ダウンロード数: 431
このアイテムのファイル:
ファイル | 記述 | サイズ | フォーマット | |
---|---|---|---|---|
j.neunet.2012.06.009.pdf | 242.04 kB | Adobe PDF | 見る/開く |
タイトル: | Canonical dependency analysis based on squared-loss mutual information. |
著者: | Karasuyama, Masayuki Sugiyama, Masashi |
著者名の別形: | 烏山, 昌幸 |
キーワード: | Canonical correlation analysis Squared-loss mutual information Direct density-ratio estimation |
発行日: | Oct-2012 |
出版者: | Elsevier Ltd. |
誌名: | Neural networks |
巻: | 34 |
開始ページ: | 46 |
終了ページ: | 55 |
抄録: | Canonical correlation analysis (CCA) is a classical dimensionality reduction technique for two sets of variables that iteratively finds projection directions with maximum correlation. Although CCA is still in vital use in many practical application areas, recent real-world data often contain more complicated nonlinear correlations that cannot be properly captured by classical CCA. In this paper, we thus propose an extension of CCA that can effectively capture such complicated nonlinear correlations through statistical dependency maximization. The proposed method, which we call least-squares canonical dependency analysis (LSCDA), is based on a squared-loss variant of mutual information, and it has various useful properties besides its ability to capture higher-order correlations: for example, it can simultaneously find multiple projection directions (i.e., subspaces), it does not involve density estimation, and it is equipped with a model selection strategy. We demonstrate the usefulness of LSCDA through various experiments on artificial and real-world datasets. |
著作権等: | © 2012 Elsevier Ltd. This is not the published version. Please cite only the published version. この論文は出版社版でありません。引用の際には出版社版をご確認ご利用ください。 |
URI: | http://hdl.handle.net/2433/159940 |
DOI(出版社版): | 10.1016/j.neunet.2012.06.009 |
PubMed ID: | 22831849 |
出現コレクション: | 学術雑誌掲載論文等 |
このリポジトリに保管されているアイテムはすべて著作権により保護されています。