ValueInfer: This function implements the inference for value of...

In muxuanliang/ITRInference: Estimation and inference of individualized treamtent rule using pooled-and-split de-correlated score

This function implements the inference for value of individualized treatment rule.

Description

This function implements the inference for value of individualized treatment rule.

Usage

ValueInfer(
  data,
  method = "ITRFit",
  trainingfrac = 0.5 * log(NROW(data$predictor)),
  resamplingIter = 2000,
  ...
)

Arguments

data	A list - list(predictor = x, treatment = trt, outcome = y)), where x is the covariate matrix, trt is 0 or 1 (1 indicates treatment), y is the outcome.
method	method can be either 'ITRFit' or 'QLearn'.
trainingfrac	determines the traning sample size by floor(total sample size * trainingfrac).
resamplingIter	determines how many resamplings. By default, it is 1000.
...	Other parameters in ITRFitInfer and QLearnFit

Value

A list

value

Estimation of the value function given the decision rule estimated by PEARL or Q-learning

se

Estimated standard error of the value function

upper

value+1.96*se

lower

value-1.96*se

Author(s)

Muxuan Liang <[e-mail]>

References

Muxuan Liang, Young-Geun Choi, Yang Ning, Maureen Smith, Yingqi Zhao (2020). Estimation and inference on high-dimensional individualized treatment rule in observational data using split-and-pooled de-correlated score.

muxuanliang/ITRInference documentation built on Aug. 17, 2022, 6:03 p.m.