plot a Fitted Point Process Model
Given a fitted point process model obtained by ppm,
create spatial trend and conditional intensity surfaces of the model,
in a form suitable for plotting, and (optionally) plot these
surfaces.
## S3 method for class 'ppm'
plot(x, ngrid = c(40,40), superimpose = TRUE,
trend = TRUE, cif = TRUE, se = TRUE, pause = interactive(),
how=c("persp","image", "contour"), plot.it = TRUE,
locations = NULL, covariates=NULL, ...)x |
A fitted point process model, typically obtained from
the model-fitting algorithm |
ngrid |
The dimensions for a grid on which to evaluate,
for plotting, the spatial trend and conditional intensity.
A vector of 1 or 2 integers. If it is of length 1,
|
superimpose |
logical flag; if |
trend |
logical flag; if |
cif |
logical flag; if |
se |
logical flag; if |
pause |
logical flag indicating whether to pause with a prompt
after each plot. Set |
how |
character string or character vector indicating the style or styles of
plots to be performed. Ignored if |
plot.it |
logical scalar; should a plot be produced immediately? |
locations |
If present, this determines the locations of the pixels
at which predictions are computed. It must be a binary pixel image
(an object of class |
covariates |
Values of external covariates required by the fitted model.
Passed to |
... |
extra arguments to the plotting functions |
This is the plot method for the class "ppm"
(see ppm.object for details of this class).
It invokes predict.ppm to compute the spatial
trend and conditional intensity of the fitted point process model.
See predict.ppm for more explanation about spatial trend
and conditional intensity.
The default action is to create a rectangular grid
of points in (the bounding box of) the observation window of
the data point pattern, and evaluate the spatial trend and
conditional intensity of the fitted spatial point process model
x at these locations. If the argument locations=
is supplied, then the spatial trend
and conditional intensity are calculated at the grid of points
specified by this argument.
The argument locations, if present, should be a
binary image mask (an object of class "owin"
and type "mask"). This determines a rectangular grid
of locations, or a subset of such a grid, at which predictions
will be computed. Binary image masks
are conveniently created using as.mask.
The argument covariates gives the values of any spatial covariates
at the prediction locations.
If the trend formula in the fitted model
involves spatial covariates (other than
the Cartesian coordinates x, y)
then covariates is required.
The argument covariates has the same format and interpretation
as in predict.ppm. It may be
either a data frame (the number of whose rows must match
the number of pixels in locations multiplied by the number of
possible marks in the point pattern), or a list of images.
If argument locations
is not supplied, and covariates is supplied, then
it must be a list of images.
If the fitted model was a marked (multitype) point process, then predictions are made for each possible mark value in turn.
If the fitted model had no spatial trend, then the default is
to omit calculating this (flat) surface, unless trend=TRUE
is set explicitly.
If the fitted model was Poisson, so that there were no spatial interactions,
then the conditional intensity and spatial trend are identical, and the
default is to omit the conditional intensity, unless cif=TRUE is set
explicitly.
If plot.it=TRUE then plot.plotppm() is called
upon to plot the class plotppm object which is produced.
(That object is also returned, silently.)
Plots are produced successively using persp,
image and contour (or only a
selection of these three, if how is given). Extra
graphical parameters controlling the display may be passed
directly via the arguments ... or indirectly reset using
spatstat.options.
An object of class plotppm. Such objects may be plotted by
plot.plotppm().
This is a list with components named trend and cif,
either of which may
be missing. They will be missing if the corresponding component
does not make sense for the model, or if the corresponding
argument was set equal to FALSE.
Both trend and cif are lists of images.
If the model is an unmarked point process, then they are lists of
length 1, so that trend[[1]] is an image of the spatial trend
and cif[[1]] is an image of the conditional intensity.
If the model is a marked point process, then trend[[i]]
is an image of the spatial trend for the mark m[i],
and cif[[1]] is an image of the conditional intensity
for the mark m[i], where m is the vector of levels
of the marks.
See warnings in predict.ppm.
Adrian Baddeley Adrian.Baddeley@curtin.edu.au, Rolf Turner r.turner@auckland.ac.nz and Ege Rubak rubak@math.aau.dk
m <- ppm(cells ~1, Strauss(0.05))
pm <- plot(m) # The object ``pm'' will be plotted as well as saved
# for future plotting.
pmPlease choose more modern alternatives, such as Google Chrome or Mozilla Firefox.