Application of Improved Particle Swarm Optimization SVM in Water Quality Evaluation of Ming Cui Lake
Round 1
Reviewer 1 Report
Accurately predicting the quality of water directly affects the living environment of the life body. In view of this, the authors’ motivation is correct and their efforts should be welcome by the science community. But to fit in the increasingly high quality of sustainability, a compulsory major revision is absolutely needed according to the following points.
1、 The writing of this paper should be improved. E.g., in line 76, .Marwah and Amr. [14] proposed to use machine learning to predict groundwater quality. In line 130, “Showed in Figure 1”.
2、 The content discussed in the introduction needs to be supported by references, such as the paragraph 1 (line 35 to 42) and paragraph 5 (line 88-102).
3、 The fonts in the paper are inconsistent.
4、 The authors need to explicitly tell the reader the difference between classe 1-3 and categories 4-5 in Table 1.
5、 In Figure 1, w needs to be transposed, and there are errors in the diagram comments (“Figure 1. The S V M overview of Fig”). Revise other similar problems in the paper.
6、 Multiple tables are used in the paper to present the results, but quantitative analysis is insufficient (such as using data tables to compare different indicators).
7、 As shown in Fig 9, the SVM achieved the best performance with an accuracy of 96%, which is higher than the optimized CPOS-SVM, So where is the innovation in the optimization model
Author Response
May 24, 2023
Dear Editor and reviewers,
Thank you very much for your kind letter concerning our manuscript ID 2410420, entitled “Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake”. We have carefully revised the manuscript in view of the constructive and helpful comments from the editors and reviewers as outlined in details below. These comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have also corrected some minor errors.
Please find attached our detailed responses to our reviewers’ comments. Their comments are highlighted in blue. Modified parts in the paper are colored with red.
Sincere thanks to the editors and the reviewers for giving us such valuable suggestions. Thank you for your consideration!
Sincerely,
Zunyang Zhang (on behalf of Cheng Yang, Qiao Qiao, Xuesheng Li, Fuping Wang, Chengcheng Li)
==========================================
Response to comments from the editor and reviewers.
Response to the Assistant Editor:
Comments to the Author:
Please use the version of your manuscript found at the above link for your revisions.
(I) Please check that all references are relevant to the contents of the manuscript.
(II) Any revisions to the manuscript should be marked up using the “Track Changes” function if you are using MS Word/LaTeX, such that any changes can be easily viewed by the editors and reviewers.
(III) Please provide a cover letter to explain, point by point, the details of the revisions to the manuscript and your responses to the referees’ comments.
(IV) If you found it impossible to address certain comments in the review reports, please include an explanation in your appeal.
(V) The revised version will be sent to the editors and reviewers.
Response: Thanks for reviewing and processing the manuscript. We have realized our mistakes and modified them in the manuscript.
We have modified and improved the whole article, such as Spaces and punctuation marks, and made a detailed check of the article.
Response to the reviewers:
Reviewer #1:
Comment: Accurately predicting the quality of water directly affects the living environment of the life body. In view of this, the authors’ motivation is correct and their efforts should be welcome by the science community. But to fit in the increasingly high quality of sustainability, a compulsory major revision is absolutely needed according to the following points.
Response: We are very grateful to the reviewer for reviewing the manuscript. And we have the following amendments and explanations for the reviewer’s comments.
Comment 1. The writing of this paper should be improved. E.g., in line 76, .Marwah and Amr. [14] proposed to use machine learning to predict groundwater quality. In line 130, “Showed in Figure 1”.
Response 1. Thank you for your excellent suggestions. We have made modifications to the places you mentioned.
Comment 2. The content discussed in the introduction needs to be supported by references, such as the paragraph 1 (line 35 to 42) and paragraph 5 (line 88-102).
Response 2. We are very grateful to the reviewer for giving such valuable suggestions. In the revised version, we have modified the manuscript according to the Reviewer's suggestions. In these two parts, we have added corresponding references to support this section.
Comment 3. The fonts in the paper are inconsistent.
Response 3. Thank you for your recommendation suggestion. We have made modifications to the font format in the manuscript.
Comment 4. The authors need to explicitly tell the reader the difference between class 1-3 and categories 4-5 in Table 1.
Response 4. Thank you for your recommendation suggestion. We have modified categories 4-5 in Table 1 to class 4-5.
Comment 5. In Figure 1, w needs to be transposed, and there are errors in the diagram comments (“Figure 1. The S V M overview of Fig”). Revise other similar problems in the paper.
