Evaluation of the indirect tensile strength (ITS) of cement-treated clayey soils using XGBoost prediction model

Statistical modeling of strength characteristics of industrial slag waste–based soilcrete

This study presents a statistical-experimental approach to optimise the proportions of Portland cement and ground granulated blast furnace slag (GGBS) for improving the strength characteristics of soft montmorillonitic clay (SMC). Response Surface Methodology (RSM) was employed to investigate the effects of three variables, cement content (10%–30% by dry soil), GGBS replacement ratio (0%–30% of cement), and curing duration (7–28 days), on unconfined compressive strength. A central composite design was adopted to minimise experimental runs while maintaining prediction accuracy. The developed quadratic model exhibited excellent correlation with experimental data (R2 = 0.977; adj. R2 = 0.969), and analysis of variance confirmed the statistical significance of all main, interaction, and quadratic effects. Microstructural investigations using Fourier transform infrared spectroscopy, X-ray diffraction, and X-ray diffraction revealed the formation of hydration products including calcium silicate hydrate, calcium aluminate silicate hydrate, and portlandite (Ca(OH)2), corroborating the strength gain. This study highlights that the incorporation of GGBS as a supplementary cementitious material in deep cement mixing columns is a technically sound and environmentally sustainable solution for ground improvement projects.

Comparative Analysis of AI and Statistical Models for Predicting Mechanical and Durability-Related Properties of Alkali-Activated Recycled Aggregate Concrete

Alkali-activated recycled aggregate concrete (AARAC) offers a sustainable alternative to traditional concrete but suffers from complex, non-linear mechanical behavior that challenges conventional prediction methods. This study develops and compares five machine learning models, linear regression (LR), M5P, Random Forest (RF), K-Nearest Neighbors (KNN) and XGBoost, for predicting the compressive strength (Cs), flexural strength (Fs), splitting tensile strength (Ss), pull-out bond strength (PT), and water absorption (Wa%) of AARAC. A dataset of 360 experimental samples, incorporating natural aggregate, recycled concrete aggregate (RCA), cement block aggregate (CBA), water-to-cement ratio (W/C), alkaline treatment status, and slump, was used. Models were evaluated via train/test split (80/20) and 10-fold cross-validation using R2, MAE, RMSE, and MAPE. Random Forest achieved the highest test R2 (0.8736) and lowest test MAPE (1.418%) and XGBoost (R2 = 0.8605, MAPE = 1.557%). KNN and M5P performed moderately, while LR was the weakest (R2 = 0.6958, MAPE = 2.147%). All tree-based models exhibited overfitting, with training R2 up to 0.98. Scatter plot analysis revealed systematic underprediction by RF for Cs (constant offset of ~2 MPa) and increasing bias for PT, Ss, and Wa% at higher values. XGBoost gave perfect predictions for PT and Wa% but underpredicted Cs and Fs. K-fold cross-validation confirmed XGBoost as the most robust (mean R2 = 0.9844). Correlation analysis showed W/C strongly increases Wa% (r = 0.80) and decreases PT (r = −0.73); RCA negatively affects mechanical properties, while CBA and alkaline treatment improve them. The study concludes that ensemble tree models, particularly Random Forest, are superior for AARAC prediction, but systematic bias requires post hoc calibration.

TabPFN-Based Prediction of Concrete Compressive Strength

The use of supplementary cementitious materials such as fly ash can reduce environmental impacts and improve the sustainability of concrete construction. However, the nonlinear interactions among mixture design parameters make accurate prediction of concrete compressive strength challenging. In this study, TabPFN, a pre-trained foundation model for tabular data, was applied to predict the compressive strength of fly ash concrete and compared with tuned Random Forest, support vector regression, an artificial neural network, LightGBM, CatBoost, Ridge regression, and Abrams empirical regression. A dataset containing 1062 samples and eight mixture-level variables was used for model development and evaluation. Predictive performance was assessed using the coefficient of determination, mean absolute error, and root mean square error over 100 repeated random splits. The results showed that TabPFN achieved the best overall performance, with an average coefficient of determination of 0.9329, a mean absolute error of 3.2758 MPa, and a root mean square error of 4.6678 MPa. Compared with the strongest tuned gradient-boosting baseline, CatBoost, TabPFN reduced the mean absolute error and root mean square error by 0.8768 MPa and 0.8560 MPa, respectively. Furthermore, repeated-split conformal prediction demonstrated reliable uncertainty quantification, with an average prediction interval coverage probability of 0.9615 and a mean prediction interval width of 23.4554 MPa. SHAP analysis identified the water-to-cement ratio, mortar strength, and water-to-binder ratio as important variables, while additional multicollinearity and feature ablation analyses indicated that correlated ratio variables should be interpreted cautiously. The results indicate that TabPFN provides an accurate, robust, and uncertainty-aware framework for preliminary prediction of 28-day fly ash concrete compressive strength.

Dual-Objective XGBoost Prediction Model for the Cementation Performance of MICP-Treated Sandy Soil in Small-Sample Scenarios

: Microbially Induced Carbonate Precipitation (MICP) is an environmentally friendly technique for sandy soil stabilization. However, the cementation performance is governed by multiple coupled factors and complex experimental procedures, making accurate prediction challenging. In this study, a dual-objective XGBoost prediction model suitable for small-sample scenarios is developed from 77 sets of laboratory data to rapidly estimate unconfined compressive strength (UCS) and calcium carbonate content (CCC) separately. A mechanism-guided feature engineering strategy is adopted to construct three cross features, including urease activity coupled with curing time, calcium carbonate content combined with curing time, and urea-calcium concentration, together with five key influencing parameters. Five-fold cross-validation is used to ensure model stability. In the UCS model, soil particle size fraction (29.94%) and urease activity (20.85%) dominate, while in the CCC model, soil particle size fraction (22.47%) and urease activity (17.63%) prevail, both align well with fundamental MICP mechanisms. The CCC model achieves a coefficient of determination (R 2 ) of 0.6342 and a mean absolute error (MAE) of 2.92%, showing reliable predictive ability. The UCS model achieved an R 2 of 0.7991 and a MAE of 844.83 kPa. However, due to the mathematical amplification of relative error, a small portion of low-strength specimens produced abnormally high MAPE (116.60%), limiting the formal engineering design of the UCS model. SHAP (SHapley Additive exPlanations) analysis is further employed to enhance model interpretability and to quantitatively clarify the marginal contributions and interaction effects of the input features. The proposed framework offers a valuable reference for parameter analysis and mechanistic interpretation of MICP-treated soils. At the same time, the larger prediction deviation for UCS highlights the intrinsic uncertainty of strength evolution in such complex multi-factor systems.

Prediction of compressive and flexural strength of modified concrete using machine learning

Results suggest that, within the present five-fold cross-validation setting and limited-sample dataset, RBF kernel ridge regression captures the nonlinear relationships more effectively than conventional linear models; however, broader generalization should be verified using larger datasets and additional validation.

Experimental investigation of the mechanical and durability performance of hybrid basalt fiber-reinforced concrete using gaussian process regression-based prediction