Model Complexity of Neural Networks in High-Dimensional Approximation

0360537 - ÚI 2013 RIV DE eng C - Conference Paper (international conference)
Kůrková, Věra
Model Complexity of Neural Networks in High-Dimensional Approximation.
Recent Advances in Intelligent Engineering Systems. Vol. 1. Berlin: Springer, 2012 - (Fodor, S.; Klempous, J.; Suárez Araujo, C.), s. 151-160. Studies in Computational Intelligence, 378. ISBN 978-3-642-23228-2. ISSN 1860-949X.
[INES 2010. International Conference on Intelligent Engineering Systems /14./. Las Palmas de Gran Canaria (ES), 05.05.2010-07.05.2010]
R&D Projects: GA MŠMT OC10047; GA MŠMT MEB040901
Institutional research plan: CEZ:AV0Z10300504
Keywords : model complexity of neural networks * Gaussian radial-basis networks * dependence on input dimension
Subject RIV: IN - Informatics, Computer Science

The role of dimensionality in approximation by neural networks is investigated. Methods from nonlinear approximation theory are used to describe sets of functions which can be approximated by neural networks with a polynomial dependence of model complexity on the input dimension. The results are illustrated by examples of Gaussian radial networks.
Permanent Link: http://hdl.handle.net/11104/0198055