Response 5. Thanks to the reviewer for pointing out the mistakes in the article. We have made modifications to w in Figure 1 and transposed it. See Figure 1 for details.
Comment 6. Multiple tables are used in the paper to present the results, but quantitative analysis is insufficient (such as using data tables to compare different indicators).
Response 6. We would like to thank the reviewer for the positive comments. We have made modifications to this issue. The specific modifications are as follows:
Figure 7, 8, 9, and 10 respectively show the specific classification results of CPOS-SVM, POS-SVM, POS-BP, and SVM algorithms. It can be seen from the figure that the SVM algorithm has the highest accuracy, which is because the SVM algorithm has not undergone data sparsity processing and retains all features of the original data. However, it can be seen from Figure 5 that its training time is greatly increased. The accuracy of the CPOS-SVM algorithm is slightly affected, but the training speed is greatly improved. This is because the characteristics of the original data are not affected when the CPOS-SVM algorithm is sparse, while the POS-SVM and POS-BP algorithm have advantages in both accuracy and training speed.
Comment 7. As shown in Fig 9, the SVM achieved the best performance with an accuracy of 96%, which is higher than the optimized CPOS-SVM, So where is the innovation in the optimization model.
Response 7. Thank you for your recommendation suggestions. In the water quality evaluation of Mingcuihu Lake, compared with SVM, POS-SVM and POS-BP, the training of the CPOS-SVM model can be completed in 2S, and does not increase with the increase of the amount of data, and the accuracy remains at about 94%, indicating that the CPOS-SVM model has good application value in the water quality evaluation.
Special thanks to you for your good comments.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Journal: Sustainability
Manuscript ID: 2410420
Title: Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake
Dear Editor,
I have read the manuscript entitled " Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake " submitted to Sustainability.
Abstract/ It is necessary to rewrite the abstract section. You should include more details. Some quantitative results should be included.
Introduction/ The introduction section of the study should be rewritten. There is no coherence between the paragraphs. Sufficient sources have not been provided, and the originality of the study has not been emphasized enough. What is the main innovative aspect of the research study? How is it different and important from similar studies?
Discussion/ Comprehensive comparison of the research findings with the relevant literature is missing.
Conclusion/ Rewrite the conclusion section. Some quantitative results should be included. The purpose of the study should be provided in this section.
The relevant citations from the journal Sustainability are strongly recommended
Line 5: China1: China
Line 15: “The quality of water quality” delete quality
Line 30: Particle Swarm Optimization
Line 37, 42, 51, 68: Should be cited
Line 43: “many factor” Please explain what these are by citing
Line 44: Delete “quality”
Line 82-87: There is no coherence between this paragraph and the preceding or following paragraphs.
Line 88-102: It should be provided under the heading of “Study Area”.
Line 175-178, 179-183, 216-220, 229-232, 267-272, 273-276, 312-315: The sentences are too long and complicated. They are quite difficult to understand. Please express them more clearly with shorter sentences.
Line 199: What is the “Particle Population Algorithm”.
Line 267: There is an incomplete sentence.
Line 273-276: It is not clear how the data is divided into training and test sets. Please include basic statistics and time series related to the data. Calculate performance statistics for the modeling results and present them as a table.
English very difficult to understand/incomprehensible
Author Response
May 24, 2023
Dear Editor and reviewers,
Thank you very much for your kind letter concerning our manuscript ID 2410420, entitled “Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake”. We have carefully revised the manuscript in view of the constructive and helpful comments from the editors and reviewers as outlined in details below. These comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have also corrected some minor errors.
Please find attached our detailed responses to our reviewers’ comments. Their comments are highlighted in blue. Modified parts in the paper are colored with red.
Sincere thanks to the editors and the reviewers for giving us such valuable suggestions. Thank you for your consideration!
Sincerely,
Zunyang Zhang (on behalf of Cheng Yang, Qiao Qiao, Xuesheng Li, Fuping Wang, Chengcheng Li)
==========================================
Response to comments from the editor and reviewers.
Response to the Assistant Editor:
Comments to the Author:
Please use the version of your manuscript found at the above link for your revisions.
(I) Please check that all references are relevant to the contents of the manuscript.
(II) Any revisions to the manuscript should be marked up using the “Track Changes” function if you are using MS Word/LaTeX, such that any changes can be easily viewed by the editors and reviewers.
(III) Please provide a cover letter to explain, point by point, the details of the revisions to the manuscript and your responses to the referees’ comments.
(IV) If you found it impossible to address certain comments in the review reports, please include an explanation in your appeal.
(V) The revised version will be sent to the editors and reviewers.
