Speed up SVM-RFE procedure using margin distribution

Yingqin Yuan, Leonid Hrebien, Moshe Kam

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, a new method is introduced to speed up the recursive feature ranking procedure by using the margin distribution of a trained SVM. The method, M-RFE, continuously eliminates features without retraining the SVM as long as the margin distribution of the SVM does not change significantly. Synthetic datasets and two benchmark microarray datasets were tested on M-RFE. Comparison with original SVM-RFE shows that our method speeds up the feature ranking procedure considerably with little or no performance degradation. Comparison of M-RFE to a similar speed up technique, E-RFE, provides similar classification performance, but with reduced complexity.

Original languageEnglish (US)
Title of host publication2005 IEEE Workshop on Machine Learning for Signal Processing
Pages297-302
Number of pages6
DOIs
StatePublished - Dec 1 2005
Externally publishedYes
Event2005 IEEE Workshop on Machine Learning for Signal Processing - Mystic, CT, United States
Duration: Sep 28 2005Sep 30 2005

Publication series

Name2005 IEEE Workshop on Machine Learning for Signal Processing

Other

Other2005 IEEE Workshop on Machine Learning for Signal Processing
CountryUnited States
CityMystic, CT
Period9/28/059/30/05

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Fingerprint Dive into the research topics of 'Speed up SVM-RFE procedure using margin distribution'. Together they form a unique fingerprint.

Cite this