Round 1

Reviewer 1 Report

 

The manuscript presented a novel GAN model for classification of Taif Rose. Although the problems being addressed are potential of interest to the readership and work seems interesting, several corrections are required as mentioned below:

1. Manuscript lacks in clarity of problem statement, experimental study, and basic details of algorithms employed.

2. Abstract is not providing the clear picture of problem being addressed and its solution.

3. GC-MS analysis may be not known to a large class of readers, its full-form and utilization in this work may be briefed in abstract.

4. Several typographical errors were observed. For e.g., “oil-bearing rose genetic”, “the proposed model (Section 4)”, etc.

5. 28: The sentence seems incomplete “As dataset contains a number of different distributions, which can be accounted for by providing the model with a sufficient number of layers.”

6. Several GANs are being presented in the manuscript: Novel GAN, cGAN, AC-GAN, Stack C-GAN? In introduction, majority of the portion is devoted to AC-GANs, but in 124 it is mentioned that Stack conditional GAN (Stack C-GAN) is being utilized to develop an efficient high-throughput method

7. R. damascena and Rosa damascena are used interchangeably without being explicitly mentioned.

8. 55: What are MS databases? Please specify in text.

9. 64: “which features are essential in a network using machine learning algorithms”. Which network authors are talking about?

10. Explanation of GAN in Introduction is not looking appropriate. It should be included in materials and methods section.

11. The objective of the work is not clear. Several statements indicating the objectives seems contradictory. For e.g., in 60: As a result, the determination of the constituents of the compound and its resolution become the primary objectives in this area of research; in 113: our work focused on the classification of small datasets in the agricultural field; in 118: A key objective of our study has been to apply generative adversarial networks (GANs) to a group of significant markers obtained through analytical instrumentation, specifically gas chromatography coupled with mass spectroscopy (GC-MS), in the context of Deep Learning (DL) techniques; in 122: The purpose of our study was to investigate classification utilizing GC-MS data structuring and fingerprint image generation based on small datasets. Finally, in conclusion, it was mentioned that “The primary aim of this research is to develop a general prediction model for Taif Rose populations using recently improved generative adversarial networks.”

12. 132: Contribution-1 indicates that authors have provided a brief introduction to GAN along with their applications. In my opinion it is not a contribution.

13. 134: Contribution-2 is not matching with abstract and introduction section. A totally different concept is addresses here. What is meant by “cognitive radio field”?

14. Paper organization at the end of the introduction section is missing.

15. Overall the introduction is written very poorly with plenty of confusing statements. Problem statement and methodology are not clear.

16. In related works, a few statements are copied as-is from sources. For e.g., in 200, it is mentioned that “. However, our work is methodologically related to the analysis of facial aging, which involves predicting faces several years into the future using cGANs”.

17. Figure 2 indicates geographical locations of rose farms which seems not informative to the readers. Already in figure 1 the map of studied area is depicted.

18. 227: For each sample, three duplicates were used. Please justify this statement.

19. 241: Citations to “NIST, WILLY library data” are missing.

20. Figure 3 consists of a figure as well as a table. Tables should not be presented as images.

21. 5th column of table in figure 3 is probabilities, but the text under this column looks something different.

22. In section 4, many equations are not numbered and cited in text.

23. Caption of table 2 (which describe its architecture) seems inappropriate. Please correct it.

24. 356, 370: sub-sections (The generator, The discriminator) are not numbered.

25. In 412, it is mentioned that a cluster model was created for the dataset, but in contributions, introduction, and methods, it was not mentioned.

26. After reading the whole manuscript, I am still confused about problem statement, methodology, and results. In conclusion, it is mentioned that “It is clear from the results of this study that the cGAN model proposed here has great potential and value”. So can I assume that the whole work was planned to assess the potential of the cGAN?

 

Author Response

The authors are grateful to the Reviewer for his valuable comments that helped us improve our manuscript. pls. find our response in the attached file

Author Response File: Author Response.pdf

Reviewer 2 Report

It will be nice if you write something about the botanical situation of this plant. Go to POWO and use the data.

