fairlens.metrics.permutation_statistic#

permutation_statistic(x, y, statistic, n_perm=100)[source]#

Performs the sampling for a two sample permutation test.

Parameters
  • x (pd.Series) – First data sample.

  • y (pd.Series) – Second data sample.

  • statistic (Callable[[pd.Series, pd.Series], float]) – Function that computes the test statistic.

  • n_perm (int) – Number of permutations.

Returns

The distribution of the statistic on a n_perm permutations of samples.

Return type

np.ndarray