R help - Apr 2003 - stepAIC/lme problem (1.7.0 only)

I can use stepAIC on an lme object in 1.6.2, but 
I get the following error if I try to do the same
in 1.7.0:

Error in lme(fixed = resp ~ cov1 + cov2, data = a, random = structure(list( : 
        unused argument(s) (formula ...)

Does anybody know why?

Here's an example:

library(nlme)
library(MASS)
a <- data.frame( resp=rnorm(250), cov1=rnorm(250),
                 cov2=rnorm(250), group=rep(letters[1:10],25) )
mod1 <- lme(resp~cov1, a, ~cov1|group, method="ML")
mod2 <- stepAIC(mod1, scope=list(upper=~(cov1+cov2)^2,
                                 lower=~cov1) )
					 
# it doesn't happen for normal linear models:

mod3 <- lm(resp~cov1, data=a)
mod4 <- stepAIC(mod3, scope=list(upper=~(cov1+cov2)^2,
                                 lower=~cov1) )

Thanks, 

Robert

On Mon, 28 Apr 2003, Robert Cuffe wrote:
> I can use stepAIC on an lme object in 1.6.2, but
> I get the following error if I try to do the same
> in 1.7.0:
>
> Error in lme(fixed = resp ~ cov1 + cov2, data = a, random = structure(list(
:
>         unused argument(s) (formula ...)
>
> Does anybody know why?
It's presumably related to the fact that the first argument to lme is
called `fixed' rather than `formula'.  It looks as though an
`addterm.lme'
may be needed.

	-thomas

There are changes to improve scoping in stepAIC that assume that update
works in ways that update.lme does not:
> mod2 <- addterm(mod1, ~(cov1+cov2)^2)
Error in lme(fixed = resp ~ cov1 + cov2, data = a, random = structure(list( : 
        unused argument(s) (evaluate ...)

The actual problem is in stepAIC which now has the line

    object$call$formula <- object$formula <- Terms

but fixing that one only opens up others (as above).

I was vaguely aware of this: I don't know of an easy fix.


On Mon, 28 Apr 2003, Robert Cuffe wrote:
> I can use stepAIC on an lme object in 1.6.2, but 
> I get the following error if I try to do the same
> in 1.7.0:
> 
> Error in lme(fixed = resp ~ cov1 + cov2, data = a, random = structure(list(
:
>         unused argument(s) (formula ...)
> 
> Does anybody know why?
> 
> Here's an example:
> 
> library(nlme)
> library(MASS)
> a <- data.frame( resp=rnorm(250), cov1=rnorm(250),
>                  cov2=rnorm(250), group=rep(letters[1:10],25) )
> mod1 <- lme(resp~cov1, a, ~cov1|group, method="ML")
> mod2 <- stepAIC(mod1, scope=list(upper=~(cov1+cov2)^2,
>                                  lower=~cov1) )
> 					 
> # it doesn't happen for normal linear models:
> 
> mod3 <- lm(resp~cov1, data=a)
> mod4 <- stepAIC(mod3, scope=list(upper=~(cov1+cov2)^2,
>                                  lower=~cov1) )
> 
> Thanks, 
> 
> Robert
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> 
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595