• Home
• Julia

svds

..  svds(A; nsv=6, ritzvec=true, tol=0.0, maxiter=1000) -> (left_sv, s, right_sv, nconv, niter, nmult, resid)

``svds`` computes largest singular values ``s`` of ``A`` using Lanczos or Arnoldi iterations.
Uses :func:`eigs` underneath.

Inputs are:

* ``A``: Linear operator. It can either subtype of ``AbstractArray`` (e.g., sparse matrix) or duck typed. For duck typing ``A`` has to support ``size(A)``, ``eltype(A)``, ``A * vector`` and ``A' * vector``.
* ``nsv``: Number of singular values.
* ``ritzvec``: Whether to return the left and right singular vectors ``left_sv`` and ``right_sv``, default is ``true``. If ``false`` the singular vectors are omitted from the output.
* ``tol``: tolerance, see :func:`eigs`.
* ``maxiter``: Maximum number of iterations, see :func:`eigs`.

**Example**::

   X = sprand(10, 5, 0.2)
   svds(X, nsv = 2)

Examples

The svds function in Julia computes the largest singular values s of a linear operator A using Lanczos or Arnoldi iterations. It uses the eigs function underneath.

Inputs:

• A: Linear operator. It can be a subtype of AbstractArray (e.g., sparse matrix) or duck-typed. For duck typing, A must support size(A), eltype(A), A * vector, and A' * vector.
• nsv: Number of singular values.
• ritzvec: Whether to return the left and right singular vectors left_sv and right_sv. Default is true. If set to false, the singular vectors are omitted from the output.
• tol: Tolerance parameter. See eigs documentation for details.
• maxiter: Maximum number of iterations. See eigs documentation for details.

Example:

X = sprand(10, 5, 0.2)
svds(X, nsv = 2)

This example computes the two largest singular values of the sparse matrix X using svds.

See Also

User Contributed Notes

Add a Note

The format of note supported is markdown, use triple backtick to start and end a code block.