Accurate prediction of concrete compressive strength is essential for mixture design, quality control, and the broader use of supplementary cementitious materials in low-carbon construction. Fly ash concrete is particularly challenging to model because its strength development is affected by nonlinear interactions among binder composition, water–binder relationships, admixture dosage, and material characteristics. To address this problem, this study proposes a Dominant Learner with Adaptive Mixing (DLAM) framework for data-driven strength prediction. DLAM uses inner cross-validation to identify the most reliable learner from a pool of machine learning models and introduces a validation-controlled Ridge calibration step to exploit complementary information among candidate predictions. The calibration branch is adopted only when it improves the inner-validation root mean squared error (RMSE), thereby reducing the risk of unnecessary model combination and performance degradation. The framework is evaluated using a leakage-free repeated outer/inner validation protocol on a fly ash concrete dataset and is further examined on an independent public concrete strength dataset. DLAM is compared with individual learners, adaptive model-averaging baselines, and Stacking. The results show that DLAM achieves the lowest mean RMSE among the focused comparators on both datasets, with a clear improvement on the external dataset and a more modest gain on the fly ash dataset. These findings demonstrate that validation-controlled calibration provides a transparent and robust way to enhance machine-learning-based concrete strength prediction, especially when different learners capture complementary aspects of the mixture–strength relationship.
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.
Zhihao Zhao, Jinjin Wang, Guohui Ma et al.· Buildings· 0 citations