Enum OnlineAlgorithm 

pub enum OnlineAlgorithm {
    Perceptron,
    PassiveAggressive {
        c: f64,
    },
    SgdMomentum {
        lr: f64,
        momentum: f64,
        l2_reg: f64,
    },
}

Online learning algorithm selection.

Variants

Perceptron

Classic Perceptron binary classifier.

Update rule (on misprediction only):

w[i] += label * x[i]
bias  += label

PassiveAggressive

Passive-Aggressive PA-I update.

loss = max(0, 1 - label * score)
tau  = loss / (||x||² + 1 / (2 * C))
w[i] += tau * label * x[i]
bias  += tau * label

Fields

c: f64

Aggressiveness parameter. Larger values → more aggressive updates.

SgdMomentum

Stochastic gradient descent with momentum and L2 regularisation.

velocity[i] = momentum * velocity[i] - lr * (grad[i] + l2_reg * w[i])
w[i]       += velocity[i]
bias       -= lr * (-label)

Fields

lr: f64

Learning rate (step size).

momentum: f64

Momentum coefficient ∈ [0, 1).

l2_reg: f64

L2 weight-decay coefficient.

Trait Implementations

impl Clone for OnlineAlgorithm

fn clone(&self) -> OnlineAlgorithm

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for OnlineAlgorithm

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

Formats the value using the given formatter.

impl PartialEq for OnlineAlgorithm

fn eq(&self, other: &OnlineAlgorithm) -> 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 StructuralPartialEq for OnlineAlgorithm