A Student Performance Prediction Model Based on Hierarchical Belief Rule Base with Interpretability

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Belief rule base (BRB) can fulfill the above requirementsThe authors effort is appreciated. However, authors need to incorporate the comments.

1.Why the authors need to measure student future performance?

2. What is the  performance in terms of Writing, Speak, Listen or soon ?

3. Introduction section is good but first line performance what basis?

4. Elaborate the problem statement instead of above requirements (refer 107 numbering)

5. Equation 5,why the authors 2M+1 conditions taken?

6.what is the outcome of optimization?

7.Sensitivity Analysis is missing 

8. Reliability of the model need to prove

9.How the proposed model overcome existing barrier?

10. Make a flow chart combined problems and solutions to easy understanding.

11. Conclusion section need to highlight the novelty and noted findings like bulletin 

12. Qualifying semester exams is a outcome of this research may constraints of readers so make wide scope

 

Overall Presentation is Good.

Comments on the Quality of English Language

Minor Editing required 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The authors proposed a new student performance prediction model based on hierarchical BRB with interpretability. The proposed method achieved comparable prediction results with some machine learning-based approaches, but the interpretability is better.

 

- Belief rule base system requires a comprehensive range of indicators based on expert knowledge. These indicators may be subjective and vary from experts to experts. The selection of rules is not justified well.

- The authors listed some machine learning based approaches and mentioned they lack enough interpretability. There is a specific field called explainable AI which allows human users to comprehend and trust the results from learning algorithms.

- The Euclidean distance of HBRB1 in Figure 8 differs greatly from the other two curves without adding interpretability constraints. However, the student performance for HBRB1 in Figure 10 looks comparable with other settings. How are Euclidean distance and student performance correlated?

- Attributes are grouped using correlation metric such as Pearson coefficient. Such correlation is calculated among primary and secondary attribute sets. Will clustering work better for grouping attributes?

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Good attempt. Compare to previous version. This version is recommended for the publication

Comments on the Quality of English Language

Minor editing required

Reviewer 2 Report

Comments and Suggestions for Authors

Thanks for the authors for addressing my previous comments.