This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Documentation error: bivariate distribution
- From: Carsten Svaneborg <svanebor at mpip-mainz dot mpg dot de>
- To: gsl-discuss at sources dot redhat dot com
- Date: Wed, 30 Apr 2003 11:38:20 +0200
- Subject: Documentation error: bivariate distribution
Hi!
The documentation states that the bivariate distribution is
(besides normalization constant)
\exp (-(x^2 + y^2 - 2 \rho x y)/2\sigma_x^2\sigma_y^2 (1-\rho^2))
I don't know what the matematical definition of a bivariate
distribution is, but for dimensional and symmetry reasons
this expression is clearly wrong.
However, the gsl_ran_bivariate_gaussian, does calculate
pairs of numbers that fullfill sigma_x = <x*x>,
sigma_y = <y*y> and <x*y> = rho*sqrt(<xx><yy>) as
desired.
I would suggest rewriting that the distribution is
N \exp( - (a x^2 + by^2 - 2 c xy) )
where a,b,c are determined to produce the desired
moments above.
--
Mvh. Carsten Svaneborg
Max Planck Institute for Polymer Science, Mainz, Germany