simulateMPP {MarkedPointProcess} | R Documentation |
simulateMPP
generates realisations of marked point processes
simulateMPP(coordmodel=c("given", "uniform", "Poisson"), coord=NULL, npoints=NULL, lambda=NULL, window=NULL, edgecorrection=0.0, repetitions=1, coordrepet=1, model=NULL, register=0, method=NULL)
coordmodel |
if coordmodel="given" then coord
are expected to be given and not simulated;
if coordmodel="uniform" then
npoints uniformly distributed points are created; if
coordmodel="Poisson" then a conditional Poisson point process is
simulated with intensity lambda |
coord |
matrix with 2 columns; coordinates of the points;
coord is given only if coordmodel="given" |
npoints |
number of coordinates;
npoints must be given if coordmodel="uniform" .
|
lambda |
intensity of the Poisson process;
lambda must be given if coordmodel="Poisson" . |
window |
= c(xlim, ylim) . window must be given
if coordmodel equals "uniform" or "Poisson" .
|
edgecorrection |
double. If edgecorrection >0
then a Poisson process is simulated
with intensity lambda in a frame of thickness edgecorrection
around the window .
If window is not given, the range of the x values and the
range of the y values are taken to define the window .
If lambda is not given, the intensity within the
window is used instead. |
repetitions |
integer; number of independent drawings of the marks for a given set of coordinates |
coordrepet |
number of independent drawing of the coordinates;
this parameter is ignored in case of coordmodel="given" |
model |
list of lists; model for the marks; see Details
and get.mpp.names .
|
register |
the register where intermediate results in the
Gaussian random field simulation are stored, see
GaussRF |
method |
the method by which the Gaussian random field is
simulated; if is.null(method) then the method is chosen
automatically, see GaussRF
|
The definition of a model is of the form
model = list(l.1, OP.1, l.2, OP.2, ..., l.n)
.
The lists l.i
are all either of the form l.i = list(model=,var=,kappas=,scale=)
or of the form l.i = list(model=,var=,kappas=,aniso=)
in case
of random field parts, or of the form l.i =
list(model=,param=)
in case of marked point process parts.
l.i$model
is a string; var
gives the variance;
scale
is a scalar whereas aniso
is a d x d matrix, which is multiplied from left to the points, and
at the transformed points the values of the random field are
calculated. The dimension d of matrix must match the
number of rows of x
. param
is vector of real values
whose length depends on the specified model
. The
models for the random field part
can be combined by OP.i="+"
or OP.i="*"
, those for the
marked point process parts only by OP.i="+"
.
coordrepet=1
the function returns list(coord, data)
,
data
contains the independent drawing of the marks (as
columns)
coordrepet>1
the function returns
list( list(coord, data), ..., list(coord, data) )
Martin Schlather, martin.schlather@math.uni-goettingen.de http://www.stochastik.math.uni-goettingen.de/institute
get.mpp.names
,
rfm.test
, simulateMPP
,
splitmodel
, MarkedPointProcess
xlim <- c(0, if (interactive()) 200 else 20) mpp <- simulateMPP(coordmodel="Poisson", lambda=1, window=c(xlim=xlim, ylim=c(20, 70)), repet=3, coordrepet=4, model=list(list(model="exp", var=1, scale=10), "+", list(model="nearest neighbour", p=1))) str(mpp)