Bounds for Approximate Solutions of Fredholm Integral Equations Using Kernel Networks

0359504 - ÚI 2012 RIV DE eng C - Konferenční příspěvek (zahraniční konf.)
Gnecco, G. - Kůrková, Věra - Sanguineti, M.
Bounds for Approximate Solutions of Fredholm Integral Equations Using Kernel Networks.
Artificial Neural Networks and Machine Learning - ICANN 2011. Part I. Berlin: Springer, 2011 - (Honkela, T.; Duch, W.; Girolami, M.; Kaski, S.), s. 126-133. Lecture Notes in Computer Science, 6791. ISBN 978-3-642-21734-0. ISSN 0302-9743.
[ICANN 2011. International Conference on Artificial Neural Networks /21./. Espoo (FI), 14.07.2011-17.07.2011]
Grant CEP: GA ČR GAP202/11/1368
Výzkumný záměr: CEZ:AV0Z10300504
Klíčová slova: radial and kernel networks * approximation of solutions of integral equations by kernel networks * model complexity
Kód oboru RIV: IN - Informatika

Approximation of solutions of integral equations by networks with kernel units is investigated theoretically. There are derived upper bounds on speed of decrease of errors in approximation of solutions of Fredholm integral equations by kernel networks with increasing numbers of units. The estimates are obtained for Gaussian and degenerate kernels.
Trvalý link: http://hdl.handle.net/11104/0197284