1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
|
\name{epi.empbayes}
\alias{epi.empbayes}
\title{Empirical Bayes estimates of observed event counts}
\description{
Computes empirical Bayes estimates of observed event counts using the method of moments.
}
\usage{
epi.empbayes(obs, pop)
}
\arguments{
\item{obs}{a vector representing the observed event counts in each unit of interest.}
\item{pop}{a vector representing the population count in each unit of interest.}
}
\details{
The gamma distribution is parameterised in terms of shape (\eqn{\alpha}) and scale (\eqn{\nu}) parameters. The mean of a given gamma distribution equals \eqn{\nu / \alpha}. The variance equals \eqn{\nu / \alpha^{2}}. The empirical Bayes estimate of event risk in each unit of interest equals \eqn{(obs + \nu) / (pop + \alpha)}.
This technique performs poorly when your data contains large numbers of zero event counts. In this situation a Bayesian approach for estimating \eqn{\alpha} and \eqn{\nu} would be advised.
}
\value{
A data frame with four elements: \code{gamma} the mean event risk across all units, \code{phi} the variance of event risk across all units, \code{alpha} the estimated shape parameter of the gamma distribution, and \code{nu} the estimated scale parameter of the gamma distribution.
}
\references{
Bailey TC, Gatrell AC (1995). Interactive Spatial Data Analysis. Longman Scientific & Technical. London, pp. 303 - 308.
Langford IH (1994). Using empirical Bayes estimates in the geographical analysis of disease risk. Area 26: 142 - 149.
Meza J (2003). Empirical Bayes estimation smoothing of relative risks in disease mapping. Journal of Statistical Planning and Inference 112: 43 - 62.
}
\examples{
## EXAMPLE 1:
data(epi.SClip)
obs <- epi.SClip$cases; pop <- epi.SClip$population
est <- epi.empbayes(obs, pop)
crude.p <- ((obs) / (pop)) * 100000
crude.r <- rank(crude.p)
ebay.p <- ((obs + est[4]) / (pop + est[3])) * 100000
dat.df01 <- data.frame(rank = c(crude.r, crude.r),
Method = c(rep("Crude", times = length(crude.r)),
rep("Empirical Bayes", times = length(crude.r))),
est = c(crude.p, ebay.p))
## Scatter plot showing the crude and empirical Bayes adjusted lip cancer
## incidence rates as a function of district rank for the crude lip
## cancer incidence rates:
\dontrun{
library(ggplot2)
ggplot(dat = dat.df01, aes(x = rank, y = est, colour = Method)) +
theme_bw() +
geom_point() +
scale_x_continuous(name = "District rank",
breaks = seq(from = 0, to = 60, by = 10),
labels = seq(from = 0, to = 60, by = 10),
limits = c(0,60)) +
scale_y_continuous(limits = c(0,30), name = "Lip cancer incidence rates
(cases per 100,000 person years)")
}
}
\keyword{univar}% at least one, from doc/KEYWORDS
\keyword{univar}% __ONLY ONE__ keyword per line
|
