gpbal_fixed: Compute posterior approximation given observed treatment...

In bvegetabile/gpbalancer: Estimates propensity scores using Gaussian processes

Description

Compute posterior approximation given observed treatment assignments and a fixed covariance matrix

Usage

gpbal_fixed(y, cov_matrix, tol = 0.01, max_iters = 20, verbose = T,
  approx_method = "ep", ep_vers = "parallel")

Arguments

y	Set of observed treatment assignments (y in (0,1))
cov_matrix	Covariance matrix; for examples, see sqexp or similar
tol	Tolerance of algorithms. Difference between the latent scores at each iteration - default 1e-2
max_iters	Maximum number of iterations of the algorithm - default 20
verbose	Decision to print progress to screen - default TRUE
approx_method	Approximation method for posterior: 'ep' or 'laplace'
ep_vers	'Sequential' or 'Parallel' EP Algorithm - default parallel, alternative sequential

Value

Object that contains the weights obtained from the balancing procedure and parameters from the optimization procedure

The object that is returned is a list that contains the following entries

• Number_Iters - Number of iterations for algorithm

• PosteriorMean - Posterior mean of latent scores

• PosteriorVar - Posterior covariance of latent scores

• tilde_nu -

• tilde_tau -

• log_Z_ep - EP Approximation to Log Likelihood

• ComputationTime - Runtime of EP algorithm for fixed covariance matrix

• ps - Probit transformed posterior mean

Examples

n_obs <- 500
X1 <- rnorm(n_obs)
X2 <- rnorm(n_obs)
p <- pnorm( 0.5 * X1 + 0.5 * X2 )
TA <- rbinom(n_obs, 1, p)
dat <- data.frame(X1 = X1, X2 = X2, TA = TA)
covmat <- sqexp(cbind(X1, X2))
system.time(res <- gpbal_fixed(TA, covmat))
plot(res$ps, p, pch = 19, col = rgb(0,0,0,0.5))

bvegetabile/gpbalancer documentation built on May 22, 2019, 1:34 p.m.