Re: Statistics/graphing programs for scientists?
On 12-Mar-99 Richard Lyon wrote:
>
> No free unix program is going to provide the sort of on-line help, user
> interface or range of analysis methods that comes with statistica. Unix
> applications like octave are not going to provide any better
> statistical analysis than a spreadsheet.
You may be right about statistica (for all I know). However, your last
sentence is so seriously misleading that it must be corrected.
First, octave is no "spreadsheet" but close to being a clone of the
MatLab core executable. Therefore it is a highly programmable matrix- and
array-oriented general-purpose numerical analysis package and it is
exremely powerful. What it lacks relative to MatLab is the range of
"toolboxes" which add specialised pre-programmed functionality. Though it
does have its own Statistics "toolbox" this is less complete than
MatLab's. However, nothing whatever prevents a knowledgeable user from
programming their own very sophisticated statistical analysis, far beyond
what any spreadsheet known to me could achieve.
Scilab is another "free unix program" with similar general capabilities
to octave and more powerful in certain respects, with excellent graphical
resources.
In addition there are some powerful "free unix programs" which specialise
in Statistics. "R" and "XLispStat" are important examples. An important
though more specialised (based on Monte Carlo approaches to Bayesian
statistics) package is BUGS, also available for Linux.
And, lest your last sentence should give the impression that only
programs with spreadsheet-like capability are available for UNIX
generally, don't forget that almost all the major programs exist
in UNIX versions (MatLab, S-plus, SPSS, Mathematica, SAS, maybe also
Statistica, and so on) and many of them have been ported to Linux
as well (MatLab being early on the Linux scene).
Regards,
Ted.
--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding@nessie.mcc.ac.uk>
Date: 13-Mar-99 Time: 15:36:55
------------------------------ XFMail ------------------------------
Reply to: