Struct TwoArmBackgroundFitResult 

pub struct TwoArmBackgroundFitResult {
    pub names: Vec<String>,
    pub amplitudes: Vec<f64>,
    pub amplitude_uncertainties: Option<Vec<f64>>,
    pub amplitude_at_bound: Vec<bool>,
    pub amplitudes_identifiable: bool,
    pub prediction: TwoArmCountPrediction,
    pub poisson_deviance: f64,
    pub deviance_per_dof: f64,
    pub n_informative: usize,
    pub converged: bool,
    pub iterations: usize,
}

Result of fitting fixed neutron signals plus background templates.

Fields

names: Vec<String>

Component names in the same order as amplitudes.

amplitudes: Vec<f64>

Fitted non-negative template amplitudes.

amplitude_uncertainties: Option<Vec<f64>>

Local one-sigma amplitude uncertainties from the expected (Fisher) information of the constrained objective at the solution.

Free amplitudes are conditioned on the constraint: the information is restricted to the free set before inversion, so an amplitude whose correlated partner is held on its bound is not reported with the marginal over a direction the constraint removed. An amplitude on its zero bound (see amplitude_at_bound) reports 1 / sqrt(I_jj), the one-sided curvature scale of the objective at the boundary, not a symmetric interval.

None when the fit did not converge, when the amplitudes are not identifiable, when the free block of the information matrix is singular, or when the caller’s PoissonConfig has compute_covariance == false. An individual entry is NaN when its variance is non-positive, or when its template is sensitive on a bin with zero expectation — the boundary of the Poisson support, where the expected information diverges and no regular estimate exists. A reported number is therefore never zero.

amplitude_at_bound: Vec<bool>

Whether each fitted amplitude sits on its non-negativity bound with the gradient still pushing it negative — a one-sided limit rather than an interior estimate.

amplitudes_identifiable: bool

Whether every named template amplitude is separately determined.

false means that at least two supplied shapes are linearly dependent: the total fitted background can be valid, but the individual amplitudes are not physically interpretable. In that case uncertainties are not reported.

prediction: TwoArmCountPrediction

Neutron signal, fitted background, and total for both arms.

poisson_deviance: f64

Poisson deviance for the concatenated open and sample arrays.

deviance_per_dof: f64

Deviance divided by n_informative - template_rank.

n_informative: usize

Concatenated bins that contribute to the deviance: the observation, the neutron signal, or at least one template is nonzero there.

This is not simply 2 * n_bins: a bin where all three are exactly zero yields identically zero deviance for any amplitude vector, so it cannot be a degree of freedom.

converged: bool

Whether the bounded optimizer converged.

iterations: usize

Joint iterations taken, including the post-convergence active-set polish; never exceeds the caller’s max_iter.

Trait Implementations

impl Clone for TwoArmBackgroundFitResult

fn clone(&self) -> TwoArmBackgroundFitResult

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TwoArmBackgroundFitResult

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

Formats the value using the given formatter.