[S] warning message from predict.gam

=?iso-8859-1?Q?Ejrn=E6s=2C_Rasmus?= (rej@DMU.dk)
Mon, 22 Jun 1998 11:18:22 +0100

Dear S-plussers

I receive the following warning message from predict.gam, when I use the
function to predict the response of a logistic regression model with
smoothed terms over an evenly spaced 4-dimensional grid ("dcaspace"):

cl27.surf <- predict(cl27.gam,dcaspace,type="response")
Warning messages:
There is a 0% discrepancy between the original and the re-fit used to
achieve 'safe'
predictions in: safe.predict.gam(object, newdata, type, se.fit,

How should I interpret this warning?

0 % discrepancy doesn't seem too serious.......or?


This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news