Self-adjustment of neuron impact width in growing and pruning RBF (GAP-RBF) neuron networks

Although the growing and pruning algorithm for RBF networks (GAP-RBF), although it has acquired better performance than other sequential learning algorithms, some parameters (including the overlapping factor) predetermined by trial-and-error may affect the performance of the algorithm and limit the algorithm to be widely conveniently used in real world applications. In this paper, a self-adjustment algorithm based on GAP-RBF is proposed for solving how to choose k, the overlap factor, an important parameter for calculating neuron impact width. Simulation results indicate that the self-adjustment algorithm has the better performance than that of GAP-RBF.