SV-NRC – Supervised New Recirculation Networks (Semeion©)
SV-NRC is the supervised version of the Auto Associative NRC. The NRC works as first networks to process the input vector. The output layer receives the NRC processing and tries to optimize the output square error. The two processing activities (NRC and final supervised error minimization) work independently each other and in parallel.
[1] P.M. Buscema
Recirculation Neural Networks
pp 383—388 in P.M.Buscema (ed), Substance use and misuse, vol.33(2). Special Issue on Artificial Neural Networks and Complex Social Systems, 1998.