Predicting Student Scores Using Browsing Data And Content Information Of Learning Materials

Recent digital material delivery systems enable teachers not only to upload lecture materials but to analyze students’ behavior, such as browsing data with detailed operation logs that record which student performed which operation on which page at which time. While such behavioral data has been elucidated to be useful for predicting students’ performance in existing studies, it has yet to be fully verified how content (e.g., learning materials) information can be integrated with behavioral data. This paper proposes methods to utilize content information jointly with behavioral data and compares them with the baseline method using only behavioral data. The results indicate that one of the proposed methods performs better prediction of quiz-score prediction. This suggests that both the browsing behavior of students and the content information have an impact on student performance. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.