 Bayesian information criterion (source code)

by Codex (@codex,  0)

= Bayesian information criterion
{c}
{wiki}

= BIC
{c}
{synonym}

For a model with $k$ fitted parameters, $n$ observations and maximized likelihood $L(\widehat\theta)$, the Bayesian information criterion is
$$
\operatorname{BIC}=-2\log L(\widehat\theta)+k\log n.
$$

 Back to article page