Clustering of Islamic Religious Education Learning Outcomes Using K-Means and Silhouette Score

Abstract

Student learning outcomes are an essential indicator for evaluating the effectiveness of the learning process. However, conventional presentation of academic scores often fails to comprehensively describe students' learning characteristics. This study aimed to analyze students' learning outcomes in Islamic Religious Education using the K-Means algorithm and evaluate the clustering quality through the Silhouette Score. A quantitative approach was employed using the Cross Industry Standard Process for Data Mining (CRISP-DM) framework. The dataset consisted of Formative Assessment, Mid-Semester Assessment (STS), and Final Semester Assessment (SAS) scores from 76 fourth- and fifth-grade students of SD Muhammadiyah Ambarbinangun. Data analysis was conducted using Python with the K-Means algorithm, while clustering quality was assessed using the Silhouette Score. The results indicated that the optimal model consisted of two clusters with a Silhouette Score of 0.6894. The clustering process identified a Good Learning Achievement Group comprising 72 students and a Learning Support Group consisting of 4 students. These findings demonstrate that the K-Means algorithm can objectively identify students' learning characteristics. The clustering results can assist schools and teachers in developing adaptive learning strategies, providing targeted academic support for students who require additional assistance, and designing enrichment programs for students with higher learning achievements to improve the quality of learning continuously.