NFMiss_se: Estimate standard error using asymptotic formula for...

In chenchi0526/SIsMiss:

Description

Estimate standard error using asymptotic formula for unregularized regression.

Usage

NFMiss_se(y, z, u, gamma_tilde_hat, beta_hat, se.method, M = 500, seed_num,
  regularize, data_pair, newcov_pair, obs_id, w_hat)

Arguments

y	Vector. The outcome variable
z	Vector. The shadow variable (fully observed).
u	Matrix. The covariate matrix U.
gamma_tilde_hat	Initial estimates of scaled γ.
beta_hat	Initial estimates of β.
se.method	Charactor. The method for estimating standard error of parameters. Three options are available. "asymp" distribution. This option is applicable to regularize=FALSE only. "perturb" This option is applicable to both regularize=FALSE and regularize=TRUE. NULL
M	Number of resampling in perturbation.
seed_num	Number of seed to control randomness of perturbation term generated.
regularize	If the objective is to fit a regularized linear regression for variable selection, regularize = TRUE. By default, regularize = FALSE.
data_pair	Matrix. The pairwise version of original data obtained from function pair_trans()$data_pair.
newcov_pair	Matrix. The pairwise transformed covariates obtained from function pair_trans()$newcov_pair.
obs_id	Vector. The indice of missingness for subjects obtained from function pair_trans.
w_hat	Matrix. The estimates of W matrix in the offset term obtained by pair_trans.

Value

A vector. The standard error of estimators in unregularized regression.

chenchi0526/SIsMiss documentation built on Dec. 8, 2020, 2:35 a.m.