[R] RES: adding points to a point pattern

[Frederico Mestre]

Hello,

Yes, I'm using spatstat.

Sorry, I forgot to mention that.

Thanks,

Frederico 



-----Mensagem original-----
De: Rolf Turner [mailto:rolf.turner at xtra.co.nz] 
Enviada em: terça-feira, 4 de Setembro de 2012 02:04
Para: Frederico Mestre
Cc: r-help at r-project.org; Adrian.Baddeley at csiro.au
Assunto: Re: [R] adding points to a point pattern

On 04/09/12 11:58, Frederico Mestre wrote:
> Hello all:
>
>   
>
> Is there any way to add points to a point pattern, while keeping a 
> given minimum distance amongst this new points and the pre-existing
points?
>

Presumably this is a question about the "spatstat" package.

If so, the answer is "yes".  At least two ways:

(1) Using rmh() --- need to specify "beta" (the "chemical" activity
parameter).  E.g.

# Generate a "test pattern" to which to add points.
set.seed(42)
X <- rSSI(0.05,100)

# Add points.
M <- rmhmodel(cif="hardcore",par=list(beta=300,hc=0.05))
Y  <-
rmh(M,start=list(x.start=X),expand=1,control=list(x.cond=as.data.frame(X)))
plot(Y,main="Pattern with points added via rmh")
plot(X,add=TRUE,chars=20,cols="red")
print(min(nndist(Y)))

(2) Using rSSI() --- need to specify the *total number* of points desired.
E.g.

# Add points to the same "test pattern", in a different way.
Z <- rSSI(0.05,npoints(X)+42,x.init=X)
plot(Z,main="Pattern with points added via rSSI")
plot(X,add=TRUE,chars=20,cols="red")
print(min(nndist(Z)))

HTH

     cheers,

         Rolf Turner