Trait AdvancedTarget 

pub trait AdvancedTarget<F>: Send + Sync
where
    F: Float + Copy + Display,
{
    // Required methods
    fn log_density(&self, x: &Array1<F>) -> F;
    fn gradient(&self, x: &Array1<F>) -> Array1<F>;
    fn dim(&self) -> usize;

    // Provided methods
    fn log_density_and_gradient(&self, x: &Array1<F>) -> (F, Array1<F>) { ... }
    fn hessian(x: &Array1<F>) -> Option<Array2<F>> { ... }
    fn fisher_information(x: &Array1<F>) -> Option<Array2<F>> { ... }
    fn riemann_metric(x: &Array1<F>) -> Option<Array2<F>> { ... }
    fn modeldimension(&self, modelid: usize) -> usize { ... }
    fn model_transition_prob(from_model: usize, _tomodel: usize) -> F { ... }
    fn batch_log_density(&self, xbatch: &Array2<F>) -> Array1<F> { ... }
}

Advanced-advanced target distribution interface

Required Methods

fn log_density(&self, x: &Array1<F>) -> F

Compute log probability density

fn gradient(&self, x: &Array1<F>) -> Array1<F>

Compute gradient of log density

fn dim(&self) -> usize

Get dimensionality

Provided Methods

fn log_density_and_gradient(&self, x: &Array1<F>) -> (F, Array1<F>)

Compute both log density and gradient efficiently

fn hessian(x: &Array1<F>) -> Option<Array2<F>>

Compute Hessian matrix (for manifold methods)

fn fisher_information(x: &Array1<F>) -> Option<Array2<F>>

Compute Fisher information matrix

fn riemann_metric(x: &Array1<F>) -> Option<Array2<F>>

Compute Riemann metric tensor (for Riemannian methods)

fn modeldimension(&self, modelid: usize) -> usize

Support for discontinuous model spaces (for Reversible Jump)

fn model_transition_prob(from_model: usize, _tomodel: usize) -> F

Model transition probability (for Reversible Jump)

fn batch_log_density(&self, xbatch: &Array2<F>) -> Array1<F>

Support parallel evaluation of multiple points

Dyn Compatibility

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors