Structure optimization algorithm for radial basis probabilistic neural networks based on the moving median center hyperspheres algorithm

Abstract

In this paper, a novel structure optimization algorithm for radial basis probabilistic neural networks (RBPNN) is proposed. Firstly, a moving median center hyperspheres (MMCH) algorithm is proposed to heuristically select the initial hidden layer centers of the RBPNN, and then a hybrid optimization algorithm is adopted to further prune the initial structure of the RBPNN. Finally, the effectiveness and efficiency of our proposed algorithm are evaluated through a plant species identification problem and a palmprint recognition task. © 2009 Springer Berlin Heidelberg.