Structural identification of the multi-layered neural networks by using revised GMDH-type neural network algorithm with a feedback loop

In this paper, the revised GMDH-type neural network algorithm with a feedback loop is proposed. This algorithm is a revised version of the conventional GMDH-type neural network algorithm with a feedback loop. The revised GMDH-type neural networks can identify nonlinear complex system very accurately because the complexity of the neural network architecture increases gradually by the feedback loop calculations.