[SciPy-user] Gram-Schmidt orthogonalization

[David Warde-Farley]

On 24-Nov-08, at 3:09 PM, nicky van foreest wrote:

> Hi David,
>
> I recall from the book numerical recipes that the Gramm Schmidt
> methods works terrible, numerically speaking. They provide some
> counterexamples too. It is better to use singular value decomposition,
> which is included in scipy too.

Hey Nicky,

You're right about Gram-Schmidt being nasty if you do it naively, but  
there is IIRC a more numerically stable variant of Gram-Schmidt, see
http://en.wikipedia.org/wiki/Gram–Schmidt_process#Algorithm

I just tend not to want to "roll my own" if I can help it, since stuff  
in SciPy is usually going to be better tested.

Cheers,

David