Comparative Analysis of Fuzzy Clustering Techniques on Students’ Self-Regulated Learning

Kistantia Elok Mumpuni, Samsul Hadi, Slamet Suyanto, Haryanto Haryanto, Yasir Sidiq, Guldana Atymtaevna Begimbetov

2025

Closed access Metadata (CSL-JSON)
DOI 10.1109/icet67257.2025.11290912
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

Self-Regulated Learning (SRL) is defined as student’s capability to manage their thoughts, emotions, and behaviors to achieve academic goals. In-depth-analysis of students’ SRL is crucial. However, common existing research on SRL has relied on descriptive analysis, which often provides insignificant description of learner differences. This study aimed to evaluate the efficacy of fuzzy clustering techniques on students’ SLR data. For this purpose, we utilized three fuzzy clustering algorithms namely Unsupervised Possibilistic Fuzzy C-Means (UPFC), Fuzzy C-Means (FCM), and Possibilistic Clustering Algorithm (PCA). The students’ SLR data were obtained through a questionnaire. We used a ten-question survey to find out three SRL dimensions from 300 students namely strategies, emotion regulation, and self-management. After the initial data screening, a multiple-start method was employed for clustering to ensure uniformity. The results showed that in case of spherical clustering structures, FCM is the best clustering option. While PCA or UPFC is better than FCM for non-spherical clustering. In addition, the findings showed that the internal validity indices indicated that the dataset was most accurately represented by two clusters. These results highlight fuzzy clustering can empower the analysis more than surface-level descriptions, providing insights that can guide the development of more flexible teaching methods and data-driven educational policies. Investigation on the application of fuzzy clustering across various educational settings with larger datasets is necessary for future study.