Presentation | 2012-11-07 Propagating Labels via Sparse Combination of Multiple Graphs Masayuki KARASUYAMA, Hiroshi MAMITSUKA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Label propagation is a widely accepted approach in graph-based semi-supervised learning that predicts labels of nodes so that they are smooth over the input graph. We address the issue of combining multiple graphs under the framework of label propagation. The most unique feature of our approach is the sparsity of graph weights which allows to eliminate graphs irrelevant to classification automatically if they are inputted, and further to improve the predictive performance and provide the interpretability of the resultant integrated graphs. We provide an optimization problem formulation, giving weights over input graphs, and an efficient algorithm for solving the problem. We demonstrate the performance advantage and the clear interpretability of our approach through various synthetic and two real-world datasets. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Semi-supervised learning / label propagation / graph / sparsity |
Paper # | IBISML2012-58 |
Date of Issue |
Conference Information | |
Committee | IBISML |
---|---|
Conference Date | 2012/10/31(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 | Information-Based Induction Sciences and Machine Learning (IBISML) |
---|---|
Language | ENG |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Propagating Labels via Sparse Combination of Multiple Graphs |
Sub Title (in English) | |
Keyword(1) | Semi-supervised learning |
Keyword(2) | label propagation |
Keyword(3) | graph |
Keyword(4) | sparsity |
1st Author's Name | Masayuki KARASUYAMA |
1st Author's Affiliation | Bioinformatics Center, Institute for Chemical Research, Kyoto University() |
2nd Author's Name | Hiroshi MAMITSUKA |
2nd Author's Affiliation | Bioinformatics Center, Institute for Chemical Research, Kyoto University |
Date | 2012-11-07 |
Paper # | IBISML2012-58 |
Volume (vol) | vol.112 |
Number (no) | 279 |
Page | pp.pp.- |
#Pages | 8 |
Date of Issue |