Interfaces for Hyper-Parameter Sets

class IProvideHyperparamsForRobustCIT

Interface specifying how to supply (customized) hyper-parameters for robust dependence testing.

class Hyperparams

Hyperparams(B: int)

B: int: block-size

__init__(B: int) → None

hyperparams_for_robust_cit(N: int, dim_Z: int) → Hyperparams

Supply hyper-parameters for robust CI testing for the given setup.

Parameters:

N (int) – sample-size
dim_Z (int) – size of the conditioning set

Returns:

Hyper-parameters to use.

Return type:

IProvideHyperparamsForRobustCIT.Hyperparams

static from_binomial_homogeneity_test(homogeneity_test_hyperparams: IProvideHyperparamsForBinomial) → IProvideHyperparamsForRobustCIT

It is runtime-efficient and simple to obtain hyper-parameters for robust CI testing compatible to those used by homogeneity-testing.

Parameters:: homogeneity_test (IProvideHyperparamsForBinomial) – homogeneity-test hyper-parameters to copy
Returns:: Hyper-parameter Provider.
Return type:: IProvideHyperparamsForRobustCIT

class IProvideHyperparamsForBinomial

Interface specifying how to supply (customized) hyper-parameters for binomial homogeneity testing.

class Hyperparams

Hyperparams(B: int, alpha: float, beta: float, max_acceptable_count: float)

B: int: block-size

alpha: float: error-control target \(\alpha\)

beta: float: binomial quantile

max_acceptable_count: float: by numerical precision, pvalue at max acceptable count may be within \(\alpha +\) tolerance (by default \(10^{-5}\))

__init__(B: int, alpha: float, beta: float, max_acceptable_count: float) → None

hyperparams_for_binomial(N: int, dim_Z: int) → Hyperparams

Supply hyper-parameters for the binomial homogeneity-test for the given setup.

Parameters:

N (int) – sample-size
dim_Z (int) – size of conditioning set

Returns:

The hyper-parameters to use.

Return type:

IProvideHyperparamsForBinomial.Hyperparams

class IProvideHyperparamsForAcceptanceInterval

Interface specifying how to supply (customized) hyper-parameters for acceptance-interval testing.

class Hyperparams

Hyperparams(B: int, alpha: float, cutoff: float)

B: int: block-size

alpha: float: error-control target

cutoff: float: cutoff

__init__(B: int, alpha: float, cutoff: float) → None

hyperparams_for_acceptance_interval(N: int, dim_Z: int) → Hyperparams

Supply hyper-parameters for acceptance-interval tests for given setup.

Parameters:

N (int) – sample-size
dim_Z (int) – size of conditioning set

Returns:

The hyper-parameters to use.

Return type:

IProvideHyperparamsForAcceptanceInterval.Hyperparams