Click or drag to resize

TruncatedSVDBlockKrylovSvd Method

Computes a truncated SVD using the Block Krylov method.

Namespace: Altaxo.Calc.LinearAlgebra.Double.Factorization
Assembly: AltaxoCore (in AltaxoCore.dll) Version: 4.8.3572.0 (4.8.3572.0)
Syntax
C#
Copy
public static (Matrix<double> U, Vector<double> S, Matrix<double> VT) BlockKrylovSvd(
	Matrix<double> A,
	int k,
	int oversampling = 10,
	int powerIterations = 2
)

Parameters

A  MatrixDouble
Input matrix.
k  Int32
Target rank (number of singular triplets to return).
oversampling  Int32  (Optional)
Additional sampling dimensions to improve accuracy.
powerIterations  Int32  (Optional)
Number of Krylov iterations (power iterations in the Block Krylov space).

Return Value

ValueTupleMatrixDouble, VectorDouble, MatrixDouble
The truncated SVD (U, S, Vt), where A ≈ U * diag(S) * Vt.
Remarks
Reference: https://people.cs.umass.edu/~cmusco/personal_site/pdfs/blockKrylov.pdf
See Also