Fast Orthogonal Identification of Nonlinear Stochastic Models and Radial Basis Function Neural Networks

摘要

A new fast orthogonal estimation algorithm is derived for a wide class of nonlinear stochastic models including training radial basis function neural networks. The selection of significant regressors and the estimation of unknown parameters in the presence of nonlinear noise sources are considered and simulated examples are included to demonstrate the efficiency of the new procedure.