Struct MetricSet 

pub struct MetricSet { /* private fields */ }

Combined metric tracker that holds both regression and classification metrics.

Useful when evaluating a model that may be used for either task, or when tracking both types of metrics simultaneously during experimentation.

Example

use irithyll::metrics::MetricSet;

let mut ms = MetricSet::new();
ms.update_regression(3.0, 2.8);
ms.update_classification(1, 1, 0.9);

assert!(ms.regression().mae() > 0.0);
assert!(ms.classification().accuracy() > 0.0);

Implementations

impl MetricSet

pub fn new() -> Self

Create a new metric set with empty regression and classification trackers.

pub fn update_regression(&mut self, target: f64, prediction: f64)

Update regression metrics with a (target, prediction) pair.

pub fn update_classification( &mut self, target: usize, predicted: usize, predicted_proba: f64, )

Update classification metrics with a single observation.

• target: true class label.
• predicted: predicted class label.
• predicted_proba: model’s predicted probability for the positive class.

pub fn regression(&self) -> &RegressionMetrics

Reference to the regression metrics tracker.

pub fn classification(&self) -> &ClassificationMetrics

Reference to the classification metrics tracker.

pub fn reset(&mut self)

Reset both regression and classification metrics to their initial states.

Trait Implementations

impl Clone for MetricSet

fn clone(&self) -> MetricSet

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MetricSet

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

Formats the value using the given formatter.

impl Default for MetricSet

fn default() -> Self

Returns the “default value” for a type.