MSS Common Information

The Minimal Sufficient Statistic Common Information is the entropy of the join of the minimal sufficient statistic of each variable about the others:

\[\M{X_{0:n}} = \H{ \join_i \left(X_i \mss X_{\overline{\{i\}}}\right) }\]

The distribution that the MSS common information is the entroy of is also known “information trim” of the original distribution, and is accessable via dit.algorithms.minimal_sufficient_statistic.info_trim().

API

mss_common_information(dist, rvs=None, crvs=None)[source]

Compute the minimal sufficient statistic common information, which is the entropy of the join of the minimal sufficent statistic of each variable about the others.

Parameters:

dist (Distribution) – The distribution for which the joint minimal sufficient statistic is computed.
rvs (list, None) – The random variables to compute the joint minimal sufficient statistic of. If None, all random variables are used.
crvs (list, None) – The random variables to condition the joint minimal sufficient statistic on. If None, then no random variables are conditioned on.