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

Eiji Takimoto, Akira Maruoka

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)


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.

Original languageEnglish
Title of host publicationDiscovery Science - 1st International Conference, DS 1998, Proceedings
EditorsSetsuo Arikawa, Hiroshi Motoda
PublisherSpringer Verlag
Number of pages12
ISBN (Print)3540653902, 9783540653905
Publication statusPublished - 1998
Externally publishedYes
Event1st International Conference on Discovery Science, DS 1998 - Fukuoka, Japan
Duration: Dec 14 1998Dec 16 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other1st International Conference on Discovery Science, DS 1998

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'On the boosting algorithm for multiclass functions based on information-theoretic criterion for approximation'. Together they form a unique fingerprint.

Cite this