Adjusting the parameters of radial basis function networks using Particle Swarm Optimization

Abstract

Particle Swarm Optimization (PSO), a new promising evolutionary optimization technique, has a wide range of application in optimization problems including training of artificial neural networks. In this paper, an attempt is made to completely train a RBF neural network architecture including the centers, optimum spreads, and the number of hidden units. The proposed method has been evaluated on some benchmark problems: Iris, Wine, Glass, New-thyroid and its accuracy was compared with other algorithms. The results show its strong generalization ability.