Identifying At-Risk Students for Early Intervention—A Probabilistic Machine Learning Approach

Round 1

Reviewer 1 Report

Check Sections 3.2.5 and 3.2.6. Section 3.2.5 is duplicated in the second one.

Author Response

"Check Sections 3.2.5 and 3.2.6. Section 3.2.5 is duplicated in the second one" Response:

• • Thank you very much. Duplication was removed 

Reviewer 2 Report

The article discusses the use of probabilistic machine learning to identify at-risk students for early intervention. The study developed a five-stage probabilistic logistic regression model to predict the probability of a student failing an upcoming assessment. The model incorporated student engagement data from Moodle, demographic and student performance data. The results showed that the significance and certainty of student engagement and demographic variables decreased after incorporating student performance variables, such as assignments and tests.

I found this paper well-written and the contribution of clear value.

It would be beneficial if the authors can use some graphs to highlight the performance of their approach against existing methods.

The literature review of this paper is good but the authors missed out an important family of approaches in  probabilistic machine learning, namely data assimilation. Some suggested references to build the related work in introduction

https://www.sciencedirect.com/science/article/pii/S0021999122003643

https://royalsocietypublishing.org/doi/full/10.1098/rsta.2020.0089?casa_token=7tANSCXC8fgAAAAA%3AtuPL0355dS5W-urw1NyfQG6FOXkFdxuHOzLgTBaqNlcCjEYMl007pSP0B_tQWDk8CBFfgRuNWN8o

 

https://www.mdpi.com/2072-4292/14/13/3228

Author Response

We truly appreciate the feedback and have taken it very seriously

Thank you for your valuable feedback on our study. We appreciate your perspective on the use of "data assimilation" in our research, and our opinion is that it may not be commonly associated with learning analytics. 

Nonetheless, we acknowledge that the suggestion of considering "data assimilation" has historically been used in other fields, such as meteorology, oceanography, and geophysics. We understand your suggestion, and we will consider including it as a potential technique for future work.

We appreciate your constructive criticism, and we will continue to explore new techniques and approaches to improve our research in the learning analytics field. Once again, thank you for taking the time to provide us with your feedback

Reviewer 3 Report

Dear authors, 

I'd like to congratulate you and your team on your excellent research work in your paper submitted for publication in this prestigious journal. The topic is very interesting, and I enjoyed it. I would like to thank you for your efforts in presenting your research work in such a professional manner. However, before your work is recommended or accepted, a few comments must be included/ addressed to improve the quality of your work as well as for future publication in this reputable journal. I have the following observations, questions, and comments that may help to improve your work. The authors must modify the following points in great detail. 

1. This is a topic very well studied in the literature and, in that sense, the paper lacks originality. 

2.In the abstract, please include 2-3 special quantitative achievements from the findings of this study in the context of the environment by combining the research objectives and problems. Check spellings for many words that are misspelt or written in haste. 

3. The introduction section needs a few more sentences to strengthen the article, and please include the research problem, objective, and novelty in the last paragraph of the Introduction section. 

4. In the evaluation performed, it would also be important to consider the most common performance metrics such as ROC, AUC, Precision, or Recall as performance metrics.

5.  Please also present the methodology section in a concise graphical format. 

6. Just after the Methodology, please mention the societal benefits of your research in terms of evaluating its key determinant. 

Author Response

Thank you very much for the Feedback

Below is how changes were effected as suggested:

Comment 1

The study relooked at the at-risk student identification problem by considering a different combination of indicators with a different modelling approach in an area that is mostly not studied. Significant indicators for predicting at-risk students may vary by geographical location due to the different demographics and student behavioural patterns present in the data. 

Most existing studies in learning analytics for solving the at-risk student problem are based on traditional machine learning models that only provide point estimate predictions and lack quantification of uncertainty in model prediction. Using Probabilistic Machine Learning, the authors introduce a different kind of predictive model that provides additional benefits compared to commonly used models in learning analytics. Thank you very much improvements were effected. 

Comment 2

An extra quantitative point got included in the abstract.

comment 3

The authors added a paragraph at the end of the introduction section to strengthen the article introduction.

Comment 4 & 5

Addressed by adding the f1-score, recall, and precision evaluation metrics and a methodology flowchart.

Comment 6

Added a paragraph noting 5 societal benefits after the methodology section.

We appreciate your constructive criticism and will continue exploring new techniques and approaches to improve our research in the learning analytics field. Once again, thank you for taking the time to provide us with your feedback.