DressRehearsalCheck#

class causalpy.checks.dress_rehearsal.DressRehearsalCheck[source]#

Pipeline-compatible dress rehearsal check for Synthetic Control.

Calls SyntheticControl.validate_design() and wraps the result as a CheckResult.

Parameters:

  • injected_effect (float) – Effect to inject (see validate_design).

  • holdout_periods (int | None) – Pseudo-post window length.

  • effect_type (Literal['relative', 'absolute']) – How the injected effect is applied.

  • sample_kwargs (dict | None) – MCMC sampling arguments for the refitted model.

Methods

DressRehearsalCheck.__init__([...])

DressRehearsalCheck.run(experiment, context)

Run the dress rehearsal and return a CheckResult.

DressRehearsalCheck.validate(experiment)

Verify the experiment is a SyntheticControl with a PyMC model.

Attributes

applicable_methods

__init__(injected_effect=0.1, holdout_periods=None, effect_type='relative', sample_kwargs=None)[source]#
Parameters:

  • injected_effect (float)

  • holdout_periods (int | None)

  • effect_type (Literal['relative', 'absolute'])

  • sample_kwargs (dict | None)

Return type:

None

classmethod __new__(*args, **kwargs)#