Struct BinomialMultiFitInputs 

pub struct BinomialMultiFitInputs<'a> {
    pub design: ArrayView2<'a, f64>,
    pub y: ArrayView2<'a, f64>,
    pub penalty: ArrayView2<'a, f64>,
    pub lambdas: ArrayView1<'a, f64>,
    pub row_weights: Option<ArrayView1<'a, f64>>,
    pub fisher_w_override: Option<ArrayView3<'a, f64>>,
    pub max_iter: usize,
    pub tol: f64,
}

Inputs for fit_penalized_binomial_multi.

Fields

design: ArrayView2<'a, f64>

Design matrix X ∈ ℝ^{N×P} (one row per observation, shared across all response columns).

y: ArrayView2<'a, f64>

Multi-column binomial response Y ∈ ℝ^{N×K}. Each column is treated as an independent binomial-logit response, so every entry must be a binomial proportion in [0, 1] (hard {0, 1} Bernoulli labels and soft proportions / probabilities alike). Entries outside [0, 1] are rejected because the per-entry log-likelihood is then unbounded in η.

penalty: ArrayView2<'a, f64>

Shared smoothing penalty S ∈ ℝ^{P×P} (symmetric, PSD).

lambdas: ArrayView1<'a, f64>

Per-response smoothing parameter λ_a (length K).

row_weights: Option<ArrayView1<'a, f64>>

Optional per-row weights (length N); None ⇒ uniform 1.0.

fisher_w_override: Option<ArrayView3<'a, f64>>

Optional per-row Fisher-block override, shape (N, K, K). The K binomial-logit columns are fit independently, so only the per-column diagonal [n, a, a] is consumed as the curvature w_n μ_a(1 − μ_a); off-diagonals must be zero (enforced at the FFI boundary) since a non-zero cross term cannot be represented by the separable per-column solve. The gradient/residual path stays analytic — this is a curvature-only override (issue #349). Diagonal entries must be finite and non-negative.

max_iter: usize

Maximum Newton iterations per response column; recommend 50.

tol: f64

Relative-step convergence tolerance; recommend 1e-7.

Trait Implementations

impl<'a> Clone for BinomialMultiFitInputs<'a>

fn clone(&self) -> BinomialMultiFitInputs<'a>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<'a> Debug for BinomialMultiFitInputs<'a>

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

Formats the value using the given formatter.