Evaluation of empirical orthogonal functions to the fuzzy inference system and artificial neural network to predict the flow

To predict the value of the streamflow, usually two methods of the Process-driven methods and Data – driven methods is used. Including Data – driven methods in river flow forecasting is artificial neural network, regression, time series and fuzzy logic. In this study, performance of another method empirical orthogonal functions than artificial neural network and fuzzy inference system to predict monthly inflow Latyan Dam reservoir was evaluated. Five models of ANN and ANFIS are similar and depends on rainfall, temperature and streamflow. Whereas five models of EOF depends on only streamflow in Latian station and Adjacent stations. Initial for all models the best combination identified, then statistical parameters of CE, MAPE, RMSE, CORR in the best combinations of all models were compared. The results showed that the fuzzy inference system, a better performance than the other way round.