Penerapan K-Means Clustering Dalam Pengelompokan Prestasi Siswa Dengan Optimasi Metode Elbow

Abstract

Student achievement is very important and meaningful in the world of education, this picture can be seen from the grouping of classrooms, determining talents and interests as well as the level of ability and willingness of students according to the abilities of each individual. This is why researchers conducted research at MIS NW 03 Pancor with data from class 4 as many as 23 students, class 5 as many as 27 students, and class 6 as many as 25 students which were related to variables in student learning achievement. In determining this matter, the school still uses conventional decision-making methods, so it is still difficult to determine student achievement according to their abilities. Thus, schools need to make changes that can help employees and institutions improve education quality by knowing student learning achievements. As for this problem, the author provides a solution that requires a calculation system with the concept of data mining. Mastery of learning material is one of the benchmarks for obtaining student grades and this is one of the policies in determining student achievement. Apart from grades, there are also attitudes and skills which are factors in determining student achievement at school. This can be solved with one of the techniques in data mining that can be used for grouping, namely clustering. The results of this research are a system that can make it easier for schools to quickly and accurately group student achievement using k-means which is optimized using the elbow method with the results obtained, namely cluster 1 "good" with 9 students, cluster 2 "very good" with 6 students, cluster 3 is "fairly good" as many as 8 students and the Davies Bouldin value is 0.711 and the AVG within centroid distance is 18.821 so it can help the school in determining student achievement groups