Multiple kernel learning by conditional entropy minimization

Hideitsu Hino, Nima Reyhani, Noboru Murata

研究成果: Conference contribution

6 被引用数 (Scopus)

抄録

Kernel methods have been successfully used in many practical machine learning problems. Choosing a suitable kernel is left to the practitioner. A common way to an automatic selection of optimal kernels is to learn a linear combination of element kernels. In this paper, a novel framework of multiple kernel learning is proposed based on conditional entropy minimization criterion. For the proposed framework, three multiple kernel learning algorithms are derived. The algorithms are experimentally shown to be comparable to or outperform kernel Fisher discriminant analysis and other multiple kernel learning algorithms on benchmark data sets.

本文言語English
ホスト出版物のタイトルProceedings - 9th International Conference on Machine Learning and Applications, ICMLA 2010
ページ223-228
ページ数6
DOI
出版ステータスPublished - 2010 12 1
イベント9th International Conference on Machine Learning and Applications, ICMLA 2010 - Washington, DC, United States
継続期間: 2010 12 122010 12 14

出版物シリーズ

名前Proceedings - 9th International Conference on Machine Learning and Applications, ICMLA 2010

Conference

Conference9th International Conference on Machine Learning and Applications, ICMLA 2010
CountryUnited States
CityWashington, DC
Period10/12/1210/12/14

ASJC Scopus subject areas

  • Computer Science Applications
  • Human-Computer Interaction

フィンガープリント 「Multiple kernel learning by conditional entropy minimization」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル