Application of K-Means Clustering for Student Class Division System

Abstract

SMP Negeri 2 Malinau Utara is a junior high school in Desa Putat, Malinau Utara, Malinau, Kalimantan Utara and has 127 students. Currently, the class division process is inefficient and random. On the other hand, the clustering process' class division must be able to provide each class a balanced number of students. This study proposes the grades of Indonesian and English languages, Mathematics, and Natural Sciences for the clustering. K-means is applied to evenly group students based on predetermined value criteria to achieve the expected class formation. K-Means Clustering is an algorithm in data analysis to group a set of data into several groups based on their similar characteristics. In the clustering process, the distance between the data and the Centroid was calculated using the Euclidean Distance. Initial centroid determination and data distance calculation with the initial centroid were performed until the centroid member remains unchanged. The initial centroid was determined using a combination of 1,081 times obtained from 47 data combinations for two clusters. This research has been successfully applied to classify students using the K-Means Clustering method and select a balanced number of students between one class and another. Next, combine some students in each cluster with other clusters, so that each class has different levels of learning ability. With the combination of two clusters in one class, it is expected that students can help each other during the learning process.