estimate_tmle2: Get TMLE estimate: E[Y | A = 1, W].

In ck37/varimpact: Variable Importance Estimation Using Targeted Causal Inference (TMLE)

Get TMLE estimate: E[Y | A = 1, W].

Description

Get TMLE estimate: E[Y | A = 1, W].

Usage

estimate_tmle2(
  Y,
  A,
  W,
  family,
  delta = rep(1, length(Y)),
  Q.lib,
  g.lib,
  id = 1:length(Y),
  Qbounds = NULL,
  gbound = 0.025,
  alpha = 0.995,
  fluctuation = "logistic",
  V = 10,
  verbose = F
)

Arguments

Y	Outcome variable
A	Treatment indicator
W	Dataframe of adjustment covariates
family	Outcome family - binomial or gaussian
delta	Indicator of missing outcome or treatment assignment. 1 - observed, 0 - missing.
Q.lib	SuperLearner library for estimating Q (potential outcome)
g.lib	SuperLearner library for estimating g (propensity score)
id	Optional subject-level identifier.
Qbounds	Bounds on Q
gbound	Bounds on G
alpha	TBD, from TMLE package
fluctuation	Only logistic is currently supported.
V	Number of folds for SuperLearner
verbose	If true output extra information during execution.

ck37/varimpact documentation built on June 23, 2022, 4:41 a.m.