order of explanatory variables in vegan CCA
Hi Matt, Would you mind sending me the data *off list* so I can see if I can reproduce. The failure is within a basic part of the R stack (Lapack matrix algebra code supplied with R) so points to something with the data and perhaps your matrix (or a subsidiary matrix computed during the CCA fitting) is close to being degenerate in some way and that when re-ordered it is slightly less degenerate. [I'm using the word degenerate loosely here). I promise to delete the data once I've reproduced the problem. G
On Wed, 2011-12-21 at 14:22 -0800, Gregory, Matthew wrote:
Hi all,
We run vegan CCA as part of an automated process. Our typical use case is to read in species and environment matrices (from CSV files) and subset the environmental matrix with a vector of field names ('vars'), e.g.
spp <- read.csv(spp_file, row.names = id_field)
env <- read.csv(env_file, row.names = id_field)
env <- env[,vars]
cca_obj <- cca(spp, env)
Today, for the first time, we received the following error:
Error in La.svd(x, nu, nv) : error code 1 from Lapack routine 'dgesdd'
which from googling implies that the SVD did not converge. If I reorder the 'vars' list and rerun, I can have it run to completion. I am completely na?ve with ordination modeling, but is there any reason why the order of environmental variables would matter to CCA? I would have thought this wouldn't matter.
Thanks for any pointers.
matt
_______________________________________________ R-sig-ecology mailing list R-sig-ecology at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~% Dr. Gavin Simpson [t] +44 (0)20 7679 0522 ECRC, UCL Geography, [f] +44 (0)20 7679 0565 Pearson Building, [e] gavin.simpsonATNOSPAMucl.ac.uk Gower Street, London [w] http://www.ucl.ac.uk/~ucfagls/ UK. WC1E 6BT. [w] http://www.freshwaters.org.uk %~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%