On the boosting algorithm for multiclass functions based on information-theoretic criterion for approximation

Eiji Takimoto, Akira Maruoka

研究成果: 書籍/レポート タイプへの寄稿会議への寄与

3 被引用数 (Scopus)

抄録

We consider the boosting technique that can be directly applied to the classification problem for multiclass functions. Although many boosting algorithms have been proposed so far, all of them are essentially developed for binary classification problems, and in order to handle multiclass classification problems, they need the problems reduced somehow to binary ones. In order to avoid such reductions, we introduce a notion of the pseudo-entropy function G that gives an information-theoretic criterion, called the conditional G-entropy, for measuring the loss of hypotheses. The conditional G-entropy turns out to be useful for defining the weakness of hypotheses that approximate, in some way, to a multiclass function in general, so that we can consider the boosting problem without reduction. We show that the top-down decision tree learning algorithm using G as its splitting criterion is an efficient boosting algorithm based on the conditional G-entropy. Namely, the algorithm intends to minimize the conditional G-entropy, rather than the classification error. In the binary case, our algorithm is identical to the error-based boosting algorithm proposed by Kearns and Mansour, and our analysis gives a simpler proof of their results.

本文言語英語
ホスト出版物のタイトルDiscovery Science - 1st International Conference, DS 1998, Proceedings
編集者Setsuo Arikawa, Hiroshi Motoda
出版社Springer Verlag
ページ256-267
ページ数12
ISBN(印刷版)3540653902, 9783540653905
出版ステータス出版済み - 1月 1 1998
外部発表はい
イベント1st International Conference on Discovery Science, DS 1998 - Fukuoka, 日本
継続期間: 12月 14 199812月 16 1998

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)
1532
ISSN(印刷版)0302-9743
ISSN(電子版)1611-3349

その他

その他1st International Conference on Discovery Science, DS 1998
国/地域日本
CityFukuoka
Period12/14/9812/16/98

!!!All Science Journal Classification (ASJC) codes

  • 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

「On the boosting algorithm for multiclass functions based on information-theoretic criterion for approximation」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