The may sound like a strange question but I was wondering if a Pseudoinverse of a matrix could be found using SVD whether there was a graphical modelling equivalent that could be used to estimate the inverse of a matrix.
Asked
Active
Viewed 53 times
1 Answers
0
If $A = U \Sigma V^{*}$ then $A^{+} = V \Sigma^{+} U^{*}$ where
$\Sigma^{+}(i,i) = \frac{1}{\sigma(i)} $ when $\sigma(i) \neq 0$ and 0 everywhere else.
Sid
- 2,617
-
Is there a name for this so I can look it up as a reference? – orbital Jul 24 '14 at 03:24
-
is sigma the standard deviation? – orbital Jul 24 '14 at 03:39
-
$\sigma$ are the singular values of the original matrix. Look up Moore-Penros Pseudoinverse – Sid Jul 24 '14 at 04:23
-
I have been using the Moore-Penros Pseudoinverse. My question was asking whether a Bayesian method existed? – orbital Jul 24 '14 at 05:09