ModelAdapter#

class causalpy.experiments.model_adapter.ModelAdapter[source]#

Experiment-agnostic wrapper around a CausalPy statistical backend.

Methods

`ModelAdapter.build`(X, y, *[, coords])	Construct the backend's graph/design state without sampling.
`ModelAdapter.coefficients`(*[, group])	Return model coefficients with canonical coefficient dimensions.
`ModelAdapter.fit`(X, y, *[, coords])	Fit the model with backend-appropriate conventions.
`ModelAdapter.predict`(X, *[, coords, ...])	Return expected outcomes with canonical prediction dimensions.
`ModelAdapter.print_coefficients`(labels[, ...])	Print model coefficients with labels.
`ModelAdapter.require_idata`()	Return fitted inference-result DataTree or raise an explicit capability error.
`ModelAdapter.sample_posterior`(**kwargs)	Sample the posterior phase with backend-appropriate conventions.
`ModelAdapter.sample_prior_predictive`(**kwargs)	Sample the prior predictive phase.
`ModelAdapter.score`(X, y, *[, coords])	Return per-unit \(R^2\) scores in the canonical container.

Attributes

`has_posterior`	Whether posterior draws are available on this backend.
`has_prior`	Whether prior draws are available on this backend.
`idata`	Return a fitted inference-result DataTree when supported, otherwise `None`.
`is_bayesian`	Whether the backend is Bayesian (PyMC or pymc-forecast).
`is_built`	Whether the backend's model graph / design state is constructed.
`is_ols`	Whether the backend is OLS/sklearn.
`kind`	Backend identifier.
`model`	The underlying model instance.
`supports_idata`	Whether the backend exposes an inference-result DataTree.
`supports_prior_predictive`	Whether this backend can sample a prior predictive phase.