Closure to “Shear Strength Prediction in Reinforced Concrete Deep Beams Using Nature-Inspired Metaheuristic Support Vector Regression” by Jui-Sheng Chou, Ngoc-Tri Ngo, and Anh-Duc Pham

Abstract

A sensitivity analysis has been added in this closure investigation to identify and rank the key parameters where L=d are removed from dataset and replaced by single parameters a, d, b w , and L. This procedure is performed by removing one or more parameters from the dataset and using the remaining dataset to train and test the model by 10-fold cross-validation using the smart firefly algorithm-based least squares support vector regression (SFA LS-SVR) model. The sensitivity analysis results also indicate that the performance of the SFA LS-SVR model with all inputs separately considered is worse than that of when inputs of d, b w , a, L have been combined into ratios of d=b w , a=d, and L=d in terms of R, RMSE, MAE, and MAPE.