Struct DistanceWeightedMetaModel 

pub struct DistanceWeightedMetaModel { /* private fields */ }

Picks the training record that maximizes evidence × w(distance), where w(d) = 1 / (d + epsilon) over z-score-normalized Euclidean distance and evidence is MetaTrainingRecord::score_lift when present, falling back to best_score for legacy records.

The distinction from NearestNeighborMetaModel: NN picks the closest record regardless of how well that record performed, so a single poorly-tuned outlier can pull predictions off. Distance-weighted folds the record’s evidence into the selection — a record is “good” if it’s both similar to the query AND demonstrated real tuner signal. Using lift instead of the raw score avoids the easy-corpus bias: best_score = 0.9 on an easy corpus is weaker evidence than best_score = 0.6 that beat its run’s mean by a wide margin. At N = 1 this degenerates to NN (same record either way).

epsilon is a smoothing floor on the distance term; at d ≈ 0 it prevents the weight from exploding and over-committing to a single near-duplicate record. Default 0.1.

Implementations

impl DistanceWeightedMetaModel

pub fn new() -> Self

pub fn with_epsilon(self, epsilon: f64) -> Self

Override the smoothing constant added to distance before inversion. Larger epsilon makes predictions smoother; smaller sharpens the preference for near-duplicate records. Must be strictly positive (silently clamped to 1e-12 if a zero or negative value is passed).

pub fn records(&self) -> &[MetaTrainingRecord]

pub fn score_candidates( &self, features: &CorpusFeatures, ) -> Vec<(usize, f64, f64)>

Per-record (weighted_score, distance) pairs for the given query features, sorted by descending weighted score. Useful for introspecting why a particular prediction was made.

Trait Implementations

impl Clone for DistanceWeightedMetaModel

fn clone(&self) -> DistanceWeightedMetaModel

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DistanceWeightedMetaModel

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

Formats the value using the given formatter.

impl Default for DistanceWeightedMetaModel

fn default() -> Self

Returns the “default value” for a type.

impl MetaModel for DistanceWeightedMetaModel

fn fit(&mut self, records: &[MetaTrainingRecord])

Fit on a training set. Replacing any prior state.

fn is_fitted(&self) -> bool

True once fit has been called with at least one usable record. Self::predict panics when this is false, so Result-returning boundaries (e.g. SphereQLPipeline::new_from_metamodel) check this first. No default impl on purpose: every model must answer for its own notion of “fitted” rather than inherit a guess.

fn predict(&self, features: &CorpusFeatures) -> PipelineConfig

Predict the config that should work best on a corpus with the given profile. Panics if fit has not been called with at least one record — callers should treat MetaModel as a trained object and front-load fit, or check Self::is_fitted when the training state isn’t statically known.

fn name(&self) -> &str

Short name for logs and model comparison.