Submit to this Journal Review for this Journal Propose a Special Issue

Article Menu

Share Help Cite Discuss in SciProfiles

Open AccessArticle

Peer-Review Record

Coupling LSTM and CNN Neural Networks for Accurate Carbon Emission Prediction in 30 Chinese Provinces

Sustainability 2023, 15(18), 13934; https://doi.org/10.3390/su151813934

by Zhonghua Han^1,†, Bingwei Cui^1,†, Liwen Xu^1,*, Jianwen Wang¹ and Zhengquan Guo²

Reviewer 1: Anonymous

Reviewer 2: Anonymous

Sustainability 2023, 15(18), 13934; https://doi.org/10.3390/su151813934

Submission received: 18 August 2023 / Revised: 9 September 2023 / Accepted: 12 September 2023 / Published: 20 September 2023

Round 1

Reviewer 1 Report

1. The abstract does not highlight the specifics of the proposed research or findings. The abstract should be more straightforward for the reader regarding the proposed method and its motivation. Especially the motivation and the novelty of the paper do not exist. The importance of the problem is not mentioned. The abstract should present some main points for the readers, such as the main contributions, the proposed method, the main problem, the benchmark tests and data, the comparative methods, and the obtained results with abstract quantitative information, etc.

2. Why LSTM-CNN model is used for the focused problem among many other hybrid models is not clear? Why hybridization is used is also not clear.

3. Keywords should be written in alphabetical order.

4. Referencing style in the main text should be corrected. “Zhang et al. [2] (2023)”, “Yan et al. [3] (2023)”, “Su et al. [4] (2023)”, “Zhang et al. [5] (2023)”, “Chen et al. [6] (2023)”, “Wang et al. [7] (2023)”, “Sun et al. [8] (2023)”, “Yan et al. [9] (2023)”, and etc. should be corrected.

5. Current Introduction section is simple and misses some content related to the problem formulation. The Introduction does not provide relevant information for the article topic and it does not provide understandable information about the problem addressed. Furthermore, it does not provide contribution and motivation/need for such contribution. Introduction section seems voluminous, broad, and heterogeneous. The authors are supposed to focus on the main topic of the study and present a Literature Review in the form of tables in order to make research gaps and innovations easy to detect. Authoritative synthesis assessing the current state-of-the-art is absent. In general, the literature review is not sufficient. It is more of the type “researcher X did Y” rather than an authoritative synthesis assessing the current state-of-the-art. Where do we stand today? What approaches are there in the literature to model the problem? What are the main differences between them? What are their weaknesses and strengths? It is recommended to the authors use a table to summarize the overall information of the related literature and to present the main features of the paper. This table can be helpful to clarify what differences there are between the proposed approach and previous works.

6. Use correct equation numbers and do not use “following..”, “…as follows”.

7. Longer sentences should be broken out into smaller ones.

8. All of the values for the parameters of all algorithms selected for comparison are not given. How control parameters of the compared algorithms are set is not clear. A sensitivity analysis with respect to control parameters may be performed.

9. Some paragraphs are too long to read. They should be divided into two or more.

10. Blank character should be correctly used in place. See for example: “Training Set(MAE)”, “Training set(RMSE)”, “Training set(R-squared)”, “Testing Set(RMSE)”, “Testing Set(R-squared)”, and etc.

11. Name of Figure 8, Figure 9 should be corrected. Capitals and blank characters should be correctly used.

12. The state-of-the-arts and future research directions should be better categorized.

13. It is not clear if experimental results were obtained under the same experimental conditions. Are the simulations performed in the same situations? How do you guarantee a fair comparison?

14. More comparative experiments and some comparisons with other up-to-date methods should be addressed or added to back your claims to expand your experiments and analysis of results further.

15. It is not clear how the prediction of carbon emission results will be validated.

16. A wider discussion about key rules and how it is explored in literature should be presented.

17. Clarifying the limitations of the study allows readers to better understand the conditions under which the results should be interpreted. A clear description of the limitations of a study also shows that the researcher has a holistic understanding of his/her study. However, the authors do not demonstrate this in their paper. The authors should clarify the advantages and disadvantages of the methods. What are the limitations of the method(s) used in this paper? Please state the practical advantages and discuss the limitations of the research.

