Multinomial distribution with negative binomial prior on the number of trials
- From: Alex Sverdlov <alsverdlov@xxxxxxxxx>
- Date: Tue, 31 Jul 2007 16:31:39 -0000
Let (X0,X1,...,Xk) given n>0 ~ Multinomial (n, p0,p1,...,pk), where
pi's sum to 1.
Let n ~ Negative Binomial (r,p), r>0, integer
What is the unconditional distribution of (X1,...,Xk) if it is known
that X0=r?
I tried to multiply the multinomial and negative binomial p.m.f's and
'integrate out' n, but this does not seem to be straightforward. Is
there any trick?
.
- Prev by Date: Re: t test for outlier detection
- Next by Date: Significance test for mean correlation
- Previous by thread: t test for outlier detection
- Next by thread: Significance test for mean correlation
- Index(es):