Optimizing K-Means Algorithm by Using Particle Swarm Optimization in Clustering for Students Learning Process

Abstract

In the implementation of learning, several factors affect the student learning process, including internal factors, external factors, and learning approach factors. For example, the physical and spiritual condition of students. Physiological aspects (body, eyes and ears and talents of students, student interests). External factors, for example, environmental conditions around students, family, teachers, community, friends) Thus, learning achievement is significant because educational institutions' success can be seen from how many students learning achievement. This research's first focus is to do student clustering based on their learning process using 11 parameters. Second, using the PSO algorithm to get maximum clustering results. The research data were obtained from vocational secondary education institutions in the city of Pasuruan. The data is obtained from the results of school reports and questionnaires as much as 100 student data. Data attributes include environmental features, social features, and related school features to group student data for learning data processing. From the classification results using the PSO method, the silhouette value is 0.97140754, very close. These results indicate that the PSO method can improve the K-Means clustering method's performance in the classification process of student learning interest.