International Journal of Artificial Intelligence and Education Technology
2835-2432ISSN (Online)
Volume 4 , Issue 2 , PP: 01–10, 2025 | Cite this article as | XML | Html | PDF | Full Length Article
Jehad Mousa 1 * , Abdallah Salama 2
Early identification of students at risk of academic failure is a persistent challenge in educational technology, with direct implications for student retention, institutional equity, and the allocation of support resources. Although supervised machine learning has been widely applied to student outcome prediction, the relative merit of competing algorithm classes and the degree to which demographic and behavioural features contribute predictive power beyond prior academic assessments remain incompletely resolved in the secondary school context. This paper presents a structured comparative evaluation of five supervised classifiers trained on a rich combination of periodic grades, attendance records, sociodemographic characteristics, and lifestyle indicators drawn from secondary school students. A dual importance analysis— combining impurity-based measures with held-out permutation importance—disentangles the distinct predictive roles of grade trajectories, absenteeism, parental background, and lifestyle variables. Ensemble methods demonstrate consistent superiority across all evaluation criteria, with prior periodic assessments and attendance emerging as the dominant predictors. Parental education level introduces a socioeconomic gradient that operates independently of student controlled factors, generating structural inequities that standard grade-monitoring systems are unlikely to address. These findings provide both a methodological benchmark for secondary school prediction tasks and practical guidance for institutions designing equitable and evidence-based early warning interventions.
Educational data mining , Machine learning , Student outcome prediction , Ensemble methods , Learning analytics , Secondary education , Early warning systems
[1] M. Adnan, A. Habib, J. Ashraf, S. Mussadiq, A. A. Raza, M. Abid, M. Bashir, and S. U. Khan. Predicting at risk students at different percentages of course length for early intervention using machine learning models. IEEE Access, 9:7519–7539, 2021. doi: 10.1109/ACCESS.2021.3049446.
[2] S. A. Albriki Balabied and H. F. Eid. Utilizing random forest algorithm for early detection of academic underperformance in open learning environments. PeerJ Computer Science, 9:e1708, 2023. doi: 10.7717/peerj-cs.1708.
[3] L. Breiman. Random forests. Machine Learning, 45(1):5–32, 2001. doi: 10.1023/A:1010933404324.
[4] P. Cortez and A. M. G. Silva. Using data mining to predict secondary school student performance. In A. Brito and J. Teixeira, editors, Proceedings of the 5th Future Business Technology Conference (FUBUTEC 2008), pages 5–12, Porto, Portugal, 2008. EUROSIS. ISBN 978-9077381-39-7.
[5] J. H. Friedman. Greedy function approximation: A gradient boosting machine. Annals of Statistics, 29(5):1189–1232, 2001. doi: 10.1214/aos/1013203451.
[6] F. Hlioui, N. Aloui, and F. Gargouri. A withdrawal prediction model of at-risk learners based on behavioural indicators. International Journal of Web-Based Learning and Teaching Technologies, 16(2):32–53, 2021. doi: 10.4018/IJWLTT.2021030103.
[7] B. Holicza and A. Kiss. Predicting and comparing students’ online and offline academic performance using machine learning algorithms. Behavioral Sciences, 13(4):289, 2023. doi: 10.3390/bs13040289.
[8] J. Kuzilek, M. Hlosta, and Z. Zdrahal. Open University Learning Analytics dataset. Scientific Data, 4:170171, 2017. doi: 10.1038/sdata.2017.171.
[9] V. Realinho, J. Machado, L. Baptista, and M. V. Martins. Predicting student dropout and academic success. Data, 7 (11):146, 2022. doi: 10.3390/data7110146.
[10] H. Waheed, S.-U. Hassan, R. Nawaz, N. R. Aljohani, G. Chen, and D. Gasevic. Early prediction of learners at risk in self-paced education: A neural network approach. Expert Systems with Applications, 213:118868, 2023. doi: 10.1016/j.eswa.2022.118868.
