Presentation | 2017-02-18 Small-sized Kernel Classifier By Support Vector Retraining Based on Minimum Classification Error Criterion Ryoma Tani, Hideyuki Watanabe, Shigeru Katagiri, Miho Osaki, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Different from the Multi-class Support Vector Machine (MSVM) that fixes Support Vectors (SVs) to training samples, the Kernel Minimum Classification Error (KMCE) training method readjusts SVs, each being treated as a sample-based prototype. Focusing on this feature of KMCE method, we aim to develop a small-sized accurate kernel classifier by reducing SVs based on the norm of the weight vector for SV and applying the KMCE training. In four classification experiments, we show that our proposed procedure can effectively reduce SVs depending on the classification difficulty of task and successfully achieve small-sized classifiers. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Multi-class Support Vector Machine / Kernel Minimum Classification Error Training / Kernel |
Paper # | PRMU2016-159,CNR2016-26 |
Date of Issue | 2017-02-11 (PRMU, CNR) |
Conference Information | |
Committee | PRMU / CNR |
---|---|
Conference Date | 2017/2/18(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Eisaku Maeda(NTT) / Michita Imai(Keio Univ.) |
Vice Chair | Seiichi Uchida(Kyushu Univ.) / Hironobu Fujiyoshi(Chubu Univ.) / Tetsuo Ono(Hokkaido Univ.) / Masayuki Kanbara(NAIST) |
Secretary | Seiichi Uchida(Kyoto Univ.) / Hironobu Fujiyoshi(NTT) / Tetsuo Ono(Univ. of Tokyo) / Masayuki Kanbara(Univ. of Tsukuba) |
Assistant | Masaki Oonishi(AIST) / Takuya Funatomi(NAIST) / Kosuke Yoshioka(Panasonic) / Daisuke Yamamoto(Toshiba) / Takahiro Matsumoto(NTT) |
Paper Information | |
Registration To | Technical Committee on Pattern Recognition and Media Understanding / Technical Committee on Cloud Network Robotics |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Small-sized Kernel Classifier By Support Vector Retraining Based on Minimum Classification Error Criterion |
Sub Title (in English) | |
Keyword(1) | Multi-class Support Vector Machine |
Keyword(2) | Kernel Minimum Classification Error Training |
Keyword(3) | Kernel |
1st Author's Name | Ryoma Tani |
1st Author's Affiliation | Doshisha University(Doshisha Univ.) |
2nd Author's Name | Hideyuki Watanabe |
2nd Author's Affiliation | Advanced Telecommunications Research Institute International(ATR) |
3rd Author's Name | Shigeru Katagiri |
3rd Author's Affiliation | Doshisha University(Doshisha Univ.) |
4th Author's Name | Miho Osaki |
4th Author's Affiliation | Doshisha University(Doshisha Univ.) |
Date | 2017-02-18 |
Paper # | PRMU2016-159,CNR2016-26 |
Volume (vol) | vol.116 |
Number (no) | PRMU-461,CNR-462 |
Page | pp.pp.41-46(PRMU), pp.41-46(CNR), |
#Pages | 6 |
Date of Issue | 2017-02-11 (PRMU, CNR) |