18. Additional comments about the reached results should be included. Graphics and charts need more explanation.

19. What are the other possible methodologies that can be used to achieve your objective in relation to this work?

20. Some more recommendations and conclusions should be discussed about the paper considering the experimental results. The Conclusion section is weak. Furthermore, there is not any discussion section about the results. The conclusion section needs revisions. It should briefly describe the findings of the study and some more directions for further research. The authors should describe academic implications, major findings, shortcomings, and directions for future research in the conclusion section. The conclusion in its current for is confused in general. Concerning Conclusion section, it would be better "Conclusions and Future Research", and it is strongly suggested to include future research of this manuscript. What will be happen next? What we supposed to expect from the future papers? So rewrite it and consider the following comments:

- Highlight your analysis and reflect only the important points for the whole paper.

- Mention the benefits.

- Mention the implication in the last of this section.

Author Response

Dear reviewer 1:

We would like to thank you for your careful reading, helpful comments, and constructive suggestions, which has significantly improved the presentation of our manuscript.

We have carefully considered all comments from your review and revised our manuscript accordingly. The manuscript has been double-checked, and the typos and grammar errors we found have been corrected. In the following section, we summarize our responses to each comments from your advice. We believe that our responses have well addressed all concerns from your suggestion. We hope our revised manuscript can be accepted for publication.

Here we will make the response point by point to the Comments and Suggestions that you have already give to us.

The abstract does not highlight the specifics of the proposed research or findings. The abstract should be more straightforward for the reader regarding the proposed method and its motivation. Especially the motivation and the novelty of the paper do not exist. The importance of the problem is not mentioned. The abstract should present some main points for the readers, such as the main contributions, the proposed method, the main problem, the benchmark tests and data, the comparative methods, and the obtained results with abstract quantitative information, etc.

Response:

Thank you for pointing out this problem in manuscript. We have rewrote Abstract in the revised manuscript. The revised abstract is as follows:

Global warming is a major environmental issue facing humanity, and the resulting climate change has severely affected the environment and daily lives of people. China attaches great importance to and actively responds to climate change issues. In order to achieve the "dual carbon" goal, it is necessary to clearly define the emission reduction path and scientifically predict future carbon emissions, which is the basis for setting emission reduction targets. To ensure the accuracy of data, this study applies the emission coefficient method to calculate the carbon emissions from energy consumption in 30 provinces, regions, and cities in China from 1997 to 2021. Considering the spatial correlation between different regions in China, we propose a new machine learning prediction model that incorporates spatial weighting, namely the LSTM-CNN combination model with spatial weighting. The spatial weighting explains the spatial correlation and the combined model is used to analyze the carbon emissions in the 30 provinces, regions, and cities of China from 2022 to 2035 under different scenarios. The results show that the LSTM-CNN combination model with 4 convolutional layers performs the best. Compared with other models, this model has the best predictive performance, with a MAE (Mean Absolute Error) of 8.0169, RMSE (Root Mean Square Error) of 11.1505, and R² (Coefficient of Determination) of 0.9661 on the test set. Based on different scenario predictions, it is found that most cities can achieve carbon peaking before 2030. Some cities need to adjust their development rates based on their specific circumstances in order to achieve carbon peaking as early as possible. This study provides a research direction for deep learning time series forecasting and proposes a new predictive method for carbon emission forecasting.

Why LSTM-CNN model is used for the focused problem among many other hybrid models is not clear? Why hybridization is used is also not clear.

Response:

We thank the reviewer for pointing out this issue. The causes of the problem are described in the following parts:

First, through the empirical verification of the LSTM model by Wang Y et al. [1], it is shown that lstm has strong nonlinear modeling ability and can deal with nonlinear relations and complex data distribution.

Secondly, BaiS et al. [2]proposed a network architecture named causal convolution based on the CNN model, which realized the introduction of longer historical information and improved prediction effect under the condition of constant time complexity.

