A Concentration Prediction-Based Crop Digital Twin Using Nutrient Co-Existence and Composition in Regression Algorithms

Round 1

Reviewer 1 Report

The article discusses the issue of precise prediction of rice nutrients. The relevance of the research results from the practical problems of monitoring crops in real time and their optimal supply of nutrients for proper development and growth of plants. The author raised the issue of the use of digital twin technology, which enables real-time viewing of the crop condition, its optimization and predictive analysis of nutrient content. In this context, the nutrient content of rice was predicted by considering the coexistence and composition of multiple nutrients. The effectiveness of six prediction models built using six regression algorithms was compared: Elastic Net, Polynomial, Stepwise, Ridge, Lasso, and Linear Regression. The research results emphasize the superiority of the model based on the polynomial regression algorithm. The conducted analyzes are a source of valuable information which, through appropriate monitoring and precise application of nutrient deficiencies, enable an increase in the volume and quality of agricultural produce production.
1.    Details regarding the implementation of regression algorithms could be more precise. Providing more detailed information on algorithm configuration and parameter settings would increase reproducibility and technical understanding.
2.   The article discusses the superiority of the polynomial regression model. A more detailed comparative analysis of the performance of all six algorithms would be beneficial. This may include deeper examination of why certain models performed better and practical implementations of these findings.
3.   The article would benefit from a more detailed description of the data set, including its limitations. It would be valuable to discuss potential biases in the data set, its representativeness, and the consequences they have for the generalizability of the results.
4.   It would be beneficial to expand the discussion on the application of this research in the context of precision agriculture and environmental sustainability.
5. The manuscript should contain at least 35-40 publications.

no specifics

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding corrections highlighted/in track changes in the re- submitted files.

Author Response File: Author Response.pdf

Reviewer 2 Report

This is a precise analysis of data on rice nutrients.

If the purpose of the analysis is to be used as crop digital twins to predict nutrient concentrations during growth, etc., it is necessary to explain in detail the attributes of the data used in the analysis. For example, attributes such as field conditions in which the rice samples used in the analysis were grown, cultivation conditions such as fertilizer application, and cultivation environmental conditions such as temperature and precipitation should be explained.

It is then necessary to indicate the range of cultivation systems to which the results of this study can be applied.

The content of the analysis is not novel, as it merely compares the results of several regression algorithms. From a practical point of view, the results have not yet been used as crop digital twins to clarify specific fertilizer application guidelines.

It is desirable to revise the paper significantly and to summarize it as a paper that discusses causal relationships among variables as the results of regression analysis on nutrients in rice.

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding corrections highlighted/in track changes in the re- submitted files.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper presents interesting research results with our main goal being to compare the effectiveness of regression models used to determine digital twins for rice farm modeling.

While reviewing the manuscript, I found much interesting information and of course same little remarks. They are listed below:

- lines 43 and 44: Please correct a reference symbol. Between a reference 2 and 3, you don't have anything else so the "-" symbol is not necessary. Correct it to the following form [2,3] and [4,5]. Check all papers in the case of similar bad's.

- in my opinion, a part introduction and Literature Review should be connected to one part. For me, this is some kind of scientific standard.

- information/words in Fig. 1 and 2, are unreadable, please correct the quality and increase the text size.   

- as above, a text in Fig. 3 is too small

- line 445, please precise the type and value of the normality test and equal variance with what you used. To use a ANOVA method we need to be sure that all assumption was passed.

- table 11, please exchange a name model 1-6 on the name of regression methods. It is not clear from the text which model 1-6 is assigned to which method

- what type of post-hoc test was used in your ANOVA analysis? 

- 

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding corrections highlighted/in track changes in the re- submitted files.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The paper is well revised and understandable.

Reviewer 3 Report

Thank you for your cooperation, and a possibility to review the fine paper