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
|
\name{huberM}
\alias{huberM}
\title{Safe (generalized) Huber M-Estimator of Location}
\concept{robust location}
\description{
(Generalized) Huber M-estimator of location with MAD scale, being
sensible also when the scale is zero where \code{\link[MASS]{huber}()}
returns an error.
}
\usage{
huberM(x, k = 1.5, weights = NULL, tol = 1e-06,
mu = if(is.null(weights)) median(x) else wgt.himedian(x, weights),
s = if(is.null(weights)) mad(x, center=mu)
else wgt.himedian(abs(x - mu), weights),
warn0scale = getOption("verbose"))
}
\arguments{
\item{x}{numeric vector.}
\item{k}{positive factor; the algorithm winsorizes at \code{k}
standard deviations.}
\item{weights}{numeric vector of non-negative weights of same length
as \code{x}, or \code{NULL}.}
\item{tol}{convergence tolerance.}
\item{mu}{initial location estimator.}
\item{s}{scale estimator held constant through the iterations.}
\item{warn0scale}{logical; if true, and \code{s} is 0 and
\code{length(x) > 1}, this will be warned about.}
}
\value{
list of location and scale parameters, and number of iterations used.
\item{mu}{location estimate}
\item{s}{the \code{s} argument, typically the \code{\link{mad}}.}
\item{it}{the number of \dQuote{Huber iterations} used.}
}
\details{
Note that currently, when non-\code{NULL} \code{weights} are
specified, the default for initial location \code{mu} and scale
\code{s} is \code{\link{wgt.himedian}}, where strictly speaking a
weighted \dQuote{non-hi} median should be used for consistency.
Since \code{s} is not updated, the results slightly differ, see the
examples below.
}
\author{Martin Maechler, building on the MASS code mentioned.}
\references{
Huber, P. J. (1981)
\emph{Robust Statistics.}
Wiley.
}
\seealso{
\code{\link[MASS]{hubers}} (and \code{huber}) in package \pkg{MASS};
\code{\link{mad}}.
}
\examples{
huberM(c(1:9, 1000))
mad (c(1:9, 1000))
mad (rep(9, 100))
huberM(rep(9, 100))
## When you have "binned" aka replicated observations:
set.seed(7)
x <- c(round(rnorm(1000),1), round(rnorm(50, m=10, sd = 10)))
t.x <- table(x) # -> unique values and multiplicities
x.uniq <- as.numeric(names(t.x)) ## == sort(unique(x))
x.mult <- unname(t.x)
str(Hx <- huberM(x.uniq, weights = x.mult), digits = 7)
str(Hx. <- huberM(x, s = Hx$s), digits = 7) ## should be ~= Hx
stopifnot(all.equal(Hx, Hx.))
str(Hx2 <- huberM(x), digits = 7)## somewhat different, since 's' differs
}
\keyword{univar}
\keyword{robust}
|
