Relationship between diversity and correlation in Multi-Classifier Systems

Kuo Wei Hsu; Jaideep Srivastava

doi:10.1007/978-3-642-13672-6_47

Relationship between diversity and correlation in Multi-Classifier Systems

Kuo Wei Hsu, Jaideep Srivastava

Computer Science and Engineering

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

6 Scopus citations

Abstract

Diversity plays an important role in the design of Multi-Classifier Systems, but its relationship to classification accuracy is still unclear from a theoretical perspective. As a step towards the solution of this probelm, we take a different route and explore the relationship between diversity and correlation. In this paper we provide a theoretical analysis and present a nonlinear function that relates diversity to correlation, which hence can be further related to accuracy. This paper contributes to connecting existing research in diversity and correlation, and also providing a proxy to the relationship between diversity and accuracy. Our experimental results reveal deeper insights into the role of diversity in Multi-Classifier Systems.

Original language	English (US)
Title of host publication	Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings
Pages	500-506
Number of pages	7
Edition	PART 2
DOIs	https://doi.org/10.1007/978-3-642-13672-6_47
State	Published - Dec 1 2010
Event	14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010 - Hyderabad, India Duration: Jun 21 2010 → Jun 24 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	6119 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010
Country/Territory	India
City	Hyderabad
Period	6/21/10 → 6/24/10

Keywords

Correlation
Diversity
Multi-Classifier System (MCS)

Publisher link

10.1007/978-3-642-13672-6_47

Find It

Find It at U of M Libraries

Cite this

Hsu, K. W., & Srivastava, J. (2010). Relationship between diversity and correlation in Multi-Classifier Systems. In Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings (PART 2 ed., pp. 500-506). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6119 LNAI, No. PART 2). https://doi.org/10.1007/978-3-642-13672-6_47

Relationship between diversity and correlation in Multi-Classifier Systems. / Hsu, Kuo Wei; Srivastava, Jaideep.

Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2. ed. 2010. p. 500-506 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6119 LNAI, No. PART 2).

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

Hsu, KW & Srivastava, J 2010, Relationship between diversity and correlation in Multi-Classifier Systems. in Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 6119 LNAI, pp. 500-506, 14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010, Hyderabad, India, 6/21/10. https://doi.org/10.1007/978-3-642-13672-6_47

Hsu KW, Srivastava J. Relationship between diversity and correlation in Multi-Classifier Systems. In Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2 ed. 2010. p. 500-506. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-642-13672-6_47

Hsu, Kuo Wei ; Srivastava, Jaideep. / Relationship between diversity and correlation in Multi-Classifier Systems. Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings. PART 2. ed. 2010. pp. 500-506 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{c7351f2fba6743219f157fd08514c813,

title = "Relationship between diversity and correlation in Multi-Classifier Systems",

abstract = "Diversity plays an important role in the design of Multi-Classifier Systems, but its relationship to classification accuracy is still unclear from a theoretical perspective. As a step towards the solution of this probelm, we take a different route and explore the relationship between diversity and correlation. In this paper we provide a theoretical analysis and present a nonlinear function that relates diversity to correlation, which hence can be further related to accuracy. This paper contributes to connecting existing research in diversity and correlation, and also providing a proxy to the relationship between diversity and accuracy. Our experimental results reveal deeper insights into the role of diversity in Multi-Classifier Systems.",

keywords = "Correlation, Diversity, Multi-Classifier System (MCS)",

author = "Hsu, {Kuo Wei} and Jaideep Srivastava",

year = "2010",

month = dec,

day = "1",

doi = "10.1007/978-3-642-13672-6_47",

language = "English (US)",

isbn = "3642136710",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

number = "PART 2",

pages = "500--506",

booktitle = "Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings",

edition = "PART 2",

note = "14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010 ; Conference date: 21-06-2010 Through 24-06-2010",

}

TY - GEN

T1 - Relationship between diversity and correlation in Multi-Classifier Systems

AU - Hsu, Kuo Wei

AU - Srivastava, Jaideep

PY - 2010/12/1

Y1 - 2010/12/1

N2 - Diversity plays an important role in the design of Multi-Classifier Systems, but its relationship to classification accuracy is still unclear from a theoretical perspective. As a step towards the solution of this probelm, we take a different route and explore the relationship between diversity and correlation. In this paper we provide a theoretical analysis and present a nonlinear function that relates diversity to correlation, which hence can be further related to accuracy. This paper contributes to connecting existing research in diversity and correlation, and also providing a proxy to the relationship between diversity and accuracy. Our experimental results reveal deeper insights into the role of diversity in Multi-Classifier Systems.

AB - Diversity plays an important role in the design of Multi-Classifier Systems, but its relationship to classification accuracy is still unclear from a theoretical perspective. As a step towards the solution of this probelm, we take a different route and explore the relationship between diversity and correlation. In this paper we provide a theoretical analysis and present a nonlinear function that relates diversity to correlation, which hence can be further related to accuracy. This paper contributes to connecting existing research in diversity and correlation, and also providing a proxy to the relationship between diversity and accuracy. Our experimental results reveal deeper insights into the role of diversity in Multi-Classifier Systems.

KW - Correlation

KW - Diversity

KW - Multi-Classifier System (MCS)

UR - http://www.scopus.com/inward/record.url?scp=79956315981&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79956315981&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-13672-6_47

DO - 10.1007/978-3-642-13672-6_47

M3 - Conference contribution

AN - SCOPUS:79956315981

SN - 3642136710

SN - 9783642136719

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 500

EP - 506

BT - Advances in Knowledge Discovery and Data Mining - 14th Pacific-Asia Conference, PAKDD 2010, Proceedings

T2 - 14th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2010

Y2 - 21 June 2010 through 24 June 2010

ER -

Relationship between diversity and correlation in Multi-Classifier Systems

Abstract

Publication series

Other

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this