lapsvdcusv#

Purpose#

Computes the singular value decomposition of a real or complex rectangular matrix, returns compact U and v.

Format#

{ u, s, v } = lapsvdcusv(x)#
Parameters:

x (MxN matrix) – real or complex rectangular matrix.

Returns:

  • u (Mxmin(M,N) matrix) – left singular vectors.

  • s (min(M,N)xN matrix) – singular values.

  • v (NxN matrix) – right singular values.

Examples#

// Assign x matrix
x = { 2.143 4.345 6.124,
      1.244 5.124 3.412,
      0.235 5.657 8.214 };

// Compute singular value decomposition
{ u,s,v } = lapsvdusv(x);

print u;

-0.55531277  0.049048431 0.83019394
-0.43090168  0.83684123 -0.33766923
-0.71130266 -0.54524400 -0.44357356

print s;

13.895868 0.0000000 0.0000000
0.0000000 2.1893939 0.0000000
0.0000000 0.0000000 1.4344261

print v;

-0.13624432  -0.62209955 -0.77099263
 0.46497296   0.64704876 -0.60425826
 0.87477862  -0.44081748  0.20110275

Remarks#

lapsvdcusv() computes the singular value decomposition of a real or complex rectangular matrix. The SVD is

x = usv'

where v is the matrix of right singular vectors. lapsvdcusv() is based on the LAPACK drivers DGESVD and ZGESVD. Further documentation of these functions may be found in the LAPACK User’s Guide.

See also

Functions lapsvds(), lapsvdusv()

On this page