[RsR] rrcov CovMest

[Valentin Todorov]

Dear Ibrahim,

As I wrote you on Friday (write me if you did not receive my mail),
for outlier detection you do not need to bother about these weights,
but rather take the distances returned by the method getDistances().

The weights wt and vt are used in the iterative computation of the
S/Constrained M estimates and you can find the necessary formulas in

David L. Woodruff, David M. Rocke, (1994) Computable Robust Estimation
of Multivariate Location and Shape in High Dimension Using Compound
Estimators, Journal of the American Statistical Association, Vol. 89,
No. 427, pp. 888-896

http://links.jstor.org/sici?sici=0162-1459(199409)89%3A427%3C888%3ACREOML%3E2.0.CO%3B2-Y


(which is referred to in the help).

Off-list I am sending you a pdf with the description of the algorithm
and I promise to add it to the help file.

Best regards,
Valentin



On 5/18/07, Ibrahim Alameddine <ima4 using duke.edu> wrote:
> I am a PhD student at Duke University working on water quality. I am
> trying to use CovMest to identify outliers in my dataset. The function
> works well, but I am confused about the difference between the two
> generated weight vectors "wt" and "vt". What is "vt" exactly?
>
> Kind regards,
> Ibrahim
>
> --
> Ibrahim Alameddine
> PhD Student
> Environmental Science and Policy
> Nicholas School of the Environment
> Box 90328, Duke University
> Durham, NC 27708-0328
> office: 919-613-8054
> Email: ima4 using duke.edu
>
> _______________________________________________
> R-SIG-Robust using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-robust
>