InstrumentalVariable.fit

InstrumentalVariable.fit(**kwargs)

Build the IV model graph and sample the posterior phase.

Returns:

The same experiment, for chaining.

Return type:

Self

Parameters:

**kwargs (Any) – Sampler overrides forwarded to sample_posterior() (e.g. draws=500), and ppc_sampler="jax" | "pymc" | None selecting the posterior-predictive backend at fit() time. Omitting ppc_sampler on a refit keeps the previous choice, so predictive groups are never left stale against a resampled posterior. The IV backend exposes no prior predictive phase, so sample_prior_predictive() raises PriorPredictiveNotSupportedException.