<![CDATA[Identifying student learning-style profiles is important for designing more adaptive instruction. However, prior studies on learning styles have mostly remained descriptive or correlational, while clustering-based mapping that integrates the VARK and Honey-Mumford models is still limited, particularly for Mathematics Study Program students. This study aimed to map student learning-style profiles using the Fuzzy C-Means (FCM) algorithm based on eight dimensions: Visual, Auditory, Read/Write, Kinesthetic, Activist, Reflector, Theorist, and Pragmatist. This study employed a descriptive quantitative approach involving 168 students of the Mathematics Study Program, FMIPA Universitas Negeri Medan. Data were collected through an online questionnaire; scores on the eight dimensions were averaged and combined as the input attributes for clustering. Cluster validity was evaluated using the Partition Coefficient (PC), Partition Entropy (PE), and Modified Partition Coefficient (MPC). The validity indices indicated that the two-cluster solution produced the best numerical values; however, the three-cluster solution was retained because it yielded more interpretable and less redundant profiles. For the three-cluster model, the values obtained were PC = 0.5222, PE = 0.8147, and MPC = 0.2834. The clustering results produced three profiles: Adaptive Multimodal (35.1%), Passive Kinesthetic (24.4%), and Practical Auditory (40.5%). These findings indicate that most students tend to learn more effectively through a combination of listening and direct practice.]]>
Copyrights © 2026
