README.md

In yimindai0521/MBalance: Mahalanobis balancing: a multivariate perspective on approximate covariate balancing

MBalance

An R package for Mahalanobis balancing.

Overview

This R package implements the Mahalanobis balancing method for estimating average treatment effect (ATE) or average treatment effect in the controlled or treated group (ATC or ATT) that is proposed in our paper: Mahalanobis balancing: a multivariate perspective on approximate covariate balancing. It includes Mahalanobis balancing and high-dimensional Mahalanobis balancing method.

• MB() produces an estimate of ATE or ATC and ATT.
• hdMB() produces an estimate of ATE or ATC and ATT in high-dimensional data.

Installation

You can install the development version of MBalance from GitHub with:

if (!require("devtools")){
    install.packages("devtools")
}
devtools::install_github("yimindai0521/MBalance")

Usage Examples

We illustrate the usage of MBalance package using simple synthetic datasets.

library(MBalance)
##MB
##estimating ATE##
##generating data
set.seed(0521)
data        <- si.data()
result1     <- MB(x = data$X, treat = data$Tr, group1 = 1, outcome = data$Y, method = "MB")
result2     <- MB(x = data$X, treat = data$Tr, group1 = 0, outcome = data$Y, method = "MB")

##an estimate of ATE
result1$AT - result2$AT

##estimating ATT
result3     <- MB(x = data$X, treat = data$Tr, group1 = 1, group2 = 0, outcome = data$Y, method = "MB")

##an estimate of ATT
result3$AT - mean(data$Y[data$Tr == 0])

##hdMB
##estimating ATE in high-dimensional data##
##generating data
set.seed(0521)
data        <- si.data(sample.size = 200, dimension = 200)
dimension   <- dim(data$X)[2]

##choosing variable
hdMB1       <- hdMB(x = data$X, treat = data$Tr, group1 = 1, outcome = data$Y, method = "MB")
hdMB1$GMIM
threshold   <- 28
index1      <- rank(hdMB1$GASMD) >= (dimension - threshold + 1)

##choosing variable
hdMB2       <- hdMB(x = data$X, treat = data$Tr, group1 = 0, outcome = data$Y, method = "MB")
hdMB2$GMIM
threshold   <- 30
index2      <- rank(hdMB2$GASMD) >= (dimension - threshold + 1)

##an estimate of ATE
result4     <- MB(x = data$X[,index1], treat = data$Tr, group1 = 1, outcome = data$Y)
result5     <- MB(x = data$X[,index2], treat = data$Tr, group1 = 0, outcome = data$Y)
result4$AT - result5$AT

yimindai0521/MBalance documentation built on May 9, 2022, 4:06 p.m.