Eco-Innovation in Construction: Forecasting Natural Fibre-Reinforced Concrete Strength Using Machine Learning

Conventional concrete faces limitations such as brittleness, low energy absorption, and a significant environmental impact due to its reliance on natural resources. Integrating natural fibers (NF) and recycled coarse aggregate (RCA) into concrete presents a promising avenue for enhancing both performance and sustainability. However, accurately predicting the strength of these innovative concrete mixtures remains challenging. This study investigates the predictive capabilities of two machine learning (ML) models: Classification and Regression Trees (CART) and Stepwise Polynomial Regression (SPR), in forecasting the compressive and splitting tensile strength of NF-reinforced concrete incorporating RCA. Results unequivocally demonstrate the superior predictive accuracy of the CART model. CART exhibited significantly higher R-squared values and lower error metrics (RMSE, MAD, MAPE, MSE) for both compressive and splitting tensile strength. For compressive strength CART achieved R² = 0.91, RMSE = 5.5686, MSE = 31.0098, MAD = 4.1076, and MAPE = 0.1055, while for splitting tensile strength, it achieved R² = 0.89, RMSE = 0.3954, MSE = 0.1563, MAD = 0.2996, and MAPE = 0.0939. These findings underscore the significant potential of ML, particularly CART, in optimizing the design of sustainable concrete structures by enabling more precise and efficient strength predictions, ultimately contributing to more sustainable and resilient infrastructure.