PredictiveCheckTest.pvalue#

PredictiveCheckTest.pvalue(statistic, name=None)[source]#

One statistic’s tail probabilities, for every variable or for one.

Parameters:

  • statistic (str) – A name from statistics.

  • name (str | None) – A variable label from names; None returns the whole (k,) row.

Return type:

NDArray[double] | float

Returns:

The (k,) row of pvalues for the statistic, or the single float for the named variable.

Raises:

SpecificationError – If the statistic or the variable is unknown. The message lists what the check does carry.

Example

>>> import numpy as np
>>> rng = np.random.default_rng(1)
>>> y = rng.standard_normal((100, 2))
>>> check = replication_check(y, rng.standard_normal((200, 100, 2)))
>>> check.pvalue("mean").shape
(2,)
>>> 0.0 <= check.pvalue("mean", "y1") <= 1.0
True
>>> check.pvalue("median")
Traceback (most recent call last):
    ...
cultivars.exceptions.SpecificationError: unknown statistic 'median'; ...