s-news
[Top] [All Lists]

Re: [S] Testreport: Math. & stat. software for data analysis

To: bates@stat.wisc.edu, stst@informatik.uni-frankfurt.de
Subject: Re: [S] Testreport: Math. & stat. software for data analysis
From: John Maindonald <john.maindonald@anu.edu.au>
Date: Wed, 21 Apr 1999 08:46:51 +1000 (EST)
Cc: s-news@wubios.wustl.edu
Sender: owner-s-news@wubios.wustl.edu
Doug Bates wrote
> >>>>> "Stefan" == Stefan Steinhaus <stst@informatik.uni-frankfurt.de> writes:
> 
>   Stefan> I would like to announce the third edition of my testreport
>   Stefan> "Comparison of mathematical programs for data analysis"
>   Stefan> which has just been finished. The testreport which is
>   Stefan> written in English is available as PDF
> 
>   Stefan> file (Acrobat Reader) at the following website :
> 
>   Stefan> http://www.informatik.uni-frankfurt.de/~stst/ncrunch/index.html
> 
.....
> By the way, it is not clear why one would want to calculate a
> determinant in the first place.  The value usually came out to be
> Infinity in my tests because it was on the order of 5^1000 for a
> 1000x1000 matrix of uniform (0,1) random numbers.

I am concerned that the comparisons should do more to distinguish the
trivial or archaic from what is consequential.  A quick scan reveals
the following:
1) An identify matrix (2.2) would be trivial to implement.  Does
anyone miss it?
2) A matrix trace (2.2) would be trivial to implement.
3) The Crout decomposition (2.2) would be trivial to implement.
4) Isn't Bernoulli (2.5) just a Binomial with n=1?
5) Surely S-PLUS does have a quantile function (2.6).
6) The sign test is trivial to implement using pbinomial.
7) A Weibull plot would be trivial to implement; is it worth adding
as a specific form of plot?

Several other functions under 2.5 fall into the same category.

In several of these cases (1, 3, 4, 6 at least) one would just be
adding clutter, in my view.  There may be a case for notes somewhere
on how to implement these functions, and even listings in the help.

John Maindonald               email : john.maindonald@anu.edu.au        
Statistical Consulting Unit,  phone : (6249)3998        
c/o CMA, SMS,                 fax   : (6249)5549  
John Dedman Mathematical Sciences Building
Australian National University
Canberra ACT 0200
Australia
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu.  To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message:  unsubscribe s-news

<Prev in Thread] Current Thread [Next in Thread>