MarginalLikelihoodSelection.compare#

MarginalLikelihoodSelection.compare(*others, names=None, prior_probabilities=None)[source]#

Rank models by evidence with posterior model probabilities.

Parameters:

  • *others (_MarginalLikelihoodSelection) – Further records on the same sample.

  • names (Sequence[str] | None) – Row labels; default each record’s source.

  • prior_probabilities (Sequence[float] | None) – Prior model probabilities, one per record; default equal.

Return type:

SummaryTable

Returns:

One row per model, best first – log marginal likelihood, its Monte Carlo error, the log Bayes factor against the best, the Kass-Raftery strength of that evidence, and the posterior model probability.

Raises:

  • DimensionError – If names or prior_probabilities do not match the number of records.

  • SpecificationError – If a prior probability is negative or all are zero.