Loess Collapsibility Prediction and Influencing Factor Analysis Using Multiple Machine Learning Algorithms in Xi’an Region
Source abstract
Abstract
Collapsibility is a fundamental geotechnical property of loess that critically affects its engineering behavior. In this study, a comprehensive dataset comprising 9041 experimental records on the physical properties and collapsibility of loess from the Xi’an region was compiled. Six parameters were selected as model inputs: sampling depth (H), water content (w), plastic limit (wP), plasticity index (IP), compression coefficient (a1–2), and compression modulus (Es). Based on these inputs, prediction models for the loess collapsibility coefficient (δs) were developed using Gaussian Process Regression (GPR), Gradient Boosting Machine (GBM), Support Vector Regression (SVR), Radial Basis Function Neural Network (RBFNN), Classification and Regression Tree (CART), and Feature Tokenizer Transformer (FT-Transformer). Among these, GPR demonstrated the best predictive performance, achieving the lowest error (RMSE = 9.88 × 10−3) and the highest accuracy (R2 = 0.844). Additionally, the coverage proportion of the 95% confidence interval of the GPR predictions reached 0.949. SHapley Additive exPlanations (SHAP) analysis for GPR further revealed that the compression coefficient exerted the greatest influence on δs (0.0149), followed by compression modulus (0.0080), water content (0.0068), plasticity index (0.0061), sampling depth (0.0061), and plastic limit (0.0052). The GPR-based prediction model offers significantly higher predictive accuracy than empirical models. The developed models provide a robust technical framework for the rapid estimation of loess collapsibility in the Xi’an region.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Soil and Unsaturated Flow · Landfill Environmental Impact Studies · Geotechnical Engineering and Soil Mechanics
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
College of Geology and Environment, Xi’an University of Science and Technology, Xi’an 710054, China
Institute of Ecological Environmental Restoration in Mine Areas of West China, Xi’an University of Science and Technology, Xi’an 710054, China
Xi’an Municipal Geologic Environment Monitoring Station, Xi’an 710054, China
Shaanxi Provincial Key Laboratory of Geological Support for Coal Green Exploitation, Xi’an 710054, China
Shaanxi Institute of Geo-Environment Monitoring, Xi’an 710054, China
Observation and Research Station of Ground Fissure and Land Subsidence, Ministry of Natural Resources, Xi’an 710054, China
Funding
National Natural Science Foundation of China · Shaanxi Province Geological Hazards Prevention and Control System Construction Project
References (51)
Show source-provided references
1. Review of collapse triggering mechanism of collapsible soils due to wetting10.1016/j.jrmge.2015.12.002 ↗
2. Collapse behavior and microstructural evolution of loess soils from the Loess Plateau of China10.1007/s11629-018-5006-2 ↗
3. Field study on the reinforcement of collapsible loess using dynamic compaction10.1016/j.enggeo.2014.12.006 ↗
4. Geological hazards in loess terrain, with particular reference to the loess regions of China10.1016/S0012-8252(01)00050-2 ↗
5. Characteristics and mechanisms of soil structure damage under salt weathering10.1016/j.still.2024.106030 ↗
6. Zheng, Z., Li, X., and Wang, L. (2021). A longitudinal analysis of collapsibility with predictions over the southeastern Loess Plateau in China. Sci. Rep., 11.10.1038/s41598-021-02623-2 ↗
7. Space Effect of Loading–Unloading Stress Path Evolution of Roof Strata and Surrounding Rock Control in Deep Mining
8. Loess collapsibility characteristics of railway engineering sites using large-scale trial immersion pit experiments10.1007/s10064-021-02124-6 ↗
9. Study on the Loess Immersion Test of Metro Line 2 in Xi’an, Shaanxi Province, China10.1155/2021/6654391 ↗
10. Collapsibility deterioration mechanism and evaluation of compacted loess with sodium sulfate under drying–wetting cycles10.1007/s11069-022-05581-8 ↗
11. A laboratory study of collapse behaviour of remoulded loess under controlled wetting and flooding10.1080/17486025.2015.1057620 ↗
12. Study on the method for determination of the maximum depth of loess collapsible under overburden pressure10.1007/s10064-019-01630-y ↗
13. Collapsible deformation evaluation of loess under tunnels tested by in situ sand well immersion experiments10.1016/j.enggeo.2021.106257 ↗
14. Evaluation of Collapsibility of Compacted Loess Based on Resistivity Index10.1155/2021/9990012 ↗
15. Li, Z., Li, X., Zhu, Y., Dong, S., Hu, C., and Fan, J. (2021). Mining and analysis of multiple association rules between the Xining loess collapsibility and physical parameters. Sci. Rep., 11.10.1038/s41598-020-78702-7 ↗
16. The evaluation of artificial filling in loess areas by in-situ tests based on statistical analysis10.1007/s12665-021-09729-w ↗
17. A new method for evaluating loess collapsibility and its application10.1016/j.enggeo.2019.105376 ↗
18. Evaluation of the collapse susceptibility of loess using machine learning10.1016/j.trgeo.2024.101327 ↗
19. Prediction of collapsibility of loess site based on artificial intelligence: Comparison of different algorithms10.1007/s12665-024-11423-6 ↗
20. Collapsible characteristics and prediction model of remodeled loess10.1007/s11069-024-06804-w ↗
21. Probabilistic prediction of uniaxial compressive strength for rocks from sparse data using Bayesian Gaussian process regression with Synthetic Minority Oversampling Technique (SMOTE)10.1016/j.compgeo.2023.105850 ↗
22. Bousquet, O., von Luxburg, U., and Rätsch, G. (2004). Gaussian Processes in Machine Learning. Advanced Lectures on Machine Learning: ML Summer Schools 2003, Canberra, Australia, February 2–14, 2003, Tübingen, Germany, August 4–16, 2003, Revised Lectures, Springer.
