Struct McmcDoSampler 

pub struct McmcDoSampler {
    pub outcome: VariableId,
    pub proposal_sd: f64,
    pub burn_in: usize,
    pub thin: usize,
}

Random-walk Metropolis–Hastings on the outcome margin.

The chain targets a Silverman (1986) Gaussian KDE fitted to a pilot batch of interventional ancestral draws — a smoothed proxy of the interventional law of outcome, not the joint mechanism density. Proposals are Gaussian random walks (proposal_sd); this is not independent MH, and is exact for the interventional law only in the large-pilot / vanishing-bandwidth limit of that KDE proxy.

Fields

outcome: VariableId

Outcome variable to record.

proposal_sd: f64

Proposal standard deviation.

burn_in: usize

Burn-in iterations.

thin: usize

Thinning.

Implementations

impl McmcDoSampler

pub fn new(outcome: VariableId) -> Self

Construct targeting outcome.

pub fn sample( &self, model: &CompiledCausalModel, interventions: &[Intervention], n_samples: usize, rng: &mut CausalRng, ws: &mut MechanismWorkspace, ctx: &ExecutionContext, ) -> Result<DoSampleResult, ModelError>

Run random-walk MH against a KDE of interventional pilot draws (see type docs).

Errors

Sampling / density failures.

Trait Implementations

impl Clone for McmcDoSampler

fn clone(&self) -> McmcDoSampler

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for McmcDoSampler

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

Formats the value using the given formatter.

impl Default for McmcDoSampler

fn default() -> Self

Returns the “default value” for a type.