mbi.estimation.mle_from_marginals

mbi.estimation.mle_from_marginals(marginals: CliqueVector, known_total: float, iters: int = 250, marginal_oracle: MarginalOracle = <PjitFunction of <function message_passing_stable>>, callback_fn=<function <lambda>>, mesh: Mesh | None = None) → MarkovRandomField[source]

Compute the MLE Graphical Model from the marginals.

Parameters:

  • marginals – The marginal probabilities.

  • known_total – The known or estimated number of records in the data.

Returns:

A MarkovRandomField object with the final potentials and marginals.