Enhanced artificial intelligence for ensemble approach to predicting high performance concrete compressive strength

Abstract

The compressive strength of high performance concrete (HPC) is a highly nonlinear function of the proportions of its ingredients. The validity of relationships between concrete ingredients and supplementary cementing materials is questionable. This work evaluates the efficacy of ensemble models by comparing individual numerical models in terms of their performance in predicting the compressive strength of HPC. The performance of support vector machines, artificial neural networks, classification and regression trees, chi-squared automatic interaction detector, linear regression, and generalized linear were applied to construct individual and ensemble models. Analytical results show that the ensemble technique combining two or more models obtained the highest prediction performance. For five experimental datasets, the ensemble models achieved 4.2–69.7% better error rates than those of prediction models in previous studies. This work confirmed the efficiency and effectiveness of the proposed ensemble approach in improving the accuracy of predicted compressive strength for HPC.