predict

SuperLSS.predict(

X: object | EagerFrame,

*,

offsets: Mapping[str, NDArray] | None = None,

) → NDArray[float64]

Return the conditional mean for each row with a built-in family.

Predictions use the response scale supplied to fitting. For example, LogNormalLS returns a mean on the original positive response scale. A Gaussian model fitted to a manually logged response returns a mean of that logged response.

Custom families define their own default prediction quantity. Use predict_parameters for the complete parameter table and predict_quantile for a chosen conditional quantile.