Between AUC based and error rate based learning

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

1 Citation (Scopus)

Abstract

Based on an earlier solution to optimize an approximated area under the ROC Curve (AUC) for binary pattern classification in [1], this paper investigates into the relationship between AUC and several error rate based classifiers. Via a generalized framework of translated scalingspace, we find that the AUC based classifier can be related to a total-error-rate (TER) classifier, an Equal Error Rate (EER) formulation, and a least-squares-error (LSE) estimator, each under a specific setting of the translated scaling-space framework. Several potential applications of the generalized framework are subsequently discussed.

Original languageEnglish
Title of host publication2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008
Pages2116-2120
Number of pages5
DOIs
Publication statusPublished - 2008 Sep 23
Event2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008 - Singapore, Singapore
Duration: 2008 Jun 32008 Jun 5

Other

Other2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008
CountrySingapore
CitySingapore
Period08/6/308/6/5

Fingerprint

Classifiers
Pattern recognition

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering
  • Industrial and Manufacturing Engineering

Cite this

Toh, K. A. (2008). Between AUC based and error rate based learning. In 2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008 (pp. 2116-2120). [4582893] https://doi.org/10.1109/ICIEA.2008.4582893
Toh, Kar Ann. / Between AUC based and error rate based learning. 2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008. 2008. pp. 2116-2120
@inproceedings{3ecc560f898f49c2a5e6353cc7c941e3,
title = "Between AUC based and error rate based learning",
abstract = "Based on an earlier solution to optimize an approximated area under the ROC Curve (AUC) for binary pattern classification in [1], this paper investigates into the relationship between AUC and several error rate based classifiers. Via a generalized framework of translated scalingspace, we find that the AUC based classifier can be related to a total-error-rate (TER) classifier, an Equal Error Rate (EER) formulation, and a least-squares-error (LSE) estimator, each under a specific setting of the translated scaling-space framework. Several potential applications of the generalized framework are subsequently discussed.",
author = "Toh, {Kar Ann}",
year = "2008",
month = "9",
day = "23",
doi = "10.1109/ICIEA.2008.4582893",
language = "English",
isbn = "9781424417186",
pages = "2116--2120",
booktitle = "2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008",

}

Toh, KA 2008, Between AUC based and error rate based learning. in 2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008., 4582893, pp. 2116-2120, 2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008, Singapore, Singapore, 08/6/3. https://doi.org/10.1109/ICIEA.2008.4582893

Between AUC based and error rate based learning. / Toh, Kar Ann.

2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008. 2008. p. 2116-2120 4582893.

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

TY - GEN

T1 - Between AUC based and error rate based learning

AU - Toh, Kar Ann

PY - 2008/9/23

Y1 - 2008/9/23

N2 - Based on an earlier solution to optimize an approximated area under the ROC Curve (AUC) for binary pattern classification in [1], this paper investigates into the relationship between AUC and several error rate based classifiers. Via a generalized framework of translated scalingspace, we find that the AUC based classifier can be related to a total-error-rate (TER) classifier, an Equal Error Rate (EER) formulation, and a least-squares-error (LSE) estimator, each under a specific setting of the translated scaling-space framework. Several potential applications of the generalized framework are subsequently discussed.

AB - Based on an earlier solution to optimize an approximated area under the ROC Curve (AUC) for binary pattern classification in [1], this paper investigates into the relationship between AUC and several error rate based classifiers. Via a generalized framework of translated scalingspace, we find that the AUC based classifier can be related to a total-error-rate (TER) classifier, an Equal Error Rate (EER) formulation, and a least-squares-error (LSE) estimator, each under a specific setting of the translated scaling-space framework. Several potential applications of the generalized framework are subsequently discussed.

UR - http://www.scopus.com/inward/record.url?scp=51949101130&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=51949101130&partnerID=8YFLogxK

U2 - 10.1109/ICIEA.2008.4582893

DO - 10.1109/ICIEA.2008.4582893

M3 - Conference contribution

AN - SCOPUS:51949101130

SN - 9781424417186

SP - 2116

EP - 2120

BT - 2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008

ER -

Toh KA. Between AUC based and error rate based learning. In 2008 3rd IEEE Conference on Industrial Electronics and Applications, ICIEA 2008. 2008. p. 2116-2120. 4582893 https://doi.org/10.1109/ICIEA.2008.4582893