Clustering and Classification Using K-Means and Random Forest for Elementary Learning

This study aims to identify elementary students’ learning abilities in Civics Education using a combination of K-Means clustering and Random Forest classification. Report card scores and supporting background variables were collected from SD Labschool students, covering knowledge, skills, and attitudes. The K-Means algorithm grouped students into three clusters based on similarity in learning profiles. The Random Forest model then classified new data into the appropriate cluster with an average accuracy of 91%. The results demonstrate that this integrated approach can help teachers adapt learning strategies to students’ diverse needs. Future work is recommended to expand the dataset across multiple schools and include non-cognitive factors for a more comprehensive analysis.