easyvvuq.comparison package

Submodules

easyvvuq.comparison.base module

Provides base class for all comparison/validation elements.

class easyvvuq.comparison.base.BaseComparisonElement[source]

Bases: BaseElement

Baseclass for all EasyVVUQ comparison elements.

compare(dataframe1, dataframe2)[source]
element_category()[source]

easyvvuq.comparison.validate module

Validation by comparing QoI distributions.

class easyvvuq.comparison.validate.ValidateSimilarity[source]

Bases: BaseComparisonElement

compare(dataframe1, dataframe2)[source]

Perform comparison between two lists or arrays of discrete distributions.

Parameters:

  • dataframe1 (NumPy array or list)

  • dataframe2 (NumPy array or list)

Returns:

  • A list of distances between two lists of discrete distributions,

  • dataframe1 and dataframe2.

dist(p, q)[source]
class easyvvuq.comparison.validate.ValidateSimilarityHellinger[source]

Bases: ValidateSimilarity

dist(p, q)[source]

Compute Hellinger distance between two discrete probability distributions (PDF). The Hellinger distance metric gives an output in the range [0,1] with values closer to 0 meaning the PDFs are more similar.

Parameters:

  • p (NumPy array)

  • q (NumPy array)

Returns:

  • Hellinger distance between distributions p and q.

  • https (//en.wikipedia.org/wiki/Hellinger_distance)

element_name()[source]
element_version()[source]
class easyvvuq.comparison.validate.ValidateSimilarityJensenShannon[source]

Bases: ValidateSimilarity

dist(p, q)[source]

Compute Jensen-Shannon distance between two discrete probability distributions (PDF). It is based on Kullback–Leibler divergence and gives an output metric un the range [0,1] with values closer to 0 meaning the PDFs are more similar.

Parameters:

  • p (NumPy array)

  • q (NumPy array)

Returns:

  • Jensen-Shannon divergence between distributions p and q.

  • https (//en.wikipedia.org/wiki/Jensen%E2%80%93Shannon_divergence)

  • https (//en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence)

element_name()[source]
element_version()[source]
class easyvvuq.comparison.validate.ValidateSimilarityWasserstein[source]

Bases: ValidateSimilarity

dist(p, q)[source]

Compute Wasserstein distance between two discrete cumulative distributions (CDF). The Wasserstein distance has an unrestricted range with a lower limit of 0. A smaller distance indicates a stronger similarity between between CFDs.

Parameters:

  • p (NumPy array)

  • q (NumPy array)

Returns:

  • Wasserstein distance between distributions p and q.

  • https (//en.wikipedia.org/wiki/Wasserstein_metric)

element_name()[source]
element_version()[source]

Module contents