[R] Matrix multiplication precision

[Steve Lianoglou]

Hi Douglas,

> And the big lesson, of course, is the first rule of numerical linear
> algebra., "Just because a formula is written in terms of the inverse
> of a matrix doesn't mean that is a good way to calculate the result;
> in fact, it is almost inevitably the worst way of calculating the
> result".  You only calculate the inverse of a matrix after you have
> investigated all other possible avenues for calculating the result and
> found them to be fruitless.

As a relative noob to the nitty gritty details of numerical linear  
algebra, could you elaborate on this a bit (even if it's just a link  
to a book/reference that explains these issues in more detail)?

Where/what else should we be looking to do that would be better? Or  
are there really no general rules, and the answer just depends on the  
situation?

Thanks,

-steve

--
Steve Lianoglou
Graduate Student: Physiology, Biophysics and Systems Biology
Weill Medical College of Cornell University

Contact Info: http://cbio.mskcc.org/~lianos/contact