estimatePi {twoStageDesignTMLE}R Documentation

estimatePi

Description

Typically not called directly by the user. Function for modeling the two-stage missingness mechanism and evaluating conditional probabilities for each observation

Usage

estimatePi(
  Y,
  A,
  W,
  condSetNames,
  W.Q,
  Delta.W,
  V.msm = NULL,
  piform,
  pi.SL.library,
  id,
  V,
  discreteSL,
  verbose,
  pi = NULL,
  obsWeights = rep(1, nrow(W))
)

Arguments

Y

outcome

A

binary treatment indicator

W

covariate matrix observed on everyone

condSetNames

Variables to include as predictors of missingness in W.stage2, any combination of Y, A, and either W (for all covariates in W) or individual covariate names in W

W.Q

additional covariates based on preliminary outcome regression

Delta.W

binary indicator of missing second stage covariates

V.msm

optional additional covariates to condition on beyond W

piform

parametric regression formula for estimating pi

pi.SL.library

super learner library for estimating pi

id

Identifier of independent units of observation, e.g., clusters

V

number of cross validation folds for estimating pi using super learner

discreteSL

Use discrete super learning when TRUE, otherwise ensemble super learning

verbose

When TRUE prints informational messages

pi

optional vector of user-specified probabilities

obsWeights

optional weights for evaluating pi

Value

list containing the predicted probabilities, estimation method coefficients in parametric regression model (if piform supplied), indicator of whether discrete or ensemble SL was used.


[Package twoStageDesignTMLE version 1.0.1.2 Index]