The Implementation of K-Means Clustering on E-Learning Feature Development for Outstanding Student Recommendation

Abstract

Schools still need to carry out the process of selecting outstanding students, which has several weaknesses. The data processing process takes a long time and tends to result in human decision-making errors. Although the selection of outstanding students is essential in giving awards and praise to students who excel, the school's current method could be more optimal. The process often takes a long time and requires a lot of human resources to collect and process student data, which can disrupt the school's daily operations. this research aims to group and select students as outstanding students by implementing the k-means clustering method and utilizing E-Learning features. The data used in this study are 30 samples of MIN 2 Malang City student grades, five criteria and grouped into 3 clusters. Experiments conducted are the best criteria weight, the best centroid, the best radius and the best number of clusters to obtain groups (clusters) of students according to the ability and assessment of students. The experimental results show that the best criterion weight is the 4th criterion weight with the percentage of criterion weights: K1 = 25%, K2 = 20%, K3 = 25%, K4 = 15% and K5 = 15%. The best centroid is the 1st test with a Percussion value of 97%, Recall of 98%, Specificity of 98% and Accuracy of 98% obtained in the 1st test. The best radius is obtained in the first and fifth tests with the farthest distance of 10.42. The best number of clusters from the trial results with division into three groups and four obtained is 3 clusters with Precision of 79%, Recall of 78%, Specificity of 89% and Accuracy of 87%. Then the implementation of the k-means method with the system resulted in grouping the highest scores (C1) in as many as 21 students, medium scores (C2) in as many as 5 students and low scores (C3) in as many as 4 students. C1 = 21 students with student data (2, 4, 6, 7, 8, 12, 13, 14, 15, 16, 17, 18, 19, 21, 24, 25, 26, 27, 28, 29, 30), C2 = 5 students with student data (9, 10, 20, 22, 23) and C3 = 4 students with student data (1, 3, 5, 11).