RandomSVDLazyMatrix: Compute randomized SVD.

In ekernf01/MatrixLazyEval: Lazy evaluation for cheap storage and fast operations with sparse matrices

Compute randomized SVD.

Description

Compute randomized SVD.

Usage

RandomSVDLazyMatrix(
  M,
  ncomp = 5,
  n_iter_spectrum_flatten = 2,
  nproj_left = min(8 * ncomp, dim(M)[1]),
  nproj_right = min(8 * ncomp, dim(M)[2])
)

Arguments

M	object of class LazyMatrix.
ncomp	number of components to compute
n_iter_spectrum_flatten	How many times to multiply by M^TM to emphasize the larger singular values when forming the projectors. If you don't know what this is, don't change it.
nproj_left	dimension of random column space to project into
nproj_right	dimension of random row space to project into

ekernf01/MatrixLazyEval documentation built on July 24, 2022, 2:42 a.m.