[R-sig-Geo] pattern simulation in R

[Rolf Turner]

On 12/05/12 04:15, Wall, Wade A ERDC-RDE-CERL-IL wrote:
> Hi all,
>
> Does anyone know if there is a pattern generation function in R? I would like to generate different patterns (random, clumped, overdispersed) with different numbers of points, but have been unable to find (and have been unsuccessful in coding) a function to do this. This is related to the power analysis question, but thought I would put it in a separate thread.
>
> Thanks for any information.

There is a wide variety of pattern simulation functions in the 
"spatstat" package
(which you appear to be using already).  Execute ?spatstat and then 
scroll down
a little to "To simulate a random point pattern:" to see a list of the 
possibilities.

I'm not sure what you mean by a pattern being "overdispersed".  Do you mean
exhibiting repulsion ("unusual regularity")?  This can be simulated 
using the Strauss
process (via rmh() or rStrauss()) or the hardcore process (using 
rmh()).  The
function rSSI() also simulates patterns with (a different version of) 
hardcore
repulsion.

Clustered patterns can be simulated via, e.g. rThomas() and rMatClust().

"Random" (I presume you mean Poisson) patterns can be simulated using
rpoispp().

Is there some form of pattern generation that you want that is *not* 
currently
provided by "spatstat"?

     cheers,

         Rolf Turner