Struct Svd 

pub struct Svd<T: Scalar, U = Vec<T>, S = Vec<<T as Scalar>::Real>> {
    pub size: [usize; 2],
    pub u: U,
    pub s: S,
    pub v: U,
    /* private fields */
}

The thin SVD of an m×n matrix: a = U diag(s) Vᴴ, where p is the smaller of m and n, U is m×p and V is n×p, both with orthonormal columns. s is real and in descending order. Without singular vectors, u and v are empty. svd returns one that owns its factors. SvdRef borrows them.

Fields

size: [usize; 2]

u: U

s: S

v: U

Implementations

impl<T: Scalar, U: AsRef<[T]>, S: AsRef<[T::Real]>> Svd<T, U, S>

pub fn new(size: [usize; 2], u: U, s: S, v: U) -> Self

The thin SVD of a matrix of size [m, n] with these factors.

pub fn as_ref(&self) -> SvdRef<'_, T>

A view of the factors, with no copies.

impl<T: Factor, U: AsRef<[T]>, S: AsRef<[T::Real]>> Svd<T, U, S>

pub fn solve(&self, b: &[T], k: usize) -> Vec<T>

The least-squares solution x of a x = b with the least norm, for the factored a and the m×k matrix b, as an n×k matrix. Singular values below ε times the larger of m and n times the largest singular value count as zero, as in NumPy’s lstsq.

pub fn pseudoinverse(&self) -> Vec<T>

The n×m pseudoinverse of the factored matrix, with the cutoff that solve uses.

Trait Implementations

impl<T: Clone + Scalar, U: Clone, S: Clone> Clone for Svd<T, U, S>

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<T: Copy + Scalar, U: Copy, S: Copy> Copy for Svd<T, U, S>

impl<T: Debug + Scalar, U: Debug, S: Debug> Debug for Svd<T, U, S>

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

Formats the value using the given formatter.

impl<T: PartialEq + Scalar, U: PartialEq, S: PartialEq> PartialEq for Svd<T, U, S>

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl<T: PartialEq + Scalar, U: PartialEq, S: PartialEq> StructuralPartialEq for Svd<T, U, S>