[R] Sampling two exponentials

[Ben Bolker]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Zhang Yanwei - Princeton-MRAm wrote:
| Thank you very much.
| Would you talk more about that? How can I use the copula package
to sample two dependent exponentials? Which function shall I use?
|

~  It's a big topic, and I'm not very familiar with it -- just
familiar enough that I can see that it might be useful.  The general
idea (Mathworld has a definition and Wikipedia has a decent
introduction with references) is that a copula is a dependence
structure among several different U(0,1) marginal distributions --
you can then transform those marginal distributions to whatever
distribution you like (e.g. if x ~ U(0,1) then -log(x) ~ Exp(1)).
It may not be easy to specify precisely what correlation you want.

~  So the bottom line/bad news is that I can't give you a canned
solution -- I just think this a good direction to look in, if
this is an important part of your problem.

~  Ben Bolker
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFIkbjvc5UpGjwzenMRAuIDAJ9sECxDVr0PUSxxv+36ZUKB+l2gvgCeJdzn
vh9GmLhJTvlXZ30qqk8vn00=
=eB7r
-----END PGP SIGNATURE-----