Enum Solver 

pub enum Solver {
    Lbfgs,
    NewtonCholesky,
}

The optimization algorithm requested for fitting a GLM, mirroring sklearn’s solver constructor parameter (sklearn/linear_model/_glm/glm.py:140-145, StrOptions({"lbfgs", "newton-cholesky"}), default "lbfgs").

Implementation note (R-DEV-7 — Rust analog). ferrolearn fits all GLMs via IRLS / Fisher-scoring (fn fit_glm_irls); the solver parameter is accepted for scikit-learn API parity (R-DEV-2) and selects the requested optimizer’s contract. The penalized GLM objective is convex, so IRLS converges to the same minimizer as both sklearn’s lbfgs (scipy L-BFGS-B) and newton-cholesky (Newton-Raphson with an inner Cholesky solve) — all three are descent methods on one convex objective. Therefore the observable contract (coef_ / intercept_) matches sklearn for either solver value, and ferrolearn does not vary the numerical path between them (verified live: PoissonRegressor(alpha=0.5) gives the same fitted attributes to ~1e-9 for lbfgs and newton-cholesky).

The type system constrains valid values to the two sklearn options, mirroring the role of sklearn’s StrOptions parameter constraint.

Variants

Lbfgs

L-BFGS-B (sklearn’s default "lbfgs"): a quasi-Newton optimizer on the penalized loss + gradient (glm.py:263-284). In ferrolearn the fit is performed by IRLS, which reaches the same convex optimum (R-DEV-7).

NewtonCholesky

Newton-Cholesky (sklearn’s "newton-cholesky"): Newton-Raphson steps with an inner Cholesky solve, equivalent in exact arithmetic to iterated reweighted least squares (glm.py:72-78, :285-296). In ferrolearn the fit is performed by IRLS, which reaches the same convex optimum (R-DEV-7).

Trait Implementations

impl Clone for Solver

fn clone(&self) -> Solver

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for Solver

impl Debug for Solver

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

Formats the value using the given formatter.

impl Eq for Solver

impl PartialEq for Solver

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for Solver