[R] Is it possible to model with Laplace's error distribution?

[Berton Gunter]

As you haven't gotten a reply, I'll make an attempt; but caveat emptor!
Hopefully others will correct my errors.

The Laplace distribution is double exponential with heavy tails and for
which the sample median, not the mean, is the mle for the location
parameter.  In the more general linear modeling context, this suggests you
might be interested in quantile regression, for which Roger Koenker's
quantreg package is the place to go. However, I doubt that the lmer package
can deal with this in the mixed model context, as special algorithms are
required. Doug Bates or others should correct me if I'm wrong on this.

HTH. And again, caveat emptor.

-- Bert Gunter
Genentech Non-Clinical Statistics
South San Francisco, CA
 
"The business of the statistician is to catalyze the scientific learning
process."  - George E. P. Box
 
 

> -----Original Message-----
> From: r-help-bounces at stat.math.ethz.ch 
> [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Petar Milin
> Sent: Tuesday, March 21, 2006 2:42 PM
> To: R-HELP
> Subject: [R] Is it possible to model with Laplace's error 
> distribution?
> 
> Hello!
> My question is stated in the Subject: Is it possible to model with
> Laplace's error distribution? For example, lmer() function have few
> families of functions, like binomial etc., but not Laplace. 
> Is there any
> other package that would allow for Laplace? Or is there a way to give
> "user-defined" family?
> 
> Sincerely,
> P. Milin
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html
>