Prev: Unzip a file
Next: Weighted Average
From: "adel F." on 26 Sep 2006 18:34
Thanks Rogerio,
Do you have any examples with S+?
Regards
Adel

rdporto1 <rdporto1(a)terra.com.br> a ?crit :
Adel,

if what you want is a PROC like S+ copula functions, I don't think
SAS has it yet. SAS uses a Normal copula in PROC MODEL but it
doesn't do any goodness of fit test or similar.

It seems that David will recomend some alternative approach but
if you HAS TO use copulas, then you will have to program it.

Best,

Rogerio.

---------- Cabe?alho original -----------

De: "SAS(r) Discussion" SAS-L(a)LISTSERV.UGA.EDU
Para: SAS-L(a)LISTSERV.UGA.EDU
C?pia:
Data: Tue, 26 Sep 2006 18:05:13 +0200
Assunto: RE : Re: copulas and SAS

> Thanks David,
>
> I have two variables X1 and X2 representing cognition scores , they are naturally associated, the corr=0.5
>
> X1 is squewed on the right and X2 is fairly normal
>
> I have also a variable representing age (continuous 40+) and educational level (3 categories)
>
> My first idea is to find the best copula, archmedian copula for (X1,X2) which fits my data, and then to try to link this copula to age and eduction, a kind of quintile regression.
>
> I am looking for goodness of fit test for archimedan copula, they are easy to interpret, because of the association paramter
>
>
> Adel
>
>
> David L Cassell a ?crit :
>
> adel_tangi(a)YAHOO.FR wrote:
> >
> >Hi,
> > I am looking for SAS procedures or examples using SAS syntax for
> >estimating a copula distribution for continuous couples of variables X and
> >Y
> >
> > Many thanks
> >
> > Adel
>
> What are you looking for? A copula is nothing more than a functional
> allowing you to link the multivariate distribution to marginal distributions
> in an invariate manner. So saying that you want to treat X and Y as
> independent is - in essence - choosing a copula that relates f(X,Y) to
> the marginals f1(X) and f2(Y). So there may be a couple dozen SAS
> procs that might apply.
>
> If you explain exactly what you are after, and why, and what the larger
> picture is, then someone here is likely to be able to assist you further.
>
> HTH,
> David
> --
> David L. Cassell
> mathematical statistician
> Design Pathways
> 3115 NW Norwood Pl.
> Corvallis OR 97330
>
> _________________________________________________________________
> Try the new Live Search today!
> http://imagine-windowslive.com/minisites/searchlaunch/?locale=en-us&FORM=WLMTAG
>
>
>
> ---------------------------------
> D?couvrez un nouveau moyen de poser toutes vos questions quelque soit le sujet ! Yahoo! Questions/R?ponses pour partager vos connaissances, vos opinions et vos exp?riences. Cliquez ici.
>




---------------------------------
D?couvrez un nouveau moyen de poser toutes vos questions quelque soit le sujet ! Yahoo! Questions/R?ponses pour partager vos connaissances, vos opinions et vos exp?riences. Cliquez ici.
 | 
Pages: 1
Prev: Unzip a file
Next: Weighted Average