Description Usage Arguments Value
View source: R/SimulationBodyPart.R
Simulation Body Part
1 2 3 4 5 | simulation(kfold, data = data, trimLowerBound.pi = 1e-05,
trimLowerBound.pibar = 1e-05, trimUpperBound.pi = 10,
trimUpperBound.pibar = 10, method = "CDML", g.method = "rf",
gps.method = "series", trimLowerBound.t = -10, trimUpperBound.t = 10,
model, variation = "whole", verbose = TRUE, detoured = FALSE)
|
kfold |
number of folds for sample splitting |
data |
data |
trimLowerBound.pi |
trimming lower bound of pi estimation |
trimLowerBound.pibar |
trimming lower bound of pibar estimation |
trimUpperBound.pi |
trimming upper bound of pi estimation |
trimUpperBound.pibar |
trimming upper bound of pibar estimation |
method |
estimation method we choose from double machine learning method ("CDML") and simple regression method ("SR"), Hirano & Imbens method ("HI") |
g.method |
method for regression estimation |
gps.method |
method for generalized propensity score estimation |
trimLowerBound.t |
trimming lower bound of treatment vector |
trimUpperBound.t |
trimming upper bound of treatment vector |
model |
model varies from "IV", "CTE" |
variation |
choose from "whole","train.data". Estimate the probability measure by empirical measure of covariates. "whole" uses every data sample, "train.data" only trainning data for estimating empirical measure. As default, we select "whole" |
verbose |
print in console step by step |
detoured |
if FALSE method will be done normally; otherwise a special simulation will be examined, where CDML is chosen method, CTE is chosen model, gps.method "series" is not allowed. |
a list of estimated value "yhat" evaluated at "t.grid", with "record" to record details of this simulation
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.