Selection of Optimized Retaining Wall Technique Using Self-Organizing Maps
Round 1
Reviewer 1 Report
Review - Selecting the optimized retaining wall technique by using a self-organizing maps (SOM) method
The paper is very interesting and meaningful but needs major improvement.
English needs to be edited by a professional editor.
Abstract should be rewritten to include the objective, scope, and result. SOM is a machine learning algorithm. It would be good to mention ML instead SOM in the abstract.
“Selecting an appropriate construction method is very important because of site conditions and the unique characteristics of a construction project. However, existing artificial intelligence (AI) techniques, in most cases, require a significant amount of data cases to get reliable results.”
-> the sentence does not make sense. Why suddenly use AI for selecting construction methods? Authors should provide a good reason to use AI. Introduction should be re-organized.
“In addition, frequent change orders during construction happen in South Korea when different participants design and construct separately. Thus, there is ...” – professional editor is needed. Why there are many CO? How does this effect on using SOM?
“SOM analysis technique is a method of visualizing data through mapping to low-dimensional maps, reducing the dimensions of multidimensional data. This can be analyzed with a relatively small amount of data. Previously, it was impossible to analyze the inherent correlation between attributes. But now it is possible to gain more insight.” -> why and how is this possible now?
Tables -> delete vertical lines.
It seems like a chapter is missing between 2. SOM and 3. Case Study. There should be a chapter that explains how SOM is used in retaining walls before applying to the case study.
“By using the SOM to map the attribute value from a multi-dimensional vector into a 2-D plane, this study can confirm the relationship between attributes visually” -> it seems that there is only 1 vector and one attribute value? Please clarify and provide number of attributes as well as the list of them.
Explanation about Figure 1 is very weak and unorganized. Meaning of color codes, sub titles in Figure 1 are not explained.
There is too much jump in the manuscript. Show some raw data to show what actual data are used and how the SOM is structured.
Author Response
First of all, thanks for the comments from the Editor and particular a very supportive from all reviewers. Please find our responses to the rest of the comments that are useful for improving the quality of the manuscript for technical note.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This is a very interesting topic; however, the manuscript needs to be carefully and thoroughly proofread. I would recommend that the authors use a professional proofreading service to correct the grammar.
The authors may provide a critical discussion about the identified relationships between attributes and retaining wall techniques with additional literature review. What is the new knowledge from this study compared with previous/existing literature?
The authors mentioned that the accuracy of the developed SOM model is not great, then what are the implication of the findings?
What are the major contributions of this research?
Author Response
First of all, thanks for the comments from the Editor and particular a very supportive from all reviewers. Please find our responses to the rest of the comments that are useful for improving the quality of the manuscript for technical note.
Author Response File: Author Response.pdf
Reviewer 3 Report
I appreciate that the authors of this technical note have done this work.
However, I would like to make the following comments:
_From my point of view, I consider that this technical note “Selecting the optimized retaining wall technique by using a self-organizing maps (SOM) method” does not sufficiently explore the scientific literature regarding the subject of study and the contributions of the research done are not sufficiently clear.
_An English review of the full text is necessary. There are mistakes.
_Please review how the work of other authors is cited.
_Is it necessary to quote phrases from other authors in quotation marks? Could a synthesis be made of what the work contributes?
_Regarding bibliographic references, there are no studies from recent years. There are only 4 references from the last 5 years.
_Table 1. Please a revision of the format of the table is necessary.
_How are figures cited in the text? Fig. 2 or Figure 2 ?. Please check in the journal instructions.
_5. Conclusions. It is necessary to clarify this section.
_5. Conclusions. "An SOM" or "a SOM"
_References: It is recommended to incorporate more references from recent years.
Author Response
First of all, thanks for the comments from the Editor and particular a very supportive from all reviewers. Please find our responses to the rest of the comments that are useful for improving the quality of the manuscript for technical note.
Author Response File: Author Response.pdf
Reviewer 4 Report
- It is needed to explain how this research conducted, how to analyze the obtained data before presenting it to make it clear to readers.
- The method used in this research not clear
- The procedure of data analysis is not clear
- The conclusion is not clear to answer the research question.
Comments for author File: Comments.pdf
Author Response
First of all, thanks for the comments from the Editor and particular a very supportive from all reviewers. Please find our responses to the rest of the comments that are useful for improving the quality of the manuscript for technical note.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
There still are many English grammar errors and many typos. The manuscript needs complete re-write for English errors and typos.
