Struct WrapperMethods 

pub struct WrapperMethods;

Wrapper-based feature selection methods

Implementations

impl WrapperMethods

pub fn forward_selection( features: &Array2<f64>, target: &Array1<f64>, config: &FeatureSelectionConfig, ) -> Result<FeatureSelectionResult>

Forward feature selection

Starts with no features and iteratively adds the best feature.

Arguments

• features - Feature matrix (samples x features)
• target - Target variable
• config - Configuration for selection

Returns

• Feature selection result

Example

use scirs2_core::ndarray::{Array1, Array2};
use scirs2_series::feature_selection::{WrapperMethods, FeatureSelectionConfig};

let features = Array2::from_shape_vec((100, 10), (0..1000).map(|x| x as f64).collect()).expect("Operation failed");
let target = Array1::from_vec((0..100).map(|x| x as f64).collect());
let config = FeatureSelectionConfig::default();

let result = WrapperMethods::forward_selection(&features, &target, &config).expect("Operation failed");
println!("Selected {} features", result.selected_features.len());

pub fn backward_elimination( features: &Array2<f64>, target: &Array1<f64>, config: &FeatureSelectionConfig, ) -> Result<FeatureSelectionResult>

Backward feature elimination

Starts with all features and iteratively removes the worst feature.

Arguments

• features - Feature matrix (samples x features)
• target - Target variable
• config - Configuration for selection

Returns

• Feature selection result

pub fn recursive_feature_elimination( features: &Array2<f64>, target: &Array1<f64>, config: &FeatureSelectionConfig, ) -> Result<FeatureSelectionResult>

Recursive feature elimination

Recursively eliminates features by fitting a model and removing the least important features.

Arguments

• features - Feature matrix (samples x features)
• target - Target variable
• config - Configuration for selection

Returns

• Feature selection result

pub fn bidirectional_selection( features: &Array2<f64>, target: &Array1<f64>, config: &FeatureSelectionConfig, ) -> Result<FeatureSelectionResult>

Bidirectional feature selection

Combines forward selection and backward elimination.

Arguments

• features - Feature matrix (samples x features)
• target - Target variable
• config - Configuration for selection

Returns

• Feature selection result

pub fn evaluate_feature_subset( features: &Array2<f64>, target: &Array1<f64>, feature_indices: &[usize], config: &FeatureSelectionConfig, ) -> Result<f64>

Evaluate the performance of a feature subset using the configured scoring method

pub fn fit_predict_linear( features: &Array2<f64>, target: &Array1<f64>, ) -> Result<Array1<f64>>

Fit a linear regression model and return predictions

pub fn fit_linear_regression( features: &Array2<f64>, target: &Array1<f64>, ) -> Result<Array1<f64>>

Fit a linear regression model and return coefficients

pub fn solve_linear_system( a: &Array2<f64>, b: &Array1<f64>, ) -> Result<Array1<f64>>

Solve a linear system using Gaussian elimination with partial pivoting