Struct PcaReducer 

pub struct PcaReducer {
    pub components: Vec<Vec<f64>>,
    pub mean: Vec<f64>,
    pub explained_variance: Vec<f64>,
    pub n_components: usize,
}

PCA dimensionality reducer using power-iteration eigenvector extraction.

Fields

components: Vec<Vec<f64>>

Principal components (each row is one eigenvector, shape n_components × n_features).

mean: Vec<f64>

Per-feature mean used for centering.

explained_variance: Vec<f64>

Variance explained by each component.

n_components: usize

Number of components retained.

Implementations

impl PcaReducer

pub fn fit( data: &[Vec<f64>], n_components: usize, ) -> Result<Self, ReductionError>

Fit PCA on data and retain n_components principal components.

Uses 50 power-iteration steps per component.

pub fn transform( &self, data: &[Vec<f64>], ) -> Result<Vec<Vec<f64>>, ReductionError>

Project data onto the PCA component space.

pub fn fit_transform( data: &[Vec<f64>], n_components: usize, ) -> Result<ReductionResult, ReductionError>

Fit PCA and immediately project the training data.

pub fn inverse_transform(&self, reduced: &[Vec<f64>]) -> Vec<Vec<f64>>

Approximately reconstruct original-space vectors from reduced-space vectors.

This is a lossy reconstruction: x̂ = mean + W * z

Trait Implementations

impl Clone for PcaReducer

fn clone(&self) -> PcaReducer

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for PcaReducer

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

Formats the value using the given formatter.