What is puzzling you is the following in stat.anova called by anova.glmlist: dev[nnas] <- 1 - pchisq(abs(dev[nnas]), abs(df[nnas])) which takes the absolute value of the change in deviance, whether j
This is a scoping issue. Assign `data' in frame 1. stepAIC is support software for a book. This is covered briefly there, and in more detail in the companion volume `S Programming'. Dear S users, I'm
Dear S users, Thanks to everyone who responded to my question (reproduced at the end of this message). As it turns out it was not a problem with stepAIC, but with my understanding of the S scoping r
Your example appears to use anova, despite two claims to use summary. Yes, it is a scoping issue, and you need to assign data1 in frame 1. See `S Programming' or another good book on that subject. De
We have a situation where we know the variance of observations for a linear regression problem and need confidence limits on the parameter estimates. The documentation in the NOTES section for lm st
You are diving into undocumented internal structures of rpart. But as a general principle, .C is not designed to pass S objects like a formula as here (except to pass them on to call_S). Your call is
I am afraid this is a tale of woe. For some reason Insightful continues to supply a function, step.glm, with show-stopping bugs. (I looked at the code in S-PLUS 7, not having 6.1 to hand.) 1) I think
data.dump() is designed for just this purpose. Use data.restore() to read the dumps on the newer system. I have some splus 4.5 (windows) data objects that I want to convert to splus 6 for windows. Wh
Thanks for Terry to reply for my qustion. Note that it is implicit in Terry's answer that the hazard is smooth. You have to make some asssumption here, and his includes how smooth it is. That comes
This is S-news, not R-help, and you have not told us your platform. Should this be R for Windows, see the rw-FAQ Q2.10. Otherwise, please ask on R-help with the details requested in the R posting gui