| Date: | Thu, 19 Feb 2009 17:29:19 -0500 |
|---|
| Reply-To: | Ligia Buia <ligia.buia@GMAIL.COM> |
|---|
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
|---|
| From: | Ligia Buia <ligia.buia@GMAIL.COM> |
|---|
| Subject: | sas/insight spline fit statistics |
|---|
| Content-Type: | text/plain; charset=ISO-8859-1 |
|---|
Hello,
I am trying to understand how to interpret the statistics provided when I do
a spline fit of my biological marker vs. time data. My specific questions
are: Does my large MSE (GCV) tell me that there really is no trend? What
constitutes an "acceptable" MSE? Should I be worried that my r-square value
is so small? Here is part of my output:
c value: 0.3673
smoothing parameter: 0.1260
df: 8.3
r-square: 0.0456
mse: 69018
mse (gcv): 71724.288
Any help would be greatly appreciated!
| 
