CatBoost–Bayesian Hybrid Model Adaptively Coupled with Modified Theoretical Equations for Estimating the Undrained Shear Strength of Clay

The undrained shear strength of clay is an important index for the calculation bearing capacity foundation soil, soil pressure pit, and analysis slope stability. Therefore, purpose this paper to conduct a comprehensive study combined use machine learning with theoretical equations estimate it. Under Bayesian framework, CatBoost algorithm (CatBoost–Bayesian) based on optimization was developed obtain feature importance level parameters affecting clay, so as adaptively couple equation K0 consolidated which derived from modified Cambridge model. Then, isotropically established critical state parameters. Finally, it illustrated verified using experimental samples Finnish clay. results indicate that by overconsolidation ratio effective overburden can well clays, parameter uncertainty be considered explicitly rigorously.