Prediction of Banana Production Using Epidemiological Parameters of Black Sigatoka: An Application with Random Forest

Round 1

Reviewer 1 Report

The scientific manuscript, entitled: "Prediction of number of banana bunches from Black Sigatoka epidemiological parameters using Random Forest" a study compiled with sufficient thoroughness and detail.

However, the title (in my opinion) does not adequately reflect the work done. Therefore, I definitely recommend changing the title to clarify the content more precisely.

The aim of the study was to predict the number of banana bunches from epidemiological parameters of Black Sigatoka (BS) disease (caused by Pseudocercospora fijiensis fungus) using a Random Forests (RF) prediction algorithm capable of predicting crop production responses to epidemiological variables. In the period 2015–2018, weekly production data (banana bunch number) from 3 commercial banana farms (in Changuinola District of the Bocas del Toro province in Panama) and the epidemiological parameters of BS were used.

The epidemiological parameters were evaluated according to the incidence and severity of BS. These were carried out in three ages of plants: plants in the acorn stage, plants seven weeks (49 days) and ten weeks (70 days) after the appearance of the acorn, in reports of identical format.

The analysis is a work compiled with sufficient thoroughness. By analyzing the epidemiological parameters and the calculated indices, the most applicable prediction model was determined in the current situation.

Based on the criteria stated above (after clarifying the title), I recommend publishing the manuscript.

Author Response

Reviewer 1 comments

The scientific manuscript, entitled: "Prediction of number of banana bunches from Black Sigatoka epidemiological parameters using Random Forest" a study compiled with sufficient thoroughness and detail.

However, the title (in my opinion) does not adequately reflect the work done. Therefore, I definitely recommend changing the title to clarify the content more precisely.

The aim of the study was to predict the number of banana bunches from epidemiological parameters of Black Sigatoka (BS) disease (caused by Pseudocercospora fijiensis fungus) using a Random Forests (RF) prediction algorithm capable of predicting crop production responses to epidemiological variables. In the period 2015–2018, weekly production data (banana bunch number) from 3 commercial banana farms (in Changuinola District of the Bocas del Toro province in Panama) and the epidemiological parameters of BS were used.

The epidemiological parameters were evaluated according to the incidence and severity of BS. These were carried out in three ages of plants: plants in the acorn stage, plants seven weeks (49 days) and ten weeks (70 days) after the appearance of the acorn, in reports of identical format.

The analysis is a work compiled with sufficient thoroughness. By analyzing the epidemiological parameters and the calculated indices, the most applicable prediction model was determined in the current situation.

Based on the criteria stated above (after clarifying the title), I recommend publishing the manuscript.

Authors: corrected. The new title is Prediction of banana production using epidemiological parameters of Black Sigatoka: an application with Random Forest

Reviewer 2 Report

This manuscript predicts the number of banana bunches based on epidemiological parameters of banana black spot using random forest algorithm to clarify the response of yield to epidemiological parameters. This method has strong usability and certain universality. However, there are still some details in the manuscript that need to be revised, and my suggestions are as follows:

1. The first paragraph in the introduction describes too much about the importance of bananas. It is suggested to merge the first and second paragraphs after simplification.

2. Lines 68~74: This part does not indicate why the author uses the RF method. In addition, there is little introduction to the research progress of banana yield prediction.

3. Line 169: "A total of 13 predictor variables were considered in this study"? Are there only 13 predicted values in the full text? Need to explain the source of the forecast quantity?

4. The serial number and format of Line 221-222 need to be paid attention to. It is suggested to add R-² as the evaluation index, and increase the calculation method of model performance evaluation index.

5. It is suggested to add the description and calculation formula of each parameter in Table2 in the material method,and the contents of Table 2 are analyzed first, so the positions of Table 2 and Figure 1 should be changed.

6. What is the criterion for dividing "High correlations" in Line 254-255? It is suggested to be expressed in terms of significance and marked as * significant and ** highly significant in Figure1.

7. Please increase the quality of the picture of Figure2, as the current clarity cannot clearly see the content of the expression.

8. It is suggested to be able to quantify productivity prediction accuracy in the conclusion.

9. The sample size for modeling and verification and the sampling period corresponding to the sample are not described in this article. Please add this part in "Materials and Methods".

10. line 303: What is the result of the PCA analysis not shown in 3.1? What is the specific variable?

