pyphi.measures.distribution.iit_4_small_phi_no_absolute_value#

pyphi.measures.distribution.iit_4_small_phi_no_absolute_value(p, q, state)[source]#

Signed intrinsic difference between two distributions at one state.

Like iit_4_small_phi(), but without the absolute value: returns the signed pointwise information density \(p_i \log_2 (p_i / q_i)\) (see information_density()) evaluated at state. The result may be negative.

Parameters:

  • p (np.ndarray) – The first probability distribution.

  • q (np.ndarray) – The second probability distribution.

  • state (State) – Index into the squeezed pointwise array selecting the value returned.

Returns:

DistanceResult – The signed information density at state, with method="IIT_4.0_SMALL_PHI_NO_ABSOLUTE_VALUE".

Return type:

float