s-news
[Top] [All Lists]

Re: [S] Loosen criteria for matrix singularity

To: Haicheng Tang <haicheng@umich.edu>
Subject: Re: [S] Loosen criteria for matrix singularity
From: Prof Brian D Ripley <ripley@stats.ox.ac.uk>
Date: Fri, 26 Feb 1999 07:26:46 +0000 (GMT)
Cc: s-news@wubios.wustl.edu, bugs@statsci.com
In-reply-to: <Pine.GSO.3.96.990226011515.20337I-100000@sunm4048at.sph.umich.edu>
Sender: owner-s-news@wubios.wustl.edu
On Fri, 26 Feb 1999, Haicheng Tang wrote:

> The following matrix looks OK for me. But Splus
> considers it singular when I tried to invert it. 
> Is there any way to loosen its criteria for 
> singularity? Any help is appreciated!
> 
>       [,1]     [,2] 
> [1,] 6.944678 6.940456
> [2,] 6.940456 6.940456

How are you trying to do this?

> A <- matrix(c(6.944678, 6.940456, 6.940456, 6.940456), 2, 2)
> solve(A)
          [,1]      [,2] 
[1,]  236.8546 -236.8546
[2,] -236.8546  236.9987
> svd(A)
$d:
[1] 13.883023321  0.002110679

It is far from singular, if somewhat ill-conditioned. 

BTW, I asked kappa for the condition number (which should be
13.883023321/0.002110679, approximately 6500) and get
> kappa(A)
[1] 3.332253

Beware that kappa() does not give sensible answers given a matrix unless it
is (upper?) triangular. That seems like a serious bug to me (still there in
5.0).

-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-----------------------------------------------------------------------
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>