Super learner machine‐learning algorithms for compressive strength prediction of high performance concrete
Abstract: Because the proportion between the compressive strength of high-performanceconcrete (HPC) and its composition is highly nonlinear, more advanced regres-sion methods are demanded to obtain better results. Super learner models,which are based on several ensemble methods including random forest regres-sion (RFR), an adaptive boosting (AdaBoost), gradient boosting machine(GBM), extreme gradient boosting (XGBoost), light gradient boosting machine(LightGBM), categorical gradient Boosting (CatBoost), are used to solve thiscomplicated problem. A grid search method is employed to determine the bestset of hyper-parameters of each ensemble algorithm. Two super learnermodels, which combine all six models or select the top three effective ones asthe base learners, are then proposed to develop an accurate approach to esti-mate the compressive strength of HPC. The results on four popular datasetsshow significant improvement of the proposed super learner models in termsof prediction accuracy. It also reveals that their trained models always performbetter than other methods since their errors (MAE, MSE, RMSE) are alwaysmuch lower and values ofR2are higher than those of the previous studies. Theproposed super learner models can be used to provide a reliable tool for mix-ture design optimization of the HPC.