Firefly optimization algorithm effecton support vector regression predictioni mprovement of amodified labyrinth side weir’s discharge coefficient

Abstract

A principal step in designing dividing hydraulic structures entails determining the side weir discharge coefficient. In this study, Firefly optimization-based Support Vector Regression (SVR-FF) is introduced and examined in terms of predicting the discharge coefficient of a modified labyrinth side weir. Ten non-dimensional parameters of various geometrical and hydraulic conditions are defined as the input parameters for the SVR-FF and the side weir discharge coefficient is defined as the output. Improvements in SVR prediction accuracy are determined by comparing SVR-FF with the traditional SVR model. The results indicate that the SVR-FF model with RMSE of 0.035 is about 10% more accurate than SVR with RMSE of 0.039. Thus, combining the Firefly optimization algorithm with SVR increases the prediction model performance.