Grade Prediction in MOOCs

Over the past few years, we have witnessed the rapid growth of Massive Open Online Courses (MOOCs). More and more researches focus on MOOCs, especially grade prediction, due to the low completion rate in MOOCs. In this paper, we proposed a grade prediction model to automatically predict students' grades based on students' previous performances. We utilized regression, back-propagation neural network methods based on feature selection comprehensively considering educational theories. The model was tested on a real MOOC data set, and the results showed that the model performed well in predicting students' grades with a high accuracy. The most important contribution of this paper is that several valuable applications based on the simple grade prediction model are discussed, including teachers, students, MOOCs, traditional class and so on.