[R] Copula Regression

[Martin Maechler]

>>>>> Janmaat, John <John.Janmaat at ubc.ca>
>>>>>     on Tue, 1 Mar 2016 04:53:43 +0000 writes:

    > Well, seem to have solved own problem.
    > The article "Enjoy the joy of copulas: with a package copula" has a nice appendix that describes precisely how to set up what I am looking for.  The only minor issue is that some of the functions have been deprecated.

    > So, if anyone is curious, here is the article info (in BibTeX):

    > @Article{YanJ_2007_CopulasR,
    > title={Enjoy the joy of copulas: with a package copula},
    > author={Yan, Jun and others},
    > journal={Journal of Statistical Software},
    > volume={21},
    > number={4},
    > pages={1--21},
    > year={2007}
    > }

Thank you. As that Journal is a free open access journal, the url may be
even more useful:
   https://www.jstatsoft.org/index.php/jss/article/view/v021i04/v21i04.pdf

and (as current maintainer of the 'copula' package)
may I ask you how exactly you solved it (R code) and
why you could not easily replace the calls to the deprecated
functions (e.g. 'dcopula') by calls to their substitutes ?

But indeed, I'm impressed by Jun Yan  who had already added such
example code to the appendix of that paper relatively long ago.

I think it would make sense to add a version of that appendix
as (Rmd) vignette to our copula package.

Martin


    > John.

    > ----------
    > Dr. John Janmaat
    > Economics (Unit 8), IK Barber School of Arts and Sciences
    > The University of British Columbia
    > 3333 University Way, Kelowna, BC


    > -----Original Message-----
    > From: R-help [mailto:r-help-bounces at r-project.org] On Behalf Of Janmaat, John
    > Sent: February-29-16 2:55 PM
    > To: 'r-help at r-project.org'
    > Subject: [R] Copula Regression

    > Hello,

    > I'm trying to figure out how to run copula regressions in R.  I see a couple of packages that do specific versions, but I am hoping to be a bit more flexible.

    > My specific problem involves two ordered regressions (four levels each) that may be correlated.  Is there a package out there that can do copula regressions with arbitrary marginals?  Alternatively, is there a way to trick functions like GLM into returning observation specific CDF values given a guess for the parameters?

    > I would prefer to avoid coding this up, as I expect the more tested code I can use, the less likelihood for my own errors to mess up the results.

    > Thanks,

    > John.

    > ----------
    > Dr. John Janmaat
    > Economics (Unit 8), IK Barber School of Arts and Sciences The University of British Columbia
    > 3333 University Way, Kelowna, BC


    > [[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.

    > ______________________________________________
    > 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.