Help System (web edition)

MAT SINGULARVALUE DECOMPOSITION OF A TO U,D,V
Alternative forms:
MAT SINGULAR VALUE DECOMPOSITION OF A TO U,D,V
MAT SVD OF A TO U,D,V

computes the singular value decomposition of A (m*n, m>=n)
A=U*D*V' where U is m*n and U'U=I, V is n*n and V'V=I and
D (n*n) is diagonal and consists of singular values.
To spare space, D will be saved as a column vector.

This decomposition has several applications. For example,
X=V*DINV(DV(D))*U' is a generalized inverse of A with properties
AXA=A, XAX=X, (AX)'=AX, (XA)'=XA.
See also MPINV?

  D = More information on matrix decompositions