tmle_estimate_q: purpose: estimate Q=E(Y |Z, A,W) data-adaptively,

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

purpose: estimate Q=E(Y |Z, A,W) data-adaptively,

Description

unless super learner not available, or user specifies initial values or a regression formula arguments:

Usage

tmle_estimate_q(
  Y,
  Z = rep(0, length(Y)),
  A,
  W,
  Delta,
  Q = NULL,
  Qbounds,
  Qform = NULL,
  maptoYstar,
  SL.library,
  cvQinit = F,
  family,
  id = 1:length(Y),
  V = 10,
  verbose = F
)

Arguments

Y	outcome
Z	intermediate variable between A and Y (default= 0 when no int. var.)
A	treatment indicator (1=treatment, 0=control)
W	baseline covariates
Delta	missingness indicator
Q	optional externally estimated values for Q
Qbounds	bounds for predicted values
Qform	optional regression formula to use for glm if
maptoYstar	if TRUE, using logistic fluctuation for bounded, continuous outcomes
SL.library	library of prediction algorithms for Super Learner
cvQinit	flag, if TRUE, cross-validate SL.
family	regression family
id	subject identifier
V	number of folds for SuperLearner
verbose	Set T for extra output

ck37/varImpact documentation built on June 11, 2025, 8:42 a.m.