aiptw: Augmented Inverse Probability of Treatment Weighted...

In tranlm/lrecCompare: Comparative methods in the estimation of a treatment specific mean outcome

Description

aiptw Estimates the parameter of interest (E[Y_d]) by directly solving the efficient influence function as an estimating equation.

Usage

aiptw(data, cum.g, Ynodes, Anodes, Cnodes = NULL, abar, Qform,
  SL.library = NULL, family = "quasibinomial", stratify = TRUE)

Arguments

data	data frame following the time-ordering of the nodes.
cum.g	a matrix of the cumulative probabilities of treatment (and being uncensored) given the parents.
Ynodes	column names or indicies in data of outcome nodes.
Anodes	column names or indicies in data of treatment nodes.
Cnodes	column names or indicies in data of censoring nodes.
abar	binary vector (numAnodes x 1) of counterfactual treatment
Qform	character vector of regression formulas for Q.
SL.library	optional character vector of libraries to pass to SuperLearner. NULL indicates glm should be called instead of SuperLearner.
stratify	if TRUE condition on following abar when estimating Q and g. If FALSE, pool over all subjects.

Details

Please refer to van der Laan and Gruben (2007) for details regarding derivation of the efficient influence function, which this estimator solves.

Value

BangRobinsDR returns a list of items as an object of class aiptw, which include

• The estimate of the parameter value under the intervention abar

• The empirical influence function for the point estimate

• The time-specific empirical influence functions

• The conditional expectation fits for Qbar

• The call to the function

tranlm/lrecCompare documentation built on May 31, 2019, 7:44 p.m.