Response: Thanks for reviewing and processing the manuscript. We have realized our mistakes and modified them in the manuscript.
We have modified and improved the whole article, such as Spaces and punctuation marks, and made a detailed check of the article.
Reviewer #2:
Comment: I have read the manuscript entitled " Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake " submitted to Sustainability.
Abstract/ It is necessary to rewrite the abstract section. You should include more details. Some quantitative results should be included.
Introduction/ The introduction section of the study should be rewritten. There is no coherence between the paragraphs. Sufficient sources have not been provided, and the originality of the study has not been emphasized enough. What is the main innovative aspect of the research study? How is it different and important from similar studies?
Discussion/ Comprehensive comparison of the research findings with the relevant literature is missing.
Conclusion/ Rewrite the conclusion section. Some quantitative results should be included. The purpose of the study should be provided in this section.
The relevant citations from the journal Sustainability are strongly recommended:
Line 5: China1: China
Line 15: “The quality of water quality” delete quality
Line 30: Particle Swarm Optimization
Line 37, 42, 51, 68: Should be cited
Line 43: “many factor” Please explain what these are by citing
Line 44: Delete “quality”
Line 82-87: There is no coherence between this paragraph and the preceding or following paragraphs.
Line 88-102: It should be provided under the heading of “Study Area”.
Line 175-178, 179-183, 216-220, 229-232, 267-272, 273-276, 312-315: The sentences are too long and complicated. They are quite difficult to understand. Please express them more clearly with shorter sentences.
Line 199: What is the “Particle Population Algorithm”.
Line 267: There is an incomplete sentence.
Line 273-276: It is not clear how the data is divided into training and test sets. Please include basic statistics and time series related to the data. Calculate performance statistics for the modeling results and present them as a table.
Response: Thank you for your recommendation and good suggestions, we have corrected the problems you pointed out. Below, we will respond to the questions raised by the reviewers one by one.
We have revised the abstract according to the reviewer's suggestions and added quantitative results. Modify to:
In the water quality evaluation of Mingcuihu Lake, compared with SVM, POS-SVM and POS-BP, the training of the CPOS-SVM model can be completed in 2S, and does not increase with the increase of the amount of data, and the accuracy remains at about 94%, indicating that the CPOS-SVM model has good application value in the water quality evaluation.
We have made significant modifications to the introduction section to increase coherence between paragraphs. In addition, the introduction adds the originality and main innovative aspects of this study. Please refer to the introduction for detailed information. Some modifications are as follows:
The innovation points of this paper are as follows: 1. Aiming at the problem of geometric increase of training time caused by excessive training set of SVM model, the concept of pareto optimal solution is put forward to sparse training set data, which can improve the training speed without affecting the prediction accuracy. 2. The selection of kernel parameters and penalty factors of SVM will affect the classification accuracy of support vector machine. To solve the problem of difficult selection of hyperparameters, particle swarm optimization algorithm is used in this paper to select hyperparameters and improve the accuracy of the model. By accurately assessing water quality classification, the degree and source of water pollution can be accurately assessed, the sources of water pollution can be discovered and eliminated in time, and the water resources and ecological environment can be effectively protected, thus promoting environmental protection. At the same time, accurate water quality assessment method can improve the utilization efficiency of water resources, ensure the supply quality of water resources, improve the green technology and cost effectiveness of environmental governance, so as to achieve sustainable development.
We have made modifications to the conclusion to include quantitative results. Please refer to the conclusion for the revised content. Modify as follows:
The CPOS-SVM algorithm proposed in this paper is to dilute samples through Pareto solution when training SVM model on large-scale sample data, eliminate a large number of sample redundant data, improve training efficiency, improve training speed and shorten training time. At the same time, after sparse processing of samples by Pareto solution, the accuracy of the model can remain unchanged, indicating that the features of samples are not lost in the process of sparse, and the algorithm can well maintain the features of original data. The penalty parameter and kernel function of SVM algorithm can guarantee the classification accuracy of SVM. Taking the water quality test data of Mingcuihu Lake from 2014 to 2021 as the research object, different model parameters were set, including four algorithms, and the POS-SVM model, POS-BP model and SVM model were compared. The results show that the training time of the proposed model is the shortest, only 2S, and does not slow down with the increase of data volume. The prediction accuracy is also about 94%, which proves that the algorithm has high application value.
Line 5: We have modified “China1” in Line 5 to “China”.
Line 15: We have removed the “The quality of water quality” from Line 15.
Line 30: We have modified the Particle Swarm Optimization in Line 30.
