CIMTx (version 0.3.0)

bart_multiTrt_att: Bayesian Additive Regression Trees (BART) for ATT estimation

Description

This function implements the BART method when estimand is ATT. Please use our main function causal_multi_treat.R.

Usage

bart_multiTrt_att(
  y,
  x,
  trt,
  k = 2,
  discard = "No",
  ntree = 100,
  ndpost = 1000,
  nskip = 1000,
  reference = parent.frame()$reference_trt
)

Arguments

y

numeric vector for the binary outcome

x

dataframe including the treatment indicator and the covariates

trt

numeric vector for the treatment indicator

k

For binary y, k is the number of prior standard deviations f(x) is away from +/-3. The bigger k is, the more conservative the fitting will be.

discard

discarding rules for BART method, inherited from causal_multi_treat.R

ntree

The number of trees in the sum

ndpost

The number of posterior draws returned

nskip

Number of MCMC iterations to be treated as burn in

reference

Reference group for ATT

Value

list with 2 elements for ATT effect. It contains

ATT12:

A dataframe containing the estimation, standard error, lower and upper 95% CI for RD/RR/OR

ATT13:

A dataframe containing the estimation, standard error, lower and upper 95% CI for RD/RR/OR

list with 3 elements for ATT effect. It contains

ATE12:

A dataframe containing the estimation, standard error, lower and upper 95% CI for RD/RR/OR

ATE13:

A dataframe containing the estimation, standard error, lower and upper 95% CI for RD/RR/OR

ATE23:

A dataframe containing the estimation, standard error, lower and upper 95% CI for RD/RR/OR

Examples

# NOT RUN {
library(CIMTx)
set.seed(3242019)
idata = data_gen(n = 5, ratio =1,scenario = 1)
trt_ind <- as.numeric(idata$trtdat$trt_ind)
all_vars <- idata$trtdat[, -1] #exclude treatment indicator
y <- idata$Yobs
causal_multi_treat(y = y, x = all_vars,
trt = trt_ind, method = "BART", estimand = "ATT", discard = "No", ndpost = 10, reference_trt = 2)
# }