ATT.DCB: ATT Estimation using Differentiated Confounder Balancing...

In wyan3/ATEDCB: ATE Estimation Using Differentiated Confounder Balancing (DCB)

Description

The function provides the estimation of ATT given the outcome Y, binary treatment T, and covariates X.

Usage

ATT.DCB(Y,T,X)

Format

The function has the following parameters

Y

Outcome

T

Binary treatment status indicator: 1=treatment; 0=control.

X

Covariate matrix

lambda

Penalty on the loss of outcome regression on covariates in the control group. Default is 10.

delta

Penalty on the L2-norm of weight for each unit. Default is 0.001

mu

Penalty on the L2-norm of coefficient for each covariate. Default is 0.001

upsilon

Penalty on the L1-norm of coefficient for each covariate. Default is 0.001

thold

Stop criteria of the update process. Default is 1e-4.

max_iter

Maximum number of iterations. Default is 100000.

References

Kuang, K., Cui, P., Li, B., Jiang, M., & Yang, S. (2017). Estimating treatment effect in the wild via differentiated confounder balancing. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 265-274).

Examples

set.seed(25)
n <- 200
Z <- matrix(rnorm(4*n),ncol=4,nrow=n)
prop <- 1 / (1 + exp(Z[,1] - 0.5 * Z[,2] + 0.25*Z[,3] + 0.1 * Z[,4]))
treat <- rbinom(n, 1, prop)
Y <- 200 + 10*treat+ (1.5*treat-0.5)*(27.4*Z[,1] + 13.7*Z[,2] + 13.7*Z[,3] + 13.7*Z[,4]) + rnorm(n)
X <- cbind(exp(Z[,1])/2,Z[,2]/(1+exp(Z[,1])), (Z[,1]*Z[,3]/25+0.6)^3,(Z[,2]+Z[,4]+20)^2)

fit <- ATT.DCB(Y,T,X)
fit$ATT

wyan3/ATEDCB documentation built on Sept. 4, 2019, 8:02 p.m.