23. Estimating the thermal conductivity of soils using six machine learning algorithms10.1016/j.icheatmasstransfer.2022.106139 ↗
24. Support-vector networks10.1023/A:1022627411411 ↗
25. LIBSVM: A library for support vector machines
26. Zhang, W., Guo, J., Li, Z., Cheng, R., Ning, C., Niu, H., and Liu, Z. (2025). Prediction of loess collapsibility coefficient using bayesian optimized random forest model. Sci. Rep., 15.10.1038/s41598-025-11121-8 ↗
27. Fast Learning in Networks of Locally-Tuned Processing Units10.1162/neco.1989.1.2.281 ↗
28. An eigenvector based center selection for fast training scheme of RBFNN10.1016/j.ins.2017.08.092 ↗
29. Integrating the LSSVM and RBFNN models with three optimization algorithms to predict the soil liquefaction potential10.1007/s00366-021-01392-w ↗
30. Greedy Function Approximation: A Gradient Boosting Machine10.1214/aos/1013203451 ↗
31. Liang, W., Luo, S., Zhao, G., and Wu, H. (2020). Predicting Hard Rock Pillar Stability Using GBDT, XGBoost, and LightGBM Algorithms. Mathematics, 8.10.3390/math8050765 ↗
32. Slope stability prediction for circular mode failure using gradient boosting machine approach based on an updated database of case histories10.1016/j.ssci.2019.05.046 ↗
33. Comparative performance of six supervised learning methods for the development of models of hard rock pillar stability prediction10.1007/s11069-015-1842-3 ↗
34. Breiman, L., Friedman, J., Olshen, R.A., and Stone, C.J. (1984). Classification and Regression Trees, Chapman and Hall/CRC. [1st ed.].
35. Comparison of machine learning algorithms for forest parameter estimations and application for forest quality assessments10.1016/j.foreco.2018.12.019 ↗
36. Revisiting deep learning models for tabular data
37. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł., and Polosukhin, I. (2017, January 4–9). Attention is all you need. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, CA, USA.
38. Transformer-based deep learning models for adsorption capacity prediction of heavy metal ions toward biochar-based adsorbents10.1016/j.jhazmat.2023.132773 ↗
39. Identification of growth years of Kudzu root by hyperspectral imaging combined with spectral–spatial feature tokenization transformer10.1016/j.compag.2023.108332 ↗
40. Characterization of groundwater contamination: A transformer-based deep learning model10.1016/j.advwatres.2022.104217 ↗
41. Optimization of site investigation program for reliability assessment of undrained slope using Spearman rank correlation coefficient10.1016/j.compgeo.2022.105208 ↗
42. Landslide susceptibility assessment using uncertain decision tree model in loess areas10.1007/s12665-017-7095-6 ↗
43. Landslide susceptibility mapping using a modified decision tree classifier in the Xanthi Perfection, Greece10.1007/s10346-015-0565-6 ↗
44. Lundberg, S.M., and Lee, S.-I. (2017, January 4–9). A unified approach to interpreting model predictions. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, CA, USA.
45. Gao, L. (2011). Study on Microstructural Effects on Collapsibility of Loess and Its Evaluation Methods. [Ph.D. Thesis, Dalian University of Technology]. (In Chinese).
46. Study on Correlation between Loess Collapsibility and Soil Physical Property Index
47. Relation between loess collapsibility and physical properties and its engineering significance
48. (2018). Code for Building Construction in Collapsible Loess Regions (Standard No. GB 50025-2018). (In Chinese).
49. Evaluation of Loess Collapsibility Based on Random Field Theory in Xi’an, China10.1155/2022/8665061 ↗
50. Spatial prediction of landslide susceptibility using data mining-based kernel logistic regression, naive Bayes and RBFNetwork models for the Long County area (China)10.1007/s10064-018-1256-z ↗
51. Risk evaluation of thaw settlement using machine learning models for the Wudaoliang-Tuotuohe region, Qinghai-Tibet Plateau10.1016/j.catena.2022.106700 ↗