Comparison of K-Means and K-Medoids Methods in Clustering Stress Levels of 6th Semester Students

Abstract

Final-year students often experience high psychological pressure due to academic demands such as thesis completion, final exams, and career uncertainty. This stress can negatively affect their academic performance and overall mental health. This study aims to compare the performance of two clustering methods, K-Means and K-Medoids, in grouping stress levels among sixth-semester students. The comparison is based on three key parameters: accuracy, consistency, and computational speed. Data were collected using psychological questionnaires reflecting students’ stress symptoms and analyzed using both clustering techniques. Preliminary results indicate that K-Medoids outperforms K-Means in terms of accuracy and result stability, particularly when dealing with datasets containing outliers, while K-Means is more efficient in processing large-scale data. These findings are expected to serve as a reference for educational institutions in developing early stress detection systems based on data mining to enhance more targeted psychological support services.