gmleb: General maximum likelihood empirical Bayes estimator of Jiang...

In sdzhao/cole: COmpound Loss Emporium

General maximum likelihood empirical Bayes estimator of Jiang and Zhang (2009)

Description

Estimates mean vector of a homoscedastic sequence of independent Gaussian observations using the nonparametric maximum likelihood procedure of Jiang and Zhang (2009)

Usage

gmleb(x1, s1, init = NULL, grid = NULL, M = 300, tol = 1e-05, maxit = 1000)

Arguments

x1	Gaussian sequence
s1	standard deviation of Gaussian sequence
init	initial values for masses of discrete prior, default is discrete uniform over support points
grid	grid of support points for discrete prior, default is M equally spaced points between min(x1) and max(x1)
M	number of grid points
tol	error tolerance of convergence of log likelihood
maxit	maximum number of iterations

Value

estimated values of means of primary Gaussian sequence

Examples

## generate data
n = 250
set.seed(1)
theta1 = rnorm(n)
x1 = theta1 + rnorm(n)
## loss of MLE
mean((theta1 - x1)^2)
## loss of GMLEB
mean((theta1 - gmleb(x1, 1)$theta_hat)^2)

sdzhao/cole documentation built on May 2, 2022, 9:42 a.m.