Abstract
Background: Large-scale professional e-learning programmes generate assessment data that can reveal variation in competency development and performance, but aggregate results may obscure heterogeneous learning needs.
Objectives: This study identifies exploratory learner profiles in a large-scale financial-reporting e-learning programme for Indonesia’s Satuan Pelayanan Pemenuhan Gizi (SPPG) to inform differentiated learning-support recommendations.
Methods: The study analysed 20,192 participant records using six standardized features. K-Means, Ward hierarchical clustering, and Louvain community detection were benchmarked on a common sample of 3,000 observations. K-Means solutions with two to ten clusters were then evaluated using internal validity indices, initialization stability, and sensitivity analyses, and the profiles were interpreted through Transformative Learning Theory.
Results: The two-cluster solution achieved the highest Silhouette coefficient (0.199) and high initialization stability (mean pairwise adjusted Rand index [ARI] = 0.999). The five-cluster solution provided more granular profiles but lower stability (mean pairwise ARI = 0.814). Supplementary checks did not independently establish an optimal number of clusters.
Conclusions: K-Means provides a practical exploratory approach for describing variation in large-scale professional e-learning. The two-cluster solution is comparatively stable, while the five-cluster solution adds less robust detail. The profiles may inform differentiated learning support but are not validated learner types.
First Page
236
Last Page
260
Recommended Citation
Agus, P.,
Riyanto, R.,
&
Zaky, A.
(2026).
K-Means Clustering for Learner Profiling in Large-Scale Nutrition Fulfillment Service Unit (SPPG) E-Learning: A Transformative Learning Perspective.
Jurnal Pendidikan dan Pembelajaran, 33(2), 236-260.
DOI: https://doi.org/https://doi.org/10.17977/2580-2313.1172
Available at:
https://citeus.um.ac.id/jpp/vol33/iss2/6
file paper
K-Means Clustering for Learner Profiling in Large-Scale Nutrition Fulfillment Service Unit (SPPG) E-Learning A Transformative Learning Perspective.pdf (526 kB)
file paper
Response_to_Reviewers_FINAL.pdf (81 kB)
response
K-Means Clustering for Learner Profiling in Large-Scale Nutrition Fulfillment Service Unit (SPPG) E-Learning A Transformative Learning Perspective.docx (855 kB)
file paper
Included in
Curriculum and Instruction Commons, Educational Methods Commons, Educational Technology Commons, Higher Education Commons