Struct RecursiveFeatureElimination 

pub struct RecursiveFeatureElimination<F>
where
    F: Fn(&Array2<f64>, &Array1<f64>) -> Result<Array1<f64>>,
{ /* private fields */ }

Recursive Feature Elimination (RFE) for feature selection

RFE works by recursively removing features and evaluating model performance. This implementation uses a feature importance scoring function to rank features.

Implementations

impl<F> RecursiveFeatureElimination<F>

where F: Fn(&Array2<f64>, &Array1<f64>) -> Result<Array1<f64>>,

pub fn new(n_features_to_select: usize, importancefunc: F) -> Self

Creates a new RFE selector

Arguments

• n_features_to_select - Number of features to select
• importance_func - Function that computes feature importance scores

pub fn with_step(self, step: usize) -> Self

Set the number of features to remove at each iteration

pub fn fit(&mut self, x: &Array2<f64>, y: &Array1<f64>) -> Result<()>

Fit the RFE selector

Arguments

• x - Training data, shape (n_samples, n_features)
• y - Target values, shape (n_samples,)

pub fn transform(&self, x: &Array2<f64>) -> Result<Array2<f64>>

Transform data by selecting features

pub fn fit_transform( &mut self, x: &Array2<f64>, y: &Array1<f64>, ) -> Result<Array2<f64>>

Fit and transform in one step

pub fn get_support(&self) -> Option<&Vec<usize>>

Get selected feature indices

pub fn ranking(&self) -> Option<&Array1<usize>>

Get feature rankings (1 is best)

pub fn scores(&self) -> Option<&Array1<f64>>

Get feature scores

Trait Implementations

impl<F> Clone for RecursiveFeatureElimination<F>

where F: Fn(&Array2<f64>, &Array1<f64>) -> Result<Array1<f64>> + Clone,

fn clone(&self) -> RecursiveFeatureElimination<F>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<F> Debug for RecursiveFeatureElimination<F>

where F: Fn(&Array2<f64>, &Array1<f64>) -> Result<Array1<f64>> + Debug,

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

Formats the value using the given formatter.