Struct TensorFeatures 

pub struct TensorFeatures {
    pub mean: f32,
    pub std: f32,
    pub min: f32,
    pub max: f32,
    pub kurtosis: f32,
    pub skewness: f32,
    pub sparsity: f32,
    pub l1_norm: f32,
    pub l2_norm: f32,
    pub inf_norm: f32,
    pub nan_count: f32,
    pub inf_count: f32,
}

12-dimensional feature vector for tensor anomaly detection.

Per Dr. Popper’s audit on Curse of Dimensionality: “Ensure you have sufficient data (n >> p) before enabling Mahalanobis distance checks, or use a regularized covariance estimator.”

Fields

mean: f32

Mean: E[x]

std: f32

Standard deviation: sqrt(Var[x])

min: f32

Minimum value: min(x)

max: f32

Maximum value: max(x)

kurtosis: f32

Kurtosis: E[(x-μ)⁴]/σ⁴ - 3 (excess kurtosis)

skewness: f32

Skewness: E[(x-μ)³]/σ³

sparsity: f32

Sparsity: |{x: x=0}| / n

l1_norm: f32

L1 norm: Σ|x|

l2_norm: f32

L2 norm: sqrt(Σx²)

inf_norm: f32

Infinity norm: max(|x|)

nan_count: f32

NaN count: |{x: isnan(x)}|

inf_count: f32

Inf count: |{x: isinf(x)}|

Implementations

impl TensorFeatures

pub fn from_data(data: &[f32]) -> TensorFeatures

Extract features from tensor data

pub fn to_vec(&self) -> Vec<f32>

Convert to feature vector for ML algorithms

pub fn has_jidoka_violation(&self) -> Option<JidokaViolation>

Check for immediate Jidoka stop conditions

Trait Implementations

impl Clone for TensorFeatures

fn clone(&self) -> TensorFeatures

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TensorFeatures

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

Formats the value using the given formatter.

impl Default for TensorFeatures

fn default() -> TensorFeatures

Returns the “default value” for a type.