Finally, Cnn-Lstm model has a remarkable effect on the fitting and prediction of time series data, and has made outstanding contributions to gold price [2], stock price [3], residential energy consumption [4] and other aspects.

In summary, compared with some classical time series models, Lstm-Cnn model not only has better fitting effect in samples, but also has higher prediction accuracy outside samples. The reason why this combined model is chosen is that LSTM can not only make use of the characteristics of LSTM for good fitting and prediction of time series data, but also make use of CNN model to extract the features of data more fully.

[1]Wang Y, Watanabe D, Hirata E, et al. Real-time management of vessel carbon dioxide emissions based on automatic identification system database using deep learning[J]. Journal of Marine Science and Engineering, 2021, 9(8): 871.

[2]Bai S, Kolter J Z, Koltun V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[J]. arXiv preprint arXiv:1803.01271, 2018.

[3]Livieris I E, Pintelas E, Pintelas P. A CNN–LSTM model for gold price time-series forecasting[J]. Neural computing and applications, 2020, 32: 17351-17360.

[4]Lu W, Li J, Li Y, et al. A CNN-LSTM-based model to forecast stock prices[J]. Complexity, 2020, 2020: 1-10.

[5]Kim T Y, Cho S B. Predicting residential energy consumption using CNN-LSTM neural networks[J]. Energy, 2019, 182: 72-81.

In response to the above suggestions, additions and adjustments have been made in the Related Work part of the article, and the contents after changes are as follows:

The reason why this study adopts the construction of an LSTM-CNN model is that the CNN-LSTM model has shown significant performance in fitting and predicting time series data, especially in areas such as gold prices [22], stock prices [23], and residential energy consumption [24]. Therefore, compared to some classical time series models, the LSTM-CNN model not only has better in-sample fitting performance but also higher prediction accuracy for out-of-sample data. The choice of this combined model allows us to take advantage of LSTM's ability to fit and predict time series data and CNN's ability to extract more comprehensive features from the data.

Keywords should be written in alphabetical order.

Response:

We thank the reviewer for pointing out this issue. We have modified the keywords accordingly and added keywords to enhance the relevance and visibility of the article. The final modification result is as follows:

Key words:Carbon emission prediction; Carbon peak; Deep Learning; LSTM-CNN model; Neural network; Scenario analysis

Referencing style in the main text should be corrected. “Zhang et al. [2] (2023)”, “Yan et al. [3] (2023)”, “Su et al. [4] (2023)”, “Zhang et al. [5] (2023)”, “Chen et al. [6] (2023)”, “Wang et al. [7] (2023)”, “Sun et al. [8] (2023)”, “Yan et al. [9] (2023)”, and etc. should be corrected.

Response:

Thank you for pointing out this problem in manuscript. We have rewrote Abstract in the revised manuscript. The revised abstract is as follows:

Current Introduction section is simple and misses some content related to the problem formulation. The Introduction does not provide relevant information for the article topic and it does not provide understandable information about the problem addressed. Furthermore, it does not provide contribution and motivation/need for such contribution. Introduction section seems voluminous, broad, and heterogeneous. The authors are supposed to focus on the main topic of the study and present a Literature Review in the form of tables in order to make research gaps and innovations easy to detect. Authoritative synthesis assessing the current state-of-the-art is absent. In general, the literature review is not sufficient. It is more of the type “researcher X did Y” rather than an authoritative synthesis assessing the current state-of-the-art. Where do we stand today? What approaches are there in the literature to model the problem? What are the main differences between them? What are their weaknesses and strengths? It is recommended to the authors use a table to summarize the overall information of the related literature and to present the main features of the paper. This table can be helpful to clarify what differences there are between the proposed approach and previous works.

Response:

We thank the reviewer for pointing out this issue. For the introduction part, we have re-written, and the final modification result is as follows:

For the Related Work, we have absorbed your suggestions and tips and divided all the literature into three parts, which are as follows:

(1) Relevant literature on the prediction of carbon emissions of individual provinces and regions;

(2) Relevant literature on carbon emission prediction for a certain spatial region;

(3) Comparison of relevant literature on energy and carbon emission prediction using machine learning algorithms

Our final modification results for the related work are as follows:

Use correct equation numbers and do not use “following..”, “…as follows”.

