Skip to content

Prior, Likelihood, and Posterior

Overview

Bayesian inference combines prior beliefs with observed data through Bayes' theorem:

\[ \pi(\theta \mid \mathbf{x}) = \frac{f(\mathbf{x} \mid \theta) \pi(\theta)}{f(\mathbf{x})} \propto f(\mathbf{x} \mid \theta) \pi(\theta) \]

Components

  • Prior \(\pi(\theta)\): encodes beliefs about \(\theta\) before seeing data
  • Likelihood \(f(\mathbf{x} \mid \theta)\): probability of the data given the parameter
  • Posterior \(\pi(\theta \mid \mathbf{x})\): updated beliefs after seeing data
  • Marginal likelihood \(f(\mathbf{x})\): normalizing constant

Bayesian Point Estimates

Estimate Definition
Posterior mean \(E[\theta \mid \mathbf{x}]\)
Posterior median Median of \(\pi(\theta \mid \mathbf{x})\)
MAP Mode of \(\pi(\theta \mid \mathbf{x})\)