Date: Thu, 21 Nov 1996 11:48:21 +0100
Reply-To: Holger Geissler <holger.geissler@URZ.UNI-HEIDELBERG.DE>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: Holger Geissler <holger.geissler@URZ.UNI-HEIDELBERG.DE>
Subject: Regression Analysis: What is c(p) for?
I'm really sorry to disturb you with such an easy question, but I
don't really understand what C(p) in Proc Reg means. I know it's
Mallow's C(p) statistic, and it's something like r2, but what is the
difference? And what does it mean, when C(p) is negative? Is a large C(p)
good or bad? Any hits or literature?
Thanks in advance