Application of Cluster Analysis and Correlation between Mathematics and Natural Sciences Subject Based on Student Test Scores Using K-Means Clustering

Abstract

Educational unit examinations are a means of supporting decisions to determine students' abilities in mastering all the material that has been taught. education unit examinations also have an impact on students' ability to continue to the next level of education, besides that they can also improve the quality of school education. Based on the background of the problem, this researcher aims to group learning outcomes for mathematics and natural sciences subjects based on educational unit exam scores using the k-means clustering algorithm, in addition to finding out the relationship between mathematics and natural sciences subject learning outcomes using Spearman rank correlation analysis. The population in this study was all 39 students in class IX of SMP Plus Berkualitas Lengkong Mandiri for the 2023/2024 academic year. The sample in this study used saturated sampling. The research results show that through clustering using the k-means clustering algorithm, it was found that cluster 1 with the high category had 21 students, cluster 2 with the medium category had 12 students and cluster 3 with the low category had 6 students. Based on the Spearman rank correlation analysis of the entire sample, a low correlation was obtained between mathematics results and natural sciences results based on educational unit exam scores of 0.608 and a coefficient of determination of 36.9%. Meanwhile, the Spearman rank correlation test on all cluster 1 samples showed a moderate relationship between mathematics learning outcomes and natural sciences subject learning outcomes of 49.5%.