Posterior probability
Updated belief after considering evidence via Bayes' rule.
Posterior probability is a conditional probability that arises from updating a prior probability using information from a likelihood function, as formalized by Bayes' rule. From an epistemological standpoint, it represents the complete state of knowledge about an uncertain proposition—such as a scientific hypothesis or a parameter value—given prior knowledge and a mathematical model of the observations available at a given time. When new data arrive, the current posterior probability can serve as the prior for a subsequent round of Bayesian updating. In Bayesian statistics, the posterior probability distribution typically describes epistemic uncertainty about statistical parameters conditional on a set of observed data. From this distribution, point and interval estimates can be derived, including the maximum a posteriori (MAP) estimate and the highest posterior density interval (HPDI). However, the posterior distribution is often not tractable and must be approximated either analytically or numerically. The posterior probability is defined as the probability distribution of the parameters given the evidence, denoted as p(θ|X). It contrasts with the likelihood function, which is the probability of the evidence given the parameters, p(X|θ). The relationship is given by Bayes’ theorem: the posterior is proportional to the product of the likelihood and the prior, divided by a normalizing constant (the marginal likelihood). For continuous parameters, this constant is an integral over all possible values; for discrete parameters, it is a sum. A classic example involves a school with 60% boys and 40% girls; all boys wear trousers, while half the girls wear trousers. Given an observation of a trouser-wearing student, the posterior probability that the student is a girl is 25%. The posterior probability is a random variable conditioned on observed data, and its uncertainty can be summarized using a credible interval. In classification, posterior probabilities reflect the uncertainty of assigning an observation to a particular class, though machine learning methods often supply membership values that do not directly correspond to probabilistic confidence.
- definition
- Conditional probability of parameters θ given evidence X, denoted p(θ|X)
- formula
- p(θ|x) = [p(x|θ) / p(x)] × p(θ)
- components
- Prior probability p(θ), likelihood p(x|θ), normalizing constant p(x)
- normalizing_constant
- p(x) = ∫ p(x|θ) p(θ) dθ for continuous θ, or sum for discrete θ
- key_relation
- Posterior probability is proportional to Likelihood × Prior probability
- example_result
- In a school with 60% boys and 40% girls, probability a trouser-wearing student is a girl is 25%
Lore & Background
Posterior probability is a type of conditional probability that results from updating the prior probability with information summarized by the likelihood via an application of Bayes' rule. From an epistemological perspective, the posterior probability contains everything there is to know about an uncertain proposition, such as a scientific hypothesis or parameter values, given prior knowledge and a mathematical model describing the observations available at a particular time. After the arrival of new information, the current posterior probability may serve as the prior in another round of Bayesian updating.
Reader's Guide
In Bayesian statistics, the posterior probability distribution describes epistemic uncertainty about statistical parameters conditional on observed data. From a given posterior distribution, various point and interval estimates can be derived, such as the maximum a posteriori (MAP) or the highest posterior density interval (HPDI). However, while conceptually simple, the posterior distribution is generally not tractable and therefore needs to be either analytically or numerically approximated. The calculation involves multiplying the prior probability distribution by the likelihood function, then dividing by the normalizing constant. The example of a school with 60% boys and 40% girls illustrates the concept: given that all boys wear trousers and half of girls do, the posterior probability that a trouser-wearing student is a girl is 25%. This demonstrates how Bayes' theorem updates initial beliefs based on observed evidence.
Did You Know?
- Posterior probability is defined as p(θ|x) = [p(x|θ) / p(x)] × p(θ).
- The normalizing constant p(x) is calculated by integrating or summing p(x|θ)p(θ) over all possible θ.
- In the school example, the posterior probability that a trouser-wearing student is a girl is 25%.
- The posterior distribution is often not tractable and requires analytical or numerical approximation.
More in Probability & Statistics 1-24
Spotted an error? Know more?
This is a living reference — every entry is fact-audited, and reader corrections feed straight into our audit queue. Suggest an edit · See this site's audit record