nnclean.Rd
\name{nnclean}
\alias{nnclean}
\alias{nnclean.ppp}
\alias{nnclean.pp3}
\title{
Nearest Neighbour Clutter Removal
}
\description{
Detect features in a 2D or 3D spatial point pattern
using nearest neighbour clutter removal.
}
\usage{
nnclean(X, k, ...)
\method{nnclean}{ppp}(X, k, ...,
edge.correct = FALSE, wrap = 0.1,
convergence = 0.001, plothist = FALSE,
verbose = TRUE, maxit = 50)
\method{nnclean}{pp3}(X, k, ...,
convergence = 0.001, plothist = FALSE,
verbose = TRUE, maxit = 50)
}
\arguments{
\item{X}{
A two-dimensional spatial point pattern (object of class
\code{"ppp"}) or a three-dimensional point pattern
(object of class \code{"pp3"}).
}
\item{k}{
Degree of neighbour: \code{k=1} means nearest neighbour,
\code{k=2} means second nearest, etc.
}
\item{\dots}{
Arguments passed to \code{\link{hist.default}} to control
the appearance of the histogram, if \code{plothist=TRUE}.
}
\item{edge.correct}{
Logical flag specifying whether periodic edge correction
should be performed (only implemented in 2 dimensions).
}
\item{wrap}{
Numeric value specifying the relative size of the margin
in which data will be replicated for the
periodic edge correction (if \code{edge.correct=TRUE}).
A fraction of window width and window height.
}
\item{convergence}{
Relative tolerance threshold for testing convergence of EM algorithm.
}
\item{maxit}{
Maximum number of iterations for EM algorithm.
}
\item{plothist}{
Logical flag specifying whether to plot a diagnostic histogram
of the nearest neighbour distances and the fitted distribution.
}
\item{verbose}{
Logical flag specifying whether to print progress reports.
}
}
\details{
Byers and Raftery (1998) developed a technique for recognising
features in a spatial point pattern in the presence of
random clutter.
For each point in the pattern, the distance to the
\eqn{k}th nearest neighbour is computed. Then the E-M algorithm is
used to fit a mixture distribution to the
\eqn{k}th nearest neighbour distances.
The mixture components represent the feature and the clutter. The
mixture model can be used to classify each point as belong to one
or other component.
The function \code{nnclean} is generic, with methods for
two-dimensional point patterns (class \code{"ppp"})
and three-dimensional point patterns (class \code{"pp3"})
currently implemented.
The result is a point pattern (2D or 3D) with two additional
columns of marks:
\describe{
\item{class}{
A factor, with levels \code{"noise"} and \code{"feature"},
indicating the maximum likelihood classification of each point.
}
\item{prob}{
Numeric vector giving the estimated probabilities
that each point belongs to a feature.
}
}
The object also has extra information stored in attributes:
\code{"theta"} contains the fitted parameters
of the mixture model, \code{"info"} contains
information about the fitting procedure, and \code{"hist"} contains
the histogram structure returned from \code{\link{hist.default}}
if \code{plothist = TRUE}.
}
\value{
An object of the same kind as \code{X},
obtained by attaching marks to the points of \code{X}.
The object also has attributes, as described under Details.
}
\references{
Byers, S. and Raftery, A.E. (1998)
Nearest-neighbour clutter removal for estimating features
in spatial point processes.
\emph{Journal of the American Statistical Association}
\bold{93}, 577--584.
}
\author{
Original by Simon Byers and Adrian Raftery.
Adapted for \pkg{spatstat} by \adrian.
}
\seealso{
\code{\link{nndist}},
\code{\link{split.ppp}},
\code{\link{cut.ppp}}
}
\examples{
data(shapley)
X <- nnclean(shapley, k=17, plothist=TRUE)
plot(X, which.marks=1, chars=c(".", "+"), cols=1:2)
plot(X, which.marks=2, cols=function(x)hsv(0.2+0.8*(1-x),1,1))
Y <- split(X, un=TRUE)
plot(Y, chars="+", cex=0.5)
marks(X) <- marks(X)$prob
plot(cut(X, breaks=3), chars=c(".", "+", "+"), cols=1:3)
}
\keyword{spatial}
\keyword{classif}