Mo logo [home] [lexicon] [problems] [tests] [courses] [auxiliaries] [notes] [staff] german flag
Mathematics-Online course: Linear Algebra - Normal Forms - Singular Value Decomposition

Pseudo-Inverse

[previous page] [next page] [table of contents][page overview]
By means of the singular value decomposition $ USV^*$ of an $ m\times n$ matrix $ A$, the solution of the minimization problem $ \Vert Ax-b\Vert _2\to\min$ can be expressed in the form

$\displaystyle x = A^+b,\quad A^+ = VS^+U^*, $

where $ A^+$ is the so-called pseudo-inverse of $ A$ (Moore-Penrose inverse), and $ S^+$ is an $ n\times m$ diagonal matrix of the form

$\displaystyle S^+ =\operatorname{diag}(1/s_1,\ldots,1/s_k,0,\ldots,0)\, , $

containing the inverse values of the singular ones.

(temporary unavailable)

(temporary unavailable)
  automatically generated 4/21/2005