bess_svm: Best subset selection on svm.

In JiangshuoZhao/StatComp21065: What the Package Does in One 'Title Case' Line

Description

Choice the best feature subset in svm model.

Usage

bess_svm(X, y, T0, alpha = 0.01, tau = 0.05, max.steps = 100)

Arguments

X	Data Matrix
y	True label
T0	Fixed nonzero feature number
alpha	Penalty coefficient in svm of L2
tau	Default param of smooth hinge loss.
max.steps	max iteration step

Value

sparse coefficient

Examples

## Not run: 
data <- gendata(1000,100,5)
bess_svm(data$x, data$y, 5)

## End(Not run)

JiangshuoZhao/StatComp21065 documentation built on Dec. 23, 2021, 10:16 p.m.