Kernel Networks with Fixed and Variable Widths

0357302 - ÚI 2012 RIV DE eng C - Konferenční příspěvek (zahraniční konf.)
Kůrková, Věra - Kainen, P.C.
Kernel Networks with Fixed and Variable Widths.
Adaptive and Natural Computing Algorithms. Part I. Vol. 1. Heidelberg: Springer, 2011 - (Dobnikar, A.; Lotrič, U.; Šter, B.), s. 12-21. Lecture Notes in Computer Science, 6593. ISBN 978-3-642-20281-0. ISSN 0302-9743.
[ICANNGA'2011. International Conference /10./. Ljubljana (SI), 14.04.2011-16.04.2011]
Grant CEP: GA MŠMT OC10047; GA MŠMT(CZ) ME10023
Výzkumný záměr: CEZ:AV0Z10300504
Klíčová slova: kernel models * Gaussian kernel networks * minimization of error functionals * regularization
Kód oboru RIV: IN - Informatika

The role of width in kernel models and radial-basis function networks is investigated with a special emphasis on the Gaussian case. Quantitative bounds are given on kernel-based regularization showing the effect of changing the width. These bounds are shown to be d-th powers of width ratios, and so they are exponential in the dimension of input data.
Trvalý link: http://hdl.handle.net/11104/0195612