Struct McdFit 

pub struct McdFit {
    pub location: Array1<f64>,
    pub scatter: Array2<f64>,
    pub distances: Array1<f64>,
    pub weights: Array1<f64>,
    pub raw_location: Array1<f64>,
    pub raw_scatter: Array2<f64>,
    pub support: Vec<usize>,
    pub coverage: usize,
    pub objective: f64,
    pub breakdown_point: f64,
}

A fitted Minimum Covariance Determinant estimate.

Like crate::regression::LtsFit, MCD carries more than the shared ScatterFit triple, so it is its own type, but it does implement RobustScatter, reporting the reweighted (RMCD) location/scatter through it. The raw min-determinant estimate and the retained subset remain available as fields for auditing.

Fields

location: Array1<f64>

Reweighted (RMCD) location: the primary estimate.

scatter: Array2<f64>

Reweighted (RMCD) covariance: the primary estimate.

distances: Array1<f64>

Robust Mahalanobis distances w.r.t. the primary (location, scatter).

weights: Array1<f64>

Reweighting weights: 1.0 for observations retained by the χ² cutoff, 0.0 for those rejected (all 1.0 when reweighting is disabled).

raw_location: Array1<f64>

Raw (consistency-corrected) MCD location: the best h-subset mean.

raw_scatter: Array2<f64>

Raw (consistency-corrected) MCD covariance: the min-determinant subset covariance times the consistency factor.

support: Vec<usize>

The retained min-determinant h-subset (ascending indices).

coverage: usize

Coverage h.

objective: f64

The MCD objective at the fit: log det of the raw (uncorrected) subset covariance.

breakdown_point: f64

Breakdown point (n − h + 1)/n.

Implementations

impl McdFit

pub fn location(&self) -> &Array1<f64>

Reweighted (RMCD) location.

pub fn scatter(&self) -> &Array2<f64>

Reweighted (RMCD) covariance.

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

The raw min-determinant h-subset (ascending indices).

pub fn breakdown_point(&self) -> f64

Breakdown point.

Trait Implementations

impl Clone for McdFit

fn clone(&self) -> McdFit

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for McdFit

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

Formats the value using the given formatter.

impl RobustScatter for McdFit

fn location(&self) -> &Array1<f64>

Robust location estimate μ̂.

fn scatter(&self) -> &Array2<f64>

Robust scatter estimate Σ̂.

fn distances(&self) -> &Array1<f64>

Robust Mahalanobis distances of the fitted rows.

fn distance_cutoff(&self, quantile: f64) -> f64

The distance cutoff √χ²_{p, quantile} (e.g. quantile = 0.975) against which robust Mahalanobis distances are compared to flag outliers.

fn outliers(&self, quantile: f64) -> Vec<bool>

Flag each fitted observation as an outlier when its robust distance exceeds distance_cutoff(quantile). Under a p-variate Gaussian the squared distances are χ²_p, so quantile = 0.975 flags ≈ 2.5% of clean data.