Ensemble Bagging Model for Predicting Flexural Strength of Geopolymer Concrete

Abstract

Waste materials, such as fly ash and lime mortar, are used in the concrete industry to create an environmentally friendly environment. However, since the experimental studies will take time, it is necessary to predict the flexural strength (FS) and properties of Geopolymer concrete (GPC) using ensemble Learning (EL) algorithms in order to shorten the experimental work process and save money and time. In this study, a new ensemble the Bagging prediction model using gradient boosting regressor estimator is proposed to predict the FS of GPC using lime mortar. The performance of the proposed model was evaluated using the performance metrics R-2, RMSE, MSE, MAE, and MAPE. The proposed model was compared using the individual learning algorithms and validated using k-fold cross-validation technique. From the SHAP plot obtained using the best proposed EL model BGR, ICE, and PDP analysis, it is seen that the blast furnace slag content has the most significant effect on the FS of GPC.