regnet (version 0.2.0)

McpLogistic: MCP logistic regression for a given lambda.

Description

This function makes predictions for MCP logistic regression for a given value of lambda. Typical usage is to have the CV.McpLogistic function compute the optimal lambda, then provide it to the McpLogistic function.

Usage

McpLogistic(X, Y, lambda, r = 5, alpha.i = 1, folds = 5)

Arguments

X

a matrix of predictors.

Y

a vector of the binary response.

lambda

the tuning parameter that imposes sparsity.

r

the regularization parameter in MCP.

alpha.i

by default, the program use the lasso penalty for choosing initial values of the coefficient vector. alpha.i is the Elastic-Net mixing parameter, with \(0 \le alpha.i \le 1\). alpha.i=1 is the lasso penalty, and alpha.i=0 is the ridge penalty. If alpha.i is assigned to be -1, the program will use zeroes as initial coefficients.

folds

the number of folds for cross-validation.

Value

the estimated coefficients vector.

References

zhang, CH. (2010). Nearly unbiased variable selection under minimax concave penalty. Annals of Statistics, 38(2):894-942.

See Also

CV.McpLogistic

Examples

# NOT RUN {
b = McpLogistic(regnet$X, regnet$Y, 0.075)
inds = which(regnet$beta != 0)
sel = which(b != 0)
tp = length(intersect(inds, sel))
fp = length(sel) - tp
list(tp=tp, fp=fp)
# }