Struct BoostingConfig 

pub struct BoostingConfig {
    pub mstop: usize,
    pub nu: f64,
    pub nbasis: usize,
    pub order: usize,
    pub lfd_order: usize,
    pub lambda: f64,
    pub ncomp_x: usize,
    pub seed: u64,
}

Configuration for component-wise gradient boosting (FOSR, FoFR, GAMLSS, stability).

All base-learners use the same nbasis, order, lfd_order, and lambda to ensure equal effective degrees of freedom, preventing selection bias toward more flexible learners (see Pitfall 4 in RESEARCH.md).

Divergence from FDboost: Fixed nu and mstop rather than CV-based early stopping; the GCV path is tracked for diagnostic purposes but not used for stopping.

Fields

mstop: usize

Number of boosting iterations (must be ≥ 1).

nu: f64

Learning rate ν ∈ (0, 1] (FDboost default: 0.1).

nbasis: usize

Number of B-spline basis functions per base-learner (must be ≥ 4).

The actual number of basis functions is nknots + order where nknots = nbasis - order. With order = 4 (cubic), nbasis = 10 gives 6 interior knots.

order: usize

B-spline order (typically 4 for cubic splines).

lfd_order: usize

Penalty derivative order (typically 2 for roughness).

lambda: f64

Smoothing parameter λ > 0 for penalized base-learners.

ncomp_x: usize

Number of predictor FPC components for FoFR base-learners (REG-06-02).

seed: u64

RNG seed (used by stability selection and future extensions; unused in pure boosting).

Trait Implementations

impl Clone for BoostingConfig

fn clone(&self) -> BoostingConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for BoostingConfig

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

Formats the value using the given formatter.

impl PartialEq for BoostingConfig

fn eq(&self, other: &BoostingConfig) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for BoostingConfig