Study on the Prediction of the Uniaxial Compressive Strength of Rock Based on the SSA-XGBoost Model
Round 1
Reviewer 1 Report
Abstract
Line 13 ‘was produced’
Line 17 ‘as the output variable. A prediction model for the uniaxial compressive strength was proposed based on the SSA-XGBoost model’
Line 20 ‘were evaluated using the root mean’
Line 21 ‘and the variance interpretation factor’
Line 23 ‘strength of rock, for the dataset tested’ (this is a suggestion)
Introduction & elsewhere
Line 29 ‘engineering stability analysis’ 0. Remove the 0 or is this reference 1???
Where is reference 1 by Chen, X (2014)?
Line 37 ‘Currently, the empirical formula methodologies are the widely used methods of assessment’ (this is a suggestion)
Line 46 ‘strength. A subsequent empirical regression equation was derived which had good correlation with the dataset used.’
Reference 7 the first author is Zhang L and not Chen et al, please check and confirm these references, thank you
Line 49 ‘which produced good, predicted results.’
Line 49 ‘The empirical formula method has the advantage of simplicity of operation when predicting the uniaxial compressive strength of rock. However it is based on statistical principles in nature, has great uncertainty, and has a limited range of application. Furthermore, it is difficult to demonstrate the reasonableness of its prediction index selection.’
After reference 7 there is a change of format e.g., [8 to 10] on line 56. There must be a consistent method of referencing within the paper. Please check the instructions for the manuscript. Please apply consistently across the paper.
Line 57 have been used to generate predict rock uniaxial compressive strength. The artificial intelligence methodology offers considerable advantages Ma et al. [10]
Line 60 ‘predicted results’
Line 63 ‘complete rock’ should this read competent rock? Please confirm
Line 67 ‘results obtained’
Line 70 ‘velocity. The analysis used a multiple linear regression 70 (MLR) and least squares support vector machine (LS-SVM) methodology.’
Line 73 ‘using a stochastic forest model and compared’
Line 76 ‘These works all provide a good reference’
Line 78 ‘accuracy and range applicability. Furthermore all of the suggested methods have high forecast risk, which has engineering implications for design looking to produce sustainable designs.’
Line 78 ‘When using artificial neural networks to forecast strength parameters, the inclusion of too many hidden layers for accuracy assessment can lead to problems such as overfitting and gradient disappearance [17].’ Please amend the reference also
Line 81 ‘To overcome the shortcomings of the current published research, the XGBoost algorithm was used to predict the uniaxial compressive strength of rock.’
Line 82 proposed by Dr. Chen et al. [18]
Line 84 ‘which overcomes the’
Line 91 ‘rockburst grade prediction. The authors the constructed the CRITIC-XGBoost prediction model’
Table 1 is an abridged data set after Wu et al. [23] please amend the title
Line 120 ‘The rational…’
Line 122 ‘were selected as’
Line 124 ‘significance when predicting the hydraulic’
Line 125 ‘The larger is the porosity,’
Line 131 ‘Generally, the higher is the uniaxial compressive strength of rock, the higher is the longitudinal wave velocity.’
Line 134 ‘significance when evaluating the uniaxial compressive’
Line 135 ‘It is both logical and credible to incorporate the index system when seeking to predict the uniaxial strength of rocks.’
Line 137 ‘A correlation analysis based’
Line 312 Should Figure 9 be Figure 7?
Line 327 ‘disadvantages of SSA-XGBoost with BPNN, RF, SVM, AdaBoost, RS, KNN prediction models and the empirical formula methods’. ‘It was necessary to evaluate the changes in algorithm performance before and after the sparrow search algorithm optimizes XGBoost parameters, the root mean square error (RMSE), correlation coefficient (R2), mean absolute error (MAE) and variance interpretation (VAF) are used to evaluate the prediction performance of all models. Thus equations (19)~(22) were used for calculations
Line 350 ‘The values are calculated’
Line 347 ‘The closer is the RMSE to the right, the smaller is the error’
Line 360 ‘After parameter optimization, the performance of the SSA-XGBoost model is better than that of XGBoost, where the accuracy of prediction results is further improved compared to XGBoost. The refined methods error is extremely low and thus for the range of materials tested, the accuracy of prediction is impressive.’
Line 365 ‘SSA-XGBoost has been shown to be a superior method’
Line 366 ‘this method can forecast the true uniaxial compressive strength of rock with a high degree of certainty’. This has significance potential for geotechnical engineering and indoor rock experiments.
Line 370 In this study, designed to evaluate of rock uniaxial compressive strength of rock types, the SSA-XGBoost prediction model was established.
