In Bayesian statistics, the posterior probability is the conditional probability of an event, sometimes called a parameter, given evidence or an observation of that event, the data.
Consider an event A and evidence of that event, another event itself, B. The posterior probability is P(A|B). It is called posterior, because event A actually occurs before event B.
In contrast, the likelihood function is the conditional probability of the evidence or observation, B, given the event A, P(B|A).
The posterior probability and the likelihood function are related through Bayes’ Theorem:
« Back to Glossary Index