Bu makalenin herhangi bir kişi veya kurum ile çıkar çatışması bulunmamaktadır.”

References

1. A. O. Hassan, I. M. Jamal, S. D. Ahmed, and A. U. Abdullahi, “Predicting student depression using machine learning: a comparative analysis of machine learning algorithms for early depression detection in students,” Aitu Scientific Research Journal, vol. 4, no. 1, pp. 28–35, Jan. 2025, doi: 10.63094/AITUSRJ.25.4.1.4.
2. R. Qasrawi, S. P. V. Polo, D. A. Al-Halawa, S. Hallaq, and Z. Abdeen, “Assessment and prediction of depression and anxiety risk factors in schoolchildren: machine learning techniques performance analysis,” JMIR Formative Research, vol. 6, no. 8, e32736, Aug. 2022, doi: 10.2196/32736.
3. X. Zhao, Y. Wang, J. Li, W. Liu, Y. Yang, Y. Qiao, J. Liao, M. Chen, D. Li, B. Wu, D. Huang, and D. Wu, “A machine-learning-derived online prediction model for depression risk in COPD patients: A retrospective cohort study from CHARLS,” Journal of Affective Disorders, vol. 377, pp. 284–293, 2025, doi: 10.1016/j.jad.2025.02.063.
4. Y. Zhong, J. He, J. Luo, J. Zhao, Y. Cen, Y. Song, Y. Wu, C. Lin, L. Pan, and J. Luo, “A machine learning algorithm-based model for predicting the risk of non-suicidal self-injury among adolescents in western China: a multicentre cross-sectional study,” Journal of Affective Disorders, vol. 345, pp. 369–377, 2024, doi: 10.1016/j.jad.2023.10.110.
5. S. S. Sara, M. A. Rahman, R. Rahman, and A. Talukder, “Prediction of suicidal ideation with associated risk factors among university students in the southern part of Bangladesh: machine learning approach,” Journal of Affective Disorders, vol. 349, pp. 502–508, 2024, doi: 10.1016/j.jad.2024.01.092.
6. Q. Li, K. Song, T. Feng, J. Zhang, and X. Fang, “Machine learning identifies different related factors associated with depression and suicidal ideation in Chinese children and adolescents,” Journal of Affective Disorders, vol. 361, pp. 24–35, 2024, doi: 10.1016/j.jad.2024.06.006.
7. Y. Kuang, X. Liao, Z. Jiang, Y. Gu, B. Liu, C. Tan, W. Zhang, and K. Li, “Federated learning-based prediction of depression among adolescents across multiple districts in China,” Journal of Affective Disorders, vol. 369, pp. 625–632, 2025, doi: 10.1016/j.jad.2024.10.027.
8. L. Li, D. Guo, C. Shi, and Y. Zheng, “The predictive role of sedentary behavior and physical activity on adolescent depressive symptoms: a machine learning approach,” Journal of Affective Disorders, vol. 378, pp. 81–89, 2025, doi: 10.1016/j.jad.2025.02.085.

