Presenter/Author Information

K. Hsu
S. Sorooshian
H. V. Gupta
X. Gao
B. Imam

Keywords

self-organizing feature map, principal component regression, rainfall-runoff process

Start Date

1-7-2002 12:00 AM

Description

Artificial neural networks (ANNs) have been broadly applied to many hydrological applicationsfor which their underlying processes are complicated nonlinear. Although many networks, such as multilayerfeedforward neural networks (MFNs), provide excellent capability in function fittings, very often, theyare referred to as black-box models. In this study, a multivariate ANN procedure, entitled SOLO (Self-Organizing Linear Output mapping network) is introduced. This model architecture has been designed forrapid estimation of network structure/parameters and system outputs. Furthermore, the SOLO providesfeatures that facilitate insight to the input-output processes, thereby extending its usefulness as a tool forinvestigations into the underlying processes through the data classification processes. A case study usingSOLO model in a hydrologic rainfall-runoff forecasting is demonstrated. Uncertainty of model estimates isalso evaluated.

Share

COinS
 
Jul 1st, 12:00 AM

Hydrologic Modelling and Analysis Using A Self- Organizing Linear Output Network

Artificial neural networks (ANNs) have been broadly applied to many hydrological applicationsfor which their underlying processes are complicated nonlinear. Although many networks, such as multilayerfeedforward neural networks (MFNs), provide excellent capability in function fittings, very often, theyare referred to as black-box models. In this study, a multivariate ANN procedure, entitled SOLO (Self-Organizing Linear Output mapping network) is introduced. This model architecture has been designed forrapid estimation of network structure/parameters and system outputs. Furthermore, the SOLO providesfeatures that facilitate insight to the input-output processes, thereby extending its usefulness as a tool forinvestigations into the underlying processes through the data classification processes. A case study usingSOLO model in a hydrologic rainfall-runoff forecasting is demonstrated. Uncertainty of model estimates isalso evaluated.