Flood forecasting using radial basis function neural networks

Author

Chang, Fi-John ; Liang, Jin-Ming ; Chen, Yen-Chang

Author_Institution

Dept. of Bioenvironmental Syst. Eng., Nat. Taiwan Univ., Taipei, Taiwan

Volume

31

Issue

4

fYear

2001

fDate

11/1/2001 12:00:00 AM

Firstpage

530

Lastpage

535

Abstract

A radial basis function (RBF) neural network (NN) is proposed to develop a rainfall-runoff model for three-hour-ahead flood forecasting. For faster training speed, the RBF NN employs a hybrid two-stage learning scheme. During the first stage, unsupervised learning, fuzzy min-max clustering is introduced to determine the characteristics of the nonlinear RBFs. In the second stage, supervised learning, multivariate linear regression is used to determine the weights between the hidden and output layers. The rainfall-runoff relation can be considered as a linear combination of some nonlinear RBFs. Rainfall and runoff events of the Lanyoung River collected during typhoons are used to train, validate,and test the network. The results show that the RBF NN can be considered a suitable technique for predicting flood flow

Keywords

disasters; fuzzy set theory; geophysics computing; learning (artificial intelligence); radial basis function networks; rain; statistical analysis; weather forecasting; Lanyoung River; RBF NN; RBF neural network; flood flow prediction; fuzzy min-max clustering; hidden layers; hybrid two-stage learning scheme; linear combination; multivariate linear regression; nonlinear RBFs; output layers; radial basis function neural networks; rainfall-runoff model; rainfall-runoff relation; supervised learning; three-hour-ahead flood forecasting; training speed; typhoons; unsupervised learning; Floods; Linear regression; Neural networks; Predictive models; Radial basis function networks; Rivers; Supervised learning; Testing; Typhoons; Unsupervised learning;

fLanguage

English

Journal_Title

Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on

Publisher

ieee

ISSN

1094-6977

Type

jour

DOI

10.1109/5326.983936

Filename

983936