double_ML: Post double selection.

In postDoubleR: Post Double Selection with Double Machine Learning

Description

Provides a convenient function to calculate the double ML estimated debiased treatment effect θ.

Usage

double_ML(X, Y, W, method = c("glmnet", "randomforest", "nn", "ols",
  "custom"), show.progress = TRUE, specify.own = NULL, k.fld = 2,
  simulations = 100, validate.inputs = TRUE, seed.use = 1071, ...)

Arguments

X	A matrix of covariates (must be all numeric)
Y	A vector of the target variable, of same length as the number of rows of Y, must be numeric
W	A vector of the treatment variable, of same length as the number of rows of X, must be numeric
method	A selection of methods to use when doing post double selection.
show.progress	Whether to display the simulation progress, defaults to TRUE.
specify.own	Allows the user to supply the method to calculate \hat{W} and \hat{Y}, please refer to custom_helper
k.fld	How many fold crossfitting to use, defaults to 2.
simulations	How many simulations to use for the final result.
validate.inputs	A safety measure indicating whether the types of inputs should be checked, defaults to TRUE (disabled for custom methods).
seed.use	The seed to use for simulations, defaults to 1071.
...	Other arguments to be passed on, see rf_helper, glmnet_helper and ols_helper for details.

Details

Custom functions are currently implemented through a function called custom_generator. For these custom functions, refer to that function and usage examples.

Value

An object of class "ML_Treatment_Effects" that can be further manipulated (ie there is a plot method implemented).

Examples

  n = 2000; p = 10
  X = matrix(rnorm(n*p), n, p)
  W = rbinom(n, 1, 0.4 + 0.2 * (X[,1] > 0))
  Y = pmax(X[,1], 0) * W + X[,2] + pmin(X[,3], 0) + rnorm(n)

double_ML(X, Y, W, method = "glmnet",
              k.fld = 2, simulations = 5,
              lambda.set.Y = 1,
              lambda.set.W = 1,
               Z.trans = F)

postDoubleR documentation built on Oct. 7, 2019, 5:05 p.m.