A Cucumber Photosynthetic Rate Prediction Model in Whole Growth Period with Time Parameters

Round 1

Reviewer 1 Report

Line 16-17, “However, photosynthetic rate prediction model for whole growth periods has been proposed yet.” Change to “However, photosynthetic rate prediction model for whole growth periods has not been proposed yet.”

 Add the detail of the SVR algorithm and MPGA, e.g. the information (name, version……) software or toolbox in the section of “materials and methods”

 For the girdsearchCV method, the value of the search step is very important for the accuracy of SVR calibration. In this work, what was the search step? How did you select a suitable step?

Delete the first conclusion (section “4. Conclusions).

Author Response

Dear reviewer,

Thank you for your comments concerning our manuscript! These comments are all valuable and helpful for revising and improving our paper. We have studied the comments carefully and have made correction which we hope meet with approval.  You can find our reply to your comments in the attached file.

Special thanks to you for your good comments to improve our research!

Should you have any questions, please contact us without hesitate.

Sincerely yours,

All authors

Author Response File: Author Response.docx

Reviewer 2 Report

This paper proposes the evaluation of a photosynthetic rate model for cucumber. This is done by evaluating the photosynthetic rate of cucumber leaves under three different parameters: photon flux density, temperature, and CO2 concentration. Unlike other works in the state of the art, the authors propose to perform data monitoring throughout the growth period, so cultivation time was also employed as an independent variable. Using this data, the authors propose a model obtained based on SVR, where the parameters of the SVR model are optimized using different optimization techniques.

The authors detail the process employed to acquire the data as well as the preprocessing employed for model learning. The article is generally well written and easy to read. Before I could recommend publication, I would like the authors to answer the following comments:

 1.       Theoretically, the model learned can obtain the photosynthetic rate of a sample, based on the photon flux employed, the temperature, CO2 concentration and time (days after planting). The authors report the effects of the environmental conditions and time on the photosynthetic rate. However, the paper lacks some insight on the model obtained. Can you comment on the optimal conditions found for protected agriculture of cucumber?

 2.       It is difficult for the non-expert to spot some effects such as photoinhibition in Figure 2.  Please add some grids to the graph or even better, include a 3D surface. The model employs four independent variables, and its difficult to plot 5D data. However, you can do this by fixing two of the variables and better illustrate these phenomena and the effect of multiple environmental/time conditions on the photosynthesis rate.

 3.       The model found using the SVR approach is evaluated against the models obtained using other approaches. Can you comment about model validation using experimental data?

 4.       Some minor comments:

a.       It is not clear from the methodology how many cucumbers/leaves were sampled. Around line 116, the text mentions that for each of the 192 environmental conditions the photosynthetic rate is measured three times per group. There is no mention of how many samples are contained in a group.

b.       References 34 is missing!

c.       Please avoid acronyms in the abstract. These are usually defined in the introduction.

d.       MLR is mentioned in the abstract,  but is never defined nor  mentioned in the rest of the paper. 

Author Response

Dear reviewer,

Thank you very much for your work! Based on your comment and request, we have revised your questions one by one and attached the necessary references. You can find our reply to your comments in the attached document.
Thank you for your time to review our articles to help us improve the quality of our manuscript.

Sincerely yours,

All authors

Author Response File: Author Response.docx

Reviewer 3 Report

Comments for authors

This study developed machine learning (ML) models to predict photosynthetic rate in cucumber using environmental conditions (light, temperature, and CO2 concentration) and growing time of the cucumbers as input variables. The study showed the SVR model optimized using multiple population genetic algorithms (MPGA) gave the best performance over other ML models. The study is interesting and presents the use of some state-of-the art machine learning tools for agricultural applications. The subject matter is thus worthy of investigation. The subject of the paper is moderately novel, as the authors introduced a predictor (growing time) which had not been considered by prior studies leading to a better performance. Overall, the methodology was conducted correctly, and the results were summarized and discussed adequately.

However, there are certain issues the authors need to address to strengthen the paper such as the need for including statistical analysis to back certain statements made in the results and discussion. There is also a need to show both training and testing results for the validation of the results using other algorithms e.g BP- neural network. There are a few grammatical errors that need to be fixed throughout the paper.

