Marginal probability and posterior probability

The front end.
Post Reply
wantida
Posts: 6
Joined: Wed Sep 06, 2017 2:13 am

Marginal probability and posterior probability

Post by wantida »

Hello,
I used GeNIe academic version, these are details:
Version 2.2.2007.0 (32-bit Academic)
Built on 2017-08-08 (1d30fba15537100)

I am a new user for both Bayesian network and GeNIe, so I confuse about the terms: marginal probability distribution, posterior probability, and posterior marginal probability distribution. Is it the same meaning? I understand that posterior probability is conditional probability when we provide an evidence/observation. I get stuck with my thesis when I try to describe the result of probability in value tab and I am not sure what term should I use.

Another thing is the marginal probability that provides in value tab, is the same calculate with posterior probability? For example, if I have P(A|B) for posterior probability, the result of marginal probability that provides in value tab is summarizing of P(A) or P(A|B). If it different how do I get posterior probability when I provide an evidence?

Please help me to clarify my confusion and anything that I might misunderstand.
Thank you,
Wantida
marek [BayesFusion]
Site Admin
Posts: 430
Joined: Tue Dec 11, 2007 4:24 pm

Re: Marginal probability and posterior probability

Post by marek [BayesFusion] »

Hi Wantida,

To make things, simple, the distributions in the Value tab are always marginal. Sometimes they are prior and sometimes they are posterior. The term posterior (or its synonym, a Latin term "a-posteriori") relates to the fact that it has been calculated based on some observations/evidence. In practice, you usually observe something and the Bayesian network calculates the new/updated marginals. They are posterior marginals or P(A|B) using your example. I hope this helps.

Marek
Post Reply