Hello there,
Here's a question regarding p-values on clusters produced by hierarchical
cluster analysis. A web search led me to the program pvclust to tackle this
problem. But when I run the problem I get strange results. The 'AU'
(approximately unbiased) p-values are very different from the 'BP' values
(ordinary boot-strap) p-values. The AUs commonly are in the 80-100 range
where the BPs are in the 0-10 range!
One clue as to what might be happening is a warning from R: "inappropriate
distance matrices are omitted in computation" for different r values.
Perhaps my data is not really well suited for this analysis? I have
presence-absence (0,1) data of 32 species in 60 different locations. It is
noticeable that the difference between AU and BP is not as bad when I try to
cluster by species, as when I try to cluster by location.
Any comments on this issue would be very much appreciated. Are there any
other programs that might be able to produce similar results? Should I try
to run the bootstrap analysis myself in R, and is there a good introductory
test to tell me how to do this?
With thanks and best wishes,
Eben
Eben Goodale
NSF International Postdoctoral Fellow
University of Colombo, Sri Lanka
eben.goodale@...
[[alternative HTML version deleted]]
______________________________________________
R-help@... mailing list
https://stat.ethz.ch/mailman/listinfo/r-helpPLEASE do read the posting guide
http://www.R-project.org/posting-guide.htmland provide commented, minimal, self-contained, reproducible code.
