Grade prediction for improved efficiency in programming education using e-Learning activities within a cloud-based development environment

Abstract

Japanese educators often face major challenges when providing programming education, including severe budget constraints and a high student-teacher ratio. This paper reports our experience of introducing e-Learning content within a cloud-based development environment, directly linked to student profile information, including past performance history. To increase educational efficiency, we attempted to predict students’ grades from information obtained through the lessons and the cloud service use history. By applying an anomaly detection algorithm to the objects, we confirmed that it is possible to forecast results that can help us anticipate ways to maximize efficiency.