Struct Detector 

pub struct Detector {
    pub recall: f64,
    pub false_positive_rate: f64,
}

A detector’s prevalence-independent behaviour.

Fields

recall: f64

Fraction of genuinely bad merges it flags. Carries across populations.

false_positive_rate: f64

Fraction of genuinely good merges it flags anyway. Carries across populations, and is the number that decides everything at low prevalence.

Implementations

impl Detector

pub fn from_reported( precision: f64, recall: f64, prevalence: f64, ) -> Result<Self, String>

Recovers the prevalence-independent behaviour from a paper’s reported precision and recall at that paper’s corpus prevalence.

Errors

Inputs outside 0.0..=1.0, a zero precision or recall (nothing to recover), or a corpus prevalence of 1.0 (no negatives to have produced the reported false positives).

pub fn precision_at(&self, prevalence: f64) -> f64

Precision this detector reaches at prevalence.

Returns 0.0 when it would flag nothing at all, which is the honest reading: a detector that never fires has no precision to speak of.

pub fn false_alarms_per_hit(&self, prevalence: f64) -> Option<f64>

False alarms per true alarm at prevalence — the number an on-call human actually experiences. None when it never fires on a true positive, so the ratio is undefined rather than infinite.

Trait Implementations

impl Clone for Detector

fn clone(&self) -> Detector

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Detector

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

Formats the value using the given formatter.

impl PartialEq for Detector

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for Detector

impl StructuralPartialEq for Detector