Enhanced support vector regression for improving accuracy in predicting shear strength in Fiber-reinforced polymer-reinforced concrete beams with FRP stirrups

Abstract

Fiber-reinforced polymer (FRP)-reinforced concrete beams have increasingly applied in construction. Accurate prediction of shear strength in FRP-reinforced concrete beams with FRP stirrups is essential for safe and efficient structural design. This study aims to improve prediction accuracy by developing an enhanced support vector regression (SVR) model with the aid of an optimization algorithm, and a reliable dataset was evaluated using a 10-fold cross-validation approach. The proposed model achieved strong predictive capability with low error values. The enhanced SVR model with 200 wolves and 100 iterations provides the best performance, with RMSE, MAE, and MAPE values of 28.14kN, 19.27kN, and 15.40%, respectively, along with a correlation coefficient of 0.955. Furthermore, the optimized models outperformed conventional SVR models with different kernel functions. These findings indicate that the enhanced SVR approach is a reliable and effective tool for predicting shear strength, thereby supporting improved structural analysis and design in engineering practice.