Assessment of water surface profile in nonprismatic compound channels using machine learning techniques

Abstract Accurate prediction of water surface profile in an open channel is the key to solving numerous critical engineering problems. The goal current research predict a compound with converging floodplains using machine learning approaches, including gene expression programming (GEP), artificial neural networks (ANNs), and support vector machines (SVMs), terms both geometric flow variables, as past studies were more focused on variables. A novel equation was also proposed profile. In order evaluate performance efficacy these models, statistical indices are used validate produced models for experimental analysis. findings demonstrate that suggested ANN model accurately predicted profile, coefficient determination (R2) 0.999, root mean square error (RMSE) 0.003, absolute percentage (MAPE) 0.107%, respectively, when compared GEP, SVM, previously developed methods. study confirms application approaches field river hydraulics, forecasting nonprismatic channels by made this unique.