[R] what does it mean when "lm.gls" says that the weight matrix has wrong dimension?

[Peter Dalgaard]

Michael <comtech.usa at gmail.com> writes:

> If first fit my data column V1 to column V2 using normal "lm" fitting,
> 
> call it "fit1",
> 
> then I used "acf(fit1$residuals, type='cov', 40) " function to obtain the
> autocovariance of the residuals,
> 
> and then constructed a autocovariance matrix, I chose it to be 40x40.
> 
> Call this autocovariance matrix B,
> 
> I then use the following "lm.gls" function to fit using the above weight
> matrix:
> 
> But what's wrong with my weight matrix?
> 
> > fit8=lm.gls(V1~V2, data=data1, W=B, inverse=TRUE);
> Error in lm.gls(V1 ~ V2, data = data1, W = B, inverse = TRUE) :
>         dim(W) is not correct

(That's in the MASS package, and the authors of the book it supports
like you to say so. The rest of us would like to be spared the
help.search step...)

I would guess that either V1 does not have length 40 or dim(B) isn't
40x40 even though you intended it to be. 

-- 
   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45) 35327907