Skip to content

question concerning nmds.min

8 messages · Ivailo, Sarah Goslee, Gian Maria Niccolò Benucci

#
2012/2/2 Gian Maria Niccol? Benucci <gian.benucci at gmail.com>:
Dear GIan,

try the following:

library(vegan)
data(dune)
sol <- metaMDS(dune)
sol

Call:
metaMDS(comm = dune)

global Multidimensional Scaling using monoMDS

Data:     dune
Distance: bray

Dimensions: 2
Stress:     0.1183195
Stress type 1, weak ties
Two convergent solutions found after 5 tries
Scaling: centring, PC rotation, halfchange scaling
Species: expanded scores based on 'dune'


Cheers,
Ivailo
#
2012/2/2 Gian Maria Niccol? Benucci <gian.benucci at gmail.com>:
...
Forgot about the R-squared - according to the vegan FAQ "1 - stress^2
transforms nonlinear stress into quantity analogous to squared
correlation coefficient. Function stressplot displays the nonlinear
fit and gives this statistic."

HTH,
Ivailo
#
Are you using ecodist?

If so, you need to look at iris.nmds to get the stress and the
r2.

min(iris.nmds$stress)
iris.nmds$r2[which.min(iris.nmds$stress)]

The advantage of nmds.min() is that it lets you choose
a particular dimension solution rather than the lowest
available.

For your example, the same thing can be achieved with
iris.nmds$conf[[which.min(iris.nmds$stress)]]

Sarah

2012/2/2 Gian Maria Niccol? Benucci <gian.benucci at gmail.com>:

  
    
#
In the next update of ecodist, nmds.min() will return the stress and
r2 for the selected configuration. Thanks for the suggestion.

Sarah

2012/2/2 Gian Maria Niccol? Benucci <gian.benucci at gmail.com>: