Self Optimising Statistical Calculator

[John Darrington <john@darrington.wattle.id.au>]

[-- Attachment #1: Type: text/plain, Size: 1555 bytes --]

I'm happy to announce version 0.0 of a hitherto unpublished Guile project.

This projectlet comprises approximately 300 lines of Scheme and
some guile interfaces to functions from the libraries of GNU PSPP 
(a statistical analysis program).

However the purpose is not to provide Guile bindings to PSPP.  Instead,
it provides a framework for the optimal computation of statistical
calculations. It draws upon PSPP's proven and efficient methods
of iterating through very large datasets, implements a cache to store the
results of intermediate calculations and uses Guile's flexibility of programatic
expression to implicitly determine dependencies between calculations.
The result is a system to automatically determine the minimum 
required number of iterations necessary to calculate a set of statistical
values and the optimal order in which to perform those calculations.

Programming input to the system is in the form of scheme expressions 
describing the calculations, referencing intermediate calculations where
necessary.  This system provides a framework to allow statisticians to write
implementations of algorithms, which can be optimally performed, without 
the implementor having to concentrate on optimisation details.


Get the source from 

git://de.cellform.com/pspp-experimental 


For examples of operation, see the file run.scm


-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://keys.gnupg.net or any PGP keyserver for public key.


[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]