Effitioned soft computing models to evaluate the impact of silicon dioxide (SiO2) to calcium oxide (CaO) ratio in fly ash on the compressive strength of concrete

Environmental issues are raised from global warming due to raised Carbon Dioxide (CO2) emissions of factories worldwide. Cement production provides about 8–10% of the total CO2 emissions to the environment. Cementitious materials, such as fly ash, are suggested as the best alternatives to cement as the main ingredient of concrete. Fly ash is a powder finer than cement, almost rich in silica and alumina. The current study investigated the effect of the ratio of SiO2/CaO in fly ash on the compressive strength of cement-based concrete modified with different fly ash contents and classes for various mix proportions. 236 fly ash-modified concrete samples were examined, evaluated, and modeled for that purpose. The study includes independent parameters such as; coarse aggregate (801–1246 kg/m3), fine aggregate (522–905 kg/m3), cement (67–356 kg/m3), fly ash (71–316 kg/m3), cement replacement (18–100%), water-to-binder ratio (0.28–0.60), silicon dioxide to calcium oxide ratio (0.984–174.0), and curing time (3–365 days). The dependent parameter is compressive strength (7.98–92.93 MPa); it is divided into three ranges; low-strength (less than 20 MPa), normal strength (20–50 MPa), and high-strength (greater than 50 MPa). This study utilized several mathematical, soft computing, and machine learning modeling tools to develop a dependable model for predicting the compressive strength of concrete. The Linear Regression (LR), Pure Quadratic (PQ), M5P-tree, and Interaction (IN), were used to prediction. To provide accurate and reliable models, multiple assessment criteria were utilized, such as correlation coefficient (R2), Root mean squared error (RMSE), Mean absolute error (MAE), Scatter Index (SI), Objective function (OBJ), and a-20 index. The IN model was the most effective and accurate model; where R2 of 0.95, RMSE of 4.33 MPa, MAE of 3.45 MPa, and SI of 0.099 were conducted from the analytical studies. The M5P-tree model produced an OBJ value of 1.31 and the lowest residual error value of −17.99 to +16.97 MPa. As well as, the PQ model maintained the best a-20% index value, which was 100%. From the IN model, increasing SiO2/CaO ratio from 0.984 to 17.0 caused a decrease in compressive strength but increased beyond the 17.0 ratio. However, the compressive strength decreased with increasing the cement replacement and fly ash content when the coarse aggregate content was used by less than 905 kg/m3. In addition, the models were applied to the dataset based on the shape