Struct BetaBernoulliObservation 

pub struct BetaBernoulliObservation {
    pub probability_variable: String,
    pub observations: Vec<bool>,
}

BetaBernoulliObservation captures a Bernoulli likelihood y ~ Bernoulli(p) where the success probability p is a BetaNP variable. It contributes (n_s, n_f) to the posterior natural parameters, i.e. adds n_s to (α − 1) and n_f to (β − 1).

A factor holds a reference to its Beta-distributed probability variable and a (possibly empty) batch of binary outcomes. Posterior inference combining prior + factor is exact in one VMP sweep because Bernoulli is conjugate to Beta.

Fields

probability_variable: String

Name of the BetaNP variable in the VMP graph.

observations: Vec<bool>

Observed binary outcomes (true = success, false = failure).

Implementations

impl BetaBernoulliObservation

pub fn new( probability_variable: impl Into<String>, observations: Vec<bool>, ) -> Self

Build a new Beta-Bernoulli observation factor from a boolean batch.

pub fn from_counts( probability_variable: impl Into<String>, successes: u64, failures: u64, ) -> Self

Convenience constructor from aggregate counts. Often you already have the sufficient statistics as (n_s, n_f) without keeping the raw batch.

pub fn num_successes(&self) -> u64

Number of successes n_s = Σ 1[y_i = 1]. Used as the α-parameter increment.

pub fn num_failures(&self) -> u64

Number of failures n_f = Σ 1[y_i = 0]. Used as the β-parameter increment.

pub fn num_observations(&self) -> usize

Total number of observations N = n_s + n_f.

Trait Implementations

impl Clone for BetaBernoulliObservation

fn clone(&self) -> BetaBernoulliObservation

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for BetaBernoulliObservation

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

Formats the value using the given formatter.