[R-sig-Geo] generate simulation data for a theoretical spatial model
Paulo Justiniano Ribeiro Jr
paulojus at c3sl.ufpr.br
Wed Feb 3 12:18:46 CET 2010
Just to add that besides unconditional simulation as discussed there are
some functionality for conditional simulations is some packages.
Is geoR the functions krige.conv() and krige.bayes() have an option for
that and in geoRglm the krige.binom(), krige.pois() or krige.glsm() also
does that. Except the first in geoR, the functions implements predictions
under the Bayesian paradigma relying on (conditional) simulations in their
algorithms. The algorithms have an argument to "keep" the simulations on
the resulting object.
They are based on the approach of assuming an hierarquical
spatial model with an underlying (latent) Gaussian field $S$
(even when the responses $Y$ are no Gaussian) and the conditional
simulatios are for such
latent field from which simulations on the scale of the response variable
can be obtained if wished under the conditional independence of the
responses $Y$ given $S$.
Paulo Justiniano Ribeiro Jr
LEG (Laboratorio de Estatistica e Geoinformacao)
Universidade Federal do Parana
Caixa Postal 19.081
CEP 81.531-990
Curitiba, PR - Brasil
Tel: (+55) 41 3361 3573
Fax: (+55) 41 3361 3141
e-mail: paulojus AT ufpr br
http://www.leg.ufpr.br/~paulojus
On Wed, 3 Feb 2010, Edzer Pebesma wrote:
> rusers.sh,
>
> demo(ugsim)
>
> in package gstat gives an example how to generate unconditional Gaussian
> simulations. Specifying the covariates in a formula and the parameter vector
> beta will add a deterministic trend to that.
>
> If, in addition to that, you want unconditionally simulated residuals added
> to a trend effect that is simulated as well, look at rmvnorm in package
> mvtnorm how to generate realisations from the multivariate normal
> distribution with given mean and covariance; finally, combine the two.
> --
> Edzer
>
> rusers.sh wrote:
>> It works. The problem is that it only generates the simulated data based
>> on our observed dataset,e.g. "meuse" here. I wonder if we can generate
>> the simulated dataset from the user-specified model with covariates
>> included, such as y~a1*x1+a2*x2+spatial effect. Y can be continuous or 0/1
>> variables. Something like this.
>> The idea is we first specify a theoretical model, and then generate the
>> simulated data based on this model. The coefficients and spatial effects
>> are fixed by users, so we may study some new methods.
>> Thanks.
>>
>> 2010/2/2 Edzer Pebesma <edzer.pebesma at uni-muenster.de
>> <mailto:edzer.pebesma at uni-muenster.de>>
>>
>>
>>
>> rusers.sh wrote:
>>
>> Hi Tomislav,
>> Thanks for your info on unconditional simulation. For conditional
>> simulations, i still cannot find any useful information.
>> I searched the R site and didnot find the possible method to do
>> conditional simulations.
>> 1. CondSimu(RandomField): trend: Not programmed yet. (used by
>> universal
>> kriging)
>> 2. grf(geoR): generates unconditional simulations of Gaussian
>> random fields
>> 3. sim.Krig(fields) #Conditonal simulation of a spatial process
>> It seems to be based on the actual dataset,not a theoretical
>> model.
>> 4. krige(gstat ):Simple, Ordinary or Universal, global or
>> local, Point or
>> Block Kriging,or simulation
>> x <- krige(log(zinc)~x+y, meuse, meuse.grid, model = m, block =
>> c(40,40),nsim=1)
>>
>> rusers.sh, please use
>>
>> x <- krige(log(zinc)~x+y, meuse, meuse.grid, model = m, nmax=40,
>> nsim=1)
>>
>> both adding the block=c(40,40) as well as omitting the nmax=40
>> tremendously increased the computing time you needed, the second
>> even more (in an O(n^2) manner) than the first.
>> --
>> Edzer
>>
>>
>>
>>
>> I used the above modified codes from krige(gstat ) example to
>> see the
>> effect of "nsim", but unfortunately, it took a longer time and
>> cannot get
>> the results. I guess it used the simulation method to test the
>> model, not
>> what i want. (My system is XP, R2.10.0, gstat09.-64.)
>> Anybody can give me further information on generating the
>> conditional
>> simulations from a theoretical model just like the
>> unconditional examples
>> that Tomislav provided?
>> Thanks a lot.
>>
>>
>> 2010/1/31 Tomislav Hengl <hengl at spatial-analyst.net
>> <mailto:hengl at spatial-analyst.net>>
>>
>>
>> Dear rusers.sh,
>>
>> Here are few simple examples of how to simulate (not-normal)
>> distributions and point processes using geoR and spatstat:
>>
>> http://spatial-analyst.net/book/node/388
>>
>> See also:
>>
>>
>> http://leg.ufpr.br/geoR/geoRdoc/vignette/geoRintro/geoRintrose8.html#x9-120008
>>
>> I guess that covariates can be also included (I guess that
>> you then need
>> to switch to conditional simulations - not sure).
>>
>> This should also work for lattice (polygon) data so that
>> you will have
>> jumps in values (but I guess you would still work in
>> gridded systems?).
>>
>> T. Hengl
>> http://home.medewerker.uva.nl/t.hengl/
>>
>>
>> rusers.sh wrote:
>>
>>
>> Hi all,
>> In classical statistics, we always need to generate a
>> theoretical model
>> such as y=a+b1*x1+b2*x2+e to study some new estimation
>> content. I am
>> wondering how to generate the similar spatial dataset
>> for a theoretical
>> model.
>> Say y is response variable, x1 and x2 are explanatory
>> variables.
>> 1. If y is a continous variable, how should we
>> generate the dataset for a
>> theoretical spatial point process model in R?
>> 2. If y is a continous variable, how should we
>> generate the dataset for a
>> theoretical spatial lattice data model in R?
>> 3. If y is 0/1 binary variable, how should we generate
>> the dataset for a
>> theoretical spatial point process model in R?
>> 4. If y is 0/1 binary variable, how should we generate
>> the dataset for a
>> ttheoretical spatial lattice data in R?
>> spatstat and other packages allow us to generate a
>> dataset of a specified
>> point process and other models, but it seems that they
>> donot allow us to
>> include possible explanatory variables into a
>> theoretical model. Maybe i
>> missed some ideas in them.
>> Anybody can express some ideas or point out some
>> useful resources on the
>> above four different situations? Small examples in R
>> are preferred.
>> Thanks a lot.
>>
>>
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at stat.math.ethz.ch
>> <mailto:R-sig-Geo at stat.math.ethz.ch>
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>>
>>
>>
>>
>>
>>
>> -- Edzer Pebesma
>> Institute for Geoinformatics (ifgi), University of Münster Weseler
>> Straße 253, 48151 Münster, Germany. Phone: +49 251 8333081, Fax:
>> +49 251 8339763 http://ifgi.uni-muenster.de
>> http://www.52north.org/geostatistics e.pebesma at wwu.de
>> <mailto:e.pebesma at wwu.de>
>>
>>
>>
>>
>> --
>> -----------------
>> Jane Chang
>> Queen's
>
> --
> Edzer Pebesma
> Institute for Geoinformatics (ifgi), University of Münster Weseler Straße
> 253, 48151 Münster, Germany. Phone: +49 251 8333081, Fax: +49 251 8339763
> http://ifgi.uni-muenster.de http://www.52north.org/geostatistics
> e.pebesma at wwu.de
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
More information about the R-sig-Geo
mailing list