Robert,

You're right about the questions to ask and the syrjala test is not
appropriate for this purpose.

Thanks

2011/11/29 Robert J. Hijmans <[email protected]>

> Mathieu,
>
> The syrjala test
> http://rgm2.lab.nig.ac.jp/RGM2/func.php?rd_id=ecespa:syrjala is, like
> other statistical tests, designed to estimate differences between
> populations from samples. I do not think such tests are relevant in this
> case as you do not have samples, you have the entire populations. Any
> difference you find is therefore not an estimate, it is exactly what it is
> (p=0 if you like). I think the question to ask in these cases is, how much,
> and where, and how systematic, are the differences, and if the differences
> are relevant for you (for your application).
>
> Robert
>
>
> On Tue, Nov 29, 2011 at 7:09 AM, Mathieu Rajerison <
> [email protected]> wrote:
>
>> Hi List,
>>
>>
>> The GDEM ASTER SRTM V2 has been released not a long time ago.
>>
>> I'd like to test whether it would be convenient to derive Terrain
>> Attributes from this Digital Elevation Model in a denoised version.
>>
>> To do that, I thought of comparing a reference paying digital terrain
>> model
>> from the French National Institute of Geography with this denoised GDEM.
>>
>> I thought of performing a syrjala test on these two data sources. Would
>> you
>> suggest other methods?
>>
>>
>> Thanks,
>>
>> Mathieu
>>
>>        [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> [email protected]
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>
>

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo

Reply via email to