Line 374 ‘models were compared and analysed,’
Line 377 ‘Based on the data, an XGBoost model was introduced, and a sparrow search algorithm was used to optimize’
In overall terms, this was an interesting and well-designed paper. There are some matters in the text (see the comments) which need addressed. The referencing needs a detailed check. The methodology and analysis are good and the paper, once amended and corrected should be send to the editor for final checking and publications. I enjoyed reading the text, thank you.
Author Response
Response to Reviewer 1 Comments
Dear Reviewer:
Thank you for your comments concerning our manuscript. Those comments are valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval. We addressed the comments of the reviewer (with green color) with responses as listed below, corresponding to necessary changes, additional sentences and short sections at proper places which are marked by blue color in the revised version of manuscript.
1-(Reviewer1): List some statements that need to be modified, such as: Line 13— ‘was produced’;Line 17—‘as the output variable. A prediction model for the uniaxial compressive strength was proposed based on the SSA-XGBoost model’;...
Response 1: According to the reviewer's suggestion, I revised the sentences listed by the reviewer in the manuscript and marked them in blue.
2-(Reviewer1): The format of a few references needs to be modified.
Response 2: According to the reviewer's suggestion, I revised the format of the references listed by the reviewer in the manuscript and marked them in blue.
We appreciate for your warm work earnestly, and hope that the correction will meet with approval. Once again, thank you very much for your comments and suggestions.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This study presented a new SSA-XGBoost model to predict the uniaxial compressive strength. The empirical formulas, XGBoost, SVM, RF, BPNN, KNN, PLSR, and other models were also adopted to verify the effectiveness of the SSA-XGBoost model by comparing RMSE, R2, MAE, and VAF. The results of this study showed that the SSA-XGBoost model could be used to predict the uniaxial compressive strength of rock based on experimental data. The following comments need to be further revised in the manuscript.
1. SSA-XGBoost model is the core of this study. It should be presented in detail, including the idea, algorithm, procedure, and parameter selection.
2 The input selection of the predicted model (SSA-XGBoost) is important to the output prediction. In this study, porosity, Schmidt rebound number, longitudinal wave velocity, and point load strength were regarded as the influence factor of UCS. Why were these factors selected? If other factors were selected, how about the result be? Please explain it briefly.
3 Experiment and experimental data are critical to building the prediction model. In the section Introduction, some key and current literature about rock tests (Zhen Li, Jiachen Liu, Huoxing Liu, Hongbo Zhao, Rongchao Xu, Filip Gurkalo. Stress distribution in direct shear loading and its implication for engineering failure analysis. International Journal of Applied Mechanics, 2023. https://doi.org/10.1142/S1758825123500369; Li, Z., Liu, J., Xu, R. et al. Study of grouting effectiveness based on shear strength evaluation with experimental and numerical approaches. Acta Geotech. 16, 3991–4005 (2021); ) should be presented and cited to enhance the integrality of the manuscript.
4. In Table 2, the value of the XGBoost model is different. Please add some explanation about the model performance.
Author Response
Response to Reviewer 2 Comments
Dear Reviewer:
Thank you for your comments concerning our manuscript. Those comments are valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval. We addressed the comments of the reviewer (with green color) with responses as listed below, corresponding to necessary changes, additional sentences and short sections at proper places which are marked by orange color in the revised version of manuscript.
1-(Reviewer2): SSA-XGBoost model is the core of this study. It should be presented in detail, including the idea, algorithm, procedure, and parameter selection.
Response 1: The SSA XGBoost model is the core of this study. The ideas, algorithms and procedures have been introduced in detail in the third part of the manuscript and provided with easy-to-understand flow charts. The parameter selection part has been related in the second part of the manuscript. Thank you!
2-(Reviewer2): The input selection of the predicted model (SSA-XGBoost) is important to the output prediction. In this study, porosity, Schmidt rebound number, longitudinal wave velocity, and point load strength were regarded as the influence factor of UCS. Why were these factors selected? If other factors were selected, how about the result be? Please explain it briefly.
Response 2: According to the suggestion of the reviewer, I revised 'Four parameters' in the text to' Consulting the availability of data acquisition, four most effective parameters are selected in this paper '. This makes the article more accurate.
3-(Reviewer2): Experiment and experimental data are critical to building the prediction model. In the section Introduction, some key and current literature about rock tests (Zhen Li, Jiachen Liu, Huoxing Liu, Hongbo Zhao, Rongchao Xu, Filip Gurkalo. Stress distribution in direct shear loading and its implication for engineering failure analysis. International Journal of Applied Mechanics, 2023. https://doi.org/10.1142/S1758825123500369; Li, Z., Liu, J., Xu, R. et al. Study of grouting effectiveness based on shear strength evaluation with experimental and numerical approaches. Acta Geotech. 16, 3991–4005 (2021); ) should be presented and cited to enhance the integrality of the manuscript.
Response 3: I carefully adopted relevant opinions and cited the listed references in the manuscript to make the article more complete.
