A theoretical study on six classifier fusion strategies

被引：506

作者：

Kuncheva, LI ^{[1
]}

机构：

[1] Univ Wales, Sch Informat, Bangor LL57 1UT, Gwynedd, Wales

来源：

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE | 2002年 / 24卷 / 02期

关键词：

classifier combination; theoretical error; fusion methods; order statistics; majority vote; independent classifiers;

D O I：

10.1109/34.982906

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We look at a single point in the feature space, two classes, and L classifiers estimating the posterior probability for class omega(1). Assuming that the estimates are independent and identically distributed (normal or uniform), we give formulas for the classification error for the following fusion methods: average, minimum, maximum, median, majority vote, and oracle.

引用

页码：281 / 286

页数：6

共 14 条

[1] Experimental evaluation of expert fusion strategies [J].

Alkoot, FM ;

Kittler, J .

PATTERN RECOGNITION LETTERS, 1999, 20 (11-13) :1361-1369

[2]

[Anonymous], 1999, Combining Artificial Neural Nets: Ensemble and Modular Multi-Net Systems

[3]

[Anonymous], COMBINING ARTICIAL N

[4]

[Anonymous], 1974, Introduction to the Theory of Statistics

[5]

Bishop C. M., 1995, NEURAL NETWORKS PATT

[6]

BREIMAN L, 1999, COMBINING ARTIFICIAL, P31

[7] On combining classifiers [J].

Kittler, J ;

Hatef, M ;

Duin, RPW ;

Matas, J .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1998, 20 (03) :226-239

[8]

KUNCHEVA L, 2001, P IEE WORKSH INT SEN

[9] Decision templates for multiple classifier fusion: an experimental comparison [J].

Kuncheva, LI ;

Bezdek, JC ;

Duin, RPW .

PATTERN RECOGNITION, 2001, 34 (02) :299-314

[10] CONCEPTUAL MODELING OF COINCIDENT FAILURES IN MULTIVERSION SOFTWARE [J].

LITTLEWOOD, B ;

MILLER, DR .

IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, 1989, 15 (12) :1596-1614

← 1 2 →