DocumentCode
313592
Title
Network complexity and generalization
Author
Park, Sangbong ; Park, Cheol Hoon
Author_Institution
Dept. of Electr. Eng., Korea Adv. Inst. of Sci. & Technol., Seoul, South Korea
Volume
1
fYear
1997
fDate
9-12 Jun 1997
Firstpage
298
Abstract
This paper explains the relationship between complexity of the neural network with sigmoidal hidden neurons and its generalization capability in function approximation. Network complexity is decided in terms of the number of degrees of freedom and their dynamic range. Computer simulation shows that dynamic range as well as degrees of freedom affects training and generalization capability
Keywords
digital simulation; function approximation; genetic algorithms; learning (artificial intelligence); mathematics computing; multilayer perceptrons; degrees of freedom; dynamic range; function approximation; generalization capability; network complexity; neural network; sigmoidal hidden neurons; Artificial neural networks; Computer simulation; Dynamic range; Electronic mail; Estimation error; Function approximation; Neural networks; Neurons; Optimization methods; Testing;
fLanguage
English
Publisher
ieee
Conference_Titel
Neural Networks,1997., International Conference on
Conference_Location
Houston, TX
Print_ISBN
0-7803-4122-8
Type
conf
DOI
10.1109/ICNN.1997.611682
Filename
611682
Link To Document