On node selection for classification in correlated data sets

R. Cristescu

doi:10.1109/CISS.2008.4558676

On node selection for classification in correlated data sets

R. Cristescu

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

Consider a system which can be in a finite number of states. Given a large number of characteristics which are measured, representing the system, we are concerned with the selection of a subset of characteristics of (small) given cardinality, for which the classification of the system according to one of the states in the state set is optimal according to the Rayleigh quotient criterion. This problem is relevant in various scenarios where a few explanatory variables have to be selected from a large set of candidates, including sensor selection in sensor networks, classification in image processing, and feature selection in data mining for bioinformatics applications. We show that the optimization amounts to finding the submatrix of the features covariance matrix for which the sum of elements of the inverse is maximized, and we present bounds which relate this optimization to a similar metric based on elements of the original covariance matrix.

Original language	English
Title of host publication	42nd Annual Conference on Proceedings of Information Sciences and Systems, 2008. CISS 2008
Place of Publication	Piscataway
Publisher	Institute of Electrical and Electronics Engineers
Pages	1064-1068
ISBN (Print)	978-1-4244-2246-3
DOIs	https://doi.org/10.1109/CISS.2008.4558676
Publication status	Published - 2008
Event	42nd Annual Conference on Information Sciences and Systems (CISS 2008), March 19-21, 2008, Princeton, NJ, USA - Princeton, NJ, United States Duration: 19 Mar 2008 → 21 Mar 2008

Conference

Conference	42nd Annual Conference on Information Sciences and Systems (CISS 2008), March 19-21, 2008, Princeton, NJ, USA
Abbreviated title	CISS 2008
Country	United States
City	Princeton, NJ
Period	19/03/08 → 21/03/08

Access to Document

10.1109/CISS.2008.4558676

Fingerprint Dive into the research topics of 'On node selection for classification in correlated data sets'. Together they form a unique fingerprint.

Cite this

@inproceedings{0914af4482e44ec9938e65c04cc10aea,

title = "On node selection for classification in correlated data sets",

abstract = "Consider a system which can be in a finite number of states. Given a large number of characteristics which are measured, representing the system, we are concerned with the selection of a subset of characteristics of (small) given cardinality, for which the classification of the system according to one of the states in the state set is optimal according to the Rayleigh quotient criterion. This problem is relevant in various scenarios where a few explanatory variables have to be selected from a large set of candidates, including sensor selection in sensor networks, classification in image processing, and feature selection in data mining for bioinformatics applications. We show that the optimization amounts to finding the submatrix of the features covariance matrix for which the sum of elements of the inverse is maximized, and we present bounds which relate this optimization to a similar metric based on elements of the original covariance matrix.",

author = "R. Cristescu",

year = "2008",

doi = "10.1109/CISS.2008.4558676",

language = "English",

isbn = "978-1-4244-2246-3",

pages = "1064--1068",

booktitle = "42nd Annual Conference on Proceedings of Information Sciences and Systems, 2008. CISS 2008",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "42nd Annual Conference on Information Sciences and Systems (CISS 2008), March 19-21, 2008, Princeton, NJ, USA, CISS 2008 ; Conference date: 19-03-2008 Through 21-03-2008",

}

Cristescu, R 2008, On node selection for classification in correlated data sets. in 42nd Annual Conference on Proceedings of Information Sciences and Systems, 2008. CISS 2008. Institute of Electrical and Electronics Engineers, Piscataway, pp. 1064-1068, 42nd Annual Conference on Information Sciences and Systems (CISS 2008), March 19-21, 2008, Princeton, NJ, USA, Princeton, NJ, United States, 19/03/08. https://doi.org/10.1109/CISS.2008.4558676

On node selection for classification in correlated data sets. / Cristescu, R.

42nd Annual Conference on Proceedings of Information Sciences and Systems, 2008. CISS 2008. Piscataway : Institute of Electrical and Electronics Engineers, 2008. p. 1064-1068.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - On node selection for classification in correlated data sets

AU - Cristescu, R.

PY - 2008

Y1 - 2008

N2 - Consider a system which can be in a finite number of states. Given a large number of characteristics which are measured, representing the system, we are concerned with the selection of a subset of characteristics of (small) given cardinality, for which the classification of the system according to one of the states in the state set is optimal according to the Rayleigh quotient criterion. This problem is relevant in various scenarios where a few explanatory variables have to be selected from a large set of candidates, including sensor selection in sensor networks, classification in image processing, and feature selection in data mining for bioinformatics applications. We show that the optimization amounts to finding the submatrix of the features covariance matrix for which the sum of elements of the inverse is maximized, and we present bounds which relate this optimization to a similar metric based on elements of the original covariance matrix.

AB - Consider a system which can be in a finite number of states. Given a large number of characteristics which are measured, representing the system, we are concerned with the selection of a subset of characteristics of (small) given cardinality, for which the classification of the system according to one of the states in the state set is optimal according to the Rayleigh quotient criterion. This problem is relevant in various scenarios where a few explanatory variables have to be selected from a large set of candidates, including sensor selection in sensor networks, classification in image processing, and feature selection in data mining for bioinformatics applications. We show that the optimization amounts to finding the submatrix of the features covariance matrix for which the sum of elements of the inverse is maximized, and we present bounds which relate this optimization to a similar metric based on elements of the original covariance matrix.

U2 - 10.1109/CISS.2008.4558676

DO - 10.1109/CISS.2008.4558676

M3 - Conference contribution

SN - 978-1-4244-2246-3

SP - 1064

EP - 1068

BT - 42nd Annual Conference on Proceedings of Information Sciences and Systems, 2008. CISS 2008

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

T2 - 42nd Annual Conference on Information Sciences and Systems (CISS 2008), March 19-21, 2008, Princeton, NJ, USA

Y2 - 19 March 2008 through 21 March 2008

ER -