Struct Lasso 

pub struct Lasso { /* private fields */ }

Lasso regression with L1 regularization.

Fits a linear model with L1 penalty on coefficient magnitudes. The optimization objective is:

minimize ||y - Xβ||² + α||β||₁

where α (alpha) controls the regularization strength.

Solver

Uses coordinate descent with soft-thresholding.

When to use Lasso

• For automatic feature selection (produces sparse models)
• When you expect only a few features to be relevant
• When interpretability through sparsity is desired

Examples

use aprender::prelude::*;
use aprender::linear_model::Lasso;

// Data with some features
let x = Matrix::from_vec(5, 2, vec![
    1.0, 2.0,
    2.0, 3.0,
    3.0, 4.0,
    4.0, 5.0,
    5.0, 6.0,
]).expect("Valid matrix dimensions");
let y = Vector::from_slice(&[5.0, 8.0, 11.0, 14.0, 17.0]);

let mut model = Lasso::new(0.1);  // alpha = 0.1
model.fit(&x, &y).expect("Fit should succeed with valid data");

let predictions = model.predict(&x);
let r2 = model.score(&x, &y);
assert!(r2 > 0.9);

Implementations

impl Lasso

pub fn new(alpha: f32) -> Self

Creates a new Lasso regression with the given regularization strength.

Arguments

• alpha - Regularization strength. Larger values = more sparsity. Must be non-negative.

pub fn with_intercept(self, fit_intercept: bool) -> Self

Sets whether to fit an intercept term.

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

Sets the maximum number of iterations.

pub fn with_tol(self, tol: f32) -> Self

Sets the convergence tolerance.

pub fn alpha(&self) -> f32

Returns the regularization strength (alpha).

pub fn coefficients(&self) -> &Vector<f32>

Returns the coefficients (excluding intercept).

Panics

Panics if model is not fitted.

pub fn intercept(&self) -> f32

Returns the intercept term.

pub fn is_fitted(&self) -> bool

Returns true if the model has been fitted.

pub fn save<P: AsRef<Path>>(&self, path: P) -> Result<(), String>

Saves the model to a binary file using bincode.

Errors

Returns an error if serialization or file writing fails.

pub fn load<P: AsRef<Path>>(path: P) -> Result<Self, String>

Loads a model from a binary file.

Errors

Returns an error if file reading or deserialization fails.

pub fn save_safetensors<P: AsRef<Path>>(&self, path: P) -> Result<(), String>

Saves the model to SafeTensors format.

SafeTensors format is compatible with:

• HuggingFace ecosystem
• Ollama (can convert to GGUF)
• PyTorch, TensorFlow
• realizar inference engine

Errors

Returns an error if:

• Model is not fitted
• Serialization fails
• File writing fails

pub fn load_safetensors<P: AsRef<Path>>(path: P) -> Result<Self, String>

Loads a model from SafeTensors format.

Errors

Returns an error if:

• File reading fails
• SafeTensors format is invalid
• Required tensors are missing

Trait Implementations

impl Clone for Lasso

fn clone(&self) -> Lasso

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Lasso

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Lasso

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Estimator for Lasso

fn fit(&mut self, x: &Matrix<f32>, y: &Vector<f32>) -> Result<()>

Fits the Lasso regression model using coordinate descent.

Errors

Returns an error if input dimensions don’t match.

fn predict(&self, x: &Matrix<f32>) -> Vector<f32>

Predicts target values for input data.

Panics

Panics if model is not fitted.

fn score(&self, x: &Matrix<f32>, y: &Vector<f32>) -> f32

Computes the R² score.

impl Serialize for Lasso

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.