Base on POWO:

Rosa × damascena Herrm.

First published in De Rosa: 14 (1762) This is an artifical hybrid The hybrid formula of this artificial cross is R. gallica × R. moschata.

Author Response

The authors are grateful to the Reviewer for his valuable comments that helped us improve our manuscript. pls. find our response in the attached file

Author Response File: Author Response.pdf

Reviewer 3 Report

The work by Abdelmigid et al reports the conditional generative adversarial neural networks (cGANs) to generate realistic metabolite data. I think this work may provide valuable information for general readers. The report may be published after some minor points given below are adressed.

·         In general the manuscript is not well organized.

·         Some Figures should be reorganized (especially Figure 4 and 5).

·         I recognized lots of grammatical errors.

 

 

Author Response

The authors are grateful to the Reviewer for his valuable comments that helped us improve our manuscript. pls. find our response in the attached file

Author Response File: Author Response.pdf

Reviewer 4 Report

 

Journal:         Applied Sciences (ISSN 2076-3417)

Manuscript ID: applsci-2125469

Title: A Novel Generative Adversarial Network Model based on GC–MS Analysis for classification of Taif Rose

Authors: Hala M. Abdelmigid *, Mohammed A. Baz , Mohammed A AlZain , Jehad F Al-Amri , Hatim Ghazi Zaini , Maissa M. Morsi , Matokah Abualnaja , Nawal Abdallah Alhuthal

 

 

Abstract

 

The appropriate research hypothesis is missing, and significant typo errors can be seen in this section. Results are not described and conclusion is also missing.

 

Introduction

 

A strong study background should provide in the beginning. Although author have provided sufficient information of GC-MS analysis and algorithms and genome scale model but still study rationale and existing gap of knowledge are lacking. Research objectives need modification in order to put them in the best interest of proposed hypothesis.

 

Material and Methods

 

Maps provided are blurred and difficult to read.

 

Result and discussion

 

Quality of the images can be improved.

Cross check your results with tables and figures.

Discussion should be more logical based on your study outcomes.

Check the statistics used in figure 5.

 

 

Conclusion

 

Don’t repeat your results in this section. It should be the synthesis of your key results. Also, cross check all the cited references.

 

 

 

Recommendation: Minor revisions

Comments for author File: Comments.pdf

Author Response

The authors are grateful to the Reviewer for his valuable comments that helped us improve our manuscript. pls. find our response in the attached file

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

1. The responses to various review comments were not provided by authors. See review comments on point no. 12 to 26 of my first review report.

2. Same for point no. 7 (usage of R. damascena and Rosa damascene) of previous review report. In authors reply, it was mentioned that they have taken the corrective measures, but in the revised manuscript the corrective measures are not there.

3. The response to the point no. 9 is not correctly mentioned. However, in the revised manuscript the sentence is corrected by author.

4. Line 94: It is mentioned that “Data modeling traditionally relies on deep learning-based generative adversarial networks (GANs) for the generation of photorealistic images rather than the priors underlying such data.”. Usage of GAN for this work seems contradictory according to this statement.

5. Line 103: What is “GAN integrates deep into all facets of deep learning”?

6. The key contributions are not written properly. For e.g., in contribution no. 2 it is mentioned that “A more accurate data analysis may be possible by integrating in silico generated samples with real observations.”.

7. Line 137: “The remainder of this p is organized as follows”. What is p here?

8. Line 138: Description of section 3 and section 4 looks same.

9. 140: What are previous techniques. Is it appropriate to use such type of vague statements here?

10. Point no. 22 of previous review report: “many equations are not numbered and cited in text.” Still several equations are not numbered and cited in text.

Author Response

The authors apologize for any inconvenience regarding the first report because a technical problem occurred while uploading our file of response.  please find attached a new file including all responses to both reports 1 and 2.  we finally thank the reviewer for his valuable comments which helped us to improve our paper.

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

All comments are addresses in the revised manuscript. Please check axis labels of figure 1 which seems blurred. Rest is fine.