Response:

Thank you for pointing out this problem in manuscript. We have revised the corresponding formula label. The following is the result of our modification:

Longer sentences should be broken out into smaller ones.

Response:

We thank the reviewer for pointing out this issue.We have modified some long sentences into multiple short sentences, and some have retained the long sentence form because only long sentences can express the full meaning of the sentence.

All of the values for the parameters of all algorithms selected for comparison are not given. How control parameters of the compared algorithms are set is not clear. A sensitivity analysis with respect to control parameters may be performed.

Response:

Thank you for pointing out this problem in manuscript. To solve this problem, we have supplemented and adjusted the parameter setting in Section 5.2 of the paper. In addition, the control parameters of the comparison algorithm are also effectively explained through the text. Finally, our adjusted results are as follows:

Some paragraphs are too long to read. They should be divided into two or more.

Response:

We thank the reviewer for pointing out this issue. We have divided some long paragraphs into smaller ones according to the meaning of the paragraph. Our adjusted results are as follows:

Blank character should be correctly used in place. See for example: “Training Set(MAE)”, “Training set(RMSE)”, “Training set(R-squared)”, “Testing Set(RMSE)”, “Testing Set(R-squared)”, and etc.

Response:

Thank you for pointing out this problem in manuscript. We have corrected the corresponding part. The following is the revised result:

Name of Figure 8, Figure 9 should be corrected. Capitals and blank characters should be correctly used.

Response:

We thank the reviewer for pointing out this issue. We have corrected the corresponding part. The following is the revised result:

The state-of-the-arts and future research directions should be better categorized.

Response:

Thank you for pointing out this problem in manuscript. The most advanced literature has been classified in the literature review, and the future research will be placed in the outlook section at the end of the article. The corresponding outlook part of the paper is as follows:

It is not clear if experimental results were obtained under the same experimental conditions. Are the simulations performed in the same situations? How do you guarantee a fair comparison?

Response:

Thank you for pointing out this problem in manuscript. The model results were all carried out under the same environment, and our models were all tested under the environment of Tesla T4 under google colab. In response to your questions, we have made corresponding adjustments in the article, and the results are as follows:

More comparative experiments and some comparisons with other up-to-date methods should be addressed or added to back your claims to expand your experiments and analysis of results further.

Response:

We thank the reviewer for pointing out this issue. In view of the above problems, the advantages of the model have been verified through relevant literature, and the corresponding part of the paper is as follows:

In addition, we put the comparison experiment of the model in the ablation experiment part, which makes the whole paper more logical and coherent and convenient for readers to compare.

15、It is not clear how the prediction of carbon emission results will be validated.

Response:

Thank you for pointing out this problem in manuscript. In view of the above problems, because we are forecasting future data, there is no relatively clear indicator for reference. However, we can infer that the model has a stronger fitting ability to the data outside the sample based on the excellent evaluation indexes of the training set and the test set in the ablation experiment mentioned above.

16、A wider discussion about key rules and how it is explored in literature should be presented.

Response:

We thank the reviewer for pointing out this issue. To be honest, I really don’t know about the content of key rules you proposed in the original text, and would you please to give me more details about it. Thanks a lot.

17、Clarifying the limitations of the study allows readers to better understand the conditions under which the results should be interpreted. A clear description of the limitations of a study also shows that the researcher has a holistic understanding of his/her study. However, the authors do not demonstrate this in their paper. The authors should clarify the advantages and disadvantages of the methods. What are the limitations of the method(s) used in this paper? Please state the practical advantages and discuss the limitations of the research.

Response:

Thank you for pointing out this problem in manuscript. In view of the above problems, a supplementary explanation has been made in the conclusion of the article.

