PENERAPAN METODE CLUSTERING DALAM IDENTIFIKASI SISWA BERMASALAH BERDASARKAN NILAI AKADEMIK

Abstract

The issue of problematic students is a common challenge faced by educational institutions, particularly at the junior secondary school level. The identification of problematic students is generally conducted subjectively based on teachers’ observations, which may introduce bias in the decision-making process. This study aims to apply a clustering method using the K-Means algorithm to identify problematic students based on academic performance and violation records at MTs Muhammadiyah Ogan Lima. This research employs a quantitative approach based on data mining techniques using two primary variables: academic scores and the number of student violations. The analytical process includes data preprocessing, determination of the optimal number of clusters, distance calculation using Euclidean Distance, and iterative centroid updates until convergence is achieved. Cluster quality evaluation is performed using the Silhouette Score and the Davies–Bouldin Index.

The results indicate that students are classified into three main clusters: (1) non-problematic students, (2) at-risk students, and (3) problematic students. The Silhouette Score of 0.61 and Davies–Bouldin Index of 0.47 demonstrate good and stable clustering performance. These findings can assist the school in developing more objective and targeted student guidance strategies.