R help - Apr 2012 - Centering (e.g. on prcomp)

Hi!

I'm doing a PCA on some testdata (only 2 features) currently to see what
happens. I read a lot of times one should center ones data, so i left the option
on.

But now i am scratching my head on how to correctly transform new data to be
comparable to the data used in the pca.

pca$x are the points (in PC system) the function returns
origpoints%*%pca$rotation

is not quite the same, because it is missing the centering

for the original data i can of course do scale(origpoints,scaled=F) beforehand,
and then the results will be the same, but for new data that option isn't
available. So how would i properly transform it?

greetings
Jessica

Ah.. forget it, something must have gone wrong while i was testing (probably
messed up some apply()). Seems you can just subtract the returned means - i hope
at least.


Am 12.04.2012 um 14:30 schrieb Jessica Streicher:
> Hi!
> 
> I'm doing a PCA on some testdata (only 2 features) currently to see
what happens. I read a lot of times one should center ones data, so i left the
option on.
> 
> But now i am scratching my head on how to correctly transform new data to
be comparable to the data used in the pca.
> 
> pca$x are the points (in PC system) the function returns
> origpoints%*%pca$rotation
> 
> is not quite the same, because it is missing the centering
> 
> for the original data i can of course do scale(origpoints,scaled=F)
beforehand, and then the results will be the same, but for new data that option
isn't available. So how would i properly transform it?
> 
> greetings
> Jessica
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

	[[alternative HTML version deleted]]