Page 1 of 1
Independent normal distributions
Posted: October 16th, 2008, 8:42 am
by Escher
If two random variables Z1 and Z2 have a joint normal distribution and have covariance zero, then how do you show that they're independent?I would guess it's quite easy, and ordinarily I'd just go look it up, but I don't have access to the library yet and google doesn't show up much (aside from a wikipedia article stating the fact but giving no proof).Cheers.
Independent normal distributions
Posted: October 16th, 2008, 12:13 pm
by pizza
No, the OP is correct, for normal multivariate distributions the converse *also* holds. (Sorry, no proof)
Independent normal distributions
Posted: October 16th, 2008, 3:35 pm
by Escher
No problem, as it happens I found the proof in a book on my desk (Vol II of Shreve's finance text).It follows from the fact that the multivariate density function factorises into seperate density functions when the correlation is zero.