Struct linfa_linalg::lobpcg::TruncatedSvd

pub struct TruncatedSvd<A: NdFloat, R: Rng> { /* private fields */ }

Truncated singular value decomposition

Wraps the LOBPCG algorithm and provides convenient builder-pattern access to parameter like maximal iteration, precision and constrain matrix.

Implementations

impl<A: NdFloat + Sum, R: Rng> TruncatedSvd<A, R>

pub fn new_with_rng(
    problem: Array2<A>,
    order: Order,
    rng: R
) -> TruncatedSvd<A, R>

Create a new truncated SVD problem

Parameters

• problem: rectangular matrix which is decomposed
• order: whether to return large or small (close to zero) singular values
• rng: random number generator

impl<A: NdFloat + Sum, R: Rng> TruncatedSvd<A, R>

pub fn precision(self, precision: f32) -> Self

Set the required precision of the solution

The precision is, in the context of SVD, the square-root precision of the underlying eigenproblem solution. The eigenproblem-precision is used to check the L2 error of each eigenvector and stops its optimization when the required precision is reached.

pub fn maxiter(self, maxiter: usize) -> Self

Set the maximal number of iterations

The LOBPCG is an iterative approach to eigenproblems and stops when this maximum number of iterations are reached

pub fn decompose(self, num: usize) -> Result<TruncatedSvdResult<A>>

Calculate the singular value decomposition

Parameters

• num: number of singular-value/vector pairs, ordered by magnitude

Example

use ndarray::{arr1, Array2};
use linfa_linalg::{Order, lobpcg::TruncatedSvd};
use rand::SeedableRng;
use rand_xoshiro::Xoshiro256Plus;

let diag = arr1(&[1., 2., 3., 4., 5.]);
let a = Array2::from_diag(&diag);

let eig = TruncatedSvd::new_with_rng(a, Order::Largest, Xoshiro256Plus::seed_from_u64(42))
   .precision(1e-4)
   .maxiter(500);

let res = eig.decompose(3);

Trait Implementations

impl<A: Clone + NdFloat, R: Clone + Rng> Clone for TruncatedSvd<A, R>

fn clone(&self) -> TruncatedSvd<A, R>

Returns a copy of the value.

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

Performs copy-assignment from source.

impl<A: Debug + NdFloat, R: Debug + Rng> Debug for TruncatedSvd<A, R>

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

Formats the value using the given formatter.