Feature selection using support vector machines

J. Brank; M. Grobelnik; N. Milić-Frayling; D. Mladenić

Feature selection using support vector machines

J. Brank^*, M. Grobelnik, N. Milić-Frayling, D. Mladenić

^*Corresponding author for this work

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

14 Citations (Scopus)

Abstract

The concept of sparsity as more approptiate characteristic of the data representation than the number of features used was discussed. A feature ranking and a feature selection method based on the linear support vector machines (SVM) that was used in conjunction with the SVM classifier was also proposed. This method can be combined with other classification algorithms. The results show that, at the same level of vector sparcity, feature selection based on SVM normals yields better classification performance than odds ratio or information gain based feature selection when linear SVM classifiers are used.

Original language	English
Title of host publication	Data Mining III
Editors	A. Zanasi, C.A. Brebbia, N.F.F.E. Ebecken, P. Melli
Publisher	WITPress
Pages	261-273
Number of pages	13
Volume	6
ISBN (Print)	1853128309
Publication status	Published - 2002
Externally published	Yes
Event	Third International Conference on Data Mining, Data Mining III - Bologna, Italy Duration: 25 Sept 2002 → 27 Sept 2002

Conference

Conference	Third International Conference on Data Mining, Data Mining III
Country/Territory	Italy
City	Bologna
Period	25/09/02 → 27/09/02

Cite this

@inproceedings{b38134d928884d6d898f131c25c3bf10,

title = "Feature selection using support vector machines",

abstract = "The concept of sparsity as more approptiate characteristic of the data representation than the number of features used was discussed. A feature ranking and a feature selection method based on the linear support vector machines (SVM) that was used in conjunction with the SVM classifier was also proposed. This method can be combined with other classification algorithms. The results show that, at the same level of vector sparcity, feature selection based on SVM normals yields better classification performance than odds ratio or information gain based feature selection when linear SVM classifiers are used.",

author = "J. Brank and M. Grobelnik and N. Mili{\'c}-Frayling and D. Mladeni{\'c}",

year = "2002",

language = "English",

isbn = "1853128309",

volume = "6",

pages = "261--273",

editor = "A. Zanasi and C.A. Brebbia and N.F.F.E. Ebecken and P. Melli",

booktitle = "Data Mining III",

publisher = "WITPress",

address = "United Kingdom",

note = "Third International Conference on Data Mining, Data Mining III ; Conference date: 25-09-2002 Through 27-09-2002",

}

TY - GEN

T1 - Feature selection using support vector machines

AU - Brank, J.

AU - Grobelnik, M.

AU - Milić-Frayling, N.

AU - Mladenić, D.

PY - 2002

Y1 - 2002

N2 - The concept of sparsity as more approptiate characteristic of the data representation than the number of features used was discussed. A feature ranking and a feature selection method based on the linear support vector machines (SVM) that was used in conjunction with the SVM classifier was also proposed. This method can be combined with other classification algorithms. The results show that, at the same level of vector sparcity, feature selection based on SVM normals yields better classification performance than odds ratio or information gain based feature selection when linear SVM classifiers are used.

AB - The concept of sparsity as more approptiate characteristic of the data representation than the number of features used was discussed. A feature ranking and a feature selection method based on the linear support vector machines (SVM) that was used in conjunction with the SVM classifier was also proposed. This method can be combined with other classification algorithms. The results show that, at the same level of vector sparcity, feature selection based on SVM normals yields better classification performance than odds ratio or information gain based feature selection when linear SVM classifiers are used.

UR - http://www.scopus.com/inward/record.url?scp=2942670930&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:2942670930

SN - 1853128309

VL - 6

SP - 261

EP - 273

BT - Data Mining III

A2 - Zanasi, A.

A2 - Brebbia, C.A.

A2 - Ebecken, N.F.F.E.

A2 - Melli, P.

PB - WITPress

T2 - Third International Conference on Data Mining, Data Mining III

Y2 - 25 September 2002 through 27 September 2002

ER -

Feature selection using support vector machines

Abstract

Conference

Other files and links

Fingerprint

Cite this