Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers

Gendo Kumoi; Hideki Yagi; Manabu Kobayashi; Masayuki Goto; Shigeichi Hirasawa

doi:10.1007/978-3-031-04819-7_37

Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers

Gendo Kumoi^*, Hideki Yagi, Manabu Kobayashi, Masayuki Goto, Shigeichi Hirasawa

^*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

Error-Correcting Output Coding (ECOC) is a method for constructing a multi-valued classifier using a combination of the given binary classifiers. ECOC is said to be able to estimate the correct category by other binary classifiers even if the output of some binary classifiers is incorrect based on the framework of the coding theory. Although it is experimentally known that this method performs well on real data, a theoretical analysis of the classification accuracy for ECOC has yet to be conducted. In this study, we evaluate the superiority of a code word table in showing the combinations of binary classifiers of ECOC that have been experimentally demonstrated. In other words, we analytically evaluate how the estimation of the categories is influenced by the estimated posterior probability, which is the output of the binary classifier, as well as by the structure of constructing the code word table.

Original language	English
Title of host publication	Information Systems and Technologies - WorldCIST 2022
Editors	Alvaro Rocha, Hojjat Adeli, Gintautas Dzemyda, Fernando Moreira
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	379-389
Number of pages	11
ISBN (Print)	9783031048180
DOIs	https://doi.org/10.1007/978-3-031-04819-7_37
Publication status	Published - 2022
Event	10th World Conference on Information Systems and Technologies, WorldCIST 2022 - Budva, Montenegro Duration: 2022 Apr 12 → 2022 Apr 14

Publication series

Name	Lecture Notes in Networks and Systems
Volume	469 LNNS
ISSN (Print)	2367-3370
ISSN (Electronic)	2367-3389

Conference

Conference	10th World Conference on Information Systems and Technologies, WorldCIST 2022
Country/Territory	Montenegro
City	Budva
Period	22/4/12 → 22/4/14

Keywords

Error-Correcting Output Coding
Estimated posterior probabilities
Multi-valued classification

ASJC Scopus subject areas

Control and Systems Engineering
Signal Processing
Computer Networks and Communications

Access to Document

10.1007/978-3-031-04819-7_37

Cite this

Kumoi, G., Yagi, H., Kobayashi, M., Goto, M., & Hirasawa, S. (2022). Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers. In A. Rocha, H. Adeli, G. Dzemyda, & F. Moreira (Eds.), Information Systems and Technologies - WorldCIST 2022 (pp. 379-389). (Lecture Notes in Networks and Systems; Vol. 469 LNNS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-04819-7_37

Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers. / Kumoi, Gendo; Yagi, Hideki; Kobayashi, Manabu et al.
Information Systems and Technologies - WorldCIST 2022. ed. / Alvaro Rocha; Hojjat Adeli; Gintautas Dzemyda; Fernando Moreira. Springer Science and Business Media Deutschland GmbH, 2022. p. 379-389 (Lecture Notes in Networks and Systems; Vol. 469 LNNS).

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

Kumoi, G, Yagi, H, Kobayashi, M , Goto, M & Hirasawa, S 2022, Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers. in A Rocha, H Adeli, G Dzemyda & F Moreira (eds), Information Systems and Technologies - WorldCIST 2022. Lecture Notes in Networks and Systems, vol. 469 LNNS, Springer Science and Business Media Deutschland GmbH, pp. 379-389, 10th World Conference on Information Systems and Technologies, WorldCIST 2022, Budva, Montenegro, 22/4/12. https://doi.org/10.1007/978-3-031-04819-7_37

Kumoi G, Yagi H, Kobayashi M , Goto M, Hirasawa S. Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers. In Rocha A, Adeli H, Dzemyda G, Moreira F, editors, Information Systems and Technologies - WorldCIST 2022. Springer Science and Business Media Deutschland GmbH. 2022. p. 379-389. (Lecture Notes in Networks and Systems). doi: 10.1007/978-3-031-04819-7_37

Kumoi, Gendo ; Yagi, Hideki ; Kobayashi, Manabu et al. / Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers. Information Systems and Technologies - WorldCIST 2022. editor / Alvaro Rocha ; Hojjat Adeli ; Gintautas Dzemyda ; Fernando Moreira. Springer Science and Business Media Deutschland GmbH, 2022. pp. 379-389 (Lecture Notes in Networks and Systems).

