[R-sig-Geo] [R-sig-eco] fitting point pattern using spatstat

[Rolf Turner]

Thanks to Marcelino for handling this question.  I would just like to
add:  Be careful!  Make sure that you really know what you are
doing.  Check that the results of your fit make sense.  The defaults
for Kest() are based on the insight of some very clever people such
as Adrian Baddeley and Brian Ripley who *do* know what they are
doing.  There are good reasons for limiting "r" to be <= 1/4 of the
shortest side.  The K function is *not* estimated very well for larger
r values --- due to edge effects.  Hence fitting by the method of minimum
contrasts may not be very reliable when based on K function fits over
larger r values.

The method of minimum contrasts is pretty much a "suck it and see" approach
anyway, as far as I can understand it.  So if you have sound reasons for 
looking
at larger values of r, then by all means go for it.  But  make sure that you
have sound reasons.

     cheers,

         Rolf Turner

On 18/10/12 20:32, Marcelino de la Cruz wrote:
> Hi Stjepan
>
> That occurs because thomas.estK calls internally Kest to compute the
> observed K and by default it computes up to 1/4 of the shortest side
> of your observational window.
>
> To "customize" the range of distances you should go in two steps:
>
> abal.K <- Kest (abal.ppp, r= seq(0,50, by = 0.5))
> fitt_pp <- thomas.estK(abal.K, c(kappa=10, sigma2=0.1), *rmin* =
> NULL, *rmax* = 50)
>
>
> By the way,  questions about spatstat are usually answered faster (or
> sometimes are only answered) if posted to the R-sig-geo list.
>
> Cheers,
>
> Marcelino
>
>
>
>
> At 18:08 17/10/2012, stjepan mikac wrote:
>> Dear all,
>> I'm trying to fit point pattern using spatstat and thomas.estK command for
>> distances r (0,50) but default values are only 0,25. Can I change this
>> values and try to fit for all range of distances.
>>
>> fitt_pp <- thomas.estK(abal.ppp, c(kappa=10, sigma2=0.1), *rmin* = NULL,
>> *rmax* = 50)
>> /Error in mincontrast(K, theoret, startpar, ctrl = list(q = q, p = p, rmin =
>> rmin,  :
>>    rmax= 50 exceeds the range of available data = [ 0 , 25 ]/
>>