[R] Confidence interval on quantile regression (quantreg)

[Cade, Brian]

Depending on the procedure used for estimating the CI, especially if the
default rankscore inversion method, then it is possible that legitimate end
points of the intervals for some quantiles with a given alpha (e.g., 0.05
for 95% CI) cannot be refined beyond plus or minus infinity.  Of course,
this typically happens for smaller sample sizes, more extreme taus, and
more complex models.  But unusual distributional characteristics of the
data distributions can also contribute to this issue.

Brian

Brian S. Cade, PhD

U. S. Geological Survey
Fort Collins Science Center
2150 Centre Ave., Bldg. C
Fort Collins, CO  80526-8818

email:  cadeb at usgs.gov <brian_cade at usgs.gov>
tel:  970 226-9326


On Wed, Oct 19, 2016 at 11:05 AM, Frank Black <frank.black.1988 at gmail.com>
wrote:

> Hi all,
>
> I am using the quantile regression package for estimating some models.
> However, in some cases the intervals' upper bounds are either abnormally
> high or low, with values such as -1.797693e+308 or 1.797693e+308. Actually,
> the number is the same in absolute terms.
>
> Does anyone know a reasonable explanation for this?
>
> Thanks.
>
> Kind regards,
> Frank
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/
> posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
>

	[[alternative HTML version deleted]]