Machine Learning Model for Analyzing Learning Situations in Programming Learning
Shota Kawaguchi, Yoshiki Sato, Hiroki Nakayama, Ryo Onuma, Shoichi Nakamura, Youzou Miyadera · 2018
In programming learning, students have individual difficulties, and teachers need to grasp those difficulties and provide appropriate support for the students. However, since it is a heavy burden for teachers, a method to automatically estimate the learning situations of students is required. In this research, we developed a method that adopts the development of a machine learning model as an approach to achieve this purpose. This machine learning model outputs the estimated learning situation when the source code editing history of new students is input. As a result of evaluating the developed method, it was possible to estimate the correct learning situations with high accuracy of 98%. The applicability of this learning situation estimation method in practical lessons was shown.