DocumentCode :
381200
Title :
A new regularization learning method for improving generalization capability of neural network
Author :
Wu, Yan ; Zhang, Liming
Author_Institution :
Dept. of Comput. Sci. & Eng., Tongji Univ., Shanghai, China
Volume :
3
fYear :
2002
fDate :
2002
Firstpage :
2011
Abstract :
When the network structure has been determined, it is very effective when regulation methods are used to improve generalization capability. However, there are some obvious drawbacks: long computation time, and difficulty in parameter control. The paper proposes a novel method that dynamically tunes the regularization coefficient by fuzzy rule inference, effectively determining the fuzzy inference rules and membership functions. Furthermore, it compares the method with the traditional BP algorithm and the fixed regularization coefficients method. The proposed method has the merits of the highest precision, rapid convergence, and the best generalization capability. Finally, it indicates that the proposed method is a very effective method by several nonlinear function approximation and pattern classification examples.
Keywords :
function approximation; fuzzy logic; generalisation (artificial intelligence); inference mechanisms; learning (artificial intelligence); neural nets; pattern classification; uncertainty handling; backpropagation; computation time; convergence; fixed regularization coefficients method; fuzzy rule inference; generalization; membership functions; neural network; nonlinear function approximation; parameter control; pattern classification; regularization coefficient; regularization learning method; regulation methods; Computer science; Convergence; Electronic mail; Function approximation; Fuzzy control; Fuzzy neural networks; Inference algorithms; Learning systems; Neural networks; Pattern classification;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Control and Automation, 2002. Proceedings of the 4th World Congress on
Print_ISBN :
0-7803-7268-9
Type :
conf
DOI :
10.1109/WCICA.2002.1021438
Filename :
1021438
Link To Document :
بازگشت