Application of K-Means Clustering to Group Islamic Boarding School Students' Kitab Kuning Reading Proficiency Based on Nahwu and Sharraf

Abstract

The ability to read classical Islamic texts (kitab kuning) is an important indicator for assessing students' mastery of Islamic sciences in Salafiyah Islamic boarding schools. However, the assessment of kitab kuning reading proficiency is often still conducted conventionally and tends to be subjective; therefore, a data-driven approach is needed to support a more objective student-grouping process. This study aims to apply the K-Means Clustering algorithm to group students based on their Nahwu and Sharraf scores as indicators of kitab kuning reading proficiency. The research methods included observation, interviews, literature review, data transformation, normalization using Min-Max Normalization, and clustering using Microsoft Excel and RapidMiner. The dataset consisted of 376 student records with the attributes of age, education level, Nahwu score, and Sharraf score. Manual grouping using Microsoft Excel produced three clusters: C1 with 57 students, C2 with 244 students, and C3 with 75 students. Meanwhile, the RapidMiner results showed C1 with 56 students, C2 with 244 students, and C3 with 76 students. The results indicate that K-Means can be used as a basis for grouping students' abilities in a more structured manner.