Hybrid Data-Driven and Explainable Modeling of Compressive Strength in GGBFS–MRP Concrete with Environmental Optimization

The production of concrete is expensive, energy-intensive, and a significant source of carbon emissions, with energy use accounting for 30–40% of production costs. Since micronised rubber powder (MRP) and ground granulated blast furnace slag (GGBFS) are industrial byproducts of the rubber and steel industries, respectively, they offer sustainable substitutes. The compressive strength (CS), a crucial measure of structural performance, of GGBFS-MRP modified concrete is examined in this work. Five predictive models Linear Regression (LR), Non-Linear Regression (NLR), Multiple Linear Regression (MLR), M5P-tree, and Artificial Neural Network (ANN) were used to analyse a total of 135 samples. CS varied from 25.01 to 59.27 MPa, while the input variables were cement content (325–425 kg/m3), water-to-binder ratio (0.35–0.45), GGBFS (0–40%), MRP (0–5%), fine aggregate (905–1105 kg/m3), coarse aggregate (711–1082 kg/m3), and curing time (7–91 days). The ANN model outperformed the M5P-tree model in terms of predictive accuracy. Sensitivity analysis revealed that the three most important factors influencing CS were curing time, MRP dosage, and GGBFS content. These findings provide reliable predictive models and insights into optimizing sustainable GGBFS-MRP concrete mixtures.