From aldo.solari@teleline.es Tue Jun 27 20:15:04 2000
Date: Thu, 22 Jun 2000 11:09:08 +0100
From: "Aldo P. Solari [APS]"
To: R. Allen Curry , fish-ecology@helios.ulpgc.es
Subject: re: stats programs
There are several packages for "data mining" which have different
features. In my opinion,
SyStat > v 5 (now on v. 9, see www.spss.com) is a very good
package for classical stats; it is easy to use, it requires no
programming knowledge, it has very good graphics, it uses Basic
for programming macros and it comes with a "state-of-the-art"
handbook; there is a distribution list for support. If I had
funds for just one stats package, I would definitely go for
Systat;
SPSS > v 6.1.3 (now on v. 10, see www.spss.com) is a good program
to handle relatively large data bases (on around 10-12 million
cells it runs _fast_ on a simple P5 at 100 MHz); graphics is
still a weak side in this package (compared to Systat); also, be
aware if you aim to run spectarl analyses: version 8 includes it
but which was taken away in versions 9 and 10; the drawback of
this package is the "add-on" marketing strategy SPSS has adopted;
S+ (see www.mathsoft.com) is a heavy duty package for data
mining; however, it requires significantly more input from the
user; an advantage is that it includes certain types of
regressions and fractional series analyses which are not included
by-default in any other stats package;
Dataplore (www.dataplore.de) appears to be a great program to run
non-linear data analyses; it is aimed to detect chaos in time
series and it includes documentation in HTML format;
For juniors, I would definitely suggest StatView (see
www.statview.com), an extremely easy to use classical stats
package with great documentation in PDF format. Stats tests are
easily explained in the docs. A very nice approach, in my
opinion. This program was originally developed for the Macintosh
and the pc version includes the highly user-friendly.
On the Linux side, there are several free data analyses packages
which should be seriously considered. You may follow the
www.linux.org link and search for SAL (Scientific Applications
for Linux).
Bottom line: Systat may be the best investment so far (pc, mac);
for junior scientists, StatView may be the way to go (pc, mac).
Intel based machines may run Linux and a series of free stats
analyses packages.
More comments are welcome.
Neither I nor fish-ecology is associated with any software
producer.
Cheers,
APS
----
Aldo P. Solari.
Fisheries Research Group, U. of Las Palmas
Box 550, 35017, Las Palmas, Canary Islands.
Mail-to[PGP key available on request]:
[academic/research matters]
[private matters]
Home page, http://www.ccbb.ulpgc.es/fish-ecology/solaris/
----
racurry@UNB.CA
RAC> What software are folks using to fit linear and non-linear curves
RAC> (including statistics on goodness of fit) to data these days?
********************************************************************
To unsubscribe: send an email to Majordomo@helios.ulpgc.es with the
command line 'unsubscribe fish-ecology your@email.here' (without the
quotations). Majordomo is case sensitive. If you have problems to
unsubscribe or other queries you may reach the FE-tream at :
fish-ecology-request@helios.ulpgc.es Thank you
********************************************************************