Prediction of Service Life of Thermoplastic Road Markings on Expressways

Round 1

Reviewer 1 Report

Overall Assessment

The manuscript presents a comprehensive study on the prediction of the service life of thermoplastic road markings on expressways in Shandong Province. The study adopts a rigorous methodological approach, leveraging both multiple linear regression and LightGBM machine learning methods. The results are promising, with good predictive performance demonstrated through real-world example data.

The use of multiple linear regression, dominance analysis, and LightGBM models provides a robust analysis. However, it would be beneficial to include further information on the model selection process, such as why these specific models were chosen.

The paper gives a brief overview of the variables considered for the study. A more detailed discussion on the selection criteria for these variables might add more depth to the research. Additionally, it would be insightful to compare the findings with previous studies in this area.

The manuscript presents the predictive validity of the models but does not delve deeply into other evaluation metrics that might provide additional insights such as OBJ criteria. Also sensitivity analysis is missing. Some good citation examples are Smart prediction of liquefaction-induced lateral spreading (https://doi.org/10.1016/j.jrmge.2023.05.017); “Prediction of the resilient modulus of compacted subgrade soils using ensemble machine learning methods”

Recommendations

Consider including more diverse evaluation metrics.

A deeper discussion on the real-world implications would add value.

Minor editing required.

Author Response

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

1.Point-by-point response to Comments and Suggestions for Authors

Comments 1: However, it would be beneficial to include further information on the model selection process, such as why these specific models were chosen.

Response 1: Thank you for pointing this out. The manuscript adds information about model selection and explains why the three methods described in the paper were chosen. It was also compared with previous studies. Please see page 4, lines 173 to 209.

Comments 2: A more detailed discussion on the selection criteria for these variables might add more depth to the research. Additionally, it would be insightful to compare the findings with previous studies in this area.

Response 2: We agree with this comment. A more detailed discussion of the selection of variables has been added to the manuscript, explaining why they were chosen. In addition, we also added feedback for industry practitioners, which is the author's real information collected during the internship experience in Shandong High-speed Group. Finally, we compare these findings with previous research in this area, presenting some of our first studies. Please see page 7, lines 283 to 296.

Comments 3: The manuscript presents the predictive validity of the models but does not delve deeply into other evaluation metrics that might provide additional insights such as OBJ criteria. Also sensitivity analysis is missing. Consider including more diverse evaluation metrics.

Response 3: Agree. We refer to and cite some excellent literature (Raja and Shukla, 2021; Raja et al., 2023), adding a more diverse set of model evaluation indicators. For the sensitivity analysis, we added the importance analysis of the model variables. Please see page 15, lines 524 to 553 and page 16, lines 562 to 580.

Comments 4: A deeper discussion on the real-world implications would add value.

Response 4: Thank you for pointing this out. The Conclusion section in the manuscript adds to the discussion of the real-world implications of this research. Includes discussions on sustainability and traffic safety. Please see page 21, lines 686 to 693.

2.Response to Comments on the Quality of English Language

Point 1: Minor editing required.

Response 1: We have corrected some grammatical errors in the manuscript. All changes are highlighted using the track changes mode in MS Word.

Reviewer 2 Report

1. The Abstract could be improved by improving the problem statement which should be defined well and also fixing some grammatical errors.

2. The introduction section is too short, elaborate on more published works (even though the current study is focussed on China province) that are relevant for a better readership; as the present version of the manuscript is not adequate.

3. The literature review section should be revised by adding more research works using artificial intelligence and machine learning approaches related to sustainability fields. For example: 1. Recurrent neural network-based model for estimating the life condition of a dry gas pipeline  2. Remaining useful life prediction of a crude oil pipeline by means of deterioration curves. 3. A feed-forward back propagation neural network approach to predict the life condition of crude oil pipeline.

4. Plenty of studies used regression and ML-based methods and overcame them with various types of datasets. So, the authors should validate with a detailed comparative analysis.

5. The influencing factors were selected based on published works. Did the authors consider any real-time industrial practioners' feedback? In finalizing the parameters?

6. Data collection/ analysis/ pre-processing can be discussed in more detail

7. Figure 4 should be explained in more detail.

8. Figure 5 is not added to the manuscript. If Figure 6 is not cited in the text, cross-check the Figure numbers again all over the manuscript.

8. Consider adding the future scope of this research work in the conclusion section, so that it would be more motivating and challenging for the researchers to conduct new research works.

9. The References section should be updated by citing more relevant recently published works in the Introduction and Literature Review sections.

