Refit subject to sparsity constraints

Source: R/refit.R

Given output of trac, solves the least squares problem with compositional constraint on the features selected by trac.

refit_trac(fit, Z, y, A, tol = 1e-05)

Arguments

fit

output of trac
Z, y, A

same arguments as passed to trac
tol

tolerance for deciding whether a gamma value is zero

Details

minimize_beta, beta0, gamma 1/(2n) || y - beta0 1_n - Z beta ||^2 subject to gamma_trac nonselected = 0, beta = A gamma, 1_p^T beta = 0