saealib.MaternKernel#

class saealib.MaternKernel(*, length_scale=None, nu=1.5)[source]#

Bases: RBFKernel

Matérn radial basis function kernel.

Strictly positive definite; does not require a polynomial term. Based on [18].

Parameters:

  • length_scale (float or None) – Length-scale parameter (denoted ell in the paper). None resolves to the median pairwise training-point distance at fit time.

  • nu (float) – Smoothness parameter. Must be 0.5, 1.5, or 2.5.

Methods

__init__

evaluate

Evaluate the Matérn kernel's nu-dependent radial profile.

pairwise

Evaluate the kernel between every row of x1 and x2.

resolve

Resolve length_scale from training data if unset.

validate

Validate length_scale and nu.

Method Details

MaternKernel.__init__(*, length_scale=None, nu=1.5)#
Parameters:

  • length_scale (float | None)

  • nu (float)

Return type:

None

MaternKernel.evaluate(r)[source]#

Evaluate the Matérn kernel’s nu-dependent radial profile.

Parameters:

r (ndarray)

Return type:

ndarray

MaternKernel.pairwise(x1, x2)#

Evaluate the kernel between every row of x1 and x2.

Parameters:

  • x1 (ndarray)

  • x2 (ndarray)

Return type:

ndarray

MaternKernel.resolve(train_x)[source]#

Resolve length_scale from training data if unset.

Parameters:

train_x (ndarray)

Return type:

Self

MaternKernel.validate()[source]#

Validate length_scale and nu.

Return type:

None