An Online Learning Prediction System for MOOCs Using Deep Learning Approaches

Using Internet, artificial intelligence and other technologies, online learning such as MOOCs provides us with new learning platforms and methods. In the paper, a transformer-based model for online learning prediction in MOOCs using deep learning approaches, namely ONLPSformer, is proposed. The model transforming the learner’s data to a sequential form and extracts feature information from students’ learning behavior using transformer-based mechanism model with their behavioral changes in online learning. After extracting feature information, the previous learning behavior data are combined to predict the students’ learning behavior and their learning effect. To verify the performances of the model, many experiments are conducted using the MOOCCubeX dataset. The experimental results show that the model achieved better effectiveness in learning prediction, compared with the traditional methods. The values of RMSE and MAE of the proposed method are 0.159 and 0.102, respectively, which are the smallest among the three models.