A method for comparison: Trans-Lasso(l1). It has the same pipeline of Trans.lasso() but with sparsity index R_k=|w^(k)-β|_1 and a naive aggregation (empirical risk minimization)

A method for comparison: Trans-Lasso(l1). It has the same pipeline of Trans.lasso() but with sparsity index R_k=|w^(k)-β|_1 and a naive aggregation (empirical risk minimization)

Usage

Trans.lasso.sp(X, y, n.vec, I.til, l1 = T)

Arguments

X

is the training data set

y

is the target value set

n.vec

PLACEHOLDER

I.til

is a subset of n.vec containing roughly half its elements

l1

is a boolean indicating l1-sparse characterization of contrast vectors (T) or l0 (F) (default: T)

Value

is a named list containing beta.sp, theta.sp, and rank.pi