est_marg_nuisance_params: Estimate nuisance parameters for each marginal mixture...

In blind-contours/CVtreeMLE: Cross Validated Decision Trees with Targeted Maximum Likelihood Estimation

Estimate nuisance parameters for each marginal mixture component

Description

For each marginal mixture component rule found, create a g estimator for the probability of being exposed to the rule thresholds, and a Q estimator for the outcome E(Y| A = a_mix, W). Get estimates of g and Q using the validation data and calculate the clever covariate used in the TMLE fluctuation step.

Usage

est_marg_nuisance_params(
  at,
  av,
  w,
  y,
  aw_stack,
  family,
  a,
  no_marg_rules,
  marg_decisions,
  parallel_cv,
  seed,
  h_aw_trunc_lvl
)

Arguments

at	Training data
av	Validation data
w	Vector of characters denoting covariates
y	Outcome variable
aw_stack	Super Learner library for fitting Q (outcome mechanism) and g (treatment mechanism)
family	Binomial or continuous
a	Vector of characters that denote the mixture components
no_marg_rules	TRUE/FALSE if no marginal rules were found across the folds
marg_decisions	List of rules found within the fold for each mixture component
parallel_cv	TRUE/FALSE if cv parallelization is used
seed	Seed number
h_aw_trunc_lvl	Truncation level of the clever covariate (induces more bias to reduce variance)

Value

marginal_data: A list of data frames for each mixture component with the baseline covariates, exposure, outcome, and nuisance parameters needed to calculate the ATE.

blind-contours/CVtreeMLE documentation built on June 22, 2024, 8:53 p.m.