 

10. Please revisit the grammatical errors in the manuscript and proofread the manuscript for better English. Also, the arrangement of the manuscript as per the journal guidelines should be cross-checked.

The manuscript needs some revision with grammar-checking

Author Response

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

1.Point-by-point response to Comments and Suggestions for Authors

Comments 1: The Abstract could be improved by improving the problem statement which should be defined well and also fixing some grammatical errors.

Response 1: Thank you for pointing this out. We have revised the abstract part of the manuscript, improved the expression of the problem and fixed some grammatical errors.

Comments 2: The introduction section is too short, elaborate on more published works.

Response 2: We agree with this comment. We have added a more detailed discussion of research motivations in the Introduction, as well as some excellent recently published literature related to manuscript research. Please see page 2, lines 58 to 90.

Comments 3: The literature review section should be revised by adding more research works using artificial intelligence and machine learning approaches related to sustainability fields.

Response 3: Thank you for providing us with excellent literature for our reference. We have combined the Introduction section with the Literature review section, and added some research on AI and machine learning methods relevant to the field of sustainability. Please see page 2, lines 75 to 90.

Comments 4: Plenty of studies used regression and ML-based methods and overcame them with various types of datasets. So, the authors should validate with a detailed comparative analysis.

Response 4: We agree with this comment. The manuscript supplements information on the application of the model. In the correlation analysis of multiple linear regression, we made a comparison with previous studies and made it clear that the previous studies only carried out a single prediction when using multiple linear regression, and the prediction effect of the model was not good enough. This study analyzed the advantages of multiple linear regression, combined with dominance analysis to identify important influencing factors, so as to make the results more rigorous. At the same time, it is also used for prediction, and the prediction results are compared with the prediction results of machine learning methods. Please see page 4, lines 173 to 205.

Comments 5: The influencing factors were selected based on published works. Did the authors consider any real-time industrial practioners' feedback? In finalizing the parameters?

Response 5: We agree with this comment. The manuscript adds a more detailed discussion on the selection of variables, explaining why they were chosen. In addition, we did refer to feedback from industry practitioners in the process of writing the manuscript, which is information authentically captured from the author's internship experience at Shandong High-Speed Group. It was not reflected in the previous writing of the manuscript, but has now been added to it. Please see page 7, lines 287 to 296.

Comments 6: Data collection/ analysis/ pre-processing can be discussed in more detail.

Response 6: Thank you for pointing this out. The manuscript adds a more detailed discussion of data analysis/preprocessing. The collection of data is supplemented. The details about data cleansing in data preprocessing are added. Please see page 10, lines 386 to 389 and page 11, lines 398 to 401.

Comments 7: Figure 4 should be explained in more detail.

Response 7: Agree. A detailed explanation about Figure 4 is added to the manuscript. The function of Figure 4 is to determine whether the requirements of linear regression are met by judging the normality of the residual. Please see page 12, lines 439 to 441.

Comments 8: Figure 5 is not added to the manuscript. If Figure 6 is not cited in the text, cross-check the Figure numbers again all over the manuscript.

Response 8: Thank you for pointing this out. Figure 5 has been added to the manuscript, and Figure 6 is also referenced in the manuscript. It is now guaranteed that all images and tables are correctly mapped and referenced. Please see page 13 and page 17.

Comments 9: Consider adding the future scope of this research work in the conclusion section, so that it would be more motivating and challenging for the researchers to conduct new research works.

Response 9: Thank you for your suggestion. The Conclusion section of the manuscript increases the future scope of this research work. Including the expansion of the data set and the discussion of the universality of this study. Please see page 21, lines 694 to 697.

Comments 10: The References section should be updated by citing more relevant recently published works in the Introduction and Literature Review sections.

Response 10: Thank you for pointing this out. The Introduction section of the manuscript is supplemented with some citations from recently published relevant literature. Please see page 2, lines 75 to 90.

Comments 11: Please revisit the grammatical errors in the manuscript and proofread the manuscript for better English. Also, the arrangement of the manuscript as per the journal guidelines should be cross-checked.

Response 11: We agree with this comment. We have corrected some grammatical errors in the manuscript. We also made corrections to the manuscript format and other issues with reference to the journal guide.

2.Response to Comments on the Quality of English Language

Point 1: The manuscript needs some revision with grammar-checking.

Response 1: We have corrected some grammatical errors in the manuscript. All changes are highlighted using the track change mode in MS Word.