@inproceedings{49499938785e4bd482c13bddea4b8c2e,

title = "Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers",

abstract = "Error-Correcting Output Coding (ECOC) is a method for constructing a multi-valued classifier using a combination of the given binary classifiers. ECOC is said to be able to estimate the correct category by other binary classifiers even if the output of some binary classifiers is incorrect based on the framework of the coding theory. Although it is experimentally known that this method performs well on real data, a theoretical analysis of the classification accuracy for ECOC has yet to be conducted. In this study, we evaluate the superiority of a code word table in showing the combinations of binary classifiers of ECOC that have been experimentally demonstrated. In other words, we analytically evaluate how the estimation of the categories is influenced by the estimated posterior probability, which is the output of the binary classifier, as well as by the structure of constructing the code word table.",

keywords = "Error-Correcting Output Coding, Estimated posterior probabilities, Multi-valued classification",

author = "Gendo Kumoi and Hideki Yagi and Manabu Kobayashi and Masayuki Goto and Shigeichi Hirasawa",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 10th World Conference on Information Systems and Technologies, WorldCIST 2022 ; Conference date: 12-04-2022 Through 14-04-2022",

year = "2022",

doi = "10.1007/978-3-031-04819-7_37",

language = "English",

isbn = "9783031048180",

series = "Lecture Notes in Networks and Systems",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "379--389",

editor = "Alvaro Rocha and Hojjat Adeli and Gintautas Dzemyda and Fernando Moreira",

booktitle = "Information Systems and Technologies - WorldCIST 2022",

}

TY - GEN

T1 - Performance Evaluation of ECOC Considering Estimated Probability of Binary Classifiers

AU - Kumoi, Gendo

AU - Yagi, Hideki

AU - Kobayashi, Manabu

AU - Goto, Masayuki

AU - Hirasawa, Shigeichi

PY - 2022

Y1 - 2022

N2 - Error-Correcting Output Coding (ECOC) is a method for constructing a multi-valued classifier using a combination of the given binary classifiers. ECOC is said to be able to estimate the correct category by other binary classifiers even if the output of some binary classifiers is incorrect based on the framework of the coding theory. Although it is experimentally known that this method performs well on real data, a theoretical analysis of the classification accuracy for ECOC has yet to be conducted. In this study, we evaluate the superiority of a code word table in showing the combinations of binary classifiers of ECOC that have been experimentally demonstrated. In other words, we analytically evaluate how the estimation of the categories is influenced by the estimated posterior probability, which is the output of the binary classifier, as well as by the structure of constructing the code word table.

AB - Error-Correcting Output Coding (ECOC) is a method for constructing a multi-valued classifier using a combination of the given binary classifiers. ECOC is said to be able to estimate the correct category by other binary classifiers even if the output of some binary classifiers is incorrect based on the framework of the coding theory. Although it is experimentally known that this method performs well on real data, a theoretical analysis of the classification accuracy for ECOC has yet to be conducted. In this study, we evaluate the superiority of a code word table in showing the combinations of binary classifiers of ECOC that have been experimentally demonstrated. In other words, we analytically evaluate how the estimation of the categories is influenced by the estimated posterior probability, which is the output of the binary classifier, as well as by the structure of constructing the code word table.

KW - Error-Correcting Output Coding

KW - Estimated posterior probabilities

KW - Multi-valued classification

UR - http://www.scopus.com/inward/record.url?scp=85131124729&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85131124729&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-04819-7_37

DO - 10.1007/978-3-031-04819-7_37

M3 - Conference contribution

AN - SCOPUS:85131124729

SN - 9783031048180

T3 - Lecture Notes in Networks and Systems

SP - 379

EP - 389

BT - Information Systems and Technologies - WorldCIST 2022

A2 - Rocha, Alvaro

A2 - Adeli, Hojjat

A2 - Dzemyda, Gintautas

A2 - Moreira, Fernando

PB - Springer Science and Business Media Deutschland GmbH

T2 - 10th World Conference on Information Systems and Technologies, WorldCIST 2022

Y2 - 12 April 2022 through 14 April 2022

ER -