multinomial distribution
Let be a random vector such that
- 1.
and
- 2.
, where is a fixed integer
Then X has a multinomial distribution if there exists a parameter vector such that
- 1.
and
- 2.
- 3.
X has a discrete probability distribution function in the form:
Remarks
- •
- •
, where
- •
When , the multinomial distribution is the same as the binomial distribution
- •
If are mutually independent Poisson random variables with parameters respectively, then the conditional
joint distribution
of given that ismultinomial with parameters , where .
Sketch of proof.Each is distributed as:
The mutual independence of the ’s shows that the joint probability distribution of the ’s is given by
where , and .Next, let . Then is Poisson distributed with parameter (which can be shown by using induction
and the mutual independence of the ’s):
The conditional probability distribution of X given that is thus given by:
where and that .