Re: Multivariate pdf of a normal distribution

[Andy Adler]

On Sun, 6 Nov 2005, Mike Miller wrote:

> When inv(A) exists,
>
> inv(A') = inv(A)'
>
> because the inverse equals the transpose of the adjoint divided by the
> determinant.  The adjoint of the transpose equals the transpose of the
> adjoint and the determinant is the same for A and A'.  The identity really
> follows from the fact that the determinant is not affected by the
> transpose operation -- the adjoint is a collection of determinants.
>
> I agree that this must be in most matrix algebra books.

Yes, that is true in theory. But the error is to assume that all
mathematical identities remain intact when done on a finite
precision computer.

Try it:
A= rand(2); inv(A')- inv(A)'
ans =

    7.1054e-15   -7.1054e-15
    8.8818e-16    0.0000e+00
--
Andy Adler <address@hidden> 1(613)562-5800x6218



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------