Reviewer 3 Report

Paper Summary

 

Currently, historical data and on-site surveys—particularly in the context of China—are heavily relied upon to determine the best time to maintain motorway road markings. This study's goal is to determine what influences the service life of thermoplastic road markings on motorways in Shandong Province, China, while taking into account both those motorways' unique characteristics and the local environment. Additionally, a scientific evaluation of the road markings' retroreflective coefficient's decay pattern will be done. With reference to the thermoplastic road markings on five motorways and potential affecting factors like the age of the marking and annual average daily traffic, authors gathered the retroreflective data for 12 consecutive months. 

 

The service life of the markings was forecast using a multiple linear regression. The dominance analysis was used to quantitatively analyse each explanatory factor's impact on the service life of the markings, and statistically significant variables were also found. Using LightGBM, a machine learning technique, a nonparametric prediction model was also created based on the examination of the relevance of influencing elements. According to the modelling results, LightGBM produces an R2 of 0.942, indicating that it is more accurate and easier to understand than the regression-based technique. Additionally, LightGBM beats MLR in terms of final validation accuracy, scoring 95.02% compared to MLR's 8%. The results are useful for motorway marker upkeep and driving safety.

Comments:

1. The motivation of the paper can be better clarified in the Introduction section. Currently, the last paragraph of Section I tries to highlight the existing research gap and motivation but it is not very clear. I suggest to extend the Introduction section to introduce the problem and the motivation. 

2. In Section 2, authors mention about the research gap that the current schemes do not consider factors such as climatic conditions. However, it is not clear if this issue is addressed in their proposed scheme. Highlight the relevant research gaps which are addressed by the proposed work. 

 

3. The motivation to use Multiple Regression in the proposed model needs clarification. In Table 1, there are several techniques that use this method. How the proposed technique differs from these techniques?

 

4. The trend in Figure 7 needs to be explained better. How much is the difference in the projected and actual value and what is the reason for it?

5. Conclusion section is too long. I suggest to move the details related to Results to the previous section by adding a subsection called Discussion on results.   

 

 

Fine.

Author Response

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

Point-by-point response to Comments and Suggestions for Authors

Thank you so much for the paper summary you provided. We have revised the abstract of the manuscript with reference to it.

Comments 1: The motivation of the paper can be better clarified in the Introduction section. I suggest to extend the Introduction section to introduce the problem and the motivation.

Response 1: Thank you for your advice. The Introduction to the manuscript adds a discussion of the motivation for the study, as well as some recently published literature related to the study of the manuscript. Please see page 2, lines 58 to 69.

Comments 2: Highlight the relevant research gaps which are addressed by the proposed work.

Response 2: Thank you for pointing this out. The manuscript adds a more detailed discussion of the variable selection process. We also highlight the gap between this study and previous studies. Two new variables are proposed in this study. In the expression of the influence of climate conditions, the more intuitive monthly average temperature and monthly average rainfall are used to express. Please see page 7, lines 283 to 296.

Comments 3: The motivation to use Multiple Regression in the proposed model needs clarification. In Table 1, there are several techniques that use this method. How the proposed technique differs from these techniques?

Response 3: Agree. The manuscript adds a description of the motivation for using a multiple linear regression model. First used in conjunction with the dominance analysis to identify important influencing factors, this can make the results more rigorous, which is an improvement on previous studies.  Secondly, comparison with machine learning predictions is also one of the motivations for applying multiple linear regression predictions. Last but not least, the important influencing factors obtained through multiple linear regression and dominance analysis provide the variable basis for machine learning methods. Please see page 4, lines 173 to 205.

Comments 4: The trend in Figure 7 needs to be explained better. How much is the difference in the projected and actual value and what is the reason for it?

Response 4: Thank you for pointing this out. A more detailed explanation of Figure 7 has been added to the manuscript, describing the difference between the predicted and actual values. The reasons for the large differences in values are discussed. Please see page 18, lines 590 to 599.

Comments 5: Conclusion section is too long. I suggest to move the details related to Results to the previous section by adding a subsection called Discussion on results.

Response 5: Thank you for the suggestion. The manuscript adds a section called "Discussion on results" and moves details related to the results to this section. In addition, a discussion of future research and the real-world implications of the research done is added to the conclusion. Please see page 19.

Round 2

Reviewer 2 Report

The authors could improve the literature review more, however, the authors addressed the main concerns raised, and now it is okay for a possible publication. 

Reviewer 3 Report

Authors have addressed comments of the last round and I suggest acceptance of the paper.