The advantages of this paper are:

(1) the uniqueness of the model. In previous studies, only traditional statistical models were used to study the influencing factors of carbon emissions, or only machine learning algorithm models were used to predict carbon emissions. In this paper, traditional spatial statistical ideas were combined with machine learning algorithms, and a new machine learning prediction model including spatial weights was proposed for the study of carbon emissions. It not only provides a new research method for carbon emission prediction, but also provides a new research direction for machine learning time prediction analysis.

(2) The universality of the research objects. Previous studies generally focus on a certain region or a single city, but the research object of this paper is 30 provinces in China, which is more comprehensive than other studies.

(3) Novelty of the model. In this paper, considering the vast territory of China, there are great differences in resource endowment, population size, economic development level and industrial structure in different regions, and there are also great differences in energy structure. There are spatial correlations among different regions, so spatial weights are introduced in the model to explain the spatial correlations.

Limitations of this paper:

This paper uses deep learning algorithm to predict carbon emissions.The internal mechanism of deep learning algorithm is very complex, and the judgment results and inference results of the model are often difficult to explain. As a result, the model can only be used to predict carbon emissions, but cannot be used to analyze the influencing factors of carbon emissions.
The independent variables in the model are all social and economic data. In scenario analysis, the change rate of each independent variable needs to be set first, the independent variable is predicted, and then the carbon emission is predicted. The setting of the rate of change of the independent variable is highly subjective, and the longer the prediction time, the greater the deviation degree between the predicted data and the real data. In this paper, according to the development plan published by the country, the change rate of each variable is carefully set to minimize the data deviation.

The corresponding part of the paper is as follows:

18、Additional comments about the reached results should be included. Graphics and charts need more explanation.

Response:

We thank the reviewer for pointing out this issue. In view of the above problems, some graphs and charts in this paper are explained. The corresponding part of the paper is as follows:

What are the other possible methodologies that can be used to achieve your objective in relation to this work?

Response:

Thank you for pointing out this problem in manuscript. In addition to traditional time series models, integrating machine learning algorithms with traditional statistical methods has also become an option. The extraction of features and the search for the relationship between input variables have also become a conventional means to improve the model effect. In view of this problem, a brief explanation is given in the outlook section at the end of the article, corresponding parts are as follows:

20、Some more recommendations and conclusions should be discussed about the paper considering the experimental results. The Conclusion section is weak. Furthermore, there is not any discussion section about the results. The conclusion section needs revisions. It should briefly describe the findings of the study and some more directions for further research. The authors should describe academic implications, major findings, shortcomings, and directions for future research in the conclusion section. The conclusion in its current for is confused in general. Concerning Conclusion section, it would be better "Conclusions and Future Research", and it is strongly suggested to include future research of this manuscript. What will be happen next? What we supposed to expect from the future papers? So rewrite it and consider the following comments:

- Highlight your analysis and reflect only the important points for the whole paper.

- Mention the benefits.

- Mention the implication in the last of this section.

Response:

We thank the reviewer for pointing out this issue. In response to your suggestions, we have made corresponding modifications to the conclusion. The conclusion is divided into three parts: conclusion, policy recommendations, deficiencies and future prospects. The corresponding parts of the original text are as follows:

Author Response File: Author Response.docx

Reviewer 2 Report

Sustainability-2592373

Article: Research on carbon emission prediction of 30 provinces in China based on LSTM-CNN neural network combination model

Dear authors,

Thank you for submitting your research article titled " Research on carbon emission prediction of 30 provinces in China based on LSTM-CNN neural network combination model" to the Sustainability Journal.

The research titled "Research on Carbon Emission Prediction of 30 Provinces in China Based on LSTM-CNN Neural Network Combination Model" focuses on utilizing advanced machine learning techniques to predict carbon emissions in various provinces of China. The study employs a combined model that integrates Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) architectures. LSTM aids in capturing sequential dependencies and temporal patterns in data, while CNN enhances feature extraction from spatial data. By applying this hybrid approach, the research aims to provide accurate and efficient predictions of carbon emissions, aiding in better understanding and managing environmental impacts across different regions in China. But the following suggestions has been made to improve it’s over all expressions.

