Struct ErrorAnalyzer 

pub struct ErrorAnalyzer {
    pub overlap_threshold: f64,
}

Analyzer for NER prediction errors.

Uses an optimized O(n + m) algorithm with spatial indexing instead of naive O(n*m) nested loops. For datasets with >1000 entities, this provides significant speedup.

Fields

overlap_threshold: f64

Overlap threshold for partial match (IoU)

Implementations

impl ErrorAnalyzer

pub fn new(overlap_threshold: f64) -> Self

Create analyzer with custom overlap threshold.

pub fn analyze( &self, predictions: &[PredictedEntity], gold: &[GoldEntity], ) -> ErrorReport

Analyze errors between predictions and gold entities.

Uses the canonical GoldEntity type from eval::datasets. Entity types are compared using their string labels.

Trait Implementations

impl Clone for ErrorAnalyzer

fn clone(&self) -> ErrorAnalyzer

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ErrorAnalyzer

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

Formats the value using the given formatter.

impl Default for ErrorAnalyzer

fn default() -> Self

Returns the “default value” for a type.