Struct GaussPointRobust 

pub struct GaussPointRobust {
    pub mu: f32,
    pub sigma: f32,
    pub mode: RobustManifoldMode,
}

A manifold point derived from sample data using a selectable estimator.

Wraps GaussPoint with explicit estimator provenance.

Examples

use dsfb_rf::fisher_geometry::{GaussPointRobust, RobustManifoldMode};
let samples = [0.1f32, 0.1, 0.1, 0.1, 5.0]; // impulsive outlier
let p = GaussPointRobust::from_samples_mad(&samples).unwrap();
assert!(p.sigma < 0.5); // outlier does not inflate scale

Fields

mu: f32

Robust location estimate (median).

sigma: f32

Robust scale estimate (1.4826 · MAD).

mode: RobustManifoldMode

Estimator mode used to produce this point.

Implementations

impl GaussPointRobust

pub fn from_samples_gaussian(samples: &[f32]) -> Option<Self>

Construct from samples using standard ML estimators (mean, std).

Returns None if samples is empty.

pub fn from_samples_mad(samples: &[f32]) -> Option<Self>

Construct from samples using MAD-regularized estimators.

Location: median of samples. Scale: 1.4826 × MAD(samples), floor 1e-9.

Robust against impulsive outliers (Alpha-stable noise, radar pulses). Returns None if samples is empty.

pub fn to_gauss_point(self) -> GaussPoint

Convert to a GaussPoint for use with fisher_rao_distance.

Trait Implementations

impl Clone for GaussPointRobust

fn clone(&self) -> GaussPointRobust

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for GaussPointRobust

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

Formats the value using the given formatter.

impl Copy for GaussPointRobust