Abstract:

 A method of intelligent optimization strategy for extreme learning machine-radial basis function (ELM-RBF) neural networks was proposed, in which the centers and impact widths of hidden neural kernels were determined by the intelligent optimization algorithms of differential evolution and particle swarm optimization. Simulation results showed that the ELM-RBF neural networks learning algorithm based on the intelligent optimization strategy could generate much better generalization performance and robustness than other algorithms with the same network architecture.

Key words:  radial basis function neural networks; intelligent optimization; differential evolution algorithm; particle swarm optimization algorithm