Reducing hubness for kernel regression

Kazuo Hara, Ikumi Suzuki, Kei Kobayashi, Kenji Fukumizu, Miloš Radovanović

研究成果: Conference contribution

抜粋

In this paper, we point out that hubness—some samples in a high-dimensional dataset emerge as hubs that are similar to many other samples—influences the performance of kernel regression. Because the dimension of feature spaces induced by kernels is usually very high, hubness occurs, giving rise to the problem of multicollinearity, which is known as a cause of instability of regression results. We propose hubnessreduced kernels for kernel regression as an extension of a previous approach for kNN classification that reduces spatial centrality to eliminate hubness.

元の言語English
ホスト出版物のタイトルSimilarity Search and Applications - 8th International Conference, SISAP 2015, Proceedings
編集者Richard Connor, Giuseppe Amato, Fabrizio Falchi, Claudio Gennaro
出版者Springer Verlag
ページ339-344
ページ数6
ISBN(印刷物)9783319250861
DOI
出版物ステータスPublished - 2015
外部発表Yes
イベント8th International Conference on Similarity Search and Applications, SISAP 2015 - Glasgow, United Kingdom
継続期間: 2015 10 122015 10 14

出版物シリーズ

名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
9371
ISSN(印刷物)0302-9743
ISSN(電子版)1611-3349

Other

Other8th International Conference on Similarity Search and Applications, SISAP 2015
United Kingdom
Glasgow
期間15/10/1215/10/14

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

フィンガープリント Reducing hubness for kernel regression' の研究トピックを掘り下げます。これらはともに一意のフィンガープリントを構成します。

  • これを引用

    Hara, K., Suzuki, I., Kobayashi, K., Fukumizu, K., & Radovanović, M. (2015). Reducing hubness for kernel regression. : R. Connor, G. Amato, F. Falchi, & C. Gennaro (版), Similarity Search and Applications - 8th International Conference, SISAP 2015, Proceedings (pp. 339-344). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 巻数 9371). Springer Verlag. https://doi.org/10.1007/978-3-319-25087-8_33