Declarations

Inside a predictor, s() declares a smooth of one numeric column, cat() a categorical effect with a reference level, re() a random effect with a coefficient for every level, term() attaches any other feature specification to a column, and ti() and interaction() declare interactions between terms already in that predictor. Those helpers return BoundTerm and BoundInteraction; the family helper wraps them in a BoundPredictor, which bind_predictor() also creates by parameter name for custom families, and Predictor is the immutable configuration underneath; the bound terms and Predictor are described on the Internals page. The families and their helper methods are on the Families page.

superglm.BoundPredictor	A predictor declaration tied to the family instance that created it.
superglm.bind_predictor	Declare a predictor by parameter name, including for custom families.
superglm.term	Attach an existing feature specification to an input column.
superglm.s	Describe a smooth effect of one numeric column.
superglm.cat	Declare a categorical effect with a reference level.
superglm.re	Declare a random effect with a coefficient for every group level.
superglm.ti	Declare a tensor interaction between two smooth effects.
superglm.interaction	Use an existing interaction specification inside a predictor.