Comparison of ANFIS and RBF models in daily stream flow forecasting

Abstract

This paper compares two expert models in daily stream flow forecasting. The adaptive-neuro fuzzy inference system (ANFIS) model, artificial neural networks with radial base function (RBF) are used to forecast daily river flow in northwest of Iran and the results of these models are compared with Observed daily values. Daily river flow data in Mahabad-Dam station on Mahabad river in northwest of Iran are used in this study. The comparison results show that the ANFIS model have better performances in forecasting of river flow from RBF. The R2, MAPE and RRMSE values of RBF model in training step are 0.829, 0.033826 and 0.981794 respectively and in testing step 0.6959, 0.018752 and 1.804379 respectively. The R2, MAPE and RRMSE values of ANFIS model in training step are 0.8378, 0.042887 and 0.874971 respectively and in testing step 0.7147, 0.042887 and 1.231364 respectively. Based on the results, using ANFIS models in river flow forecasting and water resources management results in accurate simulations.