A Data-Driven Approach to Improve Cocoa Crop Establishment in Colombia: Insights and Agricultural Practice Recommendations from an Ensemble Machine Learning Model

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The topic of the study is promising; however, it requires significant modifications. The authors need to revise specific sections with greater attention.

·         Do you mean cocoa, not coffee, in line (46-47).

·         For meteorological data, it's mentioned that data were obtained from NASA POWER, but there's no information on the quality or resolution of this data. Discussing the accuracy and potential limitations of the data source would be beneficial, in line (174).

·         This paragraph (line 179 - 183) is a repetition of lines 173-176.

·         Ensure that all major explanations for climatological analysis are backed by citations, in line (382).

·         It would be beneficial to consider potential overfitting, especially since the models perform similarly well on both datasets. Discussing the possibility of overfitting, particularly for the more complex models like ANN and RF, and how it was mitigated would enhance the credibility of the results, in line (565).

·         Additionally, the text could also explore why certain models perform better than others, perhaps by linking model structure (e.g., depth of trees in RF, number of neurons in ANN) to their performance outcomes, in line (565).

·         The technical discussion regarding the performance of the models is insufficient, necessitating a more thorough analysis and interpretation of results to substantiate the scientific merit of the study, in line (724).

·         The discussion could benefit from comparing your findings with those from previous studies in more detail. For instance, if other studies yielded different results, explore why that might be—considering factors such as experimental conditions, crop varieties, or regional differences, in line (724).

·         We note that the conclusion does not add any new information beyond what the abstract has already covered, so it is necessary to address more details.

This manuscript requires significant revisions before it can be considered complete. Therefore, I recommend a major revision. Thank you.

Author Response

Dear Reviewer,

Thank you for your thorough review and constructive feedback. We greatly appreciate the time and effort you have invested in evaluating our manuscript. Your comments have been invaluable in identifying areas where our work can be improved.

We have carefully addressed all your concerns and suggestions, making significant revisions to the manuscript to enhance its clarity, depth, and scientific rigor. Please note that this manuscript has been reviewed by a total of four reviewers. As a result, additional modifications, beyond those specifically addressing your comments, appear throughout the document. All changes are highlighted in blue for tracking purposes. We apologize if this creates any unexpected revisions and assure you that each modification was made to meet reviewer requirements and enhance the overall quality of the paper.

We hope these changes meet your expectations and address the issues raised. Please do not hesitate to share further suggestions if you believe there are additional ways to improve the manuscript.

Once again, thank you for your valuable feedback and for contributing to the improvement of our work. We look forward to hearing your thoughts on the revised version.

 

 

The Authors

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The study aims to improve cocoa crop establishment but fails to specify clear objectives and hypotheses.

 

K-Means clustering is used to group regions, but the chosen cluster range (2 to 100) needs more justification. Including additional metrics, like the Davies-Bouldin index, could strengthen this analysis. While 10-fold cross-validation is a good approach, it's essential to prevent data leakage by ensuring that test regions are not part of the training set.

 

The absence of hyperparameter tuning is a missed opportunity. Implementing techniques like grid search could optimize model performance.

 

Using a single hidden layer may limit the model's learning capacity. Testing deeper architectures or adding dropout layers could improve results.

 

The manuscript mentions correlations without providing statistical significance. Including correlation coefficients and p-values would strengthen the findings.

 

Important elements like soil quality, pest pressures, and farming practices are not addressed. Including these factors would provide a more comprehensive view.

 

The details about the Intel Xeon CPU and RAM are useful, but the significance of these specifications in relation to the complexity of the machine learning models should be emphasized. For instance, did the setup handle the data efficiently throughout the experiments?

 

While the analysis identifies Random Forest as the top performer, stating that Neural Networks have the same accuracy as Random Forest (93.62%) may confuse readers. This inconsistency should be clarified. If Neural Networks actually performed at 91.37%, it should be distinctly communicated.

