Struct BayesianEvidenceFusionOperator 

pub struct BayesianEvidenceFusionOperator {
    pub signals: Vec<Arc<dyn Operator>>,
    pub base_rate: f64,
    pub top_k: Option<usize>,
}

Exact Bayesian fusion of signed prior-free evidence.

Each present score is interpreted as a prior-free probability-like evidence value and converted to a signed EvidenceLogit. Missing signals contribute the additive identity zero. The configured corpus prior enters exactly once, after which the operator computes sigmoid(logit(base_rate) + sum(evidence_i)) without gating, confidence scaling, normalized weights, or adaptive query-pool statistics.

Fields

signals: Vec<Arc<dyn Operator>>

base_rate: f64

top_k: Option<usize>

Implementations

impl BayesianEvidenceFusionOperator

pub fn new(signals: Vec<Arc<dyn Operator>>, base_rate: f64) -> Self

pub fn with_top_k(self, top_k: usize) -> Self

Trait Implementations

impl Operator for BayesianEvidenceFusionOperator

fn execute(&self, ctx: &ExecutionContext) -> OperatorResult

fn cost_estimate(&self, stats: &IndexStats) -> f64