Presentation 2014/2/6
Improved Subspace-based Support Vector Machines by linear combination of the separating hyper-planes
Shota FUNAKI, Takuya KITAMURA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we propose the improved subspace-based SVMs (SS-SVMs) by linearly-combining the separating hyper-planes (LCS-SVMs). In this method, the discriminant function, which is determined in the training of the conventional SS-SVMs, for each class is denned as a feature quantity. With these feature quantities, the new discriminant functions are optimized by SVMs. Thus, for each class, LCS-SVMs can generate the separating hyper-planes including the information of other class feature spaces. In addition, the learning cost of LCS-SVM will be about as low as SS-SVM if the number of class is not too big. Using benchmark set, we evaluate the effectiveness of the proposed method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Support vector machine / Multi-class classifier / Pattern recognition / Subspace method
Paper # CNR2013-43,PRMU2013-135
Date of Issue

Conference Information
Committee CNR
Conference Date 2014/2/6(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 Cloud Network Robotics (CNR)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Improved Subspace-based Support Vector Machines by linear combination of the separating hyper-planes
Sub Title (in English)
Keyword(1) Support vector machine
Keyword(2) Multi-class classifier
Keyword(3) Pattern recognition
Keyword(4) Subspace method
1st Author's Name Shota FUNAKI
1st Author's Affiliation Department of Electrical Engineering, Toyama National College of Technology()
2nd Author's Name Takuya KITAMURA
2nd Author's Affiliation Department of Electrical Engineering, Toyama National College of Technology
Date 2014/2/6
Paper # CNR2013-43,PRMU2013-135
Volume (vol) vol.113
Number (no) 432
Page pp.pp.-
#Pages 6
Date of Issue