Response to the Comments on “Predicting and Understanding Student Learning Performance Using Multi-source Sparse Attention Convolutional Neural Networks”

This is the response to the reviewer of the paper of the previous manuscript ID is “TBD-2021-07-0204”. We have carefully and deeply considered all comments and suggestions from the reviewers. All of their comments are very helpful for improving this paper. Following their comments, we gave the following major revisions: 1. Adding more discussions on the motivations, including the importance of student performance prediction, the previous works by using convolutional neural networks, the reason for sparse attention, and related studies on course dependence. 2. Adding more details on model implementation, our datasets, and the experiment details. 3. Conducting more experiments on ablation studies, model parameter discussions, and feature analyses. 4. Moving the details about the proof of a theory, data descriptions, dataset splitting for experiments, the details of questionnaire, and some experiment results into the supplemental materials. 5. Double checking our paper writing and language, and correcting the mistakes and typos. In the following, we provide a point-by-point response to the reviewers’ comments. In the marked-up copy, we used the red font for reviewer 1 and the blue underline for reviewer 2. In the revised manuscript, all the modifications that are based on reviewer comments are highlighted.