Loughborough University
Leicestershire, UK
LE11 3TU
+44 (0)1509 263171
Loughborough University

Loughborough University Institutional Repository

Please use this identifier to cite or link to this item: https://dspace.lboro.ac.uk/2134/14361

Title: Sensitivity analysis for comparison, validation and physical legitimacy of neural network-based hydrological models
Authors: Dawson, Christian W.
Mount, Nick J.
Abrahart, Robert J.
Louis, J.
Keywords: Generalised linear model
Index flood
Neural network
Partial derivative
Physical legitimacy
Sensitivity analysis
Ungauged catchment
Issue Date: 2014
Publisher: © IWA Publishing
Citation: DAWSON, C.W. ... et al., 2014. Sensitivity analysis for comparison, validation and physical legitimacy of neural network-based hydrological models. Journal of Hydroinformatics, 16 (2), pp. 407–424.
Abstract: This paper addresses the difficult question of how to perform meaningful comparisons between neural network-based hydrological models and alternative modelling approaches. Standard, goodness-of-fit metric approaches are limited since they only assess numerical performance and not physical legitimacy of the means by which output is achieved. Consequently, the potential for general application or catchment transfer of such models is seldom understood. This paper presents a partial derivative, relative sensitivity analysis method as a consistent means by which the physical legitimacy of models can be evaluated. It is used to compare the behaviour and physical rationality of a generalised linear model and two neural network models for predicting median flood magnitude in rural catchments. The different models perform similarly in terms of goodness-of-fit statistics, but behave quite distinctly when the relative sensitivities of their inputs are evaluated. The neural solutions are seen to offer an encouraging degree of physical legitimacy in their behaviour, over that of a generalised linear modelling counterpart, particularly when overfitting is constrained. This indicates that neural models offer preferable solutions for transfer into ungauged catchments. Thus, the importance of understanding both model performance and physical legitimacy when comparing neural models with alternative modelling approaches is demonstrated.
Description: © IWA Publishing. The definitive peer-reviewed and edited version of this article is published in Journal of Hydroinformatics, 16 (2), pp. 407–424, doi:10.2166/hydro.2013.222 and is available at www.iwapublishing.com
Version: Accepted for publication
DOI: 10.2166/hydro.2013.222
URI: https://dspace.lboro.ac.uk/2134/14361
Publisher Link: http://dx.doi.org/10.2166/hydro.2013.222
ISSN: 1464-7141
Appears in Collections:Published Articles (Computer Science)

Files associated with this item:

File Description SizeFormat
Dawson et al SenAnal Final BA.pdfAccepted version1.35 MBAdobe PDFView/Open


SFX Query

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.