[pymvpa] Transpose dataset with PCAMapper

[John Magnotti]

Hi,

The documentation for PCAMapper (
http://www.pymvpa.org/modref/mvpa.mappers.pca.html ) shows that I can
pass in a flag for transposing the data, but I can an error when I try
to do so. It looks like the error is coming from mdp_adaptor

    node = mdp.nodes.PCANode(**kwargs)

Maybe mdp no longer accepts that argument?

I get an out of memory error if I try to run pca on the original
dataset, but I'm thinking if I can transpose the dataset first, then
it should be fine (smaller var-cov matrix).


thanks in advance for any advice,


John