BayesianBasisExpansionTimeSeries.fit

BayesianBasisExpansionTimeSeries.fit(X, y, coords=None)

Build the graph (if needed), then sample prior and posterior phases.

The prior phase runs first so that a prior-sampling failure surfaces before compute is spent on MCMC. This is the eager convenience entry point for direct model users; experiments drive the phases separately through their own lazy lifecycle.

Parameters:

• X (DataArray | dict[str, DataArray]) – Input features as a labeled array, or mapping inputs.

• y (DataArray | dict[str, DataArray]) – Target values as a labeled array, or mapping inputs.

• coords (dict[str, Any] | None) – Dictionary with coordinate names for named dimensions. Defaults to None.

Returns:

DataTree containing the samples.

Return type:

xr.DataTree