Struct ARDRegression 

pub struct ARDRegression<F> {
    pub max_iter: usize,
    pub tol: F,
    pub alpha_1: F,
    pub alpha_2: F,
    pub lambda_1: F,
    pub lambda_2: F,
    pub threshold_lambda: F,
    pub fit_intercept: bool,
}

Automatic Relevance Determination Regression.

Bayesian linear regression with per-feature precision priors. Features with high precision (small variance) are pruned, achieving sparsity.

Type Parameters

• F: The floating-point type (f32 or f64).

Fields

max_iter: usize

Maximum number of EM iterations.

tol: F

Convergence tolerance on the relative change in alpha/lambda.

alpha_1: F

Shape hyperparameter for the alpha (noise precision) Gamma prior.

alpha_2: F

Rate hyperparameter for the alpha (noise precision) Gamma prior.

lambda_1: F

Shape hyperparameter for the lambda (weight precision) Gamma prior.

lambda_2: F

Rate hyperparameter for the lambda (weight precision) Gamma prior.

threshold_lambda: F

Features with lambda_i > threshold_lambda are pruned.

fit_intercept: bool

Whether to fit an intercept (bias) term.

Implementations

impl<F: Float + FromPrimitive> ARDRegression<F>

pub fn new() -> Self

Create a new ARDRegression with default settings.

Defaults: max_iter = 300, tol = 1e-3, alpha_1 = alpha_2 = 1e-6, lambda_1 = lambda_2 = 1e-6, threshold_lambda = 1e4, fit_intercept = true.

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

Set the maximum number of iterations.

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

Set the convergence tolerance.

pub fn with_alpha_1(self, alpha_1: F) -> Self

Set the alpha shape hyperparameter.

pub fn with_alpha_2(self, alpha_2: F) -> Self

Set the alpha rate hyperparameter.

pub fn with_lambda_1(self, lambda_1: F) -> Self

Set the lambda shape hyperparameter.

pub fn with_lambda_2(self, lambda_2: F) -> Self

Set the lambda rate hyperparameter.

pub fn with_threshold_lambda(self, threshold_lambda: F) -> Self

Set the pruning threshold for feature lambda values.

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

Set whether to fit an intercept term.

Trait Implementations

impl<F: Clone> Clone for ARDRegression<F>

fn clone(&self) -> ARDRegression<F>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<F: Debug> Debug for ARDRegression<F>

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

Formats the value using the given formatter.

impl<F: Float + FromPrimitive> Default for ARDRegression<F>

fn default() -> Self

Returns the “default value” for a type.

impl<F: Float + Send + Sync + ScalarOperand + FromPrimitive + 'static> Fit<ArrayBase<OwnedRepr<F>, Dim<[usize; 2]>>, ArrayBase<OwnedRepr<F>, Dim<[usize; 1]>>> for ARDRegression<F>

fn fit( &self, x: &Array2<F>, y: &Array1<F>, ) -> Result<FittedARDRegression<F>, FerroError>

Fit the ARD model via iterative evidence maximization.

Errors

• FerroError::ShapeMismatch — sample count mismatch.
• FerroError::InsufficientSamples — fewer than 2 samples.
• FerroError::NumericalInstability — numerical failure in solver.

type Fitted = FittedARDRegression<F>

The fitted model type returned by fit.

type Error = FerroError

The error type returned by fit.

impl<F> PipelineEstimator<F> for ARDRegression<F>

where F: Float + FromPrimitive + ScalarOperand + Send + Sync + 'static,

fn fit_pipeline( &self, x: &Array2<F>, y: &Array1<F>, ) -> Result<Box<dyn FittedPipelineEstimator<F>>, FerroError>

Fit this estimator on the given data.