recursive_ice: Recursively apply the ICE algorithm

In audreyrenson/didgformula: The difference-in-differences g-formula

Recursively apply the ICE algorithm

Description

Recursively apply the ICE algorithm

Usage

recursive_ice(
  t,
  k,
  data,
  inside_formula_t,
  inside_formula_tmin1,
  outside_formula,
  inside_family,
  binomial_n = NULL,
  models = TRUE,
  tmle = FALSE,
  weights = NULL
)

Arguments

t	int. Timepoint the estimand applies to
k	int. Always 0 if called at the user level. Timepoint the outer expectation applies to
data	wide data frame
inside_formula_t	chr. glue-style formula for inside model for Yt (times indexed by t, e.g. '~Lt')
inside_formula_tmin1	chr. glue-style formula for inside model for Yt (times indexed by t, e.g. '~Lt-1')
outside_formula	chr. glue-style formula for outside models (times indexed by k, e.g. '~Lk)
inside_family	stats::families object (or chr) for glm inside models
binomial_n	vector of length nrow(data) of group sizes if Y is binomial aggregate data
models	lgl. Return all models as an attribute?
tmle	lgl. Incorporate tmle updating step to make estimator doubly robust?
weights	Matrix of IPTW weights as returned by cbind(1, calc_weights(...)). Default NULL, required if tmle=TRUE.

audreyrenson/didgformula documentation built on Oct. 9, 2022, 11:45 a.m.