Module svd 

Singular Value Decomposition (SVD)

The matrix A is decomposed as A = U * S * V^T where:

• s contains the singular values (1D vector)
• u contains the left singular vectors (matrix U)
• vt contains the transposed right singular vectors (matrix V^T)

Singular values are mathematically real. Backends choose the scalar type used to represent them through SVD::SingularValue.

// ----- Singular Value Decomposition (SVD) -----
use mdarray_linalg::svd::SVDDecomp;
use mdarray_linalg::prelude::*; // Import traits anonymously
use mdarray_linalg_backend::Backend; // Use the real backend here, Lapack, Faer, ...

let bd = Backend::default();
let SVDDecomp { s, u, vt } = bd.svd(&mut a.clone()).expect("SVD failed");
// Or the shorter ...
let SVDDecomp { s, u, vt } = bd.svd(&mut a.clone()).expect("SVD failed");

Structs

SVDDecomp

Holds the results of a singular value decomposition, including singular values and the left and right singular vectors.

Enums

SVDError

Error types related to singular value decomposition

Traits

SVD

Singular value decomposition for matrix factorization and analysis