Struct fetish_lib::inverse_schmear::InverseSchmear

pub struct InverseSchmear {
    pub mean: Array1<f32>,
    pub precision: Array2<f32>,
}

Represents a (multivariate) probability distribution by its mean and its precision matrix, which is the inverse of covariance. See also crate::schmear::Schmear, crate::func_schmear::FuncSchmear

Fields

mean: Array1<f32>

precision: Array2<f32>

Implementations

impl InverseSchmear

pub fn compress(
    &self,
    expansion_mat: ArrayView2<'_, f32>
) -> CompressedInverseSchmear

Given a matrix M representing a linear transformation from a space of smaller dimension to the dimension of this InverseSchmear, yields the CompressedInverseSchmear representing the quadratic form x -> (u - Mx)^T P (u - Mx), where u and P are this InverseSchmear’s mean vector and precision matrix, respectively

pub fn rescale_spread(&self, scale_fac: f32) -> InverseSchmear

Scales the precision matrix by the passed scale factor

pub fn sq_mahalanobis_dist(&self, vec: ArrayView1<'_, f32>) -> f32

Computes the squared Mahalanobis distance (u - x)^T P (u - x) of the argument, where u and P are this InverseSchmear’s mean vector and precision matrix, respectively, and x is the passed vector.

pub fn zero_precision_from_vec(vec: Array1<f32>) -> InverseSchmear

Creates a InverseSchmear with the given mean and an all-zero precision matrix.

Trait Implementations

impl Clone for InverseSchmear

fn clone(&self) -> InverseSchmear

Returns a copy of the value.

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

Performs copy-assignment from source.