[Openspace] Error: the inverse matrix is ill-condition!

[Roger Bivand]

On 29 Apr 2006, V.A. Ceccato wrote:

> Hello list
> 
> Every time I put all my covariates in my OLS model (around than 25, many 
> dummies), I get the following message from GeoDa: Error: the inverse matrix 
> is ill-condition! Why? The message disappears when I input fewer variables, 
> excluding the dummies.

I think you are including both the constant term and both dummies for a 
binary variable, or something like that, or some other set of variables 
that are perfectly correlated. Have you run the same regression outside 
GeoDa? Try that, and see what happens, and how to leave out at least one 
dummy category for each set of dummies if you have a constant term.

Roger

> 
> Ps. I get this message even when I don't use weight matrix.
> 
> Many thanks 
> 
> Vania Ceccato
> 
> _______________________________________________
> Openspace mailing list
> Openspace at sal.uiuc.edu
> http://sal.uiuc.edu/mailman/listinfo/openspace
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no