The article is devoted to the development of a generalized training algorithm of the evolving radial base network (ERBN). ERBN is used for a wide range of tasks of identification, control, signal and image processing. Depending on the type of problem being solved a large number of functions with different number of adjustable parameters can be used in ERBN as the basis functions, and therefore there is a need for a generalized approach to configure them. A genetic algorithm is utilized for training the neural network, which significantly accelerate the learning process. The simulation results confirm the effectiveness of the proposed method of adjusting the parameters of the ERBN with using special non-coding regions of the chromosome - introns.
neuron network, base function, genetic algorithm, radially-base network, intron
"Obobshchennыi alhorytm obuchenyia эvoliutsyonyruiushchei radyalno-bazysnoi sety" ,
Information Processing Systems,