An improved hybrid ensemble learning paradigm for estimating compressive strength of fly ash-blended green concrete

Specifically, a gradient boosting regressor (GBR) and an improved grey wolf optimiser (IGWO) were used to construct an efficient hybrid ensemble paradigm, GBR-IGWO. The outcomes of the proposed GBR-IGWO framework were compared with four ensemble learning algorithms, two standard models, and three neural network models.

The developed GBR-IGWO achieved the highest accuracy with R² of 0.9965 and 0.9632 in the training and testing stages, respectively. The influence of cement replacement on the compressive strength of green concrete was further investigated at different curing ages to evaluate the generalisation capability of the proposed framework. Separate laboratory experiments were also performed to validate the proposed model.