Struct R2 

pub struct R2 { /* private fields */ }

R² (coefficient of determination).

Uses Welford’s online algorithm for the variance of the truth, avoiding the catastrophic cancellation that sum(y²) - n · mean(y)² suffers on large-offset, small-variance data. Returns None when fewer than 2 samples have been seen or when the truth variance is zero (constant truth).

Implementations

impl R2

pub const fn new() -> Self

Create a new R² accumulator.

Trait Implementations

impl Clone for R2

fn clone(&self) -> R2

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for R2

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

Formats the value using the given formatter.

impl Default for R2

fn default() -> R2

Returns the “default value” for a type.

impl Metric for R2

type Truth = f64

The ground-truth type for this metric.

type Prediction = f64

The prediction type for this metric.

fn update(&mut self, truth: f64, prediction: f64) -> Result<(), RillError>

Incorporate a single observation.

fn value(&self) -> Option<f64>

Current metric value, or None if not enough data has been seen.

fn samples_seen(&self) -> u64

How many observations have been incorporated.

fn reset(&mut self)

Reset the metric.