Assessment of fiber-reinforced rubberized recycled aggregate concrete’s properties by optimal regression frameworks

Construction and demolition waste (C&D) and tyre garbage, in particular, are becoming urgent problems on a worldwide scale. One possible solution to this problem is to substitute man-made aggregates like recycled coarse aggregate (RCA) from C&D and crumb rubber (CR) from old tyres in newly-made building materials. The goal of this study is to determine whether machine learning and regression-based methods are best for predicting flexural strength (fs) in FRRAC, or fiber-reinforced rubberized recycled aggregate concrete. The Least Squares Support Vector Regression (LSSVR) was developed for this purpose. Hyperparameters are vital in this simulation, which use the LSSVR in conjunction with the Chimp optimisation algorithm (ChOA) and the Artificial rabbit optimisation algorithm (AROA) processes to identify the optimal set. Regression models were developed and tested to forecast fs's purpose using a portion of the study dataset (102 samples). A quarter (25 samples) were used for evaluation and seventy-five percent (77 samples) for instruction out of the 102 samples included in the database. The estimation process took into account several factors. Based on these metrics outcome numbers which provided in this study, the LSSVR(A) outperformed the LSSVR(C) in order to predicted predicting flexural strength (fs) in FRRAC.