Struct GaussianDensityProvider 

pub struct GaussianDensityProvider { /* private fields */ }

Independent Gaussian density provider with Gauss–Hermite quadrature.

probability returns the product of univariate N(μ, σ²) densities for factor variables. Self::quadrature returns product Gauss–Hermite nodes in Lebesgue measure (suitable for ∫ body(x) dx near each Gaussian mode).

Implementations

impl GaussianDensityProvider

pub fn new() -> Self

Empty provider.

pub fn set_gaussian(&mut self, var: VariableId, mean: f64, variance: f64)

Declare an independent Gaussian for var with mean mean and variance variance.

Panics

Never panics; non-positive variance is rejected by returning early no-op… actually we clamp: variance must be > 0 or the insert is skipped. Prefer validating at call sites.

Trait Implementations

impl Clone for GaussianDensityProvider

fn clone(&self) -> GaussianDensityProvider

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for GaussianDensityProvider

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

Formats the value using the given formatter.

impl Default for GaussianDensityProvider

fn default() -> GaussianDensityProvider

Returns the “default value” for a type.

impl DistributionProvider for GaussianDensityProvider

fn probability( &self, spec: &FactorSpec<'_>, assignment: &Assignment, _ctx: &EvalContext, ) -> Result<f64, EvalError>

Probability / density mass for a factor under an assignment.

fn support( &self, vars: &[VariableId], _ctx: &EvalContext, ) -> Result<Arc<[Arc<[Value]>]>, EvalError>

Discrete support for variables (cartesian rows of values aligned to vars).

fn quadrature( &self, vars: &[VariableId], _ctx: &EvalContext, ) -> Result<Option<QuadratureNodes>, EvalError>

Optional continuous quadrature nodes (assignment_row, Lebesgue weight) for crate::ExprNode::IntegralOut.

fn outcome( &self, var: VariableId, assignment: &Assignment, _ctx: &EvalContext, ) -> Result<f64, EvalError>

Outcome function value (identity: the bound value of var).

fn n_draws(&self) -> Option<usize>

Number of posterior draws, or None for a single empirical world.