Evolutionary learning of regularization networks with product kernel units

Authors: Petra Vidnerová, Roman Neruda
Publication year: 2011
Publication type: Konferenční příspěvek (mezinárodní konf.)
Periodical: Systems, Man and Cybernetics
Projects: Learning of functional relationships from high-dimensional data
Neural Networks Learning Algorithms Based on Regularization Theory
Bibliographic citation: VIDNEROVÁ, Petra, NERUDA, Roman. Evolutionary learning of regularization networks with product kernel units. In: Systems, Man and Cybernetics. Piscataway: IEEE, 2011, S. 638-643. ISBN 978-1-4577-0652-3. ISSN 1062-922X
Dostupné z: doi:10.1109/ICSMC.2011.6083783
DOI: 10.1109/ICSMC.2011.6083783
ID: 369174

Abstract

This paper deals with learning possibilities of regularization networks with product kernel units. Approximation problems formulated as regularized minimization problems with kernel-based stabilizers lead to solutions of the shape of linear combination of kernel functions. These can be expressed as one-hidden layer feed-forward neural network schemes, called regularization networks. We propose a novel evolutionary algorithm utilizing for regularization networks with product kernels. This algorithm utilizes genetic search for suitable network parameters as well as kernel functions.