Study on Landslide Displacement Prediction Considering Inducement under Composite Model Optimization

Round 1

Reviewer 1 Report

Dear Authors,

Please have my comments below

Fundamental comments

-what is the size of your population?is it continous? if yes, what algorithms are eligible to be used for these dataset to be predicted?!

Major comments

-why selecting vmd, pso-vmd, ga-vmd, goa-vmd, ga-goa-vmd?are they the conclusion and recommendation of other researchers?

-Figure 25 also you used some other models? how many models did you used? why?

-% of trining testing? how aboout validation?

-for instrumentation, what sensors type did you used? are they calibrated? what is the location of sensors and at what depth they collected data and what are the type of the sensors?

-fig 8 for rainfall is not looking ok, it seems copy paste and repeated? how do you validate it?

-archtecture of all used models are needed with input and output, layering, and etc..

-REC or ROC curve? which one and it is needed?

-r2 has no meening. MAPE needed

-comparison of the results with other past research studies needed

-compaarison of the results with field results at least three case study. becasue you are claiming that these methods can be used and others migh use them so the reliability should be confirmed

-why not cnn? or rnn? why lstm and other methods?

-for GA we have 4 types, which one you used? 

-hyper parameter table for all methods

-sensitivity analysis needed. which of the input parameter has the most impact on the final output

-ranking is needed, which method is the best

-the predictive model should be compared with analytical as well

-the detail algorithms should be step by step presented and say from input how it reaches to output. not presenting only one formula

-what is the soil type and soil characteristics impact on input parameter selection for prediction

-GWT is missing

-this prediction is for what depth and which location o the slope

- you said landslide! landslide occured or is about to occur or it is a normal slope 

English check is needed.

Author Response

Please see the attached file：Response to Reviewer 1.docx

Author Response File: Author Response.pdf

Reviewer 2 Report

The study presented provides a comprehensive approach to landslide displacement prediction, integrating various factors and employing advanced modeling techniques.

While the paper discusses the performance of the proposed model, it would benefit from further validation against independent datasets or against alternative models. This would help assess its generalizability and robustness beyond the specific case study presented.

It's essential to explicitly outline the assumptions made during modeling and any limitations of the approach. Every modeling framework involves simplifications and assumptions, and acknowledging these enhances the credibility and applicability of the study.

While the study provides numerical results, it could explore into the practical implications of these findings. How might these predictions be utilized in real-world scenarios, such as early warning systems or land management strategies?

Transparency in data availability and model code would enhance the reproducibility of the study and allow other researchers to build upon the work. Providing access to datasets and code would foster collaboration and further validation efforts.

Landslides have significant societal and environmental impacts. Discussing the broader implications of accurate displacement prediction, such as mitigating risks to human life and infrastructure, would enrich the study's context and relevance.

The paper could conclude with a discussion of potential possibilities for future research. Highlighting areas where further investigation is needed or where the proposed model could be refined would guide subsequent studies in the field.

Author Response

Please see the attached file：Response to Reviewer 2.docx

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Some of my comments were not adressed properly such as

Comments 1 (provide number)

C4 (need optimization for different % of training and test)

c5 (drawing needed)

c7 (put it in a table format)

c8 (I should disagree, REC can be developed)

c15(what parameter in each algorithm has the most impact-you need to have bar type chart for every algorithm, showing all input parameters and mentioning which one has the most impact on the output)

c19 (we should see the problem from a geotechnical stand point not just from a data science, so geotechnically it should be meaningful. landlside is a complex phenomena and environmental impact as well as slope soil body and soil layering will dictate the condition of slide, and in your study youare just focusing on rainfall and reservoir water level. The slope in reality is not homogenous, and if it is a small slope so we wont have landslide we will be having slope failure, so fundamentally the problem statement has issue )

c20(GWT mean ground water table, for slope failure and landlside GWT is a must including facto, it has a direct impact on the stability, it should be consider and this is another fundamental issue of the peroblem statement of this study)

c21 (you got data from public data and the sensors you used we dont know the configuration and plan layout of the sensors, and predicting only surface of soil slope movement without considering any data from soil layering is kind of hard to imagine and I think this is another fandamental issue )

c22(if this is landlside what is the geometry of the slope, and if it is occured, then you should be able to compare your prediction results with the actual aftermath measured movement and this means validation, and this is another issue that is missing)

Author Response

Please see the attached file: Response to Reviewer 1.docx.

Author Response File: Author Response.pdf

Reviewer 2 Report

After careful review, we find that the revised version meets the necessary criteria for publication.

Author Response

Thank you very much for your review!

Round 3

Reviewer 1 Report

The Authors addressed my comments properly.