For the most recent entries see the Petri Nets Newsletter.

Fast Transparent Neuro-Fuzzy Classifiers.

Halgamuge, S.K.; Glesner, M.

In: Proc. 15th IASTED International Conference Modelling, Identification and Control, Feb 19-21, 1996, pages 407-410. Acta Press, 1996. ISBN 0-88986-193-5.

Abstract: Despite the advantages of high generality and robustness, gradient descent learning methods are in general slow, and the problem of local minima can be disadvantageous for some applications. Therefore, new techniques for on-line training of high fan-in compact neural networks that can be interpreted as fuzzy rule based classifier systems are developed. Several novel algorithms based on supervised competitive learning are presented on-line generation of nearest prototypes. Showing that classifying radial basis function (RBFN) neural networks and the dynamic vector quantisation neural networks are functionally equivalent to fuzzy systems, the three year research project on ``Fast classifiers'' funded by the German Research Foundation (DFG) is outlined presenting application results and the hardware implementation efforts.

Keywords: fuzzy systems; neural networks; Bayes classifier; rule generation.

Do you need a refined search? Try our search engine which allows complex field-based queries.

Back to the Petri Nets Bibliography