Enhancing Concrete Properties Through the Integration of Recycled Coarse Aggregate: A Machine Learning Approach for Sustainable Construction

This study presents a novel approach to predict the compressive strength of concrete incorporating recycled coarse aggregate using machine learning models. The research begins with an extensive dataset collected from literature studies, encompassing various concrete mixtures. Statistical analyses reveal the dataset's characteristics, paving the way for meticulous preprocessing, including scaling features. Three prominent ensemble machine learning models—Random Forest Regression, Gradient Boosting Regression, and XGBoost Regression—are employed and evaluated using 5-fold cross-validation. The results demonstrate commendable performance across all models, with XGBoost emerging as the top performer. SHAP analysis provides insights into the feature importance, highlighting the substantial influence of factors such as curing age, cement, and fly ash. This study not only contributes to the understanding of concrete properties with recycled coarse aggregate (RCA) but also underscores the potential of machine learning in advancing sustainable construction practices. Future research could explore additional algorithms, diverse datasets, and real-time monitoring for comprehensive applicability in the construction industry.