posterior_bounds

SuperLSS.posterior_bounds(

X: object | EagerFrame,

quantity: str | tuple[Any, ...] | Callable[[NDArray], NDArray],

*,

sample_weight: NDArray | None = None,

**kwargs: Any,

) → DataFrame | tuple[DataFrame, NDArray[float64]]

Return per-row bounds on a quantity derived from the fitted parameters.

quantity names what to push the coefficient draws through – a parameter, a predictive quantile, an exceedance probability, an expected shortfall, or a callable of the parameter matrix – and the frame reports the plug-in estimate beside the posterior mean, standard deviation and interval. sample_weight is forwarded as the row’s prior weight, since a quantile of a row at a fifth of a year’s exposure is the quantile of its own law; under frequency semantics there is no such slot and the call is refused.