Skip to content

prcomp & eigenvectors ... ??

2 messages · Natalie Stephenson, Peter Dalgaard

1 day later
#
On May 28, 2011, at 20:08 , Natalie Stephenson wrote:

            
Not particularly, given that you may be missing some basic theory of PCA. Basically, the loadings of a PCA _are_ the eigenvectors, so e.g. plotting the loadings matrix against the variable number or whatever structure you have on them -- if they are spatial ore temporal measurements, plot against space/time. Something like

matplot(loadings(pca), type="l")
matplot(times, loadings(pca), type="l")

could prove useful (the latter case if your data are really short time series).