Abstract and introduction have been improved but still need to be improved a lot. There are too many errors.
There is a fundamental question that need to be answered.
SOM is also a black box system like any other ML systems. How the authors defend against the questioning of the results of the black box? Accuracy of the SOM. How do you define the accuracy? What did you compare with?
In this research, SOM visually showed the correlations of the variables of the construction method selection that was usually done by the expert engineers. In fact, this can be systematically done by statistical analysis which is NOT a black box procedure.
When the problem is solved by the machine learning or AI, there should be a good reason to use it because they are black boxes and the user cannot control the detail. For example, image and voice data recognition is very difficult without using AI. If there are too many data and their correlations are too much complicated, statistical analysis will not be a good option to consider. However, in this case, there are only 171 data sets(not millions or billions) with only 15 variables (wall, assist, support, shape, height, etc.) (heigh -> height. There are just too many English errors or typos in this paper). For this type of data set, the result of the SOM should be compared with traditional statistical method to verify the accuracy of the results unless there is a very good reason that cannot be done in statistical analysis.
The reviewer only comments on Abstract and introduction part as below.
Line:23: The objective of this study is developed the SOM model as
-> is to develop
Line 34: more high-rise and thus the excavation depth
-> more high rise and thus (when using the word “more” there should be “than” in the sentence) (and thus -> duplicated use of conjunctions)
Line 35: the excavation depth of the underground construction is bound to deepen.
-> what does this mean?
Line 43: The forecasting of a retaining wall technique
-> The decision-making approach for selecting the right retaining wall technique
(forecast means to predict over a time such as weather forecasting. We can forecast the technology development trend. But we can’t forecast retaining wall technique.)
Line 46-50: does not make sense. Please rewrite.
Example: The forecasting of the appropriate retaining wall technique is difficult because in an industry
domains of experience-oriented problems, experience is essential in problem solving.
-> Selecting the right method of retaining wall construction usually has been performed by the experienced engineers using their historical and empirical knowledges.
Line 52-53: to check before construction
-> to check before the construction begins
Line 52: there are many things -> explain what the many things are.
line 60: This study applies the SOM model as an assist tool in decision making at a construction task
planning.
-> This study uses the Self-Organizing Map (SOM) model as an assist tool in decision making of selecting retaining wall construction method during the designing phase of the construction project. (I do not exactly understand what authors mean. Ma be this should be rewritten to “pre-construction phase” or “during the planning phase of the construction project” or “designing phase of the construction project”, etc.)
Line 61-62: The retaining wall technique is selected, evaluate the practicability of the SOM model, and
provide a case study involving the determination of retaining wall technique.
-> The sequence of the work can be described as followings: (1) The retaining wall technique is selected; (2) evaluate the practicability of the SOM model; and (3) provide a case study involving the determination of retaining wall technique.
Line 62: Next section -> The next section
line 63: followed by the case study section that describes….
Line 65: our results -> the results of this research.
Line 133: 3*5 -> 3 by 5 (or 3 x 5)
4.1 process of selection variables -> variable section
Author Response
First of all, thanks for the comments from reviewer #1. Please find our responses to the rest of the comments that are useful for improving the quality of the manuscript for technical note. We checked English editing by Editage. The certification by Editage is embedded in attached PDF file.
Author Response File: Author Response.pdf
Reviewer 2 Report
Yes, I believe that the manuscript has been significantly improved. It is ready for publication.
Author Response
First of all, thanks for the comments from reviewer #2. We checked English editing by Editage. The certification by Editage is embedded in attached PDF file.
Author Response File: Author Response.pdf
Reviewer 3 Report
The work of the authors in improving the paper “Selecting the optimized retaining wall technique by using a self-organizing maps (SOM) method” is appreciated. The authors have considered the reviewer’s recommendations and the improvement of the paper has been important.
Some comments:
_Table 3. Please a revision of the format of the table is necessary.
_ In this version there are still errors in the English. Please review it.
Best regards.
Author Response
First of all, thanks for the comments from reviewer #3. Please find our responses to the rest of the comments that are useful for improving the quality of the manuscript for technical note. We checked English editing by Editage. The certification by Editage is embedded in attached PDF file.
Author Response File: Author Response.pdf
Reviewer 4 Report
Minor correction as the attached file
Comments for author File: Comments.pdf
Author Response
First of all, thanks for the comments from reviewer #4. Please find our responses to the rest of the comments that are useful for improving the quality of the manuscript for technical note. We checked English editing by Editage.
Author Response File: Author Response.pdf
