A Machine Learning-Based Decision Support System for Predicting and Repairing Cracks in Undisturbed Loess Using Microbial Mineralization and the Internet of Things
Round 1
Reviewer 1 Report
1. The research gap and the motivation for the proposed method should be clearly described.
2. The analysis is done over a sampled dataset. The authors should clarify the knowledge elicitation ethics and process they used in the paper.
3. A clarification is required to justify the performance of the claims.
4. The implementation environment of the stated machine learning algorithms needs to be described clearly.
5. The result and discussion section does not lead to any potential outcomes to signify the notable contributions of the authors. Major applicability enhancement is required.
Proofreading is required.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Summary/Contribution: This work proposes a decision support system (DSS) that detects, predicts, and suggests crack healing methods. The system's results are correct. Engineers working on crack repair projects in undisturbed loess receive real-time guidance on where and how to apply microbial mineralization treatments based on expected crack sites and treatment success.
Comments/Suggestions:
1. I would like to request that you consider summarizing the related work section in tabular form as a way to better identify the limitations of the related works and emphasize the originality of your proposed approach.
2. Can you provide some examples of the types of data that are collected and analyzed by the Treatment Optimization Module to determine the most suitable treatment methods, dosages, and application timings for microbial mineralization technology in crack repairing?
3. How does the Decision-Making Module integrate the outputs of the crack detection, prediction, and treatment optimization modules to make informed decisions about crack detection and repairing?
4. Can you provide some examples of how the module combines the crack detection results, crack prediction probabilities, and treatment optimization recommendations to determine the most appropriate course of action for crack repair?
5. To further enhance the quality and impact of the research presented in this paper, I would like to suggest that the authors consider including a paragraph on the use of formal methods for the verification of AI-based techniques. Formal methods, which involve the use of mathematical models and logic to analyze and verify the correctness of systems, have become increasingly important in the development and validation of AI-based techniques. By utilizing formal methods, researchers can ensure that their techniques are robust, reliable, and free from errors or biases.
6. Some relevant references related to this topic and that the authors may want to consider include:
a. https://ieeexplore.ieee.org/abstract/document/9842406
b. https://incose.onlinelibrary.wiley.com/doi/abs/10.1002/inst.12434
7. Provide specific examples of how the integration of crack detection, prediction, and treatment optimization models can lead to more efficient crack repair processes in undisturbed loess.
8. Explain how this can potentially result in cost savings in terms of materials, labor, and equipment used for crack repair, and describe some potential benefits of these cost savings in terms of resource allocation and cost-effective solutions.
9. Can you provide more details about the fitness function used in Algorithm 1 for the particle swarm based optimization algorithm? How does the fitness function balance the decision-making process and ensure efficiency in terms of resource utilization and prediction durations?
10. Additionally, can you explain how the prediction technique is used to balance the particle positions in the algorithm?
11. Provide more information about the process of restoration mentioned in the text and how it leads to an increase in mechanical strength of in-situ soil.
12. Explain how the addition of CaCO3 cement to the filled pores of soil particles contributes to the increase in strength and displacement.
13. Elaborate on the double-peak phenomenon observed in Figure 4 for groups 5 and 6 and how it affects the strength curve.
14. Additionally, provide details on the effects of clay content in the soil samples on the microbial remediation method and the remediation effects observed in the study.
15. Can you propose additional future work directions beyond exploring other deep learning approaches and proposing an aggregation mechanism to reduce the amount of trained data while maintaining accuracy? For example, are there any potential improvements or extensions to the current approach that could be explored? Are there any specific applications or domains where this approach could be further developed or optimized? Additionally, are there any limitations or challenges with the current approach that could be addressed in future work?
Can be improved.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
1. What are the contributions of the paper? Kindly highlight them.
2. What are the merits compared to the existing approach. Prepare a table.
3. References/citations must be in order i.e. calling the reference inside the paragraph must be in the order [1][2][3]........ not [7]...[15]
Minor editing of the English language required
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
A Machine Learning based Decision Support System for Predicting and Repairing Cracks in Undisturbed Loess using Microbial Mineralization and Internet of Things
Authors: Yangyang Yue and Yiqing Lv
This manuscript addresses a decision support system (DSS) that detects and predicts cracks and recommends a suitable crack repair methodology based on the microbial-induced carbonate deposition (MICP) technology using machine learning algorithm to detect and predict cracks in in undisturbed loess using various data sources, such as images captured using internet of things (IoT), devices, drones, and/or ground-based sensors. It is shown that it predicts to provide a real-time recommendation to professionals working on crack repair projects in undisturbed loess and guiding them to apply the microbial mineralization treatments based on the predicted crack locations and treatment effectiveness.
The manuscript is well generally written but needs substantial improvement by proofreading and providing a more complete exposure to the field, especially to those pertaining to the methods used (suggestions made below). With all the following edits and clarifications made satisfactorily, I would recommend this work for publication in Sustainability. Detailed comments on the manuscript are provided below.
SPECIFIC COMMENTS (MAJOR)
- The introduction is well written but needs expanding to make it clearer to a wider audience. For example, various applications of MICP and factors that impact its properties .” but do not explain in which context it is important and why, nor does it have a citation that could guide the reader to background information.
For example it might be nice to provide a comprehensive summary of the various engineering applications of MICP, like, https://doi.org/10.1016/j.bgtech.2023.100008 and https://doi.org/10.1007/s12665-023-10899-y
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The paper is well organized and well structured. It gives a detailed discussion and simulation of the ML-based Decision Support System. The paper reflects the status of your work by your fellow professionals in the field.
Overall proofreading is required.
Reviewer 2 Report
The authors considered my comments and suggestions. Good luck.
May be improved.