Line 37, 42, 51, 68: We have added references to lines 37, 42, 51, and 68.
Line 43: The “many factor” in Line 43 has been explained. Modify to: Water environmental quality is affected by many factors, such as organic pollutants, harmful byproducts, pathogenic microorganisms, iron, manganese substances, ammonia nitrogen pollutants, generally measured ammonia nitrogen content in water, dissolved oxygen, permanganate content, total phosphorus, total nitrogen to judge the grade of water quality.
Line 44: We have removed the “quality” in Line 44.
Line 82-87: We have made modifications to Line 82-87 to increase coherence between paragraphs. Modify as follows: Hítalo Tobias Lôbo Lopes used nonlinear model to predict data, but the accuracy rate was not high[22]. Firstly, he correlation indexes of factors affecting water quality were extracted, and the four indexes with the greatest correlation were found out as input of particle swarm optimization multi-classification SVM algorithm to evaluate the quality of river water environment. Although this algorithm can ignore the influencing factors with little correlation, it also ignores the essential characteristics of some original data in this process, which may cause deviations in specific working conditions.
Line 88-102: It should be provided under the heading of “Study Area”.
We have made modifications to the sentences in Lines 175-178, 179-183, 216-220, 229-232, 267-272, 273-276, 312-315. Modify as follows:
Line 175-178：SVM algorithm is very suitable for training with small sample size. However, the computational efficiency of the SVM model is problematic when large sample sizes are used. Therefore, the amount of data should be reduced as far as possible, but the characteristics of data will inevitably be lost in the process of sparse data, which poses challenges in the process of sparse data and retention of its characteristics.
Line 179-183：A Pareto solution, also called a non-dominant or undominated solution, occurs when there are multiple targets that conflict with each other or cannot be compared. In this situation, the same solution may be the best for one target but the worst for another. When any target is changed simultaneously, it will weaken at least one other target. These solutions are known as non-dominant or Pareto solutions.
Line 216-220：The general method for optimizing SVM parameters is circular validation. The basic approach involves dividing the original dataset into two parts: a test set and a validation set. The parameters and are then varied over their respective value ranges, and a classifier is trained using each combination of and . The trained classifier is then evaluated using the validation set to obtain a classification accuracy score, which serves as the performance index for evaluating the classifier's quality. The parameter combination that produces the highest performance index is identified as the best parameter combination.
Line 229-232：To speed up training, the original data is processed sparsely using the Pareto optimal solution principle. To maintain the characteristics of the original training sample, the support vector machine solves the hyperplane by placing the support vector at the edge of the dataset. This results in obtaining the Pareto optimal solution set and the worst solution set corresponding to the water quality level.
Line 267-272：To test the validity and reliability of the CPOS-SVM model for water quality prediction, a total of 480 sets of data from Mingcuihu Lake in recent 8 years are taken as research objects. In order to take into account the commonness and individuality of samples as training samples to train the cognitive ability and generalization ability of the network, sample data are obtained by using the difference between two adjacent water quality standards in the National Surface Water Environmental Quality Standards. A total of 140 groups, 540 groups, 1040 groups and 1540 groups of data were generated, and a total of 620 groups, 1020 groups, 1520 groups and 2020 groups.
Line 273-276：The sample data was randomly divided into test set and training set, in which 50 groups of monitoring data were randomly extracted and 100 groups of sample data were randomly extracted by interpolation generation as test sets to detect the accuracy of the training model, and the remaining data were used as training sets for the training model.
Line 312-315：The model's operation time varies greatly. SVM algorithm is the fastest due to its directly given parameters, eliminating the need for the pos algorithm to search. The second fastest is the CPOS-SVM algorithm, which requires data preprocessing for small datasets and a small reduction in redundant data. As data size increases, excess data is eliminated after preprocessing, reducing sample data size and resulting in similar running times. Other algorithms experience more significant increases in running time as data quantity increases.
Line 199：Modify “Particle Population Algorithm” in Line 199 to “Particle Swarm Optimization Algorithm”.
Line 267: We have made modifications to the incomplete presence of Line 267. Modify as follows: To test the validity and reliability of the CPOS-SVM model for water quality prediction, a total of 480 sets of data from Mingcuihu Lake in recent 8 years are taken as research objects.