9. B. Niu, M. Wan, and Y. Zhou, “Development of an explainable machine learning model for predicting depression in adolescent girls with non-suicidal self-injury: A cross-sectional multicenter study,” Journal of Affective Disorders, vol. 379, pp. 690–702, 2025, doi: 10.1016/j.jad.2025.03.080.
10. G. H. Al Masud, R. I. Shanto, I. Sakin, and M. R. Kabir, “Effective depression detection and interpretation: Integrating machine learning, deep learning, language models, and explainable AI,” Array, vol. 25, p. 100375, 2025, doi: 10.1016/j.array.2025.100375.
11. F. M. Basysyar, G. Dwilestari, and A. I. Purnamasari, “Analysis student emotions and mental health on cumulative GPA using machine learning and SMOTE,” JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer), vol. 10, no. 2, pp. 361–368, 2024, doi: 10.33480/jitk.v10i2.5967.
12. M. T. Aziz, T. Mahmud, M. F. B. A. Aziz, M. A. B. Siddick, M. M. Sharif, M. S. Hossain, and K. Andersson, “Textual and numerical data fusion for depression detection: A machine learning framework,” Indonesian Journal of Electrical Engineering and Computer Science, vol. 38, no. 2, pp. 1231–1244, 2025, doi: 10.11591/ijeecs.v38.i2.pp1231-1244.
13. R. Krishnan, S. Kumari, A. A. Badi, S. Jeba, and M. James, “Predictive machine learning model for mental health issues in higher education students due to COVID-19 using HADS assessment,” Arab Gulf Journal of Scientific Research, vol. 42, no. 4, pp. 1472–1490, 2024, doi: 10.1108/AGJSR-01-2023-0034.
14. T. Qiu, S. Wang, D. Hu, N. Feng, and L. Cui, “Predicting risk of bullying victimization among primary and secondary school students: based on a machine learning model,” Behavioral Sciences, vol. 14, no. 1, p. 73, 2024, doi: 10.3390/bs14010073.
15. G. Tyulepberdinova, M. Mansurova, T. Sarsembayeva, S. Issabayeva, and D. Issabayeva, “The physical, social, and mental conditions of machine learning in student health evaluation,” Journal of Computer Assisted Learning, vol. 40, no. 5, pp. 2020–2030, 2024, doi: 10.1111/jcal.12999.
16. M. H. Amirhosseini, A. L. Ayodele, and A. Karami, “Prediction of depression severity and personalised risk factors using machine learning on multimodal data,” in 2024 IEEE 12th International Conference on Intelligent Systems (IS), pp. 1–7, 2024, doi: 10.1109/IS61756.2024.10705185.
17. M. H. Amirhosseini, A. L. Ayodele, and A. Karami, “Prediction of depression severity and personalised risk factors using machine learning on multimodal data,” in 2024 IEEE 12th International Conference on Intelligent Systems (IS), pp. 1–7, 2024, doi: 10.1109/IS61756.2024.10705185.
18. B. L. Schaab, P. Ü. Calvetti, S. Hoffmann, G. B. Diaz, M. Rech, S. C. Cazella, A. T. Stein, H. M. T. Barros, P. C. Silva, and C. T. Reppold, “How do machine learning models perform in the detection of depression, anxiety, and stress among undergraduate students? A systematic review,” Cadernos de Saúde Pública, vol. 40, 2024, doi: 10.1590/0102-311XEN029323.
19. J. R. Beames, J. Han, A. Shvetcov, W. Y. Zheng, A. Slade, O. Dabash, J. O'Dea Rosenberg, B. Kasturia, L. Hoon, A. E. Whitton, H. Christensen, and J. M. Newby, “Use of smartphone sensor data in detecting and predicting depression and anxiety in young people (12–25 years): A scoping review,” Heliyon, vol. 10, no. 15, 2024, doi: 10.1016/j.heliyon.2024.e35472.
20. J. H. Rony, M. M. Syeed, R. H. Khan, K. Fatema, M. S. Hossain, and M. F. Uddin, “Predicting depression among university students: a comparative assessment of ML & DL models using XAI,” in 2024 23rd International Symposium on Communications and Information Technologies (ISCIT), pp. 175–180, 2024, doi: 10.1109/ISCIT63075.2024.10793544.
21. T. R. Shaha, M. Begum, J. Uddin, V. Y. Torres, J. A. Iturriaga, I. Ashraf, and M. A. Samad, “Feature group partitioning: An approach for depression severity prediction with class balancing using machine learning algorithms,” BMC Medical Research Methodology, vol. 24, no. 1, p. 123, 2024, doi: 10.1186/s12874-024-02249-8.
22. V. N. Dang, A. Cascarano, R. H. Mulder, C. Cecil, M. A. Zuluaga, J. Hernández-González, and K. Lekadir, “Fairness and bias correction in machine learning for depression prediction across four study populations,” Scientific Reports, vol. 14, no. 1, p. 7848, 2024, doi: 10.1038/s41598-024-58427-7.
23. A. H. Chowdhury, D. Rad, and M. S. Rahman, “Predicting anxiety, depression, and insomnia among Bangladeshi university students using tree-based machine learning models,” Health Science Reports, vol. 7, no. 4, e2037, 2024, doi: 10.1002/hsr2.2037.
24. A. R. Panda, S. Roy, M. K. Mishra, S. Sahoo, S. Mohapatra, and M. K. Gaurisaria, “Hyperparameter tuning of machine learning models: Analysing student’s depression data,” in 2025 International Conference on Emerging Systems and Intelligent Computing (ESIC), pp. 907–912, 2025, doi: 10.1109/ESIC64052.2025.10962651.
25. Y. Zhai, Y. Zhang, Z. Chu, B. Geng, M. Almaawali, R. Fulmer, Y. W. D. Lin, Z. Xu, A. D. Daniels, Y. Liu, Q. Chen, and X. Du, “Machine learning predictive models to guide prevention and intervention allocation for anxiety and depressive disorders among college students,” Journal of Counseling & Development, vol. 103, no. 1, pp. 110–125, 2025, doi: 10.1002/jcad.12543.
26. M. A. Rahman and T. Kohli, “Mental health analysis of international students using machine learning techniques,” PLOS ONE, vol. 19, no. 6, e0304132, 2024, doi: 10.1371/journal.pone.0304132.
27. A. Almadhor, S. Abbas, G. A. Sampedro, S. Alsubai, S. Ojo, A. Al Hejaili, and L. Strazovska, “Multi-class adaptive active learning for predicting student anxiety,” IEEE Access, vol. 12, pp. 58097–58105, 2024, doi: 10.1109/ACCESS.2024.3391418.
28. Y. Hong, X. Rong, and W. Liu, “Construction of influencing factor segmentation and intelligent prediction model of college students' cell phone addiction model based on machine learning algorithm,” Heliyon, vol. 10, no. 8, 2024, doi: 10.1016/j.heliyon.2024.e29245.
29. T. Vu, R. Dawadi, M. Yamamoto, J. T. Tay, N. Watanabe, Y. Kuriya, A. Oya, P. N. H. T. Tran, and M. Araki, “Prediction of depressive disorder using machine learning approaches: findings from the NHANES,” BMC Medical Informatics and Decision Making, vol. 25, no. 1, p. 83, 2025, doi: 10.1186/s12911-025-02903-1.
30. V. N. Dang, A. Cascarano, R. H. Mulder, C. Cecil, M. A. Zuluaga, J. Hernández-González, and K. Lekadir, “Fairness and bias correction in machine learning for depression prediction across four study populations,” Scientific Reports, vol. 14, no. 1, p. 7848, 2024, doi: 10.1038/s41598-024-58427-7.
31. A. Shamim, Student Depression Dataset, Kaggle, 2024. [Online]. Available: https://www.kaggle.com/datasets/adilshamim8/student-depression-dataset