Research on Concrete Strength Based on Machine Learning

Under the "Dual Carbon" strategic background, there is an increasingly urgent demand for synergistic optimization between low-carbon production and high-performance concrete. To address the limitations of traditional strength prediction methods, including insufficient analysis of multi-component interactions and high trial mix costs, this study proposes the use of Random Forest, Gradient Boosting Decision Tree and Stacking models to investigate the nonlinear effects of various components on concrete compressive strength and establish a concrete strength prediction system. Empirical analysis demonstrates that the Stacking model significantly outperforms other models. Feature importance analysis reveals that Age In Days, Cement Component and Water Component are dominant factors affecting concrete strength, while Fly Ash Component shows relatively lower influence. This model provides a data-driven decision-making tool for industrial solid waste recycling and low-carbon concrete mix design.