Machine Learning-Based Estimation of the Compressive Strength of Self-Compacting Concrete: A Multi-Dataset Study

This paper aims at performing a comparative study to investigate the predictive capability of machine learning (ML) models used for estimating compressive strength self-compacting concrete (SCC). Seven prominent ML models, including deep neural network regression (DNNR), extreme gradient boosting (XGBoost), (GBM), adaptive (AdaBoost), support vector (SVR), Levenberg–Marquardt artificial (LM-ANN), and genetic programming (GP), are employed. Four experimental datasets, compiled in previous studies, construct ML-based methods. The models’ generalization capabilities reliably evaluated by 20 independent runs. Experimental results point out superiority DNNR, which has excelled other three four datasets. XGBoost is second-best model, gained first rank one dataset. outcomes great potential utilized approaches modeling SCC. In more details, coefficient determination (R2) surpasses 0.8 mean absolute percentage error (MAPE) always below 15% all best R2 MAPE 0.93 7.2%, respectively.