Design of an unsupervised weight parameter estimation method in ensemble learning

Masato Uchida*, Yousuke Maehara, Hiroyuki Shioya

*この研究の対応する著者

研究成果: Conference contribution

2 被引用数 (Scopus)

抄録

A learning method using an integration of multiple component predictors as an ultimate predictor is generically referred to as ensemble learning. The present paper proposes a weight parameter estimation method for ensemble learning under the constraint that we do not have any information of the desirable (true) output. The proposed method is naturally derived from a mathematical model of ensemble learning, which is based on an exponential mixture type probabilistic model and Kullback divergence. The proposed method provides a legitimate strategy for weight parameter estimation under the abovementioned constraint if it is assumed that the accuracy of all multiple predictors are the same. We verify the effectiveness of the proposed method through numerical experiments.

本文言語English
ホスト出版物のタイトルNeural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers
ページ771-780
ページ数10
PART 1
DOI
出版ステータスPublished - 2008 10月 27
外部発表はい
イベント14th International Conference on Neural Information Processing, ICONIP 2007 - Kitakyushu, Japan
継続期間: 2007 11月 132007 11月 16

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)
番号PART 1
4984 LNCS
ISSN(印刷版)0302-9743
ISSN(電子版)1611-3349

Conference

Conference14th International Conference on Neural Information Processing, ICONIP 2007
国/地域Japan
CityKitakyushu
Period07/11/1307/11/16

ASJC Scopus subject areas

  • 理論的コンピュータサイエンス
  • コンピュータ サイエンス(全般)

フィンガープリント

「Design of an unsupervised weight parameter estimation method in ensemble learning」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