Classification and Causes Identification of Chinese Civil Aviation Incident Reports

Round 1

Reviewer 1 Report

A brief summary

This paper presents a technical solution to intelligently classify and successfully extract meaningful information from Chinese civil aviation incident reports using machine learning classifiers and vectorization strategies, with extensive comparative experiments and manual analysis.

Specific comments:

This paper has a clear motivation and decent experimental results.

I am leaning to give a "major revisions" based on my current knowledge and understanding of the paper. But I will be willing to revisit the decision after get feedback from the author(s).

In particular, I would be glad if the author could clarify the questions below.

*The relevant papers cited in the first chapter are too few. 5-10 additional papers are needed. More papers that try to use NLP techniques in different domains should be cited.

*The innovation point is not enough, and the experimental techniques are all methods proposed by others. Deep learning methods have penetrated into every aspect of computer, and we certainly know that they work, you just proved that on this corpus, these few techniques are effective.

*There are a lot of "manually" in the paper that are needed to show the need for these contents. What specifically is done through manpower?

What is the composition of manpower? (Member attributes)

*Experimental setup: It is puzzling that you use so many classification methods, but the detailed experimental parameters are not written out.

*On the issue of keywords, some examples should be given appropriately for illustration.

*Why wasn't a language model (e.g. GPT-3) used for the analysis?

We all know that the BERT-based language model is now the mainstream.

*The grammar of the paper still needs some enhancement. Where Chinese appears, be sure to include an English translation afterwards.

The bottom of Figure 3 is not shown completely (or maybe it's the version problem of the pdf I received).

*Some descriptions could be more concise.

*More should be described about the academic and practical significance of this study.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The abstract has problems too. It is too long and not to the point, similar to the problems in the title. The authors should only state the problem, the methods used, major results and contributions.

Authors need to add a sentence to clarify what is next or future plan in abstract section.

Authors need to add Related Work section and give a discussion/feedback of this study.

Please, kindly, recheck equation 5 reference.

The author needs to add more details of how the results are similar in the discussion section, and if there are any differences, he needs to mention them.

The author needs to add more references for the language, and libraries that are used.

The author needs to explain in details the preprocessing step in the analysis.

The quality of the presentation is low. Overall, this manuscript should be revised to improve the quality of the presentation.

The proposed method should be clearly compared with the previous works with Table in the experiment part. The previous works should be presented with reference paper in Table.

What is the originality of this manuscript? This scientific soundness should be emphasized.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors answered all the questions very carefully. Some definitions of keywords have been added, which gives the paper a certain novelty.

Reviewer 2 Report

Thank you