Struct LassoCV 

pub struct LassoCV { /* private fields */ }

Lasso with the penalty chosen by K-fold cross-validation, as scikit-learn’s LassoCV.

The candidate penalties are n_alphas values evenly spaced on a log scale from α_max = max_j |x̃_jᵀ ỹ| / n (the smallest penalty giving an all-zero solution) down to eps · α_max. For each fold the whole path is fitted with warm starts, the penalty with the lowest mean test MSE is chosen, and the Lasso is refitted on all the data. Folds are contiguous blocks (scikit-learn’s default KFold without shuffling), so results reproduce LassoCV(cv=5).

Implementations

impl LassoCV

pub fn new() -> Self

100 penalties down to 1e-3 · α_max, 5 folds (scikit-learn’s defaults).

pub fn n_alphas(self, n_alphas: usize) -> Self

Number of penalties on the automatic grid.

pub fn eps(self, eps: f64) -> Self

Ratio of the smallest to the largest penalty on the automatic grid.

pub fn alphas(self, alphas: Vec<f64>) -> Self

Explicit candidate penalties instead of the automatic grid.

pub fn cv(self, cv: usize) -> Self

Number of cross-validation folds.

pub fn tol(self, tol: f64) -> Self

Coordinate-descent tolerance, as Lasso::tol.

pub fn max_iter(self, max_iter: usize) -> Self

pub fn alpha(&self) -> Option<f64>

The chosen penalty.

pub fn alphas_path(&self) -> Option<&[f64]>

The candidate penalties, in decreasing order.

pub fn mse_path(&self) -> Option<&Array2<f64>>

Test MSE of every penalty on every fold, shape (n_alphas, cv).

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

pub fn intercept(&self) -> f64

Trait Implementations

impl Clone for LassoCV

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LassoCV

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

Formats the value using the given formatter.

impl Default for LassoCV

fn default() -> Self

Returns the “default value” for a type.

impl Regressor for LassoCV

fn fit_weighted( &mut self, x: ArrayView2<'_, f64>, y: ArrayView1<'_, f64>, w: ArrayView1<'_, f64>, ) -> Result<()>

Weighted fits and weighted test MSEs, as LassoCV.fit(sample_weight=w).

fn fit(&mut self, x: ArrayView2<'_, f64>, y: ArrayView1<'_, f64>) -> Result<()>

fn predict(&self, x: ArrayView2<'_, f64>) -> Result<Array1<f64>>

fn fresh(&self) -> Box<dyn Regressor>