Investigation of Internal Functioning of the Radial-Basis-Function Neural Network River Flow Forecasting Models

This paper deals with the challenging problem of hydrological interpretation of the internal functioning of artificial neural networks (ANNs) by extracting knowledge from their solutions. The neural network used in this study is based on the structure of the radial-basis-function neural network (RBFNN), which is considered as an alternative to the multilayer perceptron for solving complex modeling problems. This network consists of input, hidden, and output layers. The network is trained using the daily data of two catchments having different characteristics and from two different regions in the world. The present day and antecedent observed discharges are used as inputs to the network to forecast the flow one day ahead. A range of quantitative and qualitative techniques are used for hydrological interpretation of the internal functioning by examining the responses of the hidden layer nodes. The results of the study show that a single hidden layered RBFNN is an effective tool to forecast the daily flows and that the activation of the hidden layer nodes are far from arbitrary, but appear to represent flow components of the predicted hydrograph. The results of the study confirm that the three nodes in the hidden layer of this model effectively divide the input data space in such a way that the contribution from each node dominates in one of the flow domains — low, medium, or high — and form, in a crude manner, the base flow, interflow and surface runoff components of the hydrograph.

[1]  G. David Garson,et al.  Interpreting neural-network connection weights , 1991 .

[2]  K. P. Sudheer,et al.  Identification of physical processes inherent in artificial neural network rainfall runoff models , 2004 .

[3]  Kwok-wing Chau,et al.  Particle Swarm Optimization Training Algorithm for ANNs in Stage Prediction of Shing Mun River , 2006 .

[4]  Ferenc Szidarovszky,et al.  Application of Artificial Neural Networks to Complex Groundwater Management Problems , 2003 .

[5]  Daniel Rivero,et al.  A New Approach to the Extraction of ANN Rules and to Their Generalization Capacity Through GP , 2004, Neural Computation.

[6]  R. Abrahart,et al.  Detection of conceptual model rainfall—runoff processes inside an artificial neural network , 2003 .

[7]  Alan F. Murray,et al.  IEEE International Conference on Neural Networks , 1997 .

[8]  Lyle H. Ungar,et al.  Using radial basis functions to approximate a function and its error bounds , 1992, IEEE Trans. Neural Networks.

[9]  D. Broomhead,et al.  Nonlinear signal processing using radial basis functions , 1990 .

[10]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[11]  Donald F. Specht,et al.  A general regression neural network , 1991, IEEE Trans. Neural Networks.

[12]  Eric Gaume,et al.  Over-parameterisation, a major obstacle to the use of artificial neural networks in hydrology? , 2003 .

[13]  A. W. Jayawardena,et al.  Runoff Forecasting Using RBF Networks with OLS Algorithm , 1998 .

[14]  Fa-Long Luo,et al.  Applied neural networks for signal processing , 1997 .

[15]  Sheng Chen,et al.  Non-linear systems identification using radial basis functions , 1990 .

[16]  S. Jain,et al.  Radial Basis Function Neural Network for Modeling Rating Curves , 2003 .

[17]  Dimitri P. Solomatine,et al.  On the encapsulation of numerical-hydraulic models in artificial neural network , 1999 .

[18]  Joachim Diederich,et al.  The truth will come to light: directions and challenges in extracting the knowledge embedded within trained artificial neural networks , 1998, IEEE Trans. Neural Networks.

[19]  Jooyoung Park,et al.  Universal Approximation Using Radial-Basis-Function Networks , 1991, Neural Computation.

[20]  A. W. Jayawardena,et al.  Use of Radial Basis Function Type Artificial Neural Networks for Runoff Simulation , 1998 .

[21]  Jooyoung Park,et al.  Approximation and Radial-Basis-Function Networks , 1993, Neural Computation.

[22]  T. Rientjes,et al.  Constraints of artificial neural networks for rainfall-runoff modelling: trade-offs in hydrological state representation and model evaluation , 2005 .

[23]  James D. Keeler,et al.  Layered Neural Networks with Gaussian Hidden Units as Universal Approximations , 1990, Neural Computation.

[24]  R. Abrahart,et al.  Flood estimation at ungauged sites using artificial neural networks , 2006 .

[25]  Asaad Y. Shamseldin,et al.  Role of hidden neurons in a RBF type ANN in stream flow forecasting , 2007 .

[26]  R.J. Abrahart,et al.  Neural network river discharge forecasters: an empirical investigation of hidden unit processing functions based on two different catchments , 2005, Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005..

[27]  J. V. Sutcliffe,et al.  The Hydrology of the Nile , 1999 .

[28]  Jy-Shing Wu,et al.  Artificial Neural Networks for Forecasting Watershed Runoff and Stream Flows , 2005 .

[29]  Sheng Chen,et al.  Extended model set, global data and threshold model identification of severely non-linear systems , 1989 .

[30]  Yan Li,et al.  Comparison of Several Flood Forecasting Models in Yangtze River , 2005 .