Hello,
I try to find the relationships between the results of an NMDS and a
discriminant analysis.
I did a NMDS with a species dissimilarity matrix and just kept 3 axes. I
did a special discriminant analysis with a matrix of spectral reflection
(300 channels highly correlated with each other) for each plot included
in the species dissimilarity matrix. I tried to discriminate among 3
groups so I used 3 discriminant functions.
Now, I try to connect the results from both analysis.
My idea was to use a procrustes analysis and find the lowest error of
the site scores from the NMDS and the site scores from the discriminant
analysis. But I don't know whether this is valid or if this implies that
I compare apples with pears.
Does anybody else has an idea how I could combine the solution of the
previous analysis?
/Thomas
--
*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*^*
Thomas Möckel, PhD student
Department of Earth and Ecosystem Science
Lund University
Sölvegatan 12
SE-223 62 Lund
Sweden
_______________________________________________
R-sig-ecology mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
