Hybrid data-driven estimation for compressive strength of eco-friendly geopolymer concrete containing corncob ash

Bio-based geopolymer concrete, particularly corncob ash concrete, represents a transformative advancement in sustainable construction due to its renewable feedstock utilization and carbon sequestration potential. Accurately predicting the compressive strength of corncob ash concrete is crucial for enhancing design efficiency and reducing costs. This study introduces a data-driven approach to estimating the compressive strength of corncob ash concrete across various component ratios. A dataset of 256 samples was created, and three hybrid random forest (RF) models were developed using optimization methods: termite life cycle optimizer (TLCO), catch fish optimization algorithm (CFOA), and Arctic puffin optimization (APO). Eight input parameters, including corncob ash and fine aggregates, were selected, with compressive strength as the output. Comparative analyses included an unoptimized RF model and a multiple linear regression model. Sensitivity analysis assessed the influence of each input variable on strength estimation. The research results indicate the CFOA-RF model demonstrating the best performance and efficiency, and it has an RMSE of 1.1124 MPa, MAE of 0.8305 MPa, R2 of 0.9917, and VAF of 99.1669 for the training set; and RMSE of 1.2427 MPa, MAE of 0.8440 MPa, R2 of 0.9888, and VAF of 98.8868 for the testing set. CG and W are the key factors influencing the compressive strength of corncob ash concrete. The hybrid CFOA-RF model establishes a robust correlation between compositional variables of corncob ash concrete and its compressive strength, offering a decision-support tool for optimizing recycled-aggregate concrete formulations in circular construction practices.