RE: RE: [Matlisp-users] Matlisp on ACLv6 works here :)

[Ernst v. W. <ev...@in...>]

Hello,

   "Good.  You're not the first user, but there
   "are not too many windows developers using Matlisp.
OK -- had no real choice there :-)
   "
   "I'd be happy to get feedback from you as you encounter
   "problems.  Keep in mind, especially, that some
   "functions will not be available on windows, for
   "example, discrete fourier transforms and some
   "special purpose functions.
Whenever I need discrete Fourier transforms and succeed to call it form
Lisp, I will inform you.

   "
   "What are you doing that so heavily relies on SVD?
Finding orthogonal coordinate systems for non-symmetric, possibly
rectangular matrices derived from data to facilitate application of
geometric algorithms.  I would like to try these algorithms for very large
datasets to join or separate subsets of data and to find patterns on data.
A sort of data mining, if you like, but using classical multivariate
statistics combined with analysis by algorithm.  Some years ago I found out
that this can be done surprisingly well (to get into the algorithmic aspect
I downloaded Matlisp in ACL5), but due to circumstances beyond me I return
to that subject only now.  It is a sort of rainy weekend project, but it
rains a lot here.
Before coming to the decompositions, however, quite a bit of code must be
written that doesn't use Matlisp -- it will be straightforward (I hope)
Lisp.

Of course, there SAS, SPSS, S+ and friends, but these do not have the
functionality that I am looking for.

Kind regards,

Ernst