shrinkage::HybridFit

backtracks

Source: src/fit/hybrid.rs

pub fn backtracks(&self) -> usize

Number of rejected trials followed by a curvature doubling.

cluster_labels

Source: src/fit/hybrid.rs

pub fn cluster_labels(&self) -> &[usize]

Cluster label of each feature, indexing Self::cluster_values.

cluster_values

Source: src/fit/hybrid.rs

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

Optimization-scale block values, with positive zeros. Clustered lasso returns strictly increasing values. Chain fused lasso returns signed segment values in feature order, which may repeat. SLOPE returns strictly decreasing nonnegative magnitudes, with zero last when present. For SLOPE, coefficient j on the optimization scale equals coefficient_signs()[j] * cluster_values()[cluster_labels()[j]]. Equal original-scale values need not share a label when training scales differ.

coefficient_signs

Source: src/fit/hybrid.rs

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

Optimization-scale coefficient signs in feature order: -1, 0, or 1. Both signed zeros have sign zero. Signs are recorded before coefficient back-transformation; SLOPE clusters may contain both nonzero signs.

coefficients

Source: src/fit/hybrid.rs

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

Original-scale coefficients in training-column order.

coordinate_sweeps

Source: src/fit/hybrid.rs

pub fn coordinate_sweeps(&self) -> usize

Number of cluster-coordinate sweeps over the current working set.

diagnostic_checks

Source: src/fit/hybrid.rs

pub fn diagnostic_checks(&self) -> usize

Number of refreshed inner diagnostic checks. Logistic checks use the fixed-zero quadratic subproblem; original-loss checks also run initially and after every accepted outer step.

dual_certificate

Source: src/fit/hybrid.rs

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

Original-space feasible dual vector from the latest certificate check. Screening-disabled fits do not construct a certificate.

dual_objective

Source: src/fit/hybrid.rs

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

Dual objective D(theta), when a certificate is available.

duality_gap

Source: src/fit/hybrid.rs

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

Duality gap G(beta_0, beta, theta) = P(beta_0, beta) - D(theta). This diagnostic does not replace the full mapping stopping rule.

intercept

Source: src/fit/hybrid.rs

pub fn intercept(&self) -> f64

Original-scale intercept, including the centering correction.

iterations

Source: src/fit/hybrid.rs

pub fn iterations(&self) -> usize

Total committed PG transitions plus coordinate sweeps over the current working set.

objective

Source: src/fit/hybrid.rs

pub fn objective(&self) -> f64

Final primal objective P(beta_0, beta) = F(beta_0, beta) + R(beta).

outer_backtracks

Source: src/fit/hybrid.rs

pub fn outer_backtracks(&self) -> usize

Rejected logistic trials followed by a step halving. Zero for Gaussian.

outer_iterations

Source: src/fit/hybrid.rs

pub fn outer_iterations(&self) -> usize

Accepted logistic IRLS steps. Zero for Gaussian fits.

pg_steps

Source: src/fit/hybrid.rs

pub fn pg_steps(&self) -> usize

Number of committed PG transitions, excluding diagnostic proxes.

predict

Source: src/fit/hybrid.rs

pub fn predict<M: RawColumns<f64> + ?Sized>(
    &self,
    x: &M
) -> Result<Vec<f64>, FitError>

Predict linear predictors with original-scale parameters without fitting preprocessing. Binomial predictions are logits.

Errors

Reject mismatched dimensions, invalid entries, or nonfinite arithmetic.

preprocessing

Source: src/fit/hybrid.rs

pub fn preprocessing(&self) -> &Preprocessing

Training normalization retained for interpreting optimization values.

screening_diagnostics

Source: src/fit/hybrid.rs

pub fn screening_diagnostics(&self) -> &shrinkage::ScreeningDiagnostics

Work and certified zero restrictions in the original feature space.

step_policy

Source: src/fit/hybrid.rs

pub fn step_policy(&self) -> HybridStep

Curvature policy selected for this fit.

step_size

Source: src/fit/hybrid.rs

pub fn step_size(&self) -> f64

Step used by final full-space diagnostics. Gaussian fits use the final PG curvature; logistic fits use a fixed global logistic curvature bound.

stopping_criterion

Source: src/fit/hybrid.rs

pub fn stopping_criterion(&self) -> StoppingCriterion

Selected full-space convergence criterion.

stopping_threshold

Source: src/fit/hybrid.rs

pub fn stopping_threshold(&self) -> f64

Absolute mapping threshold used to certify convergence.

stopping_value

Source: src/fit/hybrid.rs

pub fn stopping_value(&self) -> f64

Full mapping infinity norm, including the free-intercept derivative, evaluated for the original loss at the returned parameters using Self::step_size. Logistic surrogate checks never certify convergence.

termination

Source: src/fit/hybrid.rs

pub fn termination(&self) -> Termination

Whether the full mapping passed or the iteration budget expired.

working_set_diagnostics

Source: src/fit/hybrid.rs

pub fn working_set_diagnostics(&self) -> &shrinkage::WorkingSetDiagnostics

Work performed by heuristic feature selection, separate from safe screening.

clone

Source: src/fit/hybrid.rs

fn clone(&self) -> Self

fmt

Source: src/fit/hybrid.rs

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

predict

Source: src/cross_validation.rs

fn predict<M: RawColumns<f64> + ?Sized>(
    &self,
    design: &M
) -> Result<Vec<f64>, FitError>

termination

Source: src/cross_validation.rs

fn termination(&self) -> Termination