Design of an unsupervised weight parameter estimation method in ensemble learning

Masato Uchida; Yousuke Maehara; Hiroyuki Shioya

doi:10.1007/978-3-540-69158-7_80

Design of an unsupervised weight parameter estimation method in ensemble learning

Masato Uchida, Yousuke Maehara, Hiroyuki Shioya

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

2 Citations (Scopus)

Abstract

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.

Original language	English
Title of host publication	Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers
Pages	771-780
Number of pages	10
Edition	PART 1
DOIs	https://doi.org/10.1007/978-3-540-69158-7_80
Publication status	Published - 2008 Oct 27
Externally published	Yes
Event	14th International Conference on Neural Information Processing, ICONIP 2007 - Kitakyushu, Japan Duration: 2007 Nov 13 → 2007 Nov 16

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	4984 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th International Conference on Neural Information Processing, ICONIP 2007
Country	Japan
City	Kitakyushu
Period	07/11/13 → 07/11/16

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-540-69158-7_80

Fingerprint Dive into the research topics of 'Design of an unsupervised weight parameter estimation method in ensemble learning'. Together they form a unique fingerprint.

Cite this

Uchida, M., Maehara, Y., & Shioya, H. (2008). Design of an unsupervised weight parameter estimation method in ensemble learning. In Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers (PART 1 ed., pp. 771-780). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4984 LNCS, No. PART 1). https://doi.org/10.1007/978-3-540-69158-7_80

Design of an unsupervised weight parameter estimation method in ensemble learning. / Uchida, Masato; Maehara, Yousuke; Shioya, Hiroyuki.

Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 1. ed. 2008. p. 771-780 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4984 LNCS, No. PART 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Uchida, M, Maehara, Y & Shioya, H 2008, Design of an unsupervised weight parameter estimation method in ensemble learning. in Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 1 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 4984 LNCS, pp. 771-780, 14th International Conference on Neural Information Processing, ICONIP 2007, Kitakyushu, Japan, 07/11/13. https://doi.org/10.1007/978-3-540-69158-7_80

Uchida M, Maehara Y, Shioya H. Design of an unsupervised weight parameter estimation method in ensemble learning. In Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 1 ed. 2008. p. 771-780. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). https://doi.org/10.1007/978-3-540-69158-7_80

Uchida, Masato ; Maehara, Yousuke ; Shioya, Hiroyuki. / Design of an unsupervised weight parameter estimation method in ensemble learning. Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers. PART 1. ed. 2008. pp. 771-780 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).

@inproceedings{b80f1cda49ae4640821a2b1d6a9633d2,

title = "Design of an unsupervised weight parameter estimation method in ensemble learning",

abstract = "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.",

author = "Masato Uchida and Yousuke Maehara and Hiroyuki Shioya",

year = "2008",

month = oct,

day = "27",

doi = "10.1007/978-3-540-69158-7_80",

language = "English",

isbn = "3540691545",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

number = "PART 1",

pages = "771--780",

booktitle = "Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers",

edition = "PART 1",

note = "14th International Conference on Neural Information Processing, ICONIP 2007 ; Conference date: 13-11-2007 Through 16-11-2007",

}

TY - GEN

T1 - Design of an unsupervised weight parameter estimation method in ensemble learning

AU - Uchida, Masato

AU - Maehara, Yousuke

AU - Shioya, Hiroyuki

PY - 2008/10/27

Y1 - 2008/10/27

N2 - 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.

AB - 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.

UR - http://www.scopus.com/inward/record.url?scp=54249116402&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=54249116402&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-69158-7_80

DO - 10.1007/978-3-540-69158-7_80

M3 - Conference contribution

AN - SCOPUS:54249116402

SN - 3540691545

SN - 9783540691549

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 771

EP - 780

BT - Neural Information Processing - 14th International Conference, ICONIP 2007, Revised Selected Papers

T2 - 14th International Conference on Neural Information Processing, ICONIP 2007

Y2 - 13 November 2007 through 16 November 2007

ER -