Re: M code for semiovariogram?

[PhilipNienhuis]

forkandwait wrote
> PhilipNienhuis 
> <pr.nienhuis <at>
>  hccnet.nl> writes:
> 
>> 
>> forkandwait wrote
> 
> (SNIP)
>  
>> There's lots of geostatistics code around in .m-file format, a few years
>> back I surveyed the web for directly usable scripts, much like you now.
> 
> I think basic variogram and kriging is simple enough we all roll it 
> ourselves, and then there is a lot "lying around".

Simple enough? All true, until you really start using it and hit the
simplifying assumptions.


>> The most accessible is probably the 'kriging toolbox' (google for it),
>> however it's a bit geared towards mining  - the 3D input requirements may 
> be
>> a bit convoluted for simple 2D analyses.
> 
> I think it also requires compiling mex files.

Now that you mention... IIRC that stuff (the source code was included) could
easily be turned into an .oct file.


>> Matlab Central probably has some.
> 
> Watch out -- JWE is going to give you a hard time on this (justifiably, 
> probably)...

Once you find something there, it's easy to contact the authors. Happened
several times before.


> (SNIP)
> 
>> I suppose there's room for an easily usable geostatistics package on
>> Octave-Forge 
> 
> Isn't that octave-gpr?

Maybe, but unless geostatistics has changed a lot in 25 years I do not
recognize the procedures in octave-gpr. That is, Jaroslav -no doubt being a
good mathematician- has generalized the easy case so much that it looks
incomprehensible to me.

Anyway, rather than trying to roll your own, you could also have a look at
e.g., gslib (gslib.com). OK it isn't m-file based but it does work.

Philip



--
View this message in context: 
http://octave.1599824.n4.nabble.com/M-code-for-semiovariogram-tp4654490p4654586.html
Sent from the Octave - General mailing list archive at Nabble.com.