Home > Research > Publications & Outputs > Multiple classifier architectures and their app...

Research

Electronic data

Keywords

View graph of relations

Multiple classifier architectures and their application to credit risk assessment

Research output: Working paper

Published

Standard

Multiple classifier architectures and their application to credit risk assessment. / Finlay, S M.
Lancaster University: The Department of Management Science, 2008. (Management Science Working Paper Series).

Research output: Working paper

Harvard

Finlay, SM 2008 'Multiple classifier architectures and their application to credit risk assessment' Management Science Working Paper Series, The Department of Management Science, Lancaster University.

APA

Finlay, S. M. (2008). Multiple classifier architectures and their application to credit risk assessment. (Management Science Working Paper Series). The Department of Management Science.

Vancouver

Finlay SM. Multiple classifier architectures and their application to credit risk assessment. Lancaster University: The Department of Management Science. 2008. (Management Science Working Paper Series).

Author

Finlay, S M. / Multiple classifier architectures and their application to credit risk assessment. Lancaster University : The Department of Management Science, 2008. (Management Science Working Paper Series).

Bibtex

@techreport{1aabc6fcccff4d37bfee28b9c7845c45,
title = "Multiple classifier architectures and their application to credit risk assessment",
abstract = "Multiple classifier systems combine several individual classifiers to deliver a final classification decision. An increasingly controversial question is whether such systems can outperform the single best classifier and if so, what form of multiple classifier system yields the greatest benefit. In this paper the performance of several multiple classifier systems are evaluated in terms of their ability to correctly classify consumers as good or bad credit risks. Empirical results suggest that many, but not all, multiple classifier systems deliver significantly better performance than the single best classifier. Overall, bagging and boosting outperform other multi-classifier systems, and a new boosting algorithm, Error Trimmed Boosting, outperforms bagging and AdaBoost by a significant margin.",
keywords = "OR in Banking, Data Mining, Classifier Combination, Classifier ensembles, Credit scoring.",
author = "Finlay, {S M}",
year = "2008",
language = "English",
series = "Management Science Working Paper Series",
publisher = "The Department of Management Science",
type = "WorkingPaper",
institution = "The Department of Management Science",

}

RIS

TY - UNPB

T1 - Multiple classifier architectures and their application to credit risk assessment

AU - Finlay, S M

PY - 2008

Y1 - 2008

N2 - Multiple classifier systems combine several individual classifiers to deliver a final classification decision. An increasingly controversial question is whether such systems can outperform the single best classifier and if so, what form of multiple classifier system yields the greatest benefit. In this paper the performance of several multiple classifier systems are evaluated in terms of their ability to correctly classify consumers as good or bad credit risks. Empirical results suggest that many, but not all, multiple classifier systems deliver significantly better performance than the single best classifier. Overall, bagging and boosting outperform other multi-classifier systems, and a new boosting algorithm, Error Trimmed Boosting, outperforms bagging and AdaBoost by a significant margin.

AB - Multiple classifier systems combine several individual classifiers to deliver a final classification decision. An increasingly controversial question is whether such systems can outperform the single best classifier and if so, what form of multiple classifier system yields the greatest benefit. In this paper the performance of several multiple classifier systems are evaluated in terms of their ability to correctly classify consumers as good or bad credit risks. Empirical results suggest that many, but not all, multiple classifier systems deliver significantly better performance than the single best classifier. Overall, bagging and boosting outperform other multi-classifier systems, and a new boosting algorithm, Error Trimmed Boosting, outperforms bagging and AdaBoost by a significant margin.

KW - OR in Banking

KW - Data Mining

KW - Classifier Combination

KW - Classifier ensembles

KW - Credit scoring.

M3 - Working paper

T3 - Management Science Working Paper Series

BT - Multiple classifier architectures and their application to credit risk assessment

PB - The Department of Management Science

CY - Lancaster University

ER -