Advanced machine learning techniques for predicting concrete mechanical properties: a comprehensive review of models and methodologies

This review presents the application of machine learning (ML) techniques for predicting the mechanical properties of concrete, addressing the limitations of traditional methods such as linear regression and empirical formulas. These conventional approaches often fail to capture the complex, nonlinear relationships inherent in concrete’s heterogeneous composition, resulting in less accurate predictions. This study analyzes eight ML models, including ensemble methods such as random forest, extreme gradient boosting, categorical boosting, natural gradient boosting, light gradient boosting machine, and adaptive boosting, as well as deep learning models like artificial neural networks and evolutionary algorithms like gene expression programming. These ML techniques demonstrate significant improvements in prediction accuracy, scalability, and generalization. The review also explores data processing techniques and performance evaluation metrics, highlighting current challenges such as data availability and the need for user-friendly graphical user interface-based prediction tools. Recommendations for future research are provided to further advance the practical application of ML in predicting concrete properties.