Student Performance Prediction by LMS Data and Classroom Videos

Abstract

This paper conducts research on predicting academic performance based on the learning management system (LMS) data and classroom videos. Except for the interactive data of the LMS, our research introduces classroom videos to expand learning behavioral variables. Through the correlation analysis with the final exam scores, we select 6 of these variables as predictors. Then, the prediction results of different predictor combinations are compared and the result shows that the prediction based on LMS predictors and video predictors achieves the best accuracy of 89.7%, which is higher than the accuracy of using LMS predictors or video predictors individually. The predictors obtained through computer processing can be used for automatic performance prediction.