[2]Li,Z.; Liu,J.; Liu,H.; Zhao,H.; Xu,R.; Gurkalo,F. Stress distribution in direct shear loading and its implication for engineering failure analysis. International Journal of Applied Mechanics. 2023, https://doi.org/10.1142/S1758825123500369
[3]Li,Z.; Liu,J.; Xu,R.; Liu,H.; Shi,W. Study of grouting effectiveness based on shear strength evaluation with experimental and numerical approaches. Acta Geotech. 2021. 16, 3991-4005.
4-(Reviewer2): In Table 2, the value of the XGBoost model is different. Please add some explanation about the model performance.
Response 4: According to the reviewer's suggestion, I added some descriptions of model performance in the upper part of Table 2 and marked them in orange.
The maximum number of iterations is a very important indicator. If it is too small, it may not converge. If it is too large, it may cause time waste. The depth of the tree will also affect the performance of the model to a certain extent. When it reaches the specified value, it will stop splitting to avoid infinite downward division. The appropriate learning rate can make the objective function converge to the local minimum in the appropriate time.
We appreciate for your warm work earnestly, and hope that the correction will meet with approval. Once again, thank you very much for your comments and suggestions.
Author Response File: Author Response.pdf
Reviewer 3 Report
In this paper, a new SSA-XGBoost optimizer prediction model is proposed to predict the uniaxial compressive strength of 290 rock samples. The study seems interesting. However, the paper also has the following shortcomings, it is recommended to modify :
1. Abstract. It is suggested to add some specific data in the abstract to make the article more convincing.
2. Introduction. Kindly, improve the literature review and diverse your literature; the literature review part is not sufficient and should be introduced more. Some latest references (in recent two years) should be cited.
3. Line 262. There may be some errors in the last paragraph in Section 3.2. The optimized variables are different with previous studies. Please clarify this.
4. Line 369. The conclusions section has to be revised by providing qualitative and quantitative aspects. The conclusion part it not justifying your work, suggest to add some specific data in the conclusion.
5. Quality of Figures 6-7 is poor, consider improving it in revision, Especially font and size.
Author Response
Response to Reviewer 3 Comments
Dear Reviewer:
Thank you for your comments concerning our manuscript. Those comments are valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval. We addressed the comments of the reviewer (with green color) with responses as listed below, corresponding to necessary changes, additional sentences and short sections at proper places which are marked by purple color in the revised version of manuscript.
1-(Reviewer3): Abstract. It is suggested to add some specific data in the abstract to make the article more convincing.
Response 1: Add the data description in the summary:
The results calculated by SSA-XGBoost model: R2=0.84, RMSE=19.85, MAE=14.79 and VAF=81.36, are the best among all prediction models.
2-(Reviewer3): Introduction. Kindly, improve the literature review and diverse your literature; the literature review part is not sufficient and should be introduced more. Some latest references (in recent two years) should be cited.
Response 2: I carefully adopted relevant opinions and cited the listed references in the manuscript to make the article more complete.
[4]Xie,S.; Han,Z.; Hu,H.; Lin,H. Application of a novel constitutive model to evaluate the shear deformation of discontinuity. Engineering Geology. 2022, 304, 106693.
[5]Xie,S.; Lin,H.; Duan,H. A novel criterion for yield shear displacement of rock discontinuities based on renormalization group theory. Engineering Geology. 2023, 314, 107008.
3-(Reviewer3): Line 262. There may be some errors in the last paragraph in Section 3.2. The optimized variables are different with previous studies. Please clarify this.
Response 3: The variable format has been modified in section 3.2 of the manuscript.
- (Reviewer3):Line 369. The conclusions section has to be revised by providing qualitative and quantitative aspects. The conclusion part it not justifying your work, suggest to add some specific data in the conclusion.
Response 4: According to the reviewer's suggestion, I added data description in the conclusion to make the article more convincing:
Compared with empirical formula methods and other machine learning prediction models, the SSA-XGBoost, XGBoost and BPNN models have good R2, RMSE, VAF and MAE values. Meanwhile, the SSA-XGBoost model (with higher R2 and VAF and lower RMSE and MAE, R2=0.84, VAF=81.36, RMSE=19.85 and MAE=14.79) can achieve the best prediction results, which indicated that the SSA-XGBoost model has the best generalization ability and more accurate prediction results and can solve the problem that other machine learning prediction models have lower accuracy in predicting different types of rocks.
5-(Reviewer3): Quality of Figures 6-7 is poor, consider improving it in revision, Especially font and size.
Response 5: The font size of some text in relevant pictures has been changed in the manuscript.
We appreciate for your warm work earnestly, and hope that the correction will meet with approval. Once again, thank you very much for your comments and suggestions.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The author revised the manuscript according to the comments. It can be accepted.