Author Response

Dear Reviewer,

Thank you for your thorough review and constructive feedback. We greatly appreciate the time and effort you have invested in evaluating our manuscript. Your comments have been invaluable in identifying areas where our work can be improved.

We have carefully addressed all your concerns and suggestions, making significant revisions to the manuscript to enhance its clarity, depth, and scientific rigor. Please note that this manuscript has been reviewed by a total of four reviewers. As a result, additional modifications, beyond those specifically addressing your comments, appear throughout the document. All changes are highlighted in blue for tracking purposes. We apologize if this creates any unexpected revisions and assure you that each modification was made to meet reviewer requirements and enhance the overall quality of the paper.

We hope these changes meet your expectations and address the issues raised. Please do not hesitate to share further suggestions if you believe there are additional ways to improve the manuscript.

Once again, thank you for your valuable feedback and for contributing to the improvement of our work. We look forward to hearing your thoughts on the revised version.

 

 

The Authors

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors:

The aim of this study is to use advanced machine learning techniques, including supervised, unsupervised, and ensemble models, to analyze various environmental datasets and provide actionable insights into the suitability of regions for cocoa cultivation. This research utilized various machine learning models but cannot be classified as ensemble models, as the predictions of individual models were not combined using techniques like Bagging, Boosting, Stacking, or Voting to enhance the robustness and reliability of the predictive outcomes. The current research content and methods in this study have the following issues:

 

What specific scientific problem is highlighted in your abstract?

 

There is too much basic information in the Intro. It could be simpler. Please focus on the scientific problem and elaborate on it.

 

There is an excessive amount of data analysis in the results; for example, what is the purpose of the correlation analysis in section 3.1.2?

 

It is recommended that the content of section 3.3, Model Performance Evaluation, be moved to the Methods section.

 

There is a lack of uncertainty analysis of the results. Please address the impact of the number of clusters (Figure 11), silhouette coefficient score (Figure 12), and model parameter analysis (Table 1) on the results.

 

The discussion section requires an in-depth analysis of the results, such as a detailed discussion based on the literature on why temperature, humidity, and wind speed are key determinants of cocoa growth. This should include a comparison with others' results to highlight your advantages, without introducing new conclusions (Table 4).

Your responses to these questions will greatly enhance the clarity and impact of your study.

Good luck!

Comments on the Quality of English Language

Overall, the expression is clear, but there is still room for improvement. Some sentences are complex in structure, which may hinder understanding. It is recommended to simplify certain sentences and ensure smoother transitions between paragraphs.

Author Response

Dear Reviewer,

Thank you for your thorough review and constructive feedback. We greatly appreciate the time and effort you have invested in evaluating our manuscript. Your comments have been invaluable in identifying areas where our work can be improved.

We have carefully addressed all your concerns and suggestions, making significant revisions to the manuscript to enhance its clarity, depth, and scientific rigor. Please note that this manuscript has been reviewed by a total of four reviewers. As a result, additional modifications, beyond those specifically addressing your comments, appear throughout the document. All changes are highlighted in blue for tracking purposes. We apologize if this creates any unexpected revisions and assure you that each modification was made to meet reviewer requirements and enhance the overall quality of the paper.

We hope these changes meet your expectations and address the issues raised. Please do not hesitate to share further suggestions if you believe there are additional ways to improve the manuscript.

Once again, thank you for your valuable feedback and for contributing to the improvement of our work. We look forward to hearing your thoughts on the revised version.

 

 

The Authors

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

1.173-175 lines of data set are collected from NASA POWER database, but most of these data are from local areas. Is it necessary to collect real-time environmental data for the regions?

2. Lines 188-214 describe the variables used to analyze the cacao cultivation environment, but only introduce the concept of each variable. The table can be added to explain the cacao cultivation environment variables in different regions of Colombia.

 

 

 

3.216-218 Dividing land into three adaptability levels of high, medium and low can add the basis for classification under the diagram description and how to classify it.

