%0 Journal Article
%T Kernel CMAC: an Efficient Neural Network for Classification and Regression
%A G¨˘bor Horv¨˘th
%J Acta Polytechnica Hungarica
%D 2006
%I ?buda University
%X Kernel methods in learning machines have been developed in the last decade asnew techniques for solving classification and regression problems. Kernel methods havemany advantageous properties regarding their learning and generalization capabilities,but for getting the solution usually the computationally complex quadratic programming isrequired. To reduce computational complexity a lot of different versions have beendeveloped. These versions apply different kernel functions, utilize the training data indifferent ways or apply different criterion functions. This paper deals with a special kernelnetwork, which is based on the CMAC neural network. Cerebellar Model ArticulationController (CMAC) has some attractive features: fast learning capability and thepossibility of efficient digital hardware implementation. Besides these attractive featuresthe modelling and generalization capabilities of a CMAC may be rather limited. The papershows that kernel CMAC ¨C an extended version of the classical CMAC networkimplemented in a kernel form ¨C improves that properties of the classical versionsignificantly. Both the modelling and the generalization capabilities are improved while thelimited computational complexity is maintained. The paper shows the architecture of thisnetwork and presents the relation between the classical CMAC and the kernel networks.The operation of the proposed architecture is illustrated using some common benchmarkproblems.
%K kernel networks
%K input-output system modelling
%K neural networks
%K CMAC
%K generalization error
%U http://uni-obuda.hu/journal/HorvathGabor_5.pdf