Struct VarianceThreshold 

#[non_exhaustive]
pub struct VarianceThreshold { /* private fields */ }

Remove features whose variance falls below a threshold.

By default, removes only constant features (threshold = 0.0). Useful as a lightweight first step before expensive feature selectors.

Examples

use scry_learn::dataset::Dataset;
use scry_learn::preprocess::Transformer;
use scry_learn::feature_selection::VarianceThreshold;

let mut data = Dataset::new(
    vec![
        vec![1.0, 2.0, 3.0],  // variable
        vec![5.0, 5.0, 5.0],  // constant — will be removed
    ],
    vec![0.0, 1.0, 0.0],
    vec!["a".into(), "b".into()],
    "target",
);

let mut vt = VarianceThreshold::new();
vt.fit_transform(&mut data).unwrap();
assert_eq!(data.n_features(), 1); // only "a" remains

Implementations

impl VarianceThreshold

pub fn new() -> Self

Create a new selector with threshold 0.0 (remove only constants).

pub fn threshold(self, t: f64) -> Self

Set the variance threshold.

Features with variance ≤ this value are removed.

pub fn variances(&self) -> &[f64]

Per-feature variances computed during fit.

Panics

Panics if called before VarianceThreshold::fit.

pub fn get_support(&self) -> &[bool]

Boolean mask of selected features.

true at index j means feature j was kept.

Trait Implementations

impl Clone for VarianceThreshold

fn clone(&self) -> VarianceThreshold

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for VarianceThreshold

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

Formats the value using the given formatter.

impl Default for VarianceThreshold

fn default() -> Self

Returns the “default value” for a type.

impl Transformer for VarianceThreshold

fn fit(&mut self, data: &Dataset) -> Result<()>

Learn parameters from the training data.

fn transform(&self, data: &mut Dataset) -> Result<()>

Apply the learned transformation to a dataset (in-place).

fn inverse_transform(&self, _data: &mut Dataset) -> Result<()>

Reverse the transformation (if invertible).

fn fit_transform(&mut self, data: &mut Dataset) -> Result<()>

Convenience: fit + transform in one call.