AveragingCausalHD: Estimate the average treatment effect while adjusting for a...

Description Usage Arguments Value Examples

View source: R/Main.R

Description

Estimate the average treatment effect while adjusting for a high dimensional set of covariates

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
AveragingCausalHD(
  y,
  t,
  x,
  estimators = c("DoublePS", "DRlasso", "Debiasing", "DML", "DMLpost_selection",
    "TMLElasso", "TMLEscreen", "HDmatching"),
  AdditionalEstimates = NULL,
  AdditionalSEs = NULL,
  trim = FALSE
)

Arguments

y

The outcome to be analyzed

t

The treatment to be analyzed

x

An n by p matrix of covariates to adjust for

estimators

A list of estimators to be included in the averaging estimator. The full list of estimators is given by c("DoublePS", "DRlasso", "Debiasing", "DML", "DMLpost_selection", "TMLElasso", "TMLEscreen", "HDC", "HDbayes", "HDmatching") The HDbayes and HDC estimators are Bayesian and may take time to run on large data sets, however, the rest should run quickly. It is recommended to use as many of these estimators in the averaging as possible

AdditionalEstimates

A vector of additional point estimates the user can include in the averaging. This should only be used if the user has fit other estimators that are not already available in the list above, and would like to include them in the averaging.

AdditionalSEs

A vector of additional standard errors that correspond to the point estimates from the estimators making up AdditionalEstimates.

trim

An indicator of whether the averaging should remove the highest and lowest point estimates from the averaging procedure

Value

A list with the estimates and standard errors from each individual estimator, as well as from the averaged estimator, which combines all of the individual estimators

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
n <- 100
p <- 100

beta.c <- c(0.75,1, 0.6, -0.8, -0.7,rep(0, p-5))

gamma <- c(0.15,0.2,0, 0, -0.4, rep(0, p-5))

beta <- 1

sigma <- matrix(0.3, p,p)
diag(sigma) <- 1

x <- mvtnorm::rmvnorm(n, sigma=sigma)
t <- as.numeric((x %*% gamma + rnorm(n)) > 0)
y <- 0 + t + x %*% beta.c + rnorm(n, sd=1)

AveragingCausalHD(y=y, t=t, x=x)

jantonelli111/AveragingCausalHD documentation built on Feb. 1, 2020, 12:54 a.m.