Enum MultinomialPosteriorRule 

pub enum MultinomialPosteriorRule {
    Exact,
    ConditionedThreeClass(usize),
    AnisotropicTensor(Vec<usize>),
    IsotropicSparse(usize),
}

Which deterministic rule produced a certified answer.

The rank of the retained posterior decides which rule can certify inside the evaluation budget, and the two are not interchangeable: a sparse grid is a saving only when the resolution a direction needs is much smaller than the number of directions, and at rank two it is strictly worse than the tensor product it is built from (#2612).

Variants

Exact

Exact reduction: the binary logistic-normal evaluator, or a covariance that is a point mass.

ConditionedThreeClass(usize)

Exact Gaussian conditioning reduces a three-class, rank-two posterior to one Gauss-Hermite direction plus the controlled scalar logistic-normal evaluator. Carries the outer rule’s node count.

AnisotropicTensor(Vec<usize>)

Tensor-product Gauss-Hermite whose one-dimensional node count is chosen per retained posterior direction. Carries those node counts.

IsotropicSparse(usize)

Isotropic Smolyak sparse grid, carrying the level that certified.

Trait Implementations

impl Clone for MultinomialPosteriorRule

fn clone(&self) -> MultinomialPosteriorRule

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for MultinomialPosteriorRule

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for MultinomialPosteriorRule

impl PartialEq for MultinomialPosteriorRule

fn eq(&self, other: &MultinomialPosteriorRule) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for MultinomialPosteriorRule