Using Stack Modelling Technique
for Student Performance Prediction:
An Educational Data Mining Approach

Educational Data Mining (EDM) and Exploratory Data Analysis (EDA) collaboratively enhance the quality of learning outcomes. Academic institutions strive to establish admission criteria that enable the selection of high-performing and deserving students. In recent years, EDM has garnered significant attention in research, facilitating the prediction of student performance and the identification of systemic inefficiencies. This study investigates the application of stacked ensemble models for student performance prediction and evaluates their performance against baseline classification models, including Decision Trees, Random Forests, Naïve Bayes, and Support Vector Machines (SVM). The findings demonstrate that the stacked model surpasses the baseline models in predictive accuracy and reliability. However, the study also reveals that the achieved prediction accuracies remain suboptimal. It highlights the need for incorporating additional parameters, alongside prior academic records, to improve predictive performance and establish a more robust merit-based admission criterion. The analysis underscores that the dataset utilized for predicting student academic performance, based on their prior or ongoing academic standing, does not yield promising outcomes. No significant correlation was observed between the variables derived from preliminary academic records (NTS scores, intermediate percentage, and matriculation percentage) and the predicted CGPAs. This indicates that the dataset exhibits stochastic behavior, suggesting that the factors currently employed in admission decision-making lack predictive utility for forecasting student performance. Consequently, the findings highlight the need for policymakers to reevaluate the admission criteria, as the existing parameters are insufficient for reliably predicting academic outcomes using historical data.

EDM, EDA, Stochastic, Demographic, Linear Regression, Decision Trees, Random Forest, Naive Bayes, SVM

1. Bhusal, A. (2021). Predicting Student’s Performance Through Data Mining. arXiv preprint arXiv:2112.01247.

2. Wakelam, E., Jefferies, A., Davey, N. and Sun, Y. (2020), The potential for student performance prediction in small cohorts with minimal available attributes. Br J Educ Technol, 51: 347-370. https://doi.org/10.1111/bjet.12836

3. Boran Sekeroglu, Kamil Dimililer, and Kubra Tuncal. 2019. Student Performance Prediction and Classification Using Machine Learning Algorithms. In Proceedings of the 2019 8th International Conference on Educational and Information Technology (ICEIT 2019). Association for Computing Machinery, New York, NY, USA, 7–11. https://doi.org/10.1145/3318396.3318419.

4. Aslam, N., Khan, I. U., Alamri, L. H., and Almuslim, R. S. (2021). An Improved Early Student’s Academic Performance Prediction Using Deep Learning. International Journal of Emerging Technologies in Learning (iJET), 16(12), pp.108–122. https://doi.org/10.3991/ijet.v16i12.20699.

5. Ashfaq, Usman and Poolan Marikannan, Booma and Raheem, Mafas. (2020). Managing Student Performance: Predictive Analytics using Imbalanced Data. International Journal of Recent Technology and Engineering. 8.2277- 2283. 10.35940/ijrteE7008.038620.

6. Science, International Journal of Scientific Research in, and Technology IJSRST. Student’s Performance Analysis with EDA and Machine Learning Models. International Journal of Scientific Research in Science and Technology, 2021. doi:10.32628/IJSRST218448.

7. Harikumar Pallathadka, Alex Wenda, Edwin RamirezAs´ıs, Maximiliano As´ıs-Lopez, Judith Flores-Albornoz, Khongdet Phasinam, ´ Classification and prediction of student performance data using various machine learning algorithms, Materials Today: Proceedings, Volume 80, Part 3, 2023, Pages 3782-3785, ISSN 2214-7853, https://doi.org/10.1016/j.matpr.2021.07.382.

8. Rodr´ıguez-Muniz LJ, Bernardo AB, Esteban M, D ˜ ´ıaz I. Dropout and transfer paths: What are the risky profiles when analyzing university persistence with machine learning techniques? PLoS One. 2019 Jun 21;14(6):e0218796. doi: 10.1371/journal.pone.0218796. PMID: 31226158; PMCID: PMC6588340.

9. V. Hegde and P. P. Prageeth, Higher education student dropout prediction and analysis through educational data mining, 2018 2nd International Conference on Inventive Systems and Control (ICISC), Coimbatore, India, 2018, pp. 694-699, doi: 10.1109/ICISC.2018.8398887.

10. M, Nirmala and Selvi, T and Saravanan, V. (2021). Student Aca demic Performance Prediction under Various Machine Learning Classification Algorithms. International Journal of Applied Science and Engineering. 9. 221. 10.22214/ijraset.2021.38786.

...