A Method to Select RBFNN´s Center Based on the SOFM Network

Author

Zheng, Mingwen ; Zhang, Yanping

Author_Institution

Sch. of Sci., ShanDong Univ. of Technol., Zibo, China

Volume

2

fYear

2012

Firstpage

87

Lastpage

89

Abstract

In order to improve the RBFNN´s central choice method, proposed one kind of optimized choice radial basis function neural network data center´s algorithm. This algorithm combined SOFM network´s pattern classification ability, preliminary results of the classification used as the initial RBFNN data center, and then used OLS algorithm to train RBFNN. Simulation experiments show that the algorithm´s approximation ability and generalization ability is better than RBFNN only used the OLS algorithm.

Keywords

least squares approximations; pattern classification; radial basis function networks; self-organising feature maps; OLS algorithm; RBFNN center selection; RBFNN central choice method; SOFM network pattern classification ability; algorithm approximation ability; generalization ability; orthogonal least squares method; radial basis function neural network data center algorithm; self-organization mapping net; Approximation algorithms; Approximation methods; Classification algorithms; Mathematical model; Radial basis function networks; Training; Vectors; Orthogonal Least Squares Method; Radial Basis Function Neural Network; SOFM Network; input samples;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Science and Electronics Engineering (ICCSEE), 2012 International Conference on

Print_ISBN

978-1-4673-0689-8

Type

conf

DOI

10.1109/ICCSEE.2012.5

Filename

6187971