Struct VectorProbabilityTransform 

pub struct VectorProbabilityTransform {
    pub mu_match: f64,
    pub mu_random: f64,
    pub sigma: f64,
    pub base_rate: f64,
}

Likelihood-ratio calibrator for vector distances (Theorem 3.1.1, Paper 5). Converts vector similarity into calibrated probability.

The transform models distances to relevant documents (f_R) and to background (random) documents (f_G) as Gaussian distributions. Calibration converts a distance into a posterior probability by applying Bayes’ rule with the configured base rate:

log f_R(d) - log f_G(d) + logit(base_rate)  ->  sigmoid

The formulation is deliberately small; downstream callers fit the means and standard deviations offline (for example, via the parameter learner) and pass the transform through. Optional per-distance weights bias the computed log-odds before the sigmoid.

Fields

mu_match: f64

Mean distance for relevant documents (numerator distribution).

mu_random: f64

Mean distance for background documents (denominator distribution).

sigma: f64

Shared standard deviation. Must be positive.

base_rate: f64

Prior probability of relevance. Default 0.5 (neutral).

Implementations

impl VectorProbabilityTransform

pub fn new( mu_match: f64, mu_random: f64, sigma: f64, base_rate: f64, ) -> ScoringResult<Self>

pub fn calibrate_one(&self, distance: f64) -> ScoringResult<f64>

Convert a single distance to a probability via the likelihood ratio + base-rate logit.

pub fn calibrate( &self, distances: &[f64], weights: Option<&[f64]>, ) -> ScoringResult<Vec<f64>>

Vectorized calibration. Optional weights bias each distance’s log-odds before the sigmoid.

Trait Implementations

impl Clone for VectorProbabilityTransform

fn clone(&self) -> VectorProbabilityTransform

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for VectorProbabilityTransform

impl Debug for VectorProbabilityTransform

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for VectorProbabilityTransform

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for VectorProbabilityTransform

fn eq(&self, other: &VectorProbabilityTransform) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for VectorProbabilityTransform

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for VectorProbabilityTransform