Thu, Jul 2, 2026
[Archive]
Volume 7, Issue 2 (JUNE ISSUE 2026)
johepal 2026, 7(2): 92-110 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Shin Y, Kang H S, Park S Y. (2026). Prediction of Admission Decisions Using Machine Learning Models: An Analysis of the Holistic Undergraduate Admissions Review Process in Korea. johepal. 7(2), 92-110. doi:10.66224/johepal.7.2.92
URL: http://johepal.com/article-1-1768-en.html
URL: http://johepal.com/article-1-1768-en.html
Abstract: (12 Views)
This study aimed to examine and validate the consistency and predictive patterns of human-led undergraduate admissions decisions through the application of machine learning models. Unlike traditional holistic evaluation processes conducted by human assessors, this study compared five machine learning algorithms − Gradient Boosting, Random Forest, Support Vector Machine, Logistic Regression, and XGBoost − to identify the most accurate prediction model. The analysis utilized a dataset of 1,554 application records from the 2024 application cycle. To further improve prediction accuracy, Latent Dirichlet Allocation (LDA) was utilized to extract relevant features from unstructured textual data. The findings revealed that the XGBoost model performed best in predicting admission outcomes. This result is attributed to the learning mechanisms of tree-based ensemble models, which is capable of capturing the complex interactions between non-linear score patterns and various others variables. Major factors influencing admission decisions encompassed interview scores, type of application, and document evaluation scores, highlighting their significance in the selection process and validating the effectiveness of the XGBoost as a supportive tool. These findings not only provide practical recommendations for improving prediction accuracy but also inform future research directions in data-driven strategies for high-stakes educational assessment.
Keywords: Holistic Undergraduate Admissions Review Process, Latent Dirichlet Allocation (LDA), XGBoost, Machine Learning, Educational Data Mining, Prediction Accuracy
Type of Study: Research |
Subject:
Special
Received: 2026/02/19 | Accepted: 2026/06/7 | Published: 2026/06/30
Received: 2026/02/19 | Accepted: 2026/06/7 | Published: 2026/06/30
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
