Student Learning Style Segmentation Using K-Means Clustering for Adaptive Learning in Higher Education

Abstract

Understanding students’ learning styles is crucial in developing adaptive instructional strategies that enhance engagement and academic achievement in higher education. This study aims to classify university students based on their dominant learning styles—visual, auditory, and kinesthetic—using the K-Means clustering algorithm. A total of 150 respondents participated in a questionnaire designed with Likert-scale items measuring learning preferences. The clustering process successfully segmented the students into three well-defined groups: visual learners (38.7%), auditory learners (31.3%), and kinesthetic learners (30%). Cluster validation using the Silhouette Score yielded an average value of 0.62, indicating a strong internal consistency and effective group separation. Further analysis revealed that 75% of students in the visual cluster showed higher comprehension using visual materials such as diagrams and videos. In the auditory cluster, 70% preferred oral explanations, discussions, and lectures, while 65% of kinesthetic learners performed better through hands-on activities and physical interaction with learning materials. These insights emphasize the importance of integrating personalized approaches into course design to support diverse learning needs. The findings suggest that adaptive learning models based on clustering techniques like K-Means can assist educators in tailoring instruction, thereby improving the overall learning experience. This approach supports more inclusive and effective pedagogy in higher education environments.