4.A variety of advanced machine learning classification models are mentioned in lines 238-278, and picture frames can be added to illustrate them.

5.In lines 332-346, formulas can be added to describe the evaluation indicators.

Author Response

Dear Reviewer,

Thank you for your thorough review and constructive feedback. We greatly appreciate the time and effort you have invested in evaluating our manuscript. Your comments have been invaluable in identifying areas where our work can be improved.

We have carefully addressed all your concerns and suggestions, making significant revisions to the manuscript to enhance its clarity, depth, and scientific rigor. Please note that this manuscript has been reviewed by a total of four reviewers. As a result, additional modifications, beyond those specifically addressing your comments, appear throughout the document. All changes are highlighted in blue for tracking purposes. We apologize if this creates any unexpected revisions and assure you that each modification was made to meet reviewer requirements and enhance the overall quality of the paper.

We hope these changes meet your expectations and address the issues raised. Please do not hesitate to share further suggestions if you believe there are additional ways to improve the manuscript.

Once again, thank you for your valuable feedback and for contributing to the improvement of our work. We look forward to hearing your thoughts on the revised version.

 

 

The Authors

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors improved the manuscript according to my comments. It can be accepted for publication.

Author Response

Dear Reviewer,

Thank you very much for your kind feedback and for acknowledging the improvements made to the manuscript based on your valuable comments. We truly appreciate your time and effort in reviewing our work, as your insights have been instrumental in enhancing the quality and clarity of the study.

We are delighted that the revised manuscript meets your expectations and is deemed suitable for publication. Your encouragement motivates us to continue striving for excellence in our research endeavors.

Thank you once again for your support and guidance throughout this process.

 

The Authors 

Reviewer 2 Report

Comments and Suggestions for Authors

The revised version is appropriate for acceptance.

Author Response

Dear Reviewer,

Thank you very much for your kind feedback and for acknowledging the improvements made to the manuscript based on your valuable comments. We truly appreciate your time and effort in reviewing our work, as your insights have been instrumental in enhancing the quality and clarity of the study.

We are delighted that the revised manuscript meets your expectations and is deemed suitable for publication. Your encouragement motivates us to continue striving for excellence in our research endeavors.

Thank you once again for your support and guidance throughout this process.

 

The Authors 

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors:

After the author's revisions, this article has been significantly improved. Thank you for the detailed responses and the additional analyses provided in response to my previous comments. After reviewing the revised manuscript, I have the following further questions and suggestions:

It is recommended to streamline the discussion section and reorganize it into distinct subsections, including a summary of the main findings, a comparison of the similarities and differences between the current results and previous studies, and a presentation of the study's limitations (uncertainties).

Good luck!

Comments on the Quality of English Language

Can be further refined.

Author Response

Dear Reviewer,

Thank you once again for your valuable feedback and commitment to improving the quality of our manuscript. We truly appreciate your acknowledgment of our efforts and the thoughtful suggestions for further enhancing the discussion section.

We have reorganized the discussion section in this revised version into distinct subsections to improve readability and flow. The new subsections include:

1. Summary of Main Findings – Highlighting the key results and their significance.
2. Comparison with Previous Studies – Discussing similarities and differences with existing research, providing context and clarity.
3. Study Limitations and Uncertainties – Addressing our study's limitations, including the inherent uncertainties in our approach.
4. Future Directions – Outlining potential avenues for further research, including incorporating additional data sources, advanced machine learning techniques, and broader geographical applications to enhance the generalizability of our findings.

Additionally, we streamlined the discussion by reducing redundant phrases and improving overall clarity. At the same time, we retained essential information based on the feedback from all four reviewers, ensuring that the discussion remains comprehensive while adhering to your suggestion to reorganize the section.

We hope this revised version meets your expectations and effectively addresses all concerns. Thank you again for your insightful recommendations and for supporting us in refining our manuscript. Please do not hesitate to share any further suggestions.

Best regards,