pyphi.measures.distribution#

Metrics on probability distributions.

Functions

absolute_information_density(p, q)

Return the absolute information density function of two distributions.

absolute_intrinsic_difference(p, q)

Compute the absolute intrinsic difference (AID) between two distributions.

absolute_pointwise_mutual_information(p, q, ...)

Computes the state-specific absolute pointwise mutual information between two distributions.

approximate_specified_state(repertoire, ...)

Estimate the purview state that maximizes the AID between the repertoires.

effect_emd(p, q)

Compute the EMD between two effect repertoires.

emd(p, q[, direction])

Compute the EMD between two repertoires for a given direction.

entropy_difference(p, q)

Return the difference in entropy between two distributions.

generalized_intrinsic_difference(...[, state])

Generalized intrinsic difference (GID) between two repertoires.

hamming_emd(p, q)

Return the Earth Mover's Distance between two distributions (indexed by state, one dimension per node) using the Hamming distance between states as the transportation cost function.

iit_4_small_phi(p, q, state)

Absolute intrinsic difference between two distributions at one state.

iit_4_small_phi_no_absolute_value(p, q, state)

Signed intrinsic difference between two distributions at one state.

information_density(p, q)

Return the information density of p relative to q, in base 2.

intrinsic_difference(p, q)

Compute the intrinsic difference (ID) between two distributions.

intrinsic_differentiation(p, state)

Smallest positive surprisal in a repertoire slice.

intrinsic_information(forward_repertoire, ...)

Intrinsic information (ii) between two repertoires.

intrinsic_specification(forward_repertoire, ...)

Generalized intrinsic difference (GID) between two repertoires.

kld(p, q)

Return the Kullback-Leibler Divergence (KLD) between two distributions.

l1(p, q)

Return the L1 distance between two distributions.

mp2q(p, q)

Compute the MP2Q measure.

pointwise_intrinsic_differentiation(p)

Pointwise surprisal \(-\log_2 p_i\) of a distribution.

pointwise_mutual_information(p, q)

Compute the pointwise mutual information (PMI).

pointwise_mutual_information_vector(p, q)

Elementwise pointwise mutual information \(\log_2(p_i / q_i)\).

psq2(p, q)

Compute the PSQ2 measure.

repertoire_distance(r1, r2[, direction, ...])

Compute the distance between two repertoires for the given direction.

resolve_actual_causation_measure(name)

Look up a measure registered in actual_causation_measures.

resolve_distribution_measure(name)

Look up a distribution measure for EMD ground distance and IIT-side dispatch.

resolve_mechanism_measure(name[, alphabet_sizes])

Look up a measure usable at the mechanism level.

resolve_system_measure(name)

Look up a measure usable at the system level.

weighted_pointwise_mutual_information(p, q)

Compute the weighted pointwise mutual information (WPMI).

Classes

ActualCausationMeasureRegistry()

Storage for distance functions used in pyphi.actual.

CompositeMeasureRegistry()

Storage for composite measures of shape (forward, partitioned, selectivity, *, state) -> DistanceResult.

DistanceResult(value, **kwargs)

A float that carries auxiliary data about its computation.

DistributionMeasureRegistry()

Storage for (p, q) -> float | DistanceResult distance functions.

OptionalEMD()

Class to handle EMD computations via POT (Python Optimal Transport).

StateAwareMeasureRegistry()

Storage for (p, state) -> float | DistanceResult measures.

StatefulDistributionMeasureRegistry()

Storage for (p, q, state) -> float | DistanceResult measures.

np_suppress()

Decorator to suppress NumPy warnings about divide-by-zero and multiplication of NaN.