Abstract:
This paper presents a new support vector machine for simultaneous gene selection and microarray classification. By introducing the adaptive elastic net penalty which is a...Show MoreMetadata
Abstract:
This paper presents a new support vector machine for simultaneous gene selection and microarray classification. By introducing the adaptive elastic net penalty which is a convex combination of weighted 1-norm penalty and weighted 2-norm penalty, the proposed support vector machine can encourage an adaptive grouping effect and reduce the shrinkage bias for the large coefficients. According to a reasonable correlation between the two regularization parameters, the optimal coefficient paths are shown to be piecewise linear and the corresponding solving algorithm is developed. Experiments are performed on leukaemia data that verify the research results.
Published in: 2009 American Control Conference
Date of Conference: 10-12 June 2009
Date Added to IEEE Xplore: 10 July 2009
ISBN Information: