Of all the matrix routines to be built-in, among the most useful would be
the computation of eigenvectors and eigenvalues. Better yet, of course,
would be generalised, singular value decomposition (SVD). With SVD and a
good matrix inversion routine built-in, the remaining routines (and full,
multivariate stats computations) could easily be coded as simple scripts
without much of a performance hit.
> on 11/12/2000 4:28 PM, "Monte Goulding" <[EMAIL PROTECTED]> wrote:
>
>>
>> OK I'll sent the stats library as soon as I put it all together. MetaCard
>> corp. has proved itself to me as a company that will do it's best to
>> accomodate users requests so try emailing [EMAIL PROTECTED] if you have
>> specific algorithmns you would like to see as part of the engine.
>>
>> PS It sounds like you won't be learning much about stats functions from me
>> but the prob F approximation was pretty hard to find (most references to
>> this function actualy calculate it the long way which is extremley
>> complicated with only very minor improvements in accuracy.
>>
>> Regards
>>
>> Monte
--
This was too good not to pass along...
My wife recently gave an intro stat quiz which some questions on tables
showing whether or not a convicted murder was given a death penalty, broken
down by the race of the perpetrator and race of the victim (Data are on page
187 of Moore & McCabe's book). The final question asked students to discuss
how the data provided an example of Simpson's paradox.
You've probably guessed it by now -- she had one student who wrote a long
paragraph about OJ Simpson's legal difficulties as a black man being accused
of murdering a white woman. Will future generations of statistics students
automatically thinks that this is the Simpson in Simpson's paradox?
- Robin Lock
John R. Vokey, Ph.D.
Chair
Department of Psychology and Neuroscience
University of Lethbridge
Archives: http://www.mail-archive.com/[email protected]/
Info: http://www.xworlds.com/metacard/mailinglist.htm
Please send bug reports to <[EMAIL PROTECTED]>, not this list.
