[ADMB Users] inverse of a bivariate normal CDF (question for Dave)

[Ben Bolker]

  googling "admb bivariate normal" brings up links to

http://admb-project.org/documentation/api/vbivnorm_8cpp_source.html
http://admb-project.org/documentation/api/cbivnorm_8cpp_source.html

http://admb-project.org/documentation/api/cbivnorm_8cpp.html

 but these just compute cumulative bivariate normals, not quantiles
(inverse CDFs).

  The syntax seems fairly straightforward (arguments x,y,rho or xl, yl,
xu, yu, rho).

  Looked through the rest of the source tree and didn't find anything
else, so if there is an inverse CDF out there it might be something
custom-built by DF (who I think is on vacation/holiday right now, so you
might have to wait a bit ...)

  Hope that helps, sorry if it was all already obvious.


On 13-01-03 10:12 AM, Hai Che wrote:
> Hi all (and especially Dave),
> 
> Happy 2013. I am working on a project and need to have the following 2
> functions in AD Model Builder,
> 
> Say I have a bivariate normal CDF (cumulative density function), with
> means 0, variances 1, and correlation coefficient \pho. I need to get
> the inverse of this bivariate normal CDF.
> 
> I did not find the statistical functions in AD Model Builder manuals,
> but I remember Dave wrote something on this (I contacted you earlier and
> you send me the syntax). Would you please help me with this again by
> letting me and others know the syntax of bivariate normal CDF and its
> inverse? Posting this to the forum will also ensure that I won't forget
> this again (or rather, could not locate the email anymore due to moving...).
> 
> Many thanks,
> 
> Hai
> 
> 
> 
> 
> _______________________________________________
> Users mailing list
> Users at admb-project.org
> http://lists.admb-project.org/mailman/listinfo/users
>