Assessment of Machine Learning Models in Liquefaction Prediction with Emphasis on ROC Curve and AUC Index Differences
Subject Areas : Soil-Structure Interaction
shima aghakasiri
1
*
,
Mehdi Shahraki
2
,
sanaz aghakasiri
3
1 - Department of Civil Engineering, ST.C , Islamic Azad University, Teharn, Iran.
2 - Department of Civil Engineering, Zah.C., Islamic Azad University, Zahedan, Iran.
3 - M.Sc., Grad., Department of Civil Engineering, Babol Noshirvani University of Technology, Babol, Iran.
Keywords: Soil liquefaction potential, Artificial Neural Network, Logistic Regression, Neuro-Fuzzy Network, Cone Penetration Test (CPT),
Abstract :
This study presents a novel approach for predicting soil liquefaction potential, a critical concern in geotechnical engineering. Liquefaction refers to the behavior of soil under dynamic loading or transient shear wave excitation, during which the soil completely loses its shear strength and temporarily transforms into a fluid-like state. By integrating empirical geotechnical relationships with advanced machine learning techniques, the research offers a modern perspective on evaluating the likelihood of liquefaction occurrence. The analysis is based on data derived from Cone Penetration Test (CPT) records. Three soft computing models were implemented: Artificial Neural Networks (ANN), Logistic Regression (LR), and Neuro-Fuzzy Network. Their performance was evaluated using Receiver Operating Characteristic (ROC) curves. Among the models compared, Logistic Regression demonstrated superior performance, with the Area Under the Curve (AUC) from the “All” dataset reaching approximately 0.975, indicating high reliability in classification accuracy. In this study, the logistic regression model achieved an AUC of 0.975 on the full dataset, followed by the artificial neural network (AUC = 0.925) and the fuzzy logic system (AUC = 0.71).
.
1. Galli, P., New empirical relationships between magnitude and distance for liquefaction. Tectonophysics, 2000. 324(3): p. 169-187. https://doi.org/10.1016/S0040-1951(00)00118-9
2. Kokusho, T., Innovative earthquake soil dynamics. 2017: CRC Press. https://doi.org/10.1201/9781315645056
3. Zhou, J., et al., Feasibility of stochastic gradient boosting approach for evaluating seismic liquefaction potential based on SPT and CPT case histories. Journal of Performance of Constructed Facilities, 2019. 33(3): p. 04019024. https://doi.org/10.1061/(ASCE)CF.1943-5509.0001292
4. Khanbabazadeh, H., Nonlinearity effect on the dynamic behavior of the clayey basin edge. Geomechanics and Engineering, 2024. 36(4): p. 367-380. https://doi.org/10.12989/gae.2024.36.4.367.
5. Khanbabazadeh, H., R. Iyisan, and B.Ozaslan, Seismic behavior of the shallow clayey basins subjected to obliquely incident wave. Geomech. Eng, 2022. 31(2): p. 183-195. https://doi.org/10.12989/gae.2022.31.2.183
6. Khanbabazadeh, H., R. Iyisan, and B. Ozaslan, 2D seismic response of shallow sandy basins subjected to obliquely incident waves. Soil Dynamics and Earthquake Engineering, 2022. 153: p.107080.
https://doi.org/10.1016/j.soildyn.2021.107080
7. Matsuoka, M., et al., Evaluation of liquefaction potential for large areas based on geomorphologic classification. Earthquake Spectra, 2015. 31(4): p. 2375-2395. https://doi.org/10.1193/072313EQS211M
8. Bahrainy, H. and A. Bakhtiar, Manjil Earthquake of June 20, 1990, The Lessons Learned, in Urban Design in Seismic-Prone Regions. 2022, Springer International Publishing: Cham. p. 49-95. DOIhttps://doi.org/10.1007/978-3-031-08321-1
9. Uyanık, O., Soil liquefaction analysis based on soil and earthquake parameters. Journal of Applied Geophysics, 2020. 176: p. 104004 https://doi.org/10.1016/j.jappgeo.2020.104004.
10. Ahmad, M., et al., Application of machine learning algorithms for the evaluation of seismic soil liquefaction potential. Frontiers of Structural and Civil Engineering, 2021. 15(2): p. 490-505. https://doi.org/10.1007/s11709-020-0669-5
11. García, S., M. Romo, and E. Ovando- Shelley, Machine learning for assessing liquefaction potential of soils.
12. Liu, C. and J. Macedo, Machine learning- based models for estimating liquefaction- induced building settlements. Soil Dynamics and Earthquake Engineering, 2024. 182: p. 108673.
https://doi.org/10.1016/j.soildyn.2024.108673
13. Abbasimaedeh, P., Soil liquefaction in seismic events: pioneering predictive models using machine learning and advanced regression techniques. Environmental Earth Sciences, 2024. 83(7): p. 189. https://doi.org/10.1007/s12665-024-11480-x
14. Ozsagir, M., et al., Machine learning approaches for prediction of fine-grained soils liquefaction. Computers and Geotechnics, 2022. 152: p. 105014. https://doi.org/10.1016/j.compgeo.2022.105014
15. Obaidullah ,S., Preliminary Liquefaction Susceptibility Using Different Machine Learning Techniques. 2024, (SCEE), NUST.
