est_g: Estimate the Generalized Propensity Score (Treatment...
In txshift: Efficient Estimation of the Causal Effects of Stochastic Interventions

Description Usage Arguments Details Value

Estimate the Generalized Propensity Score (Treatment Mechanism)

est_g(
  A,
  W,
  delta = 0,
  ipc_weights = rep(1, length(A)),
  fit_type = c("sl", "hal"),
  sl_learners_density = NULL,
  haldensify_args = list(n_bins = c(5, 10), grid_type = c("equal_range", "equal_mass"),
    lambda_seq = exp(seq(-1, -13, length = 300)), use_future = FALSE)
)

`A`	A `numeric` vector of observed treatment values.
`W`	A `numeric` matrix of observed baseline covariate values.
`delta`	A `numeric` value identifying a shift in the observed value of the treatment under which observations are to be evaluated.
`ipc_weights`	A `numeric` vector of observation-level weights, as produced by the internal procedure to estimate the censoring mechanism `estimate-ipc_weights`.
`fit_type`	A `character` specifying whether to use Super Learner (from sl3) or the Highly Adaptive Lasso (from hal9001) to estimate the conditional treatment density.
`sl_learners_density`	Object containing a set of instantiated learners from sl3, to be used in fitting an ensemble model.
`haldensify_args`	A `list` of argument to be directly passed to `haldensify` when `fit_type` is set to `"hal"`. Note that this invokes the Highly Adaptive Lasso instead of Super Learner and is thus only feasible for relatively small data sets.

Compute the propensity score (treatment mechanism) for the observed data, including the shift. This gives the propensity score for the observed data (at the observed A) and the shift (at A - delta).

A data.table with four columns, containing estimates of the generalized propensity score at a downshift (g(A - delta | W)), no shift (g(A | W)), an upshift (g(A + delta) | W), and an upshift of magitudie two (g(A + 2 delta) | W).

txshift documentation built on Oct. 23, 2020, 8:27 p.m.