Enum Scoring 

pub enum Scoring {
    Accuracy,
    NegLogLoss,
    F1,
}

Scoring rule for ml_cross_val_score; every rule is higher-is-better.

Hard predictions come from SimpleClassifier::predict (by default predict_proba >= 0.5). Labels are expected to be 0.0 or 1.0. AFML §7.5 and Chapter 9 recommend log loss for anything sized by probability.

Variants

Accuracy

Share of test samples whose hard prediction equals the label (compared within 1e-12, so a label other than 0.0 or 1.0 never matches).

NegLogLoss

Negative mean log loss, mean(y ln p + (1 - y) ln(1 - p)), with p clipped to [1e-15, 1 - 1e-15]. Always <= 0.

F1

F1 score of the positive class (label > 0.5); 0.0 when precision and recall are both 0, including when there are no positive predictions. NaN on an empty test set, like the other rules.

Trait Implementations

impl Clone for Scoring

fn clone(&self) -> Scoring

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for Scoring