Does Machine Learning Offer Added Value Vis-à-Vis Traditional Statistics? An Exploratory Study on Retirement Decisions Using Data from the Survey of Health, Ageing, and Retirement in Europe (SHARE)

Round 1

Reviewer 1 Report

The manuscript presents an interesting comparison of two variants of parsing data to obtain information on retirement decisions from a relative vast dataset.

However, I would propose that the authors attempt to add/clarify several aspects of the manuscript:

1. Section 2.2 in my opinion is too succinct. The manuscript attempts to study Machine Learning algorithms so I would expect a more comprehensive presentation on the subject.
2. Please give the reader more details for the Cox regressions, Section 3.2, and Machine learning algorithms, Section 3.3. I feel that even if it is considered as public knowledge they should be more fine explained as a concept before how they are used in this example.
3. I would consider a different way of representing Figure 2. It is very hard to read and understand.
4. I would propose the authors to add a Conclusion section after the Discussion sections. The title of the manuscript is “Does Machine Learning offer an added value vis-à-vis traditional statistics?” so a reader would expect a conclusion to this.

Author Response

Pleas see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

The article entitled “Does Machine Learning offer an added value vis-à-vis traditional statistics? An exploratory study on retirement decisions using data from the Survey of Health, Ageing and Retirement in Europe (SHARE)” is well-written and, from my point of view, would be of interest for the readers of Mathematics. In spite of these, and before its publication, certain changes are required. Please see below the changes suggested:

• Authors adresses are not complete. Please, complete them.
• State of the art. The content of this section is not enough. More bibliographical references are required. Please make a better description of the Figure introduced in this section.
• Check the use of bold letters in tables.
• Line 250: introduce any bibliographical reference to SHARE.
• Line 265 it is said [see 2]. What do you mean? Why do you use brackets?
• Table 3. Please improve its layout as it is difficult to read.
• Figure 2. Think about if it would be possible to delete some or them or if you could split in two figures to enlarge them.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

The authors of this article presented An exploratory study on retirement decisions 3 based on data from the European Survey of Health, Ageing, and Retirement 4. Additionally, the research addresses the question "Do machine learning algorithms outperform statistical survival analysis in predicting retirement decisions?" The following Major Concern must be addressed before resubmission: 

1. The literature review seems poor, it shall be revised by including some recently published works. 
2. How can you include context into the model? Many social, and cultural factors may affect the performance. 
3. Please write the research gap and highlight your contribution to this work in bullet format. 
4. Explain eqn 2 and eqn 3 in detail and demonstrate the relationship between these equations and the outcome.
5. Check all the symbols. Some symbols are used in the italic form in the equation, but use the non-italic form in the text. 
6. There are typos in the article, please read the article carefully to eliminate these. 
7. Compare these results with the other state-of-the-art. 
8. Why do you have used AUC only? There are other metrics to evaluate the performance of the model. 
9. How do you justify time-dependency to the modeling of retirement decisions? 
10. Add future work at the end of the paper. 

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The authors have incorporated all the suggestions. Thank you.