ShrinkageEstimates: Shrinkage Estimation of Small Domain Parameters

Description Usage Arguments Value Author(s) Examples

View source: R/ShrinkageEstimates.R

Description

Computes shrinkage estimates for small domain parameters using different choices for the regression weights.

Usage

1
2
ShrinkageEstimates(Y, X, sd.resid, weights=c("cure", "cure.plugin", 
                  "cure.multitau", "mle","bpe"), vc.est=c("mle", "reml","ure"))

Arguments

Y

The K x 1 vector of responses.

X

The K x p design matrix.

sd.resid

The K x 1 vector sampling standard deviations.

weights

The method by which to choose regression weights.

vc.est

When the mle method is used for regression weights, this determines which method is used to estimate the variance component. This can be either maximum likelihood estimation, restricted maximum likelihood estimation, or minimization of an unbiased risk estimate.

Value

A list with the following components

Author(s)

Nicholas Henderson

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
K <- 20
X <- cbind(rep(1, K), rnorm(K))
beta.coef <- rnorm(2)
nn <- 1 + rpois(K, lambda=4)
sig.sq <- 1/nn
ran.ef <- rnorm(K)

Y <- X%*%beta.coef + ran.ef + sqrt(sig.sq)*rnorm(K)
sd.resid <- sqrt(sig.sq)


cbp.est <- ShrinkageEstimates(Y, X, sd.resid, weights="cure")
eblup.est <- ShrinkageEstimates(Y, X, sd.resid, weights="mle")
obp.est <- ShrinkageEstimates(Y, X, sd.resid, weights="bpe")

par(mfrow=c(1,2))
plot(cbp.est$shrinkage.estimate, eblup.est$shrinkage.estimate)
plot(cbp.est$shrinkage.estimate, obp.est$shrinkage.estimate)
abline(0, 1)

nchenderson/shrinkcbp documentation built on March 11, 2021, 2:52 p.m.