Struct DataNormalizer 

pub struct DataNormalizer {
    pub mean: Vec<f32>,
    pub std_dev: Vec<f32>,
}

Feature-wise Z-score normalizer.

Stores per-feature mean and standard deviation fitted on a training corpus.

Fields

mean: Vec<f32>

Per-feature mean.

std_dev: Vec<f32>

Per-feature standard deviation.

Implementations

impl DataNormalizer

pub fn fit(data: &[Vec<f32>]) -> Self

Fit normalizer statistics from a collection of sample vectors.

Panics

Panics if data is empty or if sample vectors have inconsistent lengths.

pub fn transform(&self, x: &[f32]) -> Vec<f32>

Standardise a single sample: (x - mean) / std.

pub fn inverse_transform(&self, x: &[f32]) -> Vec<f32>

Invert standardisation: x * std + mean.

Trait Implementations

impl Clone for DataNormalizer

fn clone(&self) -> DataNormalizer

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DataNormalizer

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

Formats the value using the given formatter.