Struct StackingEnsemble 

pub struct StackingEnsemble<M: Model, Meta: Model> { /* private fields */ }

Stacking ensemble with a meta-learner.

Uses base models’ predictions as features for a meta-model.

Implementations

impl<M: Model, Meta: Model> StackingEnsemble<M, Meta>

pub fn new(base_models: Vec<M>, meta_model: Meta) -> TrainResult<Self>

Create a new stacking ensemble.

Arguments

• base_models - First-level base models
• meta_model - Second-level meta-learner

pub fn generate_meta_features( &self, input: &Array2<f64>, ) -> TrainResult<Array2<f64>>

Generate meta-features from base model predictions.

Arguments

• input - Input data

Returns

Meta-features [batch_size, num_base_models * num_classes]

Trait Implementations

impl<M: Debug + Model, Meta: Debug + Model> Debug for StackingEnsemble<M, Meta>

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

Formats the value using the given formatter.

impl<M: Model, Meta: Model> Ensemble for StackingEnsemble<M, Meta>

fn predict(&self, input: &Array2<f64>) -> TrainResult<Array2<f64>>

Predict using the ensemble.

fn num_models(&self) -> usize

Get the number of models in the ensemble.