Stepwise Iterative Maximum Likelihood Clustering Based on Kompetisi Sains Madrasah’ Scores for Identifying Quality of Junior High School Grading Distribution

Kompetisi Sains Madrasah (KSM) is a science competition organized by the Ministry of Religious Affairs Indonesia. KSM scores can help to determine the quality distribution of students based on KSM data. Clustering is a method that can group data based on the similarity of data. One method of student clustering that has never been done is Stepwise Iterative Maximum Likelihood. SIML does clustering based on the average value and standard deviation of each cluster. SIML grouping is done several times by creating several different clusters. The best grouping experimental results are in the number of clusters 3. Cluster data is divided into 3 groups consisting of very good, good, and fair students. Silhouette values indicate the best grouping using 3 clusters. The silhouette value is 0.423745 for natural sciences subject data. Social sciences data has a silhouette value of 0.415654, and a silhouette value of 0.487071 for mathematics subjects. Three silhouette values are the best silhouette values when compared to other experiments.