[RsR] MM-scale

[Christian Hennig]

Hi there,

I have finally decided to include the MM-estimator in my Data Analysis 
course and so I again had to think about it and its implementation a bit.

Apparently the scale estimator given out by lmrob is the S-scale. One could 
wonder whether the optimal value of the final MM-sum multiplied by a 
suitable factor to get consistency under normality would be a more 
appropriate scale estimator because it is directly associated to the 
MM-estimator (which we eventually use in most situations) rather than
to the S-estimator.

Wouldn't it be good to give this out as default scale estimator (or both)?
There could also be a control parameter specifying which scale is given 
out.

Any thoughts?

Cheers,
Christian


*** --- ***
Christian Hennig
University College London, Department of Statistical Science
Gower St., London WC1E 6BT, phone +44 207 679 1698
chrish using stats.ucl.ac.uk, www.homepages.ucl.ac.uk/~ucakche