Issues and Challenges of Intelligent Systems and Computational Intelligence

0398349 - ÚI 2014 RIV DE eng M - Část monografie knihy
Kůrková, Věra
Accuracy of surrogate solutions of integral equations by feedforward networks.
Issues and Challenges of Intelligent Systems and Computational Intelligence. Cham: Springer, 2014 - (Kóczy, L.; Pozna, C.; Kacprzyk, J.), s. 91-102. Studies in Computational Intelligence, 530. ISBN 978-3-319-03205-4
Grant CEP: GA ČR GAP202/11/1368; GA MŠMT(CZ) LD13002
Institucionální podpora: RVO:67985807
Klíčová slova: surrogate modeling by neural networks * approximate solutions of integral equations * feedforward neural networks * model complexity * rates of approximation
Kód oboru RIV: IN - Informatika

Surrogate solutions of Fredholm integral equations by feedforward neural networks are investigated theoretically. Convergence of surrogate solutions computable by networks with increasing numbers of computational units to theoretically optimal solutions is proven and upper bounds on rates of convergence are derived. The results hold for a variety of computational units, they are illustrated by examples of perceptrons and Gaussian radial units.
Trvalý link: http://hdl.handle.net/11104/0225861