Presentation 2005-07-30
Characteristics of BP Neural Networks for Pattern Classification
Hirohito SHINTANI, Masatake AKUTAGAWA, Hirofumi NAGASHINO, Yohsuke KINOUCHI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) It is difficult to find that the neural network use where part of patterns for the recognition. Cause of that is nonlinearity of networks. This paper describes a recognition mechanism of a four layer backpropagation neural network using Alopex algorithm. We have developed a small scale four-layered neural network model for simple character recognition, which can recognize the patterns transformed by affine conversion. Alopex algorithm is an iterative and stochastic processing to minimize or maximize a cost function. By these methods the receptive fields of the units in the output layers are obtained.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) neural network / characteristics / back propagation / Alopex algorithm
Paper # MBE2005-53
Date of Issue

Conference Information
Committee MBE
Conference Date 2005/7/23(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To ME and Bio Cybernetics (MBE)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Characteristics of BP Neural Networks for Pattern Classification
Sub Title (in English)
Keyword(1) neural network
Keyword(2) characteristics
Keyword(3) back propagation
Keyword(4) Alopex algorithm
1st Author's Name Hirohito SHINTANI
1st Author's Affiliation Faculty of Engineering, The Univ. of Tokushima()
2nd Author's Name Masatake AKUTAGAWA
2nd Author's Affiliation Faculty of Engineering, The Univ. of Tokushima
3rd Author's Name Hirofumi NAGASHINO
3rd Author's Affiliation School of health sciences, The Univ. of Tokushima
4th Author's Name Yohsuke KINOUCHI
4th Author's Affiliation Faculty of Engineering, The Univ. of Tokushima
Date 2005-07-30
Paper # MBE2005-53
Volume (vol) vol.105
Number (no) 222
Page pp.pp.-
#Pages 4
Date of Issue