<![CDATA[Optimal utilization of academic data is an important requirement in supporting data-based learning decision making. One approach that can be used is Educational Data Mining (EDM) through clustering techniques to map students' academic abilities. This study aims to apply the K-Means Clustering algorithm in grouping students based on exam score patterns in one subject at MAN 2 Labuhanbatu Utara. The data used consists of daily scores, midterm scores, and final exam scores of 11th grade students, which were processed through pre-processing, data normalization, and clustering analysis stages. The determination of the optimal number of clusters was carried out using the Elbow method with the Within Cluster Sum of Squares (WCSS) indicator. The results showed that the three-cluster configuration was the most representative grouping structure, which could be interpreted as groups of students with high, medium, and low academic performance, respectively. The differences in centroid values between clusters indicate significant and structured variations in academic achievement. These findings prove that the K-Means algorithm is effective for mapping student learning groups objectively without requiring initial labels. The clustering results are expected to serve as a basis for teachers and schools in designing more adaptive learning strategies tailored to students' ability characteristics.]]>
Copyrights © 2025
