Predicting E-learning Course Final Average-Grade using Machine Learning Techniques : A Case Study in Shaqra University

Abstract

It is critical to understand the factors that may influence students’ performance in an e-learning course delivered through a Learning Management System (LMS). The conditions affecting students are unique to every e-learning course. With wide adoption of using Machine-learning for making decisions in many areas of research. In this research, we apply machine-learning algorithms using regression analysis to predict final average grades of an e-learning course based on number of factors: total activities, total time-spent on the LMS, number of course views, and number of enrolled students. We use deep learning, decision tree, linear regression, bayesian ridge regression, and random forest techniques.The results show that the deep learning model presents the best mean absolute error, mean squared error, and R-squared for predicting the courses’ final average grade. In addition, the results reveal that the relationships between various input course features (total activities, total time-spent on the LMS, number of course views, and number of enrolled students) and the e-learning course final average grade of students is weak and required considering more features.