Runoff prediction using hydro-meteorological variables and a new hybrid ANFIS-GPR model

Abstract Precise and credible runoff forecasting is extraordinarily vital for various activities of water resources deployment implementation. The neoteric contribution the current article to develop a hybrid model (ANFIS-GPR) based on adaptive neuro-fuzzy inference system (ANFIS) Gaussian process regression (GPR) monthly in Beiru river China, optimal input schemes models are discussed detail. Firstly, variables related selected from precipitation, soil moisture content, evaporation as first set according correlation analysis (CA). Secondly, principal component (PCA) used eliminate redundant information between original forming second schemes. Finally, predicted different models, prediction performance compared comprehensively. results show that jointly established by CA PCA (CA-PCA) can greatly improve accuracy. ANFIS-GPR displays best among all peer models. In single GPR better than ANFIS.