Struct PercentileEstimator 

pub struct PercentileEstimator { /* private fields */ }

P² algorithm for online quantile estimation with O(1) memory per quantile.

Estimates a specific percentile p (0 < p < 1) without storing individual samples. After at least 5 observations the estimate is maintained via the P² marker update rule.

Accuracy is typically within a few percent for stationary distributions with thousands of samples.

References

Jain, R., & Chlamtac, I. (1985). The P² algorithm for dynamic calculation of quantiles and histograms without storing observations. Communications of the ACM, 28(10), 1076–1085.

Implementations

impl PercentileEstimator

pub fn new(p: f64) -> Self

Creates an estimator for the given quantile p (e.g. 0.95 for p95).

Panics

Panics in debug builds if p is not in (0, 1).

pub fn update(&mut self, x: f64)

Updates the estimator with a new observation.

pub fn estimate(&self) -> Option<f64>

Returns the current percentile estimate, or None before 5 observations.

pub fn count(&self) -> u64

Returns the number of observations processed.

pub fn quantile(&self) -> f64

Returns the target quantile.

Trait Implementations

impl Clone for PercentileEstimator

fn clone(&self) -> PercentileEstimator

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for PercentileEstimator

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

Formats the value using the given formatter.