Line 273-276: We have made modifications to the issues mentioned in Line 273-276 and added Table 2. The specific modifications are as follows: To test the validity and reliability of the CPOS-SVM model for water quality prediction, a total of 480 sets of data from Mingcuihu Lake in recent 8 years are taken as research objects. In order to take into account the commonness and individuality of samples as training samples to train the cognitive ability and generalization ability of the network, sample data are obtained by using the difference between two adjacent water quality standards in the National Surface Water Environmental Quality Standards. A total of 140 groups, 540 groups, 1040 groups and 1540 groups of data were generated, and a total of 620 groups, 1020 groups, 1520 groups and 2020 groups. The sample data was randomly divided into test set and training set, in which 50 groups of monitoring data were randomly extracted and 100 groups of sample data were randomly extracted by interpolation generation as test sets to detect the accuracy of the training model, and the remaining data were used as training sets for the training model. The specific distribution of data is shown in Table 2.
Table 2 Training model data partitioning
|
Number of data sets |
Number of training sets |
Number of test set groups |
|
620 groups |
470 |
150 |
|
1020 groups |
870 |
150 |
|
1520 groups |
1370 |
150 |
|
2020 groups |
1870 |
150 |
Once again, thank you very much for your comments and suggestions.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The manuscript is in good shape for publication. I have no other comments.
minor editing of English language required.
Author Response
Dear Editor and reviewers,
Thank you very much for your kind letter concerning our manuscript ID 2410420, entitled “Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake”. We have carefully revised the manuscript in view of the constructive and helpful comments from the editors and reviewers as outlined in details below. These comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have also corrected some minor errors.
Please find attached our detailed responses to our reviewers’ comments. Their comments are highlighted in blue. Modified parts in the paper are colored with red.
Sincere thanks to the editors and the reviewers for giving us such valuable suggestions. Thank you for your consideration!
Sincerely,
Zunyang Zhang (on behalf of Cheng Yang, Qiao Qiao, Xuesheng Li, Fuping Wang, Chengcheng Li)
==========================================
Author Response File: Author Response.pdf
Reviewer 2 Report
Journal: Sustainability
Manuscript ID: 2410420
Title: Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake
Dear Editor,
I have read the revised manuscript entitled " Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake " submitted to Sustainability.
The corrections that I mentioned in the initial evaluation report and provided below in bullet points have not been satisfactorily implemented.
Abstract/ It is necessary to rewrite the abstract section. You should include more details. Some quantitative results should be included.
Discussion/ Comprehensive comparison of the research findings with the relevant literature is missing.
Conclusion/ Rewrite the conclusion section. Some quantitative results should be included. The purpose of the study should be provided in this section.
Line 88-102: It should be provided under the heading of “Study Area”.
The sentences are too long and complicated. They are quite difficult to understand. Please express them more clearly with shorter sentences.
Extensive editing of English language required
Author Response
Dear Editor and reviewers,
Thank you very much for your kind letter concerning our manuscript ID 2410420, entitled “Application of improved particle swarm optimization SVM in water quality evaluation of Mingcuihu Lake”. We have carefully revised the manuscript in view of the constructive and helpful comments from the editors and reviewers as outlined in details below. These comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have also corrected some minor errors.
Please find attached our detailed responses to our reviewers’ comments. Their comments are highlighted in blue. Modified parts in the paper are colored with red.
Sincere thanks to the editors and the reviewers for giving us such valuable suggestions. Thank you for your consideration!
Sincerely,
Zunyang Zhang (on behalf of Cheng Yang, Qiao Qiao, Xuesheng Li, Fuping Wang, Chengcheng Li)
==========================================
Response to the reviewers:
Reviewer #1:
Comment 1: Abstract/ It is necessary to rewrite the abstract section. You should include more details. Some quantitative results should be included.
Response: I have rewritten this part and it has been highlighted in red.
Comment 2:Discussion/ Comprehensive comparison of the research findings with the relevant literature is missing.
Response: In the introduction I have discussed the recent literature and contrasted it. In the part of experimental verification, I compared three other algorithms, among which POS-BP model and POS-SVM model are both models proposed by other scholars recently, and I have made a comparative evaluation of these models.
Comment 3:Conclusion/ Rewrite the conclusion section. Some quantitative results should be included. The purpose of the study should be provided in this section.
Response: I have rewritten this part and it has been highlighted in red.
Comment 4：Line 88-102: It should be provided under the heading of “Study Area”.
Response: The location of this section has been changed to the data sources section and is marked in the article.
Comment 5. The sentences are too long and complicated. They are quite difficult to understand. Please express them more clearly with shorter sentences.
Response：I have polished the full text in English, and the changes have been marked in red font.
Round 3
Reviewer 2 Report
Dear Editor,
The authors have made the necessary revisions as deemed appropriate. I am pleased to accept the article for publication.
Yours sincerely,
Minor editing of English language required
