[R] what is mvrCv PRESS?

[Markus Mühlbacher]

   Hello!
   I have a problem with the multivariante regression function mvr (PLS
   package) with leave-one-out validation.
   In the value ...$validation there are the $PRESSs listed. I found in the
   literature PRESS is the sum of squares of observed minus predicted. Well, I
   don't understand how PRESS can grow with a better model (which indicated by
   more the use of more latent variables and a higher R-squared).
   In fact my curve plotting the PRESS against the number of latent variables
   has a minimum, which is another point I don't really understand. I apologize
   if my questions are silly, but I am a newbie in this context.
   I just don't know get it how the PRESS value is calculated and I would
   kindly appreciate it if somebody could help me with this topic.
   Best wishes,
   Markus
     _________________________________________________________________

   Be a better friend, newshound, and know-it-all with Yahoo! Mobile. [1]Try it
   now.

References

   1. http://us.rd.yahoo.com/evt=51733/*http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