16. Kumar, D., et al., A novel methodology to classify soil liquefaction using deep learning. Geotechnical and Geological Engineering, 2021. 39 :p. 1049-1058
https://doi.org/10.1007/s10706-020-01544-7.
17. Raja, M.N.A., T. Abdoun, and W. El-Sekelly, Smart prediction of liquefaction-induced lateral spreading. Journal of Rock Mechanics and Geotechnical Engineering, 2024. 16(6): p. 2310-2325.
https://doi.org/10.1016/j.jrmge.2023.05.017
18. Mohammadikish, S., et al., Soil liquefaction assessment by CPT and VS data and incomplete-fuzzy C-means clustering. Geotechnical and Geological
Engineering, 2024. 42(3): p. 2205-2220. https://doi.org/10.1007/s10706-023-02669-1
19. Şehmusoğlu, E.H., T.F. Kurnaz, and C. Erden, Estimation of soil liquefaction using artificial intelligence techniques: an extended comparison between machine and deep learning approaches. Environmental Earth Sciences, 2025.84(5): p. 1-22. https://doi.org/10.1007/s12665-025-12116-4
20. Niu, F., et al., Ultra-high performance concrete: A review of its material properties and usage in shield tunnel segment. Case Studies in Construction Materials, 2025: p. e04194. https://doi.org/10.1016/j.cscm.2024.e04194
21. Agha Kasiri, Sh., Agha Kasiri, S., Ghazawi,M., Farrokhzad, F. Investigation of Sandy Soil Settlement Due to Liquefaction Under Earthquake in PILE Group. International Conference on Architecture, Urban Planning, Art, Industrial Design, Construction and Technology of Hikmat-e-Bonyan. 2025. https://civilica.com/doc/2235759
22. Agha Kasiri, Sh., Emami Korandeh, M., Mohammadi, Gh., Taban, A. Deep Learning-Based Fluidity Data Evaluation. International Conference on Architecture, Urbanism, Art, Industrial Design, Construction and Technology Hekmat-Bonyan, 2025.
https://civilica.com/doc/2235758
23. Agha Kasiri, Sh., Agha Kasiri, S., Farrokhzad, F,. Qadawi, M. Comparison of settlement in single pile and pile group under dynamic loading. Fourth International Congress of Civil Engineering, Architecture and Urban Development, 2016. https://civilica.com/doc/617967
24. Agha Kasiri, Sh., Farrokhzad, F., Qadawi, M. Determination of bearing capacity and settlement of piles in sandy soil in static and dynamic mode. International Conference on New Horizons in Civil Engineering, Architecture and Urban Planning and Cultural Management of Cities, 2016. https://civilica.com/doc/567742
25. Marzouk, I., et al., A case study on advanced CPT data interpretation: from stratification to soil parameters. Geotechnical and Geological Engineering,
2024. 42(5): p. 4087-4113. https://doi.org/10.1007/s10706-024-02774-9
26. Seed, H.B. and I.M. Idriss, Simplified procedure for evaluating soil liquefaction potential. Journal of the Soil Mechanics and Foundations division, 1971. 97(9): p. 1249-1273. https://doi.org/10.1061/JSFEAQ.0001662
27. Liao, S.S. and R.V. Whitman, Overburden correction factors for SPT in sand. Journal of geotechnical engineering, 1986. 112(3): p. 373-377. https://doi.org/10.1061/(ASCE)0733-9410(1986)112:3(373)
28. Robertson, P.K. and C. Wride, Evaluating cyclic liquefaction potential using the cone penetration test. Canadian geotechnical journal, 1998. 35(3): p. 442-459. https://doi.org/10.1139/t98-017
29. Robertson, P.K., Soil classification using the cone penetration test. Canadian geotechnical journal, 1990. 27(1): p. 151-158. https://doi.org/10.1139/t90-014
30. ÖNorm, E., 1 [1996]: Eurocode 7: Entwurf. Berechnung und Bemessung in der Geotechnik–Teil, 1997. 1.
31. Robertson, P.K., Interpretation of cone penetration tests—a unified approach. Canadian geotechnical journal, 2009. 46(11): p. 1337-1355.
32. Robertson, P.K. Soil behaviour type from the CPT: an update. in 2nd International symposium on cone penetration testing. 2010. Cone Penetration Testing Organizing Committee Huntington Beach.
33. Robertson, P.K., Cone penetration test (CPT)-based soil behaviour type (SBT) classification system—an update. Canadian Geotechnical Journal, 2016. 53(12): p. 1910-1927.
34. Robertson, P.K. and K. Cabal, Guide to cone penetration testing for geotechnical engineering. Signal Hill, CA: Gregg Drilling & Testing, 2015.
35. Niu, F., et al., Ultra-high performance concrete: A review of its material properties and usage in shield tunnel segment. Case Studies in Construction Materials, 2025: p. e04194.
36. Boulanger, R.W. and I.M. Idriss, CPT and SPT based liquefaction triggering procedures. Report No. UCD/CGM.-14, 2014. 1: p. 134.
37. Shen, M., et al., Predicting liquefaction probability based on shear wave velocity: an update. Bulletin of Engineering Geology and the Environment, 2016. 75: p. 1199-1214. https://doi.org/10.1007/s10064-016-0880-8
38. Haykin, S., Neural networks and learning machines, 3/E. 2009: Pearson Education India
39. Ross, T.J., Fuzzy logic with engineering applications. 2005: John Wiley & Sons.