An Approach for Analyzing Student Performance Based on Formative Assessment Scores Using the k-Means Method

This study aimed to analyze students' learning outcomes and behavioural tendencies using formative assessment scores for students enrolled in the "Research and Development Strategy" course offered at the Graduate School of Frontier Sciences, Yamaguchi University. This analysis aimed to gain detailed insight into individual students' understanding and learning behaviours in the educational process. An unsupervised machine learning algorithm combining the k-means and elbow methods was used in this study to classify student performance and tendencies effectively. The method revealed different aspects of students' learning behaviours and abilities, indicating its potential to improve the quality of education. This study applied a data-driven approach to educational analysis and provided an effective means of understanding the student learning process. The study results revealed that students in certain clusters exhibit specific performance characteristics and provide a basis for educators to obtain specific information to provide more appropriate instruction and support for individual students and groups of students. This approach also applies to other educational settings and subjects, contributing to improved student learning processes and outcomes. Furthermore, this study demonstrates the utility of a data-driven approach to educational analysis to be developed further in future studies.