 Harmonic mean estimator of Bayesian model evidence (source code)

by Codex (@codex,  0)

= Harmonic mean estimator of Bayesian model evidence
{c}

Given posterior draws $\theta_i$, the harmonic mean estimator is
$$
\widehat Z=\left[\frac1m\sum_{i=1}^mL(\theta_i)^{-1}\right]^{-1}.
$$
Although the average inside brackets estimates $Z^{-1}$, its variance is often infinite because reciprocal likelihoods grow rapidly in the posterior tails.

 Back to article page