Journal article
Study of the approximation capabilities of two four-layered neural networks
Optical Engineering, Vol.37(3), pp.889-897
1998
DOI: 10.1117/1.601923
Abstract
The approximation capabilities of two different four-layered neural networks are studied. First, a network with the backpropagation algorithm is analyzed, and its error surface, convergence properties, and network design are considered. An alternative to the backpropagation approach is presented, namely, we construct a network that uses a one-pass algorithm. We show that the proposed network can correctly classify
different patterns with
hidden units. We also show that an arbitrarily small approximation error can be obtained for this network by adjusting the appropriate parameters. ©
Details
- Title: Subtitle
- Study of the approximation capabilities of two four-layered neural networks
- Creators
- Claudia Mello-Thoms - Rutgers, The State University of New JerseyStanley M Dunn - Rutgers, The State University of New Jersey
- Resource Type
- Journal article
- Publication Details
- Optical Engineering, Vol.37(3), pp.889-897
- DOI
- 10.1117/1.601923
- ISSN
- 0091-3286
- eISSN
- 1560-2303
- Language
- English
- Date published
- 1998
- Academic Unit
- Roy J. Carver Department of Biomedical Engineering; Radiology
- Record Identifier
- 9984318814902771
Metrics
8 Record Views