[R] Odp: save a regression model that can be used later

Petr PIKAL petr.pikal at precheza.cz
Tue Nov 23 10:13:18 CET 2010

r-help-bounces at r-project.org napsal dne 22.11.2010 16:02:20:

> Hi everyone
> I have a question about how to save a regression model in R and how to 
> retrieve it for making predictions in a new session.
> To be more specific, I fitted a multilevel logistic regression model 
using the
> lmer  from the "lme4" package. I then successfully make predictions 
> fitted(mymodel).
> Since data are complex (three levels, nested, numerous categorical and 
> continuous data describing types of laparoscopic surgery), the computer 
> quite a while to fit the MLM model.  I wonder whether it's possible to 
> the fitted model so that I don't have to fit it again for making 
> every time I start a new R session.

When you quit R session with option save =Yes you get a file .Rdata. 
Whenever you start R with this file you get saved environment back, 
together with your "mymodel".

I recommend for each bigger task to create separate directory in which you 
can save your .Rhelp and .Rdata file without interfering other tasks.


> I searched the mailing-list archive. Suggestions include using save () 
to save
> the model as "mymodel.rda" and then use load(mymodel.rda) into the 
> I tried without success (in Windows), returning the error message: 
"Error in 
> object$fitted : $ operator is invalid for atomic vectors"
> Did I do anything wrong?  Any help on this topic is much appreciated

1.11 Data permanency and removing objects

from R-intro


> BW, Melody
> --
> Dr Melody Ni
> Imperial College
> Department of Surgery and Cancer
> 10th floor, QEQM Building
> St. Mary's Hospital
> London W2 1NY
> Tel/Fax: +44 (0) 20 331 27657/26309
> z.ni at imperial.ac.uk<mailto:z.ni at imperial.ac.uk>
>    [[alternative HTML version deleted]]
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
> and provide commented, minimal, self-contained, reproducible code.

More information about the R-help mailing list