Citation

Hashim, Nik Mohd Zarifie and Mohd Sam, Mohd Fazli and Abd Rahman, Noor Ziela and Wan Nahar, Wan Nur Azah (2026) Interpretable Intelligence: Benchmarking Traditional Statistics and Machine Learning for Engineering Course’s Performance Evaluation. Studies in Big Data, 192. pp. 521-530. ISSN 2197-6503

Text Interpretable Intelligence_ Benchmarking Traditional Statistics and Machine Learning for Engineering Course’s Performance Evaluation _ Springer Nature Link.pdf - Published Version Restricted to Repository staff only Download (660kB)

Abstract

The most significant problem in the age of data-driven education is the method of assessing student performance. This issue is especially vital for academic settings where students are separated into small groups. The current study presents a comparison of several conventional statistical techniques and machine-learning models involving linear regression, decision trees, random forests in measuring Test, Assignment, Final exam, and TotalScore for 112 students. Statistically, R-score, t-test, and chi-square analysis show strong linear patterns and group effects, while the final exam score has the highest predictive power of TotalScore, which is r = 0.95, p < 0.00001. In addition, Machine-learning models, including linear regression, decision trees, and random forests, are performed. Classification success in the decision tree model is 82.6%, and the average success in cross-validation is 87%, which remained stable. Consolidation and importance of functions from feature importance analysis enabled us to identify the most crucial feature. Finally, K-Means clustering allows us to reveal patterns of consolidation between students that accurately indicate groups. The comparative results provide the specific explanatory analytics validation power of conventional methods on the one hand, while on the other, prediction and genetic strength, power generalization ability, machine learning-based models. The study can aid researchers and teachers who want to utilize data-driven approaches to assess academic performance.

Item Type:	Article
Uncontrolled Keywords:	Machine learning in education
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Business (FOB) Faculty of Engineering and Technology (FET)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	30 Oct 2026 00:50
Last Modified:	30 Oct 2026 00:50
URII:	http://shdl.mmu.edu.my/id/eprint/17013

Downloads

Edit (login required)