Boosting the Performance of RBF Networks with Dynamic Decay Adjustment

Cite This

Files in this item

Checksum: MD5:d432561ea7a2f03822d8ccac67d4e63b

BERTHOLD, Michael R., Jay DIAMOND, 1995. Boosting the Performance of RBF Networks with Dynamic Decay Adjustment. In: Advances in Neural Information Processing. 7, pp. 8

@article{Berthold1995Boost-5427, title={Boosting the Performance of RBF Networks with Dynamic Decay Adjustment}, year={1995}, volume={7}, journal={Advances in Neural Information Processing}, author={Berthold, Michael R. and Diamond, Jay} }

<rdf:RDF xmlns:dcterms="" xmlns:dc="" xmlns:rdf="" xmlns:bibo="" xmlns:dspace="" xmlns:foaf="" xmlns:void="" xmlns:xsd="" > <rdf:Description rdf:about=""> <dcterms:hasPart rdf:resource=""/> <dc:format>application/pdf</dc:format> <dcterms:abstract xml:lang="eng">Radial Basis Function (RBF) Networks, also known as networks of locally-tuned processing units (see [6]) are well known for their ease of use. Most algorithms used to train these types of networks, however, require a fxed architecture, in which the number of units in the hidden layer must be determined before training starts. The RCE training algorithm, introduced by Reilly, Cooper and Elbaum (see [8]), and its probabilistic extension, the P-RCE algorithm, take advantage of a growing structure in which hidden units are only introduced when necessary. The nature of these algorithms allows training to reach stability much faster than is the case for gradient-descent based methods. Unfortunately P-RCE networks do not adjust the standard deviation of their prototypes individually, using only one global value for this parameter. This paper introduces the Dynamic Decay Adjustment (DDA) algorithm which utilizes the constructive nature of the P-RCE algorithm together with independent adaptation of each prototype's decay factor. In addition, this radial adjustment is class dependent and distinguishes between different neighbours. It is shown that networks trained with the presented algorithm perform substantially better than common RBF networks.</dcterms:abstract> <dspace:isPartOfCollection rdf:resource=""/> <dcterms:issued>1995</dcterms:issued> <dspace:hasBitstream rdf:resource=""/> <dcterms:bibliographicCitation>First publ. in: Advances in Neural Information Processing 7 (1995), pp. 8</dcterms:bibliographicCitation> <dc:creator>Berthold, Michael R.</dc:creator> <dcterms:available rdf:datatype="">2011-03-24T15:55:18Z</dcterms:available> <dc:contributor>Diamond, Jay</dc:contributor> <dcterms:rights rdf:resource=""/> <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/> <foaf:homepage rdf:resource="http://localhost:8080/jspui"/> <dc:rights>Attribution-NonCommercial-NoDerivs 2.0 Generic</dc:rights> <bibo:uri rdf:resource=""/> <dc:language>eng</dc:language> <dc:creator>Diamond, Jay</dc:creator> <dc:contributor>Berthold, Michael R.</dc:contributor> <dcterms:isPartOf rdf:resource=""/> <dcterms:title>Boosting the Performance of RBF Networks with Dynamic Decay Adjustment</dcterms:title> <dc:date rdf:datatype="">2011-03-24T15:55:18Z</dc:date> </rdf:Description> </rdf:RDF>

Downloads since Oct 1, 2014 (Information about access statistics)

BeDi95_dda_nips7.pdf 268

This item appears in the following Collection(s)

Attribution-NonCommercial-NoDerivs 2.0 Generic Except where otherwise noted, this item's license is described as Attribution-NonCommercial-NoDerivs 2.0 Generic

Search KOPS


My Account