Incorporation of ARMA models into flow forecasting by artificial neural networks
- 10 April 2003
- journal article
- research article
- Published by Wiley in Environmetrics
- Vol. 14 (4) , 417-427
Abstract
A frequently encountered problem during the application of artificial neural networks (ANNs) to various water resource problems is the limitation of the data sets required for the training stage of ANNs. This prevents ANNs from learning input and output data sets within different ranges, thus decreasing the prediction capability during the testing stage. In this article the well known ARMA models are used to generate synthetic series, and these series are incorporated into the training data sets of ANNs. The method is applied to the monthly mean river flow data of a station in the East Mediterranean region of Turkey. The forecasting accuracy of the future monthly flows carries significance because a water reservoir is planned for the downstream of this station. Because the available data record length is limited studies should be carried out to extend the training data set of ANNs. It is seen that the extension of input and output data sets in the training stage improves the accuracy of forecasting using ANNs. The introduction of periodicity components in the input layer also increases the forecasting accuracy of ANNs. Copyright © 2003 John Wiley & Sons, Ltd.Keywords
This publication has 15 references indexed in Scilit:
- Bivariate stochastic modelling of ephemeral streamflowHydrological Processes, 2002
- Multivariate Reservoir Inflow Forecasting Using Temporal Neural NetworksJournal of Hydrologic Engineering, 2001
- Performance Evaluation of Artificial Neural Networks for Runoff PredictionJournal of Hydrologic Engineering, 2000
- A Method Based on Taking the Average of Probabilities to Compute the Flow Duration CurveHydrology Research, 2000
- A generalized seasonal model for flow duration curveHydrological Processes, 2000
- Forecasting river flow rate during low‐flow periods using neural networksWater Resources Research, 1999
- River flood forecasting with a neural network modelWater Resources Research, 1999
- Application of Gamma Autoregressive Model to Analysis of Dry PeriodsJournal of Hydrologic Engineering, 1998
- Runoff Forecasting Using RBF Networks with OLS AlgorithmJournal of Hydrologic Engineering, 1998
- Artificial Neural Network Modeling of the Rainfall‐Runoff ProcessWater Resources Research, 1995