Student Performance Predictions for Advanced Engineering Mathematics Course With New Multivariate Copula Models

Nguyen-Huy, Thong ORCID: https://orcid.org/0000-0002-2201-6666 and Deo, Ravinesh C. ORCID: https://orcid.org/0000-0002-2290-6749 and Khan, Shahjahan ORCID: https://orcid.org/0000-0002-0446-086X and Devi, Aruna and Adeyinka, Adewuyi Ayodele ORCID: https://orcid.org/0000-0001-9668-220X and Apan, Armando A. ORCID: https://orcid.org/0000-0002-5412-8881 and Yaseen, Zaher Mundher (2022) Student Performance Predictions for Advanced Engineering Mathematics Course With New Multivariate Copula Models. IEEE Access, 10. 45112 -45136.

[img]
Preview
Text (Published Version)
Student_Performance_Predictions_for_Advanced_Engineering_Mathematics_Course_with_New_Multivariate_Copula_Models.pdf
Available under License Creative Commons Attribution 4.0.

Download (4MB) | Preview

Abstract

Engineering Mathematics requires that problem-solving should be implemented through ongoing assessments; hence the prediction of student performance using continuous assessments remains an important task for engineering educators, mainly to monitor and improve their teaching practice. This paper develops probabilistic models to predict weighted scores ( WS , or the overall mark leading to a final grade) for face-to-face (on-campus) and web-based (online) Advanced Engineering Mathematics students at an Australian regional university over a 6-year period (2013–2018). We fitted parametric and non-parametric D-vine copula models utilizing multiple quizzes, assignments and examination score results to construct and validate the predicted WS in independently test datasets. The results are interpreted in terms of the probability of whether a student’s continuous performance (i.e., individually or jointly with other counterpart assessments) is likely to lead to a passing grade conditional upon joint performance in students’ quizzes and assignment scores. The results indicate that the newly developed D-vine model, benchmarked against a linear regression model, can generate accurate grade predictions, and particularly handle the problem of low or high scores (tail dependence) compared with a conventional model for both face-to-face, and web-based students. Accordingly, the findings advocate the practical utility of joint copula models that capture the dependence structure in engineering mathematics students’ marks achieved. This therefore, provide insights through learning analytic methods to support an engineering educator’s teaching decisions. The implications are on better supporting engineering mathematics students’ success and retention, developing evidence-based strategies consistent with engineering graduate requirements through improved teaching and learning, and identifying/addressing the risk of failure through early intervention. The proposed methods can guide an engineering educator’s practice by investigating joint influences of engineering problem-solving assessments on their student’s grades.


Statistics for USQ ePrint 48383
Statistics for this ePrint Item
Item Type: Article (Commonwealth Reporting Category C)
Refereed: Yes
Item Status: Live Archive
Faculty/School / Institute/Centre: Current - Institute for Life Sciences and the Environment - Future Drought Fund Hub (1 Jun 2021 - )
Faculty/School / Institute/Centre: Current - Institute for Life Sciences and the Environment - Centre for Applied Climate Sciences (1 Aug 2018 -)
Date Deposited: 10 May 2022 06:27
Last Modified: 10 May 2022 06:27
Uncontrolled Keywords: Engineering mathematics performance prediction, D-vine copula, multivariate probability model, academic performance, education decision-making, statistical model
Fields of Research (2020): 39 EDUCATION > 3904 Specialist studies in education > 390402 Education assessment and evaluation
39 EDUCATION > 3999 Other Education > 399999 Other education not elsewhere classified
49 MATHEMATICAL SCIENCES > 4905 Statistics > 490501 Applied statistics
46 INFORMATION AND COMPUTING SCIENCES > 4601 Applied computing > 460105 Applications in social sciences and education
Identification Number or DOI: https://doi.org/10.1109/ACCESS.2022.3168322
URI: http://eprints.usq.edu.au/id/eprint/48383

Actions (login required)

View Item Archive Repository Staff Only