Module fetish_lib::model

Structs

Model

A representation of the information known via Bayesian regression of a nonlinear model from the compressed space of the given argument TypeId to the compressed space of the given return TypeId. The wrapped NormalInverseWishart linear model is defined on mappings from the feature space of the input to the compressed space of the output, and the whole Model is assumed to exist in the given Context.

Functions

to_features	Convenience method to get all ordered features for a given collection of FeatureCollections and a given input vector.
to_features_mat	Convenience method to get all features for a given collection of FeatureCollections and a given input matrix whose rows are vectors to featurize.
to_jacobian	Convenience method to get the jacobian of the concatenation of the given collection of FeatureCollections evaluated at the given input vector.