srd {MarkedPointProcess}R Documentation

Simulation study for marked point processes

Description

The function reproduces the simulation study published by Schlather, Ribeiro and Diggle (2004)

Usage

srd.jrssb(input=NULL, repet=500, dev=2, PrintLevel=2, readlines=TRUE,
          ps.path="ps/", data.save.path="data/", simu.path="simu/",
          tex.path="tex/", biondi.etal=NULL, final=TRUE)

Arguments

input vector of values for automatic choice of the submenues
repet number of realisations investigated for each parameter set. 2 * repet gives also roughly the number of hours needed to do all the simulations!
dev device, see Dev
PrintLevel If PrintLevel>1 some information on the course of the algorithm is printed
readlines logical. Only used if is.numeric(dev). If the operating system is unix and readlines=FALSE then the system waits a second before the next plot is calculated or shown. Otherwise the system waits for return.
ps.path path used for all postscript files in srd.jrssb
data.save.path directory where intermediate results in the analysis of the data are stored
simu.path directory used to store intermediate results for the simulations
tex.path path used for all created tex files containing tables
biondi.etal list of two components:
    coord:
    a matrix of two columns for the coordinates and
    diameter:
    a vector for the marks
The data set by Biondi et al., used in Schlather, Ribeiro and Diggle (2004), is not freely available.
final logical. If FALSE then several parameters are changed so that the algorithms run faster. Especially repet is limited to 3, optim is very sloppy, and the parameter sets for the simulation study are remarkably reduced. final=FALSE is used for debugging or demonstration, but the results are not reliable.

Details

The simulation study is written in a way that srd.jrssb() might be called several times on a parallel CPU system or PVD from the same directory. The non-interaction of the different instances is managed by means of lock files. It is ensured that no simulation study for a specific parameter set is performed twice. However, the programming has been kept simple so that a lock files might be created but the simulation is not performed by any process because of strange interaction. In this case all remaining lock files have to be deleted after the first run of the simulation study, and the study has to be re-run. Then, srd.jrssb() does the simulation study only for remaining cases.

For both the simulation studies and the data analyses, intermediate stored results are re-used if run again.

Value

NULL. Side effect is the creation of postscript files and tex files.

Author(s)

Martin Schlather, martin.schlather@math.uni-goettingen.de http://www.stochastik.math.uni-goettingen.de/institute

References

Schlather, M., Ribeiro, P. and Diggle, P. (2004) Detecting Dependence Between Marks and Locations of Marked Point Processes. J. R. Statist. Soc., Ser. B 66, 79-83.

Biondi, F., Myers, D.E., and Avery, C.C. (1994) Geostatistically modeling stem size and increment in an old-growth forest. Can. J. Forest Res. 24, 1354-1368.

See Also

rfm.test, simulateMPP, MarkedPointProcess

Examples

# repet=2 is for demonstration of functionality
# repet=50 gives more reliable results
# repet=500 and final=TRUE have been used in the simulation study
  
printlevel <- 1 + !interactive()
srd.jrssb(input=if(!interactive()) c(1:2,0), repet=2, dev=TRUE,
          final=FALSE, Pr=printlevel)

[Package MarkedPointProcess version 0.2.9 Index]