Application of the MLR and MLP models in prediction of daily river flow

Recently, several studies concentrated to investigations of the river flow forecasting in the throughout the world. However, it is important to finding variability of river flow especially in reservoir inflow due to current and future demand and supply management. This paper focused on application of the MLR (Multiple Linear Regression) and MLP (Multi-Layer Perceptrons) in prediction of daily river flow time series in Jelogir station, west of Iran. The results of two methods based on the MSE, RMSE as well as correlation coefficient showed good performances in daily river flow forecasting. Based on the MLR model, the results determined that the model 4 was found the best model among anthers, in comparison, the MLR model showed the scenario 4 with Qt-1, Qt-2, Qt-3, Qt-4 inputs had the lowest error between observed and estimated datasets. Moreover, the results showed that in some cases the outputs of these models were the same.

کلیدواژه ها:

MLR ، Reservoir ، MLP ، River flow ، Karkheh Dam