11. Why is the research content of "3.3. Identification of the main variables" placed at the end of the paper？ Why not in the first part (3.1) of the results for analysis?

Author Response

Reviewer 2 comments

This manuscript predicts the number of banana bunches based on epidemiological parameters of banana black spot using random forest algorithm to clarify the response of yield to epidemiological parameters. This method has strong usability and certain universality. However, there are still some details in the manuscript that need to be revised, and my suggestions are as follows:

1. The first paragraph in the introduction describes too much about the importance of bananas. It is suggested to merge the first and second paragraphs after simplification.

Authors: Corrected.

2. Lines 68~74: This part does not indicate why the author uses the RF method. In addition, there is little introduction to the research progress of banana yield prediction.

Authors: Corrected. Line 81. The following phrase was added: Today, RF regression applications in crop science remain lacking, with few exceptions. Numerous studies have pointed out various promising advantages of RF as a regression tool compared to traditional regression models (21,22,23), it is for this reason that the initiative to use the RF algorithm arises in this study focused on its usefulness as a prediction tool in banana production.

3. Line 169: "A total of 13 predictor variables were considered in this study"? Are there only 13 predicted values in the full text? Need to explain the source of the forecast quantity?

Authors: Thank you for the question. In our study, 13 epidemiological variables were used. It is precisely these variables that are obtained in commercial farms in Panama according to the international protocol used for the evaluation of Sigatoka in bananas. The RF model can select the best variables that best explain or predict banana production based on our data.

Stover, R.H. A proposed international scale for estimating intensity of Banana leaf spot (Mycosphaerella musicola Leach). Tropical Agriculture. Trinidad and Tobago 1971, 48(3), 185-196. https://n9.cl/h2ic4

Fouré, E. Black Leaf Streak Disease of Bananas and Plantains (Mycosphaerella fijiensis Morelet). Study of the symptoms and stages of the disease in Gabon. IRFA-CIRAD. Paris. France. 1985. p.20.

4. The serial number and format of Line 221-222 need to be paid attention to. It is suggested to add R-² as the evaluation index, and increase the calculation method of model performance evaluation index.

Authors: Corrected. Line 144: The variable related to production was the total number of bunches harvested weekly (52 weeks per year) during the 2015-2018 period for the three sites evaluated (S1, S2 and S3).

In the table 3: the column with the values of R2 was added.

5. It is suggested to add the description and calculation formula of each parameter in Table2 in the material method, and the contents of Table 2 are analyzed first, so the positions of Table 2 and Figure 1 should be changed.

Authors: Corrected. All the epidemiological parameters (except the incidence (INC) and state of the symptom (SIN) [Stover and Fouré] do not require a formula or equation, only identification of the person observing the position and the number of the leaf on the banana plant to obtain: YLWS, YLS, LEAF. The incidence (INC) and state of the symptom (SIN) through a spreadsheet are obtained according to the protocol of Stover (1971).

6. What is the criterion for dividing "High correlations" in Line 254-255? It is suggested to be expressed in terms of significance and marked as * significant and ** highly significant in Figure1.

Authors: Corrected. Line: 395: whose Pearson´s r was greater than ∣ 0.85 ∣ are

7. Please increase the quality of the picture of Figure2, as the current clarity cannot clearly see the content of the expression.

Authors: Corrected.

8. It is suggested to be able to quantify productivity prediction accuracy in the conclusion.

Authors:  Line 774: with the accuracy of the model's predictions of 0.71

9. The sample size for modeling and verification and the sampling period corresponding to the sample are not described in this article. Please add this part in "Materials and Methods".

Authors: Corrected. Line 198: The following phrase was added: The data matrix X was constituted by the set of vectors of the observations X_[ij], j=1,...,p and where each vector X_[ij] presented the j-th variable for all the observations and where X, was the data matrix formed by "n=623" observations with "p=13" variables.

10. line 303: What is the result of the PCA analysis not shown in 3.1? What is the specific variable?

Authors: Corrected. Line 428: The following phrase was added: e.g., (YLS10 and YLS7); (LEAF, LEAF 7, and LEAF10); (YLWS, YLWS7, and YLWS10).

11. Why is the research content of "3.3. Identification of the main variables" placed at the end of the paper？ Why not in the first part (3.1) of the results for analysis?

