how to calculate "axis variance" in metaMDS, pakage vegan?

headers Victor Landeiro | 1 Dec 15:36

Dear all,
What about a r2 statistic in this case? I have seen some people using the r2
of the relationship (a matrix regression) between te original distance and
the final distance obtained with the NMDS.
This r2 is not similar to the stress measure, as we can see above:

library(vegan)
data(varespec)
original.dist<-vegdist(varespec)
r2<-numeric(10)
stress<-numeric(10)
for(i in 1:10){
nmds.resu<-metaMDS(varespec,k=i)
nmds.scores<-scores(nmds.resu)
nmds.dist<-dist(nmds.scores)
r2[i]<-summary(lm(original.dist~nmds.dist))[[8]]
stress[i]<-nmds.resu$stress
}
r2     ## r2 initially increase and as more axis are added it begins to
decrease.
stress
plot(r2,stress)

I have read somewhere something like "dont stress about stress", however I
really dont know what is best, stress or r2. People say that this r2 is
something like "how good you recovered the original distances with the NMDS
ordination".
What about this r2 measure?

victor

2009/12/1 Jari Oksanen <jari.oksanen@...>

> On 1/12/09 11:12 AM, "Gian Maria Niccolò Benucci" <gian.benucci <at> gmail.com>
> wrote:
>
> > Hi Hi there,
> >
> > I am trying to use funcion metaMDS (vegan pakage) for Community Ecology
> > data, but I find no way to calculate the "expressed variance" of the
> first 2
> > axis? is there a way to do that?
> > Thanks a lot in advance,
> >
> Gian,
>
> Why nobody asks how to calculate the non-linear stress of PCA or RDA?
> Non-linear stress would make just as little sense for PCA or RDA as the
> axis
> variance makes for NMDS. The axis variance is purely a measure of PCA
> family
> of methods (even CA family has a bit different measure), and there is no
> meaningful way of calculating "expressed variance" for NMDS axes (and
> indeed, a separate axis is not meaningful for NMDS: it is the configuration
> spanned by all axes together that makes sense).
>
> Cheers, Jari Oksanen
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology@...
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>

-- 
victor
_______________________________

Victor Lemes Landeiro
Instituto Nacional de Pesquisas da Amazônia - INPA - CPEc (Ecologia)
Av. André Araujo, 2936, Petrópolis
CEP: 69067-375
Manaus, Amazonas, Brasil
Telefones: INPA (92) 3643 1912
Casa (92) 3646 3942 ou 3304 3942
Celular (92) 88311121
www.inpa.gov.br
http://ppbio.inpa.gov.br
http://ppbio.inpa.gov.br/Port/public/disciplinas2/Introducao%20ao%20R%202009.pdf

	[[alternative HTML version deleted]]

Re: how to calculate "axis variance" in metaMDS, pakage vegan?