Abstract: In this article, the combined cascade radial basis network is proposed and a hybrid learning method is presented, in which the optimal values for network’s parameters are calculated in two stages. In the first stage, the nonlinear parameters of radial basis functions are optimized and, in the second stage, the connection weights are optimized.

Keywords: radial basis function network, neural network, cascade neural network, learning algorithm, neural network model.

ACM Classification Keywords: I.2.6 Connectionism and neural nets.

Link:

МОДЕЛЬ КОМБИНИРОВАННОЙ КАСКАДНОЙ РАДИАЛЬНО БАЗИСНОЙ НЕЙРОННОЙ СЕТИ И АЛГОРИТМ ЕЕ ОБУЧЕНИЯ (The Model for the Combined Cascade Radial Basis Neural Network and Its Learning Algorithm)

Елена Зайченко, Екатерина Малышевская

http://www.foibg.com/ijitk/ijitk-vol09/ijitk09-03-p06.pdf