Authors: Thank you for the question. The authors consider it correct that in methodology, the exploratory result (PCA and correlations) is described first, followed by the prediction with RF.

Reviewer 3 Report

Dear Authors,

The subject of the study is interesting and topical, with scientific and practical importance.

The introduction is presented correctly, in accordance with the subject. Numerous scientific articles, in concordance to the topic of the study, were consulted.

Methodology of the study was clearly presented, and appropriate to the proposed objectives.

The obtained results are important and have been analyzed and interpreted correctly, in accordance with the current methodology.

The discussions are appropriate, in the context of the results, and was conducted compared to other studies in the field.

The scientific literature, to which the reporting was made, is recent and representative in the field.

Some suggestions and corrections were made in the article.

The following aspects are brought to the attention of the authors.

1.

Please check the number according to English

Page 1, row 30

Most likely "1107.93" instead of “1107,93”

2.

Please check the Geographic coordinates in table 1, to be written the same.

Eg

“9° 25' 43.0'' N” with space

“82°32'57.1''W” without of space

3.

According to Instructions for Authors and Microsoft Word template, it is recommended that the presentation of a table be made as close as possible to the reference in the text to the respective table.

Eg Table 2

It is more appropriate for Table 2 to be presented on page 6, and the current content of page 6 to move to page 7.

4.

Page 8, Figure 2

The text in some figures is hard to read

Eg. Figure 2c

It would be appropriate to revise and if possible a readable text for the content

5.

If it is possible to harmonize the notaries, with upper or lower case letters

eg

Page 10, figure 4

In the text there are references to figure 4 a), b), and c), and in the figure the notaries are A), B), and C)

Similarly in figure 5, page 11

6.

Check text setting for figures title

eg Figure 4, Figure 5

It is recommended to check the title in figure 4, figure 5, to be in accordance with the indicated text style

Style: MDPI_5.1_figure_caption

7.

References

The entire References chapter needs to be revised in accordance with the Instructions for Authors and Microsoft Word template, Sustainability journal.

“Author 1, A.B.; Author 2, C.D. Title of the article. Abbreviated Journal Name Year, Volume, page range.”

Comments for author File: Comments.pdf

Author Response

Dear Authors,

The subject of the study is interesting and topical, with scientific and practical importance.

The introduction is presented correctly, in accordance with the subject. Numerous scientific articles, in concordance to the topic of the study, were consulted.

Methodology of the study was clearly presented, and appropriate to the proposed objectives.

The obtained results are important and have been analyzed and interpreted correctly, in accordance with the current methodology.

The discussions are appropriate, in the context of the results, and was conducted compared to other studies in the field.

The scientific literature, to which the reporting was made, is recent and representative in the field.

Some suggestions and corrections were made in the article.

The following aspects are brought to the attention of the authors.

1.

Please check the number according to English

Page 1, row 30

Most likely "1107.93" instead of “1107,93”

Authors: Corrected

2.

Please check the Geographic coordinates in table 1, to be written the same.

Eg

“9° 25' 43.0'' N” with space

“82°32'57.1''W” without of space

 Authors: Corrected

3.

According to Instructions for Authors and Microsoft Word template, it is recommended that the presentation of a table be made as close as possible to the reference in the text to the respective table.

Eg Table 2

It is more appropriate for Table 2 to be presented on page 6, and the current content of page 6 to move to page 7.

 Authors: Corrected

4.

Page 8, Figure 2

The text in some figures is hard to read

Eg. Figure 2c

It would be appropriate to revise and if possible a readable text for the content

 Authors: Corrected

5.

If it is possible to harmonize the notaries, with upper or lower case letters

eg

Page 10, figure 4

In the text there are references to figure 4 a), b), and c), and in the figure the notaries are A), B), and C)

Similarly in figure 5, page 11

Authors: Corrected

6.

Check text setting for figures title

eg Figure 4, Figure 5

It is recommended to check the title in figure 4, figure 5, to be in accordance with the indicated text style

Style: MDPI_5.1_figure_caption

 Authors: Corrected

7.

References

The entire References chapter needs to be revised in accordance with the Instructions for Authors and Microsoft Word template, Sustainability journal.

“Author 1, A.B.; Author 2, C.D. Title of the article. Abbreviated Journal Name Year, Volume, page range.”

Authors: Corrected