Title:

Authors are advised to change the title like “Coupling LSTM and CNN Neural Networks for Accurate Carbon Emission Prediction in 30 Chinese Provinces” instead of Research on

Carbon emission ..................................model.

Abstract:

The abstract provides an overview of the study on “carbon emission prediction through LSTM CNN Model in China”. Here are some suggestions to enhance the abstract:

The abstract currently lacks sufficient materials and methods description. To enhance the clarity of the methodology, we recommend incorporating additional details. Additionally, consider augmenting the abstract with pertinent keywords to improve its relevance and visibility.

Introduction: Please incorporate the following suggestions

· The paper lacks novelty. It would be beneficial to highlight the original contributions of the research in the concluding paragraphs of the introduction section.

· An explanation regarding the rationale for selecting the current study area would enhance the clarity of the introduction.

· To enhance the significance of the study, it is recommended to allocate at least half a page to explaining how carbon emission prediction can play role through LSTM CNN Model in China to achieving sustainability goals.

· Provide context and relevance?

· Specify research gap or objective?

· Streamline references?

Methodology:

· The present software and analysis techniques used are considered inadequate. Thus, it is suggested to incorporate more detailed methodology that outlines the proposed study's approach.

· All equations, Fig and tables as well as graph in the methodology section and results do not have references in the text. It is suggested to include the relevant references to these equations to enhance the clarity and accuracy of the manuscript.

Results:

· Tables, Figs and the discussion section seem lack of proper citation of literature in many places. It is recommended to include relevant references to strengthen the arguments and findings presented in the manuscript.

· Equations should be appropriately referenced within the text to ensure their clear integration and understanding.

· Some of the references used in the paper are outdated. It is suggested to update the references to include recent publications from 2010 to the present.

· The limitations of the study are not sufficiently explained. It is suggested to dedicate at least half a page to discussing the limitations of the work before the conclusion.

· The conclusion could be improved with additional information to provide a more comprehensive summary of the study's findings and implications.

· Please ensure that all numbered tables in the manuscript are referenced in the text.

· Kindly update the references in the manuscript according to the latest standards.

· Please ensure that the images and figures used in the manuscript are of high quality and have high pixel density. The use of low-quality images may compromise the visual impact of the figures and result in difficulties in interpreting the data. Hence, it is recommended to use high-resolution images that can be clearly viewed and interpreted

· Please carefully review the language and grammar of the manuscript to ensure clarity and coherence.

Please carefully review the language and grammar of the manuscript to ensure clarity and coherence.

Author Response

Dear reviewer 2:

We would like to thank you for your careful reading, helpful comments, and constructive suggestions, which has significantly improved the presentation of our manuscript.

Here we will make the response point by point to the Comments and Suggestions that you have already give to us.

1、Title:

Authors are advised to change the title like “Coupling LSTM and CNN Neural Networks for Accurate Carbon Emission Prediction in 30 Chinese Provinces” instead of Research on

Carbon emission ..................................model.

Response:

Thank you for pointing out this problem in manuscript. We agreed and adjusted accordingly. The final title will be changed to “Coupling LSTM and CNN Neural Networks for Accurate Carbon Emission Prediction in 30 Chinese Provinces”.

2、Abstract:

The abstract provides an overview of the study on “carbon emission prediction through LSTM CNN Model in China”. Here are some suggestions to enhance the abstract:

Response:

We thank the reviewer for pointing out this issue. We modified the abstract and added more details on the basis of the original description:

Added the description of the background and motivation of the research method.

(2) Data description was added to the research results.

(3) Relevant keywords are added. The final modification result is as follows:

Key words:Carbon emission prediction; Carbon peak; Deep Learning; LSTM-CNN model; Neural network; Scenario analysis

3、Introduction: Please incorporate the following suggestions

The paper lacks novelty. It would be beneficial to highlight the original contributions of the research in the concluding paragraphs of the introduction section.
An explanation regarding the rationale for selecting the current study area would enhance the clarity of the introduction.
To enhance the significance of the study, it is recommended to allocate at least half a page to explaining how carbon emission prediction can play role through LSTM CNN Model in Chinato achieving sustainability goals.
Provide context and relevance?
Specify research gap or objective?
Streamline references?

