講演抄録/キーワード |
講演名 |
2009-03-12 15:40
Independent Component Analysis by Direct Density-Ratio Estimation Taiji Suzuki(Univ. of Tokyo)・○Masashi Sugiyama(Tokyo Inst. of Tech.) NC2008-136 |
抄録 |
(和) |
Accurately evaluating statistical independence
among random variables is a key component of
Independent Component Analysis (ICA). In this
paper, we employ a squared-loss variant of mutual
information as an independence measure and give
its estimation method. Our basic idea is to estimate
the ratio of probability densities directly without
going through density estimation, by which a hard
task of density estimation can be avoided. In this
density-ratio approach, a natural cross-validation
procedure is available for model selection. Thanks to
this, all tuning parameters such as the kernel width
or the regularization parameter can be objectively
optimized. This is a highly useful property in unsupervised
learning problems such as ICA. Based
on this novel independence measure, we develop a
new ICA algorithm named Least-squares Independent
Component Analysis (LICA). |
(英) |
Accurately evaluating statistical independence
among random variables is a key component of
Independent Component Analysis (ICA). In this
paper, we employ a squared-loss variant of mutual
information as an independence measure and give
its estimation method. Our basic idea is to estimate
the ratio of probability densities directly without
going through density estimation, by which a hard
task of density estimation can be avoided. In this
density-ratio approach, a natural cross-validation
procedure is available for model selection. Thanks to
this, all tuning parameters such as the kernel width
or the regularization parameter can be objectively
optimized. This is a highly useful property in unsupervised
learning problems such as ICA. Based
on this novel independence measure, we develop a
new ICA algorithm named Least-squares Independent
Component Analysis (LICA). |
キーワード |
(和) |
独立成分解析 / 密度比 / 交差確認法 / 相互情報量 / 二乗ロス / / / |
(英) |
independent component analysis / density ratio / cross validation / mutual information / squared loss / / / |
文献情報 |
信学技報, vol. 108, no. 480, NC2008-136, pp. 195-199, 2009年3月. |
資料番号 |
NC2008-136 |
発行日 |
2009-03-04 (NC) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
NC2008-136 |