A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles

研究成果: 著書/レポートタイプへの貢献会議での発言

抄録

Ensembles of classifiers have been studied for some time. It is widely known that weak learners should be accurate and diverse. However, in the real world there are many constraints and few have been said about the robustness of ensembles and how to develop it. In the context of ran- dom subspace methods, this paper addresses the question of developing ensembles to face problems under time con- straints. Experiments show that selecting weak learners based on their accuracy can be used to create robust en- sembles. Thus, the selection pressure in ensembles is a key technique to create not just effective ensembles but also robust ones. Moreover, the experiments motivate further research on ensembles made of low dimensional classifiers which achieve general accurate results.

元の言語英語
ホスト出版物のタイトルGECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion
ページ1755-1756
ページ数2
DOI
出版物ステータス出版済み - 2013
イベント15th Annual Conference on Genetic and Evolutionary Computation, GECCO 2013 - Amsterdam, オランダ
継続期間: 7 6 20137 10 2013

その他

その他15th Annual Conference on Genetic and Evolutionary Computation, GECCO 2013
オランダ
Amsterdam
期間7/6/137/10/13

Fingerprint

Ensemble
Classifiers
Experiments
Classifier
Subspace Methods
Experiment
Robustness

All Science Journal Classification (ASJC) codes

  • Computational Mathematics

これを引用

Vargas, D. V., Takano, H., & Murata, J. (2013). A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles. : GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion (pp. 1755-1756) https://doi.org/10.1145/2464576.2480775

A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles. / Vargas, Danilo Vasconcellos; Takano, Hirotaka; Murata, Junichi.

GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion. 2013. p. 1755-1756.

研究成果: 著書/レポートタイプへの貢献会議での発言

Vargas, DV, Takano, H & Murata, J 2013, A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles. : GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion. pp. 1755-1756, 15th Annual Conference on Genetic and Evolutionary Computation, GECCO 2013, Amsterdam, オランダ, 7/6/13. https://doi.org/10.1145/2464576.2480775
Vargas DV, Takano H, Murata J. A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles. : GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion. 2013. p. 1755-1756 https://doi.org/10.1145/2464576.2480775
Vargas, Danilo Vasconcellos ; Takano, Hirotaka ; Murata, Junichi. / A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles. GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion. 2013. pp. 1755-1756
@inproceedings{f525a1639e2747c885043b6ab53d4cdd,
title = "A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles",
abstract = "Ensembles of classifiers have been studied for some time. It is widely known that weak learners should be accurate and diverse. However, in the real world there are many constraints and few have been said about the robustness of ensembles and how to develop it. In the context of ran- dom subspace methods, this paper addresses the question of developing ensembles to face problems under time con- straints. Experiments show that selecting weak learners based on their accuracy can be used to create robust en- sembles. Thus, the selection pressure in ensembles is a key technique to create not just effective ensembles but also robust ones. Moreover, the experiments motivate further research on ensembles made of low dimensional classifiers which achieve general accurate results.",
author = "Vargas, {Danilo Vasconcellos} and Hirotaka Takano and Junichi Murata",
year = "2013",
doi = "10.1145/2464576.2480775",
language = "English",
isbn = "9781450319645",
pages = "1755--1756",
booktitle = "GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion",

}

TY - GEN

T1 - A study on the importance of selection pressure and low dimensional weak learners to produce robust ensembles

AU - Vargas, Danilo Vasconcellos

AU - Takano, Hirotaka

AU - Murata, Junichi

PY - 2013

Y1 - 2013

N2 - Ensembles of classifiers have been studied for some time. It is widely known that weak learners should be accurate and diverse. However, in the real world there are many constraints and few have been said about the robustness of ensembles and how to develop it. In the context of ran- dom subspace methods, this paper addresses the question of developing ensembles to face problems under time con- straints. Experiments show that selecting weak learners based on their accuracy can be used to create robust en- sembles. Thus, the selection pressure in ensembles is a key technique to create not just effective ensembles but also robust ones. Moreover, the experiments motivate further research on ensembles made of low dimensional classifiers which achieve general accurate results.

AB - Ensembles of classifiers have been studied for some time. It is widely known that weak learners should be accurate and diverse. However, in the real world there are many constraints and few have been said about the robustness of ensembles and how to develop it. In the context of ran- dom subspace methods, this paper addresses the question of developing ensembles to face problems under time con- straints. Experiments show that selecting weak learners based on their accuracy can be used to create robust en- sembles. Thus, the selection pressure in ensembles is a key technique to create not just effective ensembles but also robust ones. Moreover, the experiments motivate further research on ensembles made of low dimensional classifiers which achieve general accurate results.

UR - http://www.scopus.com/inward/record.url?scp=84882432200&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84882432200&partnerID=8YFLogxK

U2 - 10.1145/2464576.2480775

DO - 10.1145/2464576.2480775

M3 - Conference contribution

SN - 9781450319645

SP - 1755

EP - 1756

BT - GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference Companion

ER -