Message-ID: <Pine.LNX.4.61.0505141942390.17431@gannet.stats>
Date: 2005-05-14T18:50:03Z
From: Brian Ripley
Subject: pmvnorm
In-Reply-To: <000c01c558b0$ade17710$c800a8c0@thesahajamach>
On Sat, 14 May 2005, Dimitri Joe wrote:
> pmvnorm(lo=c(-Inf,-Inf), up=c(Inf,Inf), mean=c(0,0) )
>
> should give me "1", right? But it doens't - it giver me "0".
> Would someone help me, please?
Yes, but pmvnorm is in package 'mvtnorm' (uncredited here) so please
discuss this with the package maintainer (Cc:ed here).
Extreme cases are often overlooked: I have this past week corrected
several in R itself (so please do report them to the appropriate person).
For now, please use something like
> In <- 1e100
> pmvnorm(lo=c(-In,-In), up=c(In,In), mean=c(0,0) )
[1] 1
--
Brian D. Ripley, ripley at stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595