Physics-guided explainable machine learning and uncertainty quantification for predicting unconfined compressive strength of nano shale–quicklime stabilized kaolin clay
Abstract
Abstract Data-driven prediction of stabilized-soil strength is often constrained by small experimental datasets, leakage between related specimens, and limited physical interpretability. This study develops an explainable, uncertainty-aware framework to predict the unconfined compressive strength (UCS) of nano shale–quicklime stabilized kaolin. A specimen-level database of 180 measurements from 36 mixture–curing conditions was analyzed using multiple linear regression (MLR), support vector regression (SVR), random forest (RF), XGBoost, CatBoost, and backpropagation artificial neural network (BP-ANN), with nonlinear models optimized by particle swarm optimization and evaluated through nested GroupKFold cross-validation. Mechanism-inspired features represented binder dosage, binder–curing interactions, nonlinear curing behavior, relative plasticity, and compaction state. Sequential feature ablation identified a compact CatBoost model with six physics-guided descriptors, achieving R 2 = 0.841 ± 0.123 and RMSE = 151.11 ± 34.94 kPa. SHAP analysis, sensitivity assessment, and two-dimensional partial dependence mapping identified plasticity, curing duration, and binder–curing interactions as dominant controls and revealed nonlinear stabilizer effects and favorable binder combinations. Monte Carlo input perturbation and group-bootstrap resampling quantified input-driven and model-related uncertainty, with the highest 95% prediction-interval coverage reaching 0.894. The framework…
How strong is the evidence?
Evidence Hub reads this as a experimental / empirical study, done in computer models (in silico), not in people. Its evidence level is 5 of 5 (Very low · case / narrative / animal / in vitro / in silico). Case reports, narrative reviews, and animal, cell or computer-model studies give the weakest evidence for human health.
Read the paper
Related studies
- Preliminary Toxicity Analysis of 3DCRT versus IMRT on the High Dose Arm of the RTOG 0126 Prostate Cancer Trial
- Adaptive Radiotherapy for Head and Neck Cancer: Initial Clinical Outcomes from a Prospective Trial
- American Association of Physicists in Medicine Task Group 263: Standardizing Nomenclatures in Radiation Oncology
- Motion Management Strategies and Technical Issues Associated with Stereotactic Body Radiotherapy of Thoracic and Upper Abdominal Tumors: A Review from NRG Oncology
- High-dose-rate interstitial brachytherapy as a monotherapy for localized prostate cancer: treatment description and preliminary results of a phase I/II clinical trial
- Highly efficient training, refinement, and validation of a knowledge-based plan quality control system for radiotherapy clinical trials
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.