Presentation 2006/2/16
Electric Network Classifier for Semi-Supervised Learning on Graphs
Masaki RIKITOKU, Hiroshi HIRAI, Kazuo MUROTA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a new classifier, named electric network classifiers, for semi-supervised learning on graphs. Our classifier is based on nonlinear electric network theory and classifies data set with respect to the sign of electric potential. Close relationships to Support Vector Machine and graph kernel methods are revealed. Unlike other graph kernel methods, our classifier does not require heavy kernel computations and obtain the potential directly using efficient network flow algorithms. Therefore, our classifier has the potential to tackle large complex real world problems. Experimental results for the 20 newsgroups data set and MNIST handwritten character recognition data show that the performance is good compared with the other standard methods.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) semi-supervised learning text classification / svm / network / graph
Paper # TL2005-54,PRMU2005-159
Date of Issue

Conference Information
Committee PRMU
Conference Date 2006/2/16(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 Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Electric Network Classifier for Semi-Supervised Learning on Graphs
Sub Title (in English)
Keyword(1) semi-supervised learning text classification
Keyword(2) svm
Keyword(3) network
Keyword(4) graph
1st Author's Name Masaki RIKITOKU
1st Author's Affiliation Justsystem Corporation Innovative Technology R & D Strategy Dept.()
2nd Author's Name Hiroshi HIRAI
2nd Author's Affiliation Research Institute for Mathematical Sciences, Kyoto University
3rd Author's Name Kazuo MUROTA
3rd Author's Affiliation Department of Mathematical Informatics, Graduate School of Information Science and Technology. University of Tokyo
Date 2006/2/16
Paper # TL2005-54,PRMU2005-159
Volume (vol) vol.105
Number (no) 614
Page pp.pp.-
#Pages 6
Date of Issue