Response:

Thank you for pointing out this problem in manuscript. We have modified the introduction part and sorted out the writing ideas. (1) In the last paragraph of the introduction, the original contribution of this study is systematically expounded. (2) The research background and motivation are enriched, and the content is simplified and modified. (3) We have revised the writing ideas of the literature review. The revised introduction is as follows:

The revised related work is as follows:

4、Methodology:

The present software and analysis techniques used are considered inadequate. Thus, it is suggested to incorporate more detailed methodology that outlines the proposed study's approach.
All equations, Fig and tables as well as graphin the methodology section and results do not have references in the text. It is suggested to include the relevant references to these equations to enhance the clarity and accuracy of the manuscript.

Response:

We thank the reviewer for pointing out this issue. We have revised the methodology section.

We have described the various methods in more detail. The following is the result of our modification:

We have added the original references cited in the methodology section.The following is the result of our modification:

5、Results:

Tables, Figs and the discussion section seem lack of proper citation of literature in many places. It is recommended to include relevant references to strengthen the arguments and findings presented in the manuscript.
Equations should be appropriately referenced within the text to ensure their clear integration and understanding.
Some of the references used in the paper are outdated. It is suggested to update the references to include recent publications from 2010 to the present.
The limitations of the studyare not sufficiently explained. It is suggested to dedicate at least half a page to discussing the limitations of the work before the conclusion.
The conclusioncould be improved with additional information to provide a more comprehensive summary of the study's findings and implications.
Please ensure that all numbered tables in the manuscript are referenced in the text.
Kindly update the references in the manuscript according to the latest standards.
Please ensure that the imagesand figures used in the manuscript are of high quality and have high pixel density. The use of low-quality images may compromise the visual impact of the figures and result in difficulties in interpreting the data. Hence, it is recommended to use high-resolution images that can be clearly viewed and interpreted
Please carefully review the language and grammarof the manuscript to ensure clarity and coherence.

Response:

Thank you for pointing out this problem in manuscript. In response to your suggestions, we have made the following adjustments:

(1) References are added where references can be cited. The following is the result of our modification:

(2)We have updated the references, but have retained some of the pre-2010 literature, which is the original literature (book) of some research methods (such as IPCC guidelines and spatial econometrics), for easy access by readers. The following is the result of our modification:

We add the shortcomings of this paper to the last part of the conclusion: shortcomings and future prospects.

(a) Deep learning algorithm is applied to predict carbon emissions in this paper. The internal mechanism of deep learning algorithm is very complex, and the judgment results and inference results of the model are often difficult to explain. As a result, the model can only be used to predict carbon emissions, but cannot be used to analyze the influencing factors of carbon emissions.

(b) The independent variables in the model are all socio-economic data. In scenario analysis, the change rate of each independent variable should be set first, the independent variable should be predicted, and then the carbon emission should be predicted. The setting of the rate of change of the independent variable is highly subjective, and the longer the prediction time, the greater the deviation degree between the predicted data and the real data. In this paper, according to the development plan published by the country, the change rate of each variable is carefully set to minimize the data deviation.

The limitations of this study are as follows:

In order to make the conclusion more clear and rich, we have rewritten the conclusion. The conclusion is divided into three parts: conclusion, policy recommendations, deficiencies and future prospects.

The corresponding parts of the original text are as follows:

It has been amended to ensure that all numbered tables in the manuscript are referenced in the text.
References in the manuscript have been updated according to the latest standards.

The following is the revised result:

(7) The clearest high-resolution images available have been uploaded to ensure that the images and graphics used in the manuscript are of high quality and have high pixel density.

The following is the revised result:

(8) The language and grammar in the manuscript have been more refined.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The paper is updated upon the reviewer comments. All of the required changes have been performed and concerns have been covered. This reviewer recommends that this revision be accepted in this form.

Article Menu

Coupling LSTM and CNN Neural Networks for Accurate Carbon Emission Prediction in 30 Chinese Provinces

Further Information

Guidelines

MDPI Initiatives

Follow MDPI