Title :
On representation and generalization capability of pyramid neural networks
Author :
Hoshino, Miho ; Chao, Jinhui
Author_Institution :
Dept. of Electr. & Electron. Eng., Chuo Univ., Tokyo, Japan
fDate :
6/24/1905 12:00:00 AM
Abstract :
This paper presents a study of a recently proposed network called pyramid network theoretically and experimentally. First we show a lower boundary for the VC dimension of the network and also an analysis on its generalization performance. Then simulation experiments are presented, and they agree with the theoretical analysis
Keywords :
data structures; generalisation (artificial intelligence); multilayer perceptrons; VC dimension; generalization capability; pyramid neural networks; representation capability; Analytical models; Chaotic communication; Multi-layer neural network; Network topology; Neural networks; Neurons; Nonhomogeneous media; Pattern recognition; Performance analysis; Virtual colonoscopy;
Conference_Titel :
Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
0-7803-7278-6
DOI :
10.1109/IJCNN.2002.1007659
