Bayesian Inference with Conjugate Priors

Watch the posterior update as data arrives

Model

n: 0   successes: 0
MAP:
95% CI:

Conjugate Priors

A conjugate prior is a prior distribution that, when combined with a given likelihood, produces a posterior in the same family. This enables closed-form Bayesian updating.

Beta-Binomial: The Beta(α,β) prior is conjugate to the Binomial likelihood.

Prior: Beta(α, β)   Likelihood: Bin(n, θ)
Posterior: Beta(α + k, β + n − k)

Gamma-Poisson: Gamma(α,β) prior conjugate to Poisson(λ).

Posterior: Gamma(α + Σxᵢ, β + n)

As more data arrives, the posterior narrows and shifts toward the true parameter value. The prior's influence shrinks — it contributes "pseudo-counts" α and β that become negligible for large n.

The prior (blue), likelihood (orange), and posterior (green) are shown. Click observations to update!