Skip to content

This repository is associated with interpretable/explainable ML model for liquefaction potential assessment of gravelly soils. This model is developed using LightGBM and SHAP.

Notifications You must be signed in to change notification settings

Kaushikjas10/Liquefaction-gravel-eml-2023

Repository files navigation

Liquefaction-gravel-eml-2023

This repository is for explainable/interpretable model to assess gravelly soil liquefaction potential. The DPT-Vs database is used to develop the model. This model is developed by using LightGBM and SHAP.

This repository contains these files: processed dataset, a csv file containing IPs of test sites, and a python script.

This repository is part of the below project. If you use this in your work, dont forget to cite as below:

Jas, K., Mangalathu, S., & Dodagoudar, G. R. (2024). Evaluation and analysis of liquefaction potential of gravelly soils using explainable probabilistic machine learning model. Computers and Geotechnics, 167, 106051. https://doi.org/10.1016/j.compgeo.2023.106051.

Diect link to this article: https://www.sciencedirect.com/science/article/pii/S0266352X2300808X?via%3Dihub.

Researchgate paper link: https://www.researchgate.net/publication/377261175_Evaluation_and_analysis_of_liquefaction_potential_of_gravelly_soils_using_explainable_probabilistic_machine_learning_model.

You can request full text. Read and recommend also.

In case any queries about this work, feel free to contact with this email id: ce19d202@smail.iitm.ac.in. // kaushik.jas.2013@gmail.com

Data and Code Availability Statement

Name of the code: Gravel_EML_KJSMGRD_23.

Developer name: Mr. Kaushik Jas.

Contact address: Department of Civil Engineering, Indian Institute of Technology Madras, Chennai, Tamil Nadu 600036, India.

Email: ce19d202@smail.iitm.ac.in.

First available year: 2023

Program language: Python 3.7; MATLAB R2022b

Downloadable link of source code: https://github.com/Kaushikjas10/Liquefaction-gravel-eml-2023.git

File descriptions (Beta version)

(1) [Processed dataset]: It is available with the article in the supplemental material at "Appendix A". The file is now available in this repository as file name "Jas et al. 2024_Supply_EML_Gravel_Liq_COGE.pdf". This is the same dataset used for the model training and testing. This same file is attached to the COMGE journal section.

(2) [TestSitesGravel.csv]: This is the spreadsheet template of the test sites where you want to assess the liquefaction potential of gravelly soils using the developed model.

(3) [MODEL_A_KJSMGRD_Gravel_COMGE.json; MODEL_A_KJSMGRD_Gravel_COMGE.txt]: These are the LightGBM model files (Initial trial) with two different extensions. This file can be loaded in the Python platform to predict the liquefaction potential assessment of gravelly soils for site(s) of interest having all IPs in an excel spreadsheet. Also you can find the model details/parameters (Table 5) in the manuscript itself. The final/best trial (MODEL_B_KJSMGRD_Gravel_COMGE_1.0.txt) will be uploaded soon.

(4) [Model_Train_Test_KJ_SM_GRD.ipynb]: This Python script is provided to give a demonstration of the LightGBM model training and testing in the Jupyter notebook platform.

(5) [Implement_Gravel_MODEL_v0.2_KJSMGRD_COMGE.ipynb]: This Python script is provided to give a hands-on demonstration of using the developed/trained DPT Vs-based model in the Jupyter Notebook platform.

(6) Interactive online tutorial: Will be uploaded soon....................

Release notes:

Update 26-03-2024:

1)All the necessary files are uploaded except (2) and (5). Those files will be uploaded soon..........

Update 28-03-2024:

The file number (2) and (5) are uploaded to this repository. See the "Implement_Gravel_MODEL_v0.2_KJSMGRD_COMGE.ipynb". The interactive online tutorial will be uploaded soon.