Identify elementary student distribution based on Kompetisi Sains Madrasah data using probabilistic distance clustering

Indonesia is a developing country. The quality of human resources also influences the development of a country. The quality of education is one of the benchmarks for the quality of human resources. The quality of human resources can be seen from the quality of education. Improving the quality of education can be done in various ways. One effort designed to improve the quality of education in Indonesia is the provision of educational competitions in each region in Indonesia, and Kompetisi Sains Madrasah (KSM) is one of the pre-eminent competitions that have been designed. From the KSM Competition, a set of student scores is obtained which is a sample of the quality of education in each province. The number of students in educational institutions, especially elementary schools, is increasing which causes the data of students in the system to improve. The data can be grouped based on ability. Grouping is done using PD clustering. This method is one of the hierarchical grouping methods that have good performance. The cluster of students’ abilities is very helpful in finding out educational information in the regions making it easier for parties to do special handling. The results of clustering using PD clustering show that three clusters represent the distribution of student’s abilities with a silhouette coefficient of 0.5384 and a standard deviation of 0.3506 for mathematics subjects. Silhouette coefficient is 0.4351 and standard deviation is 0.4688 for science subjects.