Presentation | 2016-10-22 Constructing Multi-Class Classifiers by Combination of 2-Class Layered Neural Networks Shota Utsumi, Keisuke Kameyama, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Exhaustive code is one of the error correcting codes usable inimproving the accuracy of multi-class classifier by manipulating themisclassifications. When the subdigits of exhaustive code are chosen, conventionally, onlythe Hamming distance between codewords were considered, regardless of theperformances of the 2-class classifiers that generate the digit of anexhaustive code. In this paper, we propose a new method for choosingthe subdigits of exhaustive code, by considering the accuracy of the2-class classifiers and their expected outputs. In the experiments, weconfirmed that the accuracy of the multi-class classification usingthe proposed methods give higherperformances when compared with the conventional method. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | neural networks / error correcting code / two-class classifier / exhaustive code |
Paper # | BioX2016-18,MBE2016-32,NC2016-17 |
Date of Issue | 2016-10-15 (BioX, MBE, NC) |
Conference Information | |
Committee | MBE / BioX / NC |
---|---|
Conference Date | 2016/10/22(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Osaka Electro-Communication University |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Yutaka Fukuoka(Kogakuin Univ.) / Masakatsu Nishigaki(Shizuoka Univ.) / Shigeo Sato(Tohoku Univ.) |
Vice Chair | Kazuki Nakajima(Univ. of Toyama) / Akira Otsuka(AIST) / Hiroshi Takano(Toyama Pref. Univ.) / Masafumi Hagiwara(Keio Univ.) |
Secretary | Kazuki Nakajima(Kogakuin Univ.) / Akira Otsuka(Toyama Pref. Univ.) / Hiroshi Takano(NEC) / Masafumi Hagiwara(AIST) |
Assistant | Ryota Horie(Shibaura Inst. of Tech.) / Kim Juhyon(Univ. of Toyama) / Masatsugu Ichino(Univ. of Electro-Comm.) / Naoyuki Takada(Secom) / Takahiro Aoki(Fujitsu Labs.) / Hisanao Akima(Tohoku Univ.) / Yoshihisa Shinozawa(Keio Univ.) |
Paper Information | |
Registration To | Technical Committee on ME and Bio Cybernetics / Technical Committee on Biometrics / Technical Committee on Neurocomputing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Constructing Multi-Class Classifiers by Combination of 2-Class Layered Neural Networks |
Sub Title (in English) | |
Keyword(1) | neural networks |
Keyword(2) | error correcting code |
Keyword(3) | two-class classifier |
Keyword(4) | exhaustive code |
1st Author's Name | Shota Utsumi |
1st Author's Affiliation | University of Tsukuba(Univ. of Tsukuba) |
2nd Author's Name | Keisuke Kameyama |
2nd Author's Affiliation | University of Tsukuba(Univ. of Tsukuba) |
Date | 2016-10-22 |
Paper # | BioX2016-18,MBE2016-32,NC2016-17 |
Volume (vol) | vol.116 |
Number (no) | BioX-263,MBE-264,NC-265 |
Page | pp.pp.7-12(BioX), pp.7-12(MBE), pp.7-12(NC), |
#Pages | 6 |
Date of Issue | 2016-10-15 (BioX, MBE, NC) |