Struct FairLoss 

pub struct FairLoss {
    pub c: Option<f64>,
}

Fair Loss objective. c * |y - yhat| - c^2 * ln(|y - yhat| / c + 1).

Fields

c: Option<f64>

The threshold constant c.

Implementations

impl FairLoss

pub fn loss_single(&self, y: f64, yhat: f64, sample_weight: Option<f64>) -> f32

Trait Implementations

impl Clone for FairLoss

fn clone(&self) -> FairLoss

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for FairLoss

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

Formats the value using the given formatter.

impl Default for FairLoss

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for FairLoss

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl ObjectiveFunction for FairLoss

fn loss( &self, y: &[f64], yhat: &[f64], sample_weight: Option<&[f64]>, _group: Option<&[u64]>, ) -> Vec<f32>

Per-sample loss.

fn gradient( &self, y: &[f64], yhat: &[f64], sample_weight: Option<&[f64]>, _group: Option<&[u64]>, ) -> (Vec<f32>, Option<Vec<f32>>)

Per-sample gradient and (optional) hessian.

fn default_metric(&self) -> Metric

Default evaluation metric for this objective.

fn requires_batch_evaluation(&self) -> bool

Whether this objective requires batch evaluation (e.g. for Python custom objectives).

fn initial_value( &self, y: &[f64], sample_weight: Option<&[f64]>, _group: Option<&[u64]>, ) -> f64

Initial prediction (base score) before any trees are added.

fn gradient_and_loss( &self, y: &[f64], yhat: &[f64], sample_weight: Option<&[f64]>, group: Option<&[u64]>, ) -> (Vec<f32>, Option<Vec<f32>>, Vec<f32>)

Gradient, hessian and loss in a single pass over the data.

fn gradient_and_loss_into( &self, y: &[f64], yhat: &[f64], sample_weight: Option<&[f64]>, group: Option<&[u64]>, grad: &mut [f32], hess: &mut Option<Vec<f32>>, loss: &mut [f32], )

In-place version of gradient_and_loss.

impl Serialize for FairLoss

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.