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
|
\name{glmrob..control}
\title{Controlling Robust GLM Fitting by Different Methods}
\alias{glmrobMqle.control}
\alias{glmrobMT.control}
\alias{glmrobBY.control}
\description{
These are auxiliary functions as user interface for \code{\link{glmrob}} fitting
when the different methods, \code{"Mqle"}, \code{"BY"}, or
\code{"MT"} are used. Typically only used when calling \code{\link{glmrob}}.
}
\usage{
glmrobMqle.control(acc = 1e-04, test.acc = "coef", maxit = 50, tcc = 1.345)
glmrobBY.control (maxit = 1000, const = 0.5, maxhalf = 10)
glmrobMT.control (cw = 2.1, nsubm = 500, acc = 1e-06, maxit = 200)
}
\arguments{
\item{acc}{positive convergence tolerance;
the iterations converge when ???}
\item{test.acc}{Only "coef" is currently implemented}
\item{maxit}{integer giving the maximum number of iterations. }
\item{tcc}{tuning constant c for Huber's psi-function}
\item{const}{for "BY", the normalizing constant ..}% FIXME
\item{maxhalf}{for "BY"; the number of halving steps when the gradient
itself no longer improves. We have seen examples when increasing
\code{maxhalf} was of relevance.}
\item{cw}{tuning constant c for Tukey's biweight psi-function}
\item{nsubm}{the number of subsamples to take for finding an initial
estimate for \code{method = "MT"}.}
}
%% \details{
%% }
\value{
A \code{\link{list}} with the arguments as components.
}
\author{Andreas Ruckstuhl and Martin Maechler}
\seealso{\code{\link{glmrob}}}
\examples{
str(glmrobMqle.control())
str(glmrobBY.control())
str(glmrobMT.control())
}
\keyword{robust}
\keyword{regression}
\keyword{nonlinear}
|