Please use this identifier to cite or link to this item:
|Title:||A Bayesian method to improve the extrapolation ability of ANNs|
|Citation:||Proceedings of the 14th IASTED International Conference on Applied Simulation and Modelling, June 15-17, 2005, Benalmadena, Spain / M.H. Hamza (ed.): pp.126-131|
|Publisher Place:||Anaheim, CA, USA|
|Conference Name:||International Conference on Applied Simulation and Modelling (14th : 2005 : Benalmadena, Spain)|
|G.B. Kingston, H.R. Maier, and M.F. Lambert|
|Abstract:||Although artiﬁcial neural networks have been shown to be superior prediction models in many hydrology-related areas, their known lack of extrapolation capability has limited the wider use and acceptance of ANNs as forecasting models. This problem lies mainly with the fact that a single 'most likely' weight vector, which is determined by calibration with a ﬁnite set of data, is used to deﬁne the function modelled by the ANN. There are, in fact, many different weight vectors that result in approximately equal model performance; however, standard ANN development approaches do not allow for any weight vectors, other than that which provides the best ﬁt to the calibration data, to impact on the predictions made. In this paper, a Bayesian method is presented that enables the entire range of plausible weight vectors to be accounted for in the model predictions. In doing so, the relationship modelled by the ANN is more general and less dominated by the information contained in the calibration data. The method is applied to a real-world case study known to require extrapolation and the resulting ANN is shown to perform signiﬁcantly better than an ANN developed using standard approaches.|
|Keywords:||Artiﬁcial neural networks; hydrological forecasting; Bayesian estimation; Markov chain Monte Carlo|
|Rights:||Copyright © 2005 ACTA Press|
|Appears in Collections:||Civil and Environmental Engineering publications|
Environment Institute publications
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.