Educational Data Mining to Improve Decision Support on the Ratio of students and Study Groups in Elementary Schools in Indonesia using K-Means Method

Abstract

In the world of education, the process of processing space in learning activities is very important. This is in order to prevent unwanted behavior and can direct student activities. The purpose of the study was to analyze whether the mapping of regions (provinces) on the ratio of students to study groups (abbreviated as rombel) could be done by utilizing artificial intelligence techniques. The data
source was obtained from the Ministry of Education and Culture which was processed by the Central Statistics Agency (abbreviated as BPS) for the academic year 2018/2019 which consisted of 34 records. The research is aimed at student to class ratios at the primary school level. The technique used is the kmeans method which is part of data mining. The analysis process was carried out with the help of Rapid Miner software. Two clusters of mapping labels were used, namely the largest student-to-class ratio cluster (K1) and the smallest student-to-class ratio cluster (K2). The analysis results show that 14 provinces are in the largest cluster (K2) with an average per class = 24.2 and 20 provinces are in the smallest cluster (K2) with an average per class = 19.1. The validity test was carried out by testing the cluster results (k = 2) with Davies Bouldin was 0.570. The cluster results are optimal. The validity test was also conducted on the results of the cluster with Performance (Classification) with the results of classification error: 0.00%. The results of the analysis can be used as input for the government in making policies so that the quality of human resources is increasingly competitive and can compete regionally and internationally.