ctmle: ctmle

In jucheng1992/ctmlejl: ctmlejl: An interface to use TargetedLearning package in julia

Description

A function compute ctmle estimator and its standard error based on julia package TargetedLearning.jl.

Usage

ctmle(logitQnA1, logitQnA0, W, A, Y, v = 10, gbounds = c(0, 1),
  searchstrategy = "ForwardStepwise", patience = 100,
  penalize_risk = TRUE, param = "ATE", order = vector())

Arguments

logitQnA1	numeric vector containing initial estimates for treatment group in logit scale.
logitQnA0	numeric vector containing initial estimates for control group in logit scale.
W	numeric matrix or dataframe, stand for covariates.
A	numeric 1-0 vector, stand for treatment.
Y	numeric vector, stand for outcome.
v	integer, stand for number of folds for cross-validation in cTMLE.
gbounds	a vector of numeric of length two (with both of them in [0,1]), stand for the truncation boud for gn.
searchstrategy	Strategy for adding covariates to estimates of g. Default is "ForwardStepwise". Other options: "LogisticOrdering", "PartialCorrOrdering", "SuperLearner"
patience	Integer, stand for how many steps should CV continue after finding a local optimum. Defaults to typemax(Int).
param	a string stand for different target parameter. "ATE" stands for additive treatment effect, Mean1 and Mean0 stand for the mean outcome in treatment or control group.
order	an integer vector for the manual ordering option. Notice, the first column is intercept, so the indexes start from 2.

Value

A list for targeted estimator and standard error.

Author(s)

Cheng JU

jucheng1992/ctmlejl documentation built on May 20, 2019, 4:20 a.m.