The following are some specific comments for the authors to consider for improving their paper.

Abstract

1.     Line 14: Remove "however" from the beginning of the sentence

2.     Line 17: Do you mean "has not been proposed yet"?

3.     Line 25: Please define "BP"

4.     Line 26: Define "MLR"

5.     Line 27: "Coefficient of determination" not "determination coefficient"

Introduction

1.     The authors should try breaking the introduction into multiple (at least 4 paragraphs) whereby each paragraph talks about a different subject but all paragraphs are linked.

2.     Line 46: "Conditions" not "condition"

3.     Line 67: Should be "photosynthesis in cucumber" not "photosynthetic of cucumber"

4.     Lines 76-79: Please break this sentence into two sentences.

Materials and Methods

1.      Line 93-94: I assume seeds of the cucumber were sown and not the whole cucumber, but this sentence makes it sound like the whole cucumber was sown. Please clarify this sentence

2.      Experimental methods: Please include a summarized flow chart showing the steps of how you went from data collection to model evaluation. It makes it easier for the reader to quickly understand your methodology.

3.      Line 103: What were the growth parameters?

4.      Line 170: Please define what C and G parameters are.

5.      Line 179: Again, coefficient of determination

6.      Line 183: Does the G here refer to the gamma hyperparameter or something else? Please clarify

 Results and Discussion

1.      Figure 2 caption: After conditions, perhaps you can add "(changing CO2 concentration and temperature)" to make it specific.

2.      Line 224: Indeed, there seems to be differences in the photosynthetic rate between the CO2 concentrations, especially between the 600 and the other three (900, 1200, 1500). However, I am not convinced there was any significant difference between the 900, 1200, and 1500 as the points are too close.

Did the authors perform any statistical analysis between the groups since the authors mention at line 116-118 that the measurements were replicated 3 times?

It may be helpful to show error bars for each mean point on Figure 2 (a-d) so the reader knows the spread of the data. Or alternatively mention some statistical results to support this sentence.

 3.      Line 228-229: Again, some statistical results will be beneficial to support this statement.

And here I am thinking something like a Two-way ANOVA where CO2 concentration and temperature are main effect factors and photosynthetic rate is the response. In which case you can also test the interaction between CO2 concentration and temperature on photosynthetic rate.

4.      Figure 3: Why was CO2 concentration of 1500 specifically chosen to depict the effect of growing time on photosynthetic rate? Please clarify

5.      Line 279: It can be "seen" not "known"

6.      Line 280: What is "E"?

7.      Line 297-298: Which part of the input samples were used here? Is it the same 30% sample used to train the models in Table 3 or another portion of the original data? Please clarify

8.      Line 311-313" The authors mention that their model's performance was compared to models developed using other algorithms e.g. BP- neural network. However, in Table 5.  only validation results are presented.

 It is suggested that the authors include both training and testing results for the other algorithms so the reader can compare how good the other models were able to generalize to test data compared to the MPGA-SVR model.

Currently, it is not clear whether Table 5 shows results for training or testing results.

 9.      Line 313: Please define "BP"

10.   Line 312: Please define "GS" since it’s the first time you are using it.

11.   Line 319: Correct to "chosen"

12.   Line 322-324: It will be beneficial to include at least one plot of observed vs predicted values for the test dataset for the best model among all the models.

13.   Line 336: What is BPNN?

Conclusions

1.     Line 351-353: Since this is one of your main conclusions, you need to back this statement up with statistical analysis as pointed out in the results and discussions.

2.     Line 356: Maybe use a simple word than "precocious"

Author Response

Dear reviewer,
Thank you very much for your suggestions on this manuscript to help us improve the quality of our manuscript, and we highly appreciate your time and consideration. We have carefully considered all comments from the reviewer and revised our manuscript accordingly. Your comments are valuable and very helpful! The manuscript has also been double-checked, and the typos and grammar errors we found have been corrected.

We used the official reply template to reply, and marked the content in red font. You can find our reply to your comments in the attached file.

Sincerely yours,

All authors

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The authors have thoroughly addressed all comments and suggestions and improved the quality of their paper. I therefore approve acceptance of their paper.