I'm not sure the idea of having a generic 'sd' extract the residual
standard
deviation from a linear model quite jives with me. Personally, I think a
generic function with name like 'residSD' or something similar would be
better.
Either way, I think it would be nice to have some function like this.
-roger
Rolf Turner wrote:> I have from time to time seen inquiries on r-help in respect of
> how to obtain the estimated standard deviation from the output of
> fitting a linear model. And have had occasion to want to do this
> myself.
>
> The way I currently do it is something like summary(fit)$sigma
> (where fit is returned by lm()).
>
> It strikes me that it might be a good idea to have an extractor
> function, analogous with coef() and resid(), to dig out this value.
> This would be in keeping with the R philosophy of ``don't muck
> about with the internal components of objects returned by functions,
> because the internal structure of such objects might change in future
> releases''.
>
> I'm not sure what the name of the extractor function should be.
> One idea would be to make sd() generic, and create a function
> sd.lm() which could currently have code
>
> sd.lm <- function(x,...) {
> summary(x)$sigma
> }
>
> The sd.default() function would have the code of the current sd().
>
> Does this idea have any merit?
>
> cheers,
>
> Rolf Turner
>
> ######################################################################
> Attention:\ This e-mail message is privileged and confid...{{dropped:9}}
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
--
Roger D. Peng | http://www.biostat.jhsph.edu/~rpeng/