Skip to content

Ridge

class Ridge(M:Expression<Real[_,_]>, Σ:Expression<Real[_,_]>, α:Expression<Real>, β:Expression<Real[_]>) < Distribution<(Real[_,_], Real[_])>

Opaque distribution over parameters for Bayesian ridge regression. See Regression for synopsis.

Internally, it keeps parameters for a ridge regression with multiple outputs. Only a single covariance parameter is necessary despite multiple outputs, providing significant computational gains over multiple ridge regressions with a single output.

Factory Functions

Name Description
Ridge Create ridge prior.
Ridge Create ridge prior.

Member Variables

Name Description
M:Expression<Real[_,_]> Weight means. Each column gives the mean of a weight vector.
Σ:Expression<Real[_,_]> Weight covariance scale.
α:Expression<Real> Common weight and likelihood covariance shape.
β:Expression<Real[_]> Weight and likelihood covariance scales.

Factory Function Details

function Ridge(M:Expression<Real[_,_]>, Σ:Expression<Real[_,_]>, α:Expression<Real>, β:Expression<Real[_]>) -> Ridge

Create ridge prior.

function Ridge(M:Real[_,_], Σ:Real[_,_], α:Real, β:Real[_]) -> Ridge

Create ridge prior.