<![CDATA[Identifying student learning characteristics is essential for supporting evidence-based academic evaluation and improving educational quality in higher education. This study proposes the application of the K-Means clustering algorithm to classify students according to attendance and learning outcomes. The research utilized academic records comprising attendance percentage, assignment scores, midterm examination scores, final examination scores, and final course scores. Data preprocessing included cleaning and Min–Max normalization prior to clustering. The optimal number of clusters was identified using the Elbow Method, while clustering quality was assessed through the Silhouette Score. Experimental results revealed that the dataset was optimally partitioned into three clusters, corresponding to high-performing, moderate-performing, and academically at-risk student groups. The obtained Silhouette Score of 0.72 demonstrates good cluster compactness and separation, indicating that the selected variables effectively represent student academic characteristics. The proposed clustering model provides meaningful insights into student learning profiles and offers practical support for lecturers and study program administrators in implementing data-driven academic evaluation, targeted learning interventions, and the development of an Early Warning System (EWS). These findings demonstrate the potential of Educational Data Mining techniques to enhance academic decision-making and improve student success in higher education.]]>
Copyrights © 2026
