This function performs post-selection estimation on a given solution.
Usage
Logistic_FAR_Path_Further_Improve(
x_mat,
y_vec,
h,
k_n,
p,
delta_vec_init,
eta_stack_init,
mu1_vec_init,
mu2,
a = 1,
lam = 0.1,
weight_vec = 1,
logit_weight_vec = 1,
weight_already_combine = FALSE,
tol = 10^(-5),
max_iter = 1000,
fast_glm = FALSE
)Arguments
- x_mat
covariate matrix, consists of two parts. dim(x_mat) = (n, h + p * kn) First h columns are for demographical covariates(can include an intercept term) Rest columns are for p functional covariates, each being represented by a set of basis functions resulting kn covariates.
- y_vec
response vector, 0 for control, 1 for case. n = length(y_vec) is the number of observations.
- h, k_n, p
dimension information for the dataset(
x_mat).- delta_vec_init, eta_stack_init, mu1_vec_init
Initial values for the algorithm. This function uses these initial values to find out the active functional covariates. And the post-selection estimation begins with these initial values.
- mu2
quadratic term in the ADMM algorithm
- a
parameters for the algorithm. The 1st term in the loss function is
1 / a * loglik. See Algorithm_Details.pdf for more information.- lam
A scalar for the regularize in ridge penalty form in case of model saturation.
- tol, max_iter
convergence tolerance and max number of iteration of the algorithm.