Undesirable effects of output normalization in multiple classifier systems

被引：9

作者：

Altinçay, H ^{[1
]}

Demirekler, M

机构：

[1] Eastern Mediterranean Univ, Dept Comp Engn, KKTC, Mersin 10, Turkey

[2] Middle E Tech Univ, Dept Elect & Elect Engn, TR-06531 Ankara, Turkey

来源：

PATTERN RECOGNITION LETTERS | 2003年 / 24卷 / 9-10期

关键词：

output score normalization; dimensionality reduction; class separability; output post-processing; measurement level classifier combination;

D O I：

10.1016/S0167-8655(02)00286-6

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Incomparability of the classifier output scores is a major problem in the combination of different classification systems. In order to deal with this problem, the measurement level classifier outputs are generally normalized. However, empirical results have shown that output normalization may lead to some undesirable effects. This paper presents analyses for some most frequently used normalization methods and it is shown that the main reason for these undesirable effects of output normalization is the dimensionality reduction in the output space. An artificial classifier combination example and a real-data experiment are provided where these effects are further clarified. (C) 2002 Elsevier Science B.V. All rights reserved.

引用

页码：1163 / 1170

页数：8

共 50 条

[1] Why does output normalization create problems in multiple classifier systems?
Altinçay, H
Demirekler, M
16TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOL II, PROCEEDINGS, 2002, : 775 - 778
[2] Serial multiple classifier systems exploiting a coarse to fine output coding
Kittler, J
Ahmadyfard, A
Windridge, D
MULTIPLE CLASSIFIER SYSTEMS, PROCEEDING, 2003, 2709 : 106 - 114
[3] On output independence and complementariness in rank-based multiple classifier decision systems
Saranli, A
Demirekler, M
PATTERN RECOGNITION, 2001, 34 (12) : 2319 - 2330
[4] The inclusion of correlation effects in the performance of multiple sensor and classifier systems
Schubert, Christine M.
Oxley, Mark E.
Bauer, Kenneth W., Jr.
2005 IEEE Aerospace Conference, Vols 1-4, 2005, : 2259 - 2269
[5] New measure of classifier dependency in multiple classifier systems
Ruta, D
Gabrys, B
MULTIPLE CLASSIFIER SYSTEMS, 2002, 2364 : 127 - 136
[6] Implementation of multiple outpatient formularies: Undesirable effects
Stein, CM
Wood, AJJ
Pincus, T
CLINICAL PHARMACOLOGY & THERAPEUTICS, 1997, 61 (01) : 1 - 7
[7] Multiple classifier combination methodologies for different output levels
Suen, CY
Lam, L
MULTIPLE CLASSIFIER SYSTEMS, 2000, 1857 : 52 - 66
[8] A survey of multiple classifier systems as hybrid systems
Wozniak, Michal
Grana, Manuel
Corchado, Emilio
INFORMATION FUSION, 2014, 16 : 3 - 17
[9] Adaptive Score Normalization for Output Integration in Multiclassifier Systems
Pirlo, Giuseppe
Impedovo, Donato
IEEE SIGNAL PROCESSING LETTERS, 2012, 19 (12) : 837 - 840
[10] Multiple classifier systems for robust classifier design in adversarial environments
Battista Biggio
Giorgio Fumera
Fabio Roli
International Journal of Machine Learning and Cybernetics, 2010, 1 : 27 - 41

← 1 2 3 4 5 →