Machine learning-based compressive strength estimation in nano silica-modified concrete

This study investigated the efficacy of advanced machine learning (ML) algorithms for predicting the compressive strength (CS) of concrete modified with nano-silica and supplementary cementitious materials. Utilizing datasets with 1143 samples with a CS rage of 4-129 MPa derived from established experimental literature, the predictive performance of these models was quantitatively evaluated via statistical measures. The outcomes revealed that the Random Forest (R2 = 0.93) and Artificial Neural Networks (R2 = 0.92) models excelled in accuracy, indicating the potential of ML techniques to enhance mixture designs, thus providing substantial savings in both time and fiscal resources related to experimental evaluations.