Struct MetricDivergence 

pub struct MetricDivergence {
    pub muc_f1: f64,
    pub b3_f1: f64,
    pub ceaf_e_f1: f64,
    pub muc_ceaf_divergence: f64,
    pub muc_b3_divergence: f64,
    pub b3_ceaf_divergence: f64,
}

Divergence between coreference metrics.

Research Context

At book scale, different metrics diverge significantly:

• MUC tends to be inflated (favors link-based evaluation)
• CEAF-e tends to collapse (entity alignment struggles)
• B³ falls between but is more stable

Large divergence (>0.20) indicates potential metric unreliability.

Example

use anno_eval::eval::MetricDivergence;

let divergence = MetricDivergence::from_scores(0.90, 0.65, 0.45);
assert!(divergence.has_high_divergence());
println!("MUC-CEAF divergence: {:.0}%", divergence.muc_ceaf_divergence * 100.0);

Fields

muc_f1: f64

MUC F1 score.

b3_f1: f64

B³ F1 score.

ceaf_e_f1: f64

CEAF-e F1 score.

muc_ceaf_divergence: f64

Divergence between MUC and CEAF-e (absolute difference).

muc_b3_divergence: f64

Divergence between MUC and B³.

b3_ceaf_divergence: f64

Divergence between B³ and CEAF-e.

Implementations

impl MetricDivergence

pub fn from_scores(muc_f1: f64, b3_f1: f64, ceaf_e_f1: f64) -> Self

Compute divergence from raw scores.

pub fn has_high_divergence(&self) -> bool

Check if divergence is high (>0.20), indicating unreliable metrics.

pub fn muc_likely_inflated(&self) -> bool

Check if MUC is likely inflated (MUC >> CEAF-e).

pub fn ceaf_likely_collapsed(&self) -> bool

Check if CEAF-e is likely collapsed (CEAF-e << others).

pub fn most_reliable_metric(&self) -> &'static str

Get most reliable metric recommendation.

Trait Implementations

impl Clone for MetricDivergence

fn clone(&self) -> MetricDivergence

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for MetricDivergence

impl Debug for MetricDivergence

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

Formats the value using the given formatter.

impl Default for MetricDivergence

fn default() -> MetricDivergence

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for MetricDivergence

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for MetricDivergence

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

where __S: Serializer,

Serialize this value into the given Serde serializer.