Design of multi-layer neural networks for accurate identification of nonlinear mappingsGuidelines for the design of multilayer neural networks for the identification of nonlinear mappings are considered. Since nonlinear mappings can be approximated by a one-hidden-layer neural network, an approach to determine the sufficient number of hidden layer nodes to achieve a global minima of the identification error function is considered.
Document ID
19920046406
Acquisition Source
Legacy CDMS
Document Type
Conference Paper
Authors
Teixeira, Edilberto (NASA Lewis Research Center Cleveland, OH, United States)
Loparo, Kenneth (Case Western Reserve University Cleveland, OH, United States)
Gomide, Fernando A. C. (Campinas, Universidade Estadual, Brazil)