The following module was proposed for inclusion in the Module List:
modid: Statistics::LSNoHistory
DSLIP: RdpOa
description: Least-squares regression w/o data history
userid: PLIAM (John Pliam)
chapterid: 6 (Data_Type_Utilities)
communities:
comp.lang.perl.announce
similar:
Statistics::OLS, Statistics::Descriptive, Statistics::GaussHelmert,
Statistics::Regression
rationale:
There is an inevitable trade-off between memory and roundoff
resistance for software which computes classic/standard
least-squares regression.
In one extreme, one must make an initial pass over the data in
order to compute the average, and then another pass, rescaling and
computing squares which might otherwise have blown-up. With this
approach, one must retain the entire data history, costing memory.
(see Press et al ref in pod).
In the other extreme, one must make intuitive guesses about the
averages, updating preliminary statistics once for each data point.
With this approach, one saves memory at the cost of increased risk
of roundoff error.
All existing CPAN modules for simple 2-dimensional least-squares
fall in the first extreme, while this module falls in the second.
Both have their uses...
enteredby: PLIAM (John Pliam)
enteredon: Sun Mar 2 20:01:35 2003 GMT
The resulting entry would be:
Statistics::
::LSNoHistory RdpOa Least-squares regression w/o data history PLIAM
Thanks for registering,
The Pause Team
PS: The following links are only valid for module list maintainers:
Registration form with editing capabilities:
https://pause.perl.org/pause/authenquery?ACTION=add_mod&USERID=95300000_9b06faf94d7c4c47&SUBMIT_pause99_add_mod_preview=1
Immediate (one click) registration:
https://pause.perl.org/pause/authenquery?ACTION=add_mod&USERID=95300000_9b06faf94d7c4c47&SUBMIT_pause99_add_mod_insertit=1
