Implementation of the Supervised Growing Cell Structure on the CNAPS Neurocomputer

IntroductionThe growing cell structure is a generalization of Kohonens feature map [1][3]. Inthe case of supervised learning, the supervised growing cell structure (SGCS) is a neuralnetwork type which combines the properties of the classical feed forward with the selforganizing networks. During training it shows a dynamical structure: the number of nodesis increasing continuously. The network performance is superior to common networkmodels as tests with classical benchmarks have shown...