[R-sig-Geo] back with cutoff/variogram question

[Edzer Pebesma]

Erin, it is in

https://r-forge.r-project.org/scm/viewvc.php/pkg/src/sem.c?view=markup&root=gstat

file sem.c, line 692

defaults.h defines:

#define DEF_fraction    0.33333 /* fraction of max_dist for def. cutoff */

which is the default value for gl_fraction.

The default number of lag classes is 15 (gl_n_intervals). You can change
these defaults to e.g. 0.5 and 25 by:

variogram(gstat(NULL, "zinc", zinc~1, meuse,
    set=list(fraction=.5,intervals=25)))


On 11/24/2013 04:53 PM, Hodgess, Erin wrote:
> Hi!
> 
> This is from the gstat variogram.
> 
> I tried to track this down via the vario.c function, but I'm having some trouble with it.
> 
> ________________________________________
> From: b.rowlingson at gmail.com [b.rowlingson at gmail.com] on behalf of Barry Rowlingson [b.rowlingson at lancaster.ac.uk]
> Sent: Sunday, November 24, 2013 3:08 AM
> To: Hodgess, Erin
> Cc: r-sig-geo at r-project.org
> Subject: Re: [R-sig-Geo] back with cutoff/variogram question
> 
> On Sun, Nov 24, 2013 at 5:46 AM, Hodgess, Erin <HodgessE at uhd.edu> wrote:
>> Hello again.
>>
>> I'm back with the variogram question.  Here is the info from the help file:
>>
>>   cutoff: spatial separation distance up to which point pairs are
>>           included in semivariance estimates; as a default, the length
>>           of the diagonal of the box spanning the data is divided by
>>           three.
>>
>> By the box spanning the data, does that mean the bounding box, please?  If so, wouldn't that just be 2, please?
> 
>  There's quite a few 'variogram' functions in various R packages so it
> might help if you told us which one!
> 
>  I should think that "the box spanning the data" is indeed the
> bounding box. I don't see what you think would be '2' in this case
> though, unless the diagonal length of the bounding box happened to be
> 6.
> 
>  For data that included points at (0,0) and (1,1) the diagonal would
> have length sqrt(2) and so the cutoff default should be sqrt(2)/3.
> Note you have to actually have points at those coordinates and not
> just generate 1000 random points in the unit square.
> 
>  A few simple experiments with simulated point data sets with various
> bounding boxes should confirm this, or you could just look at the
> code?
> 
> Barry
> 
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> 

-- 
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of Münster
Heisenbergstraße 2, 48149 Münster, Germany. Phone: +49 251
83 33081 http://ifgi.uni-muenster.de GPG key ID 0xAC227795

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 555 bytes
Desc: OpenPGP digital signature
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20131124/6716903c/attachment.bin>