pm4py.algo.evaluation.earth_mover_distance.variants.pyemd module#

class pm4py.algo.evaluation.earth_mover_distance.variants.pyemd.Parameters[source]#

Bases: object

STRING_DISTANCE = 'string_distance'#
USE_FAST_EMD = 'use_fast_emd'#
pm4py.algo.evaluation.earth_mover_distance.variants.pyemd.normalized_levensthein(s1, s2)[source]#
pm4py.algo.evaluation.earth_mover_distance.variants.pyemd.get_act_correspondence(activities, parameters=None)[source]#
pm4py.algo.evaluation.earth_mover_distance.variants.pyemd.encode_two_languages(lang1, lang2, parameters=None)[source]#
class pm4py.algo.evaluation.earth_mover_distance.variants.pyemd.EMDCalculator[source]#

Bases: object

A class that provides an EMD (Earth Mover’s Distance) computation similar to what pyemd offers. It uses linear programming via scipy.optimize.linprog to solve the underlying flow problem.

Usage:#

emd_value = EMDCalculator.emd(first_histogram, second_histogram, distance_matrix)

static emd(first_histogram: ndarray, second_histogram: ndarray, distance_matrix: ndarray) → float[source]#

Compute the Earth Mover’s Distance given two histograms and a distance matrix.

Parameters:

  • first_histogram (np.ndarray) – The first distribution (array of nonnegative numbers).

  • second_histogram (np.ndarray) – The second distribution (array of nonnegative numbers).

  • distance_matrix (np.ndarray) – Matrix of distances between points of the two distributions.

Returns:

The computed EMD value.

Return type:

float

class pm4py.algo.evaluation.earth_mover_distance.variants.pyemd.POTEMDCalculator[source]#

Bases: object

A faster implementation of EMD using the POT (Python Optimal Transport) library. Falls back to the SciPy implementation if POT is not available.

Usage:#

emd_value = POTEMDCalculator.emd(first_histogram, second_histogram, distance_matrix)

static is_pot_available()[source]#

Check if POT (Python Optimal Transport) is available.

static emd(first_histogram: ndarray, second_histogram: ndarray, distance_matrix: ndarray) → float[source]#

Compute the Earth Mover’s Distance using POT if available, otherwise fall back to SciPy.

Parameters:

  • first_histogram (np.ndarray) – The first distribution (array of nonnegative numbers).

  • second_histogram (np.ndarray) – The second distribution (array of nonnegative numbers).

  • distance_matrix (np.ndarray) – Matrix of distances between points of the two distributions.

Returns:

The computed EMD value.

Return type:

float

pm4py.algo.evaluation.earth_mover_distance.variants.pyemd.apply(lang1: Dict[List[str], float], lang2: Dict[List[str], float], parameters: Dict[str | Parameters, Any] | None = None) → float[source]#