Higher education institutions generate academic records that can reveal student performance patterns, yet these data remain underutilized for unsupervised analysis. This study compared K-Means, Agglomerative Clustering, and Density-Based Spatial Clustering of Applications with Noise to identify academic performance profiles using records extracted from a researcher-developed academic management system. A total of 1,073 subject-enrollment records representing 120 students were transformed into four standardized academic indicators. The algorithms were evaluated using Silhouette Score, Davies–Bouldin Index, Calinski–Harabasz Index, and cluster coverage. K-Means and Agglomerative Clustering achieved identical Silhouette Scores of 0.6916 and Davies–Bouldin Indices of 0.6536, forming two uneven clusters. The density-based approach identified two clusters with 80% coverage and 24 noise points. These findings demonstrate that clustering methods reveal different academic group structures. The study provides a foundation for interpreting student performance patterns and integrating unsupervised analytics into existing academic information systems to support institutional decision-making.
 

Academic Performance Clustering, Educational Data Mining, K-Means, Agglomerative Clustering, Density-Based Spatial Clustering of Applications with Noise