Clustering of Student Status in the Spring Semester of the First Year Using Predicted GPA Trends

Abstract

First-year spring semester grades are an important factor in whether or not a student will graduate within the standard academic year. It has been mentioned that the GPA in the spring semester of the first year is related to variables known before admission, placement tests, absence rate, etc. However, it is late in the process of determining grades and it is difficult to determine which students should be intervened upon and how many students should be intervened upon because the status of students within the semester is unknown. There have been few studies to clarify how many students and how their status transitions during the spring semester of the first year. In this study, we propose a method to predict how students' GPA changes week by week during the semester, capture it as a time series vector, and use the transition results of the predicted GPA to classify students. By identifying the differences between the typed groups, we can clarify how much and from which times students decline during the semester, and where and when interventions should be made.