Multinomial and Poisson Distributions (4c)

Video Lecture from the course INST 414: Advanced Data Science at UMD's iSchool.
(I stupidly wore a greenish shirt in front of a green screen.)
Full course information here:
www.umiacs.umd....

[@daneshj4013]

Hope Lee makes an appearance on the actual Bake Off

[@NB-lx2fp]

Sir please explain what is standard distribution?

[@danieldeychakiwsky1928]

Is it safe to say that the multinomial distribution is a multivariate distribution?

[@JordanBoydGraber]

Yes, otherwise you'd probably better call it a Bernoulli distribution.

[@JordanBoydGraber]

Yes, otherwise it would be a Bernoulli distribution.

[@danieldeychakiwsky1928]

@@JordanBoydGraber
I'm looking at this link: docs.scipy.org/doc/numpy-1.15.0/reference/generated/numpy.random.multinomial.html
A quick and dirty analogy:
Flipping a coin one time = Bernoulli
Flipping a coin more than one time = Binomial
Rolling a die one time = Categorical
Rolling a die more than one time = Multinomial
How about the case of flipping five separate coins at the same time but only once each w/ p(heads) = 0.8. Then you have a Multivariate Bernoulli right (a joint distribution of 5 separate distributions) and each of your variables is independent.
What confuses me is how we can say the Multinomial is a Multivariate when Multinomial assumes mutual exclusivity of its categories, e.g., in the case of a die, if the probability of the outcome 2 is 1.0 then that means it will always land on 2 and nothing else. A Multivariate distribution doesn't have this constraint. I fear that my confusion sprouts from terminology? Thanks in advance!