Predicting Concrete Compressive Strength Using an Ensemble Machine Learning Model

Abstract

Accurate prediction of concrete compressive strength (CCS) is essential for ensuring structural safety, optimizing mix design, and promoting sustainable construction practices. Traditional empirical models often fail to capture the complex nonlinear relationships among concrete constituents, leading to limited predictive reliability for modern concrete mixtures. This study proposes an advanced ensemble machine learning framework for robust and accurate prediction of CCS. Multiple individual learning models—Artificial Neural Network, Support Vector Regression, Random Forest, and Extreme Gradient Boosting—were developed and integrated using bagging, boosting, and stacking ensemble strategies. The performance of each model was evaluated using standard statistical metrics, including the coefficient of determination (R²), root mean square error, mean absolute error, and mean absolute percentage error. Results demonstrate that ensemble models significantly outperform individual learners, with the stacking-based ensemble achieving the highest predictive accuracy and lowest error values. Feature importance analysis further revealed that curing age, cement content, and water–cement ratio are the most influential parameters governing strength development. The proposed framework provides a reliable, cost-effective, and interpretable solution for concrete mix optimization and quality control, offering substantial potential for practical implementation in sustainable construction engineering.