[R] [Q] specifying weights for robust regression with rlm

Prof Brian Ripley ripley at stats.ox.ac.uk
Tue Feb 28 19:14:34 CET 2006


On Tue, 28 Feb 2006, Young-Jin Lee wrote:

> Hello, R-listers
>
> I am relatively new to R and looking for some help on the "rlm" command.
>
> Instead of using provided weight functions such as huber or hampel, I
> want to specify weights to be applied to each data point. After reading
> rlm help files, I found a "weight" option. According to the help
> document, the "weight" option

It is an argument, not an option: it is listed under arguments!

> allows me to specify prior weights for
> each case. However, the help documents do not give me enough information
> and I have the following two questions:
> 1. Does it mean that rlm command will be using only these prior weights
> in doing a robust regression task when they are supplied? Please correct
> me if I misunderstand anything.
> 2. How can I specify prior weights? I guess I should provide a vector of
> weights. Can someone show me how to do it?

rlm is part of the support software for a book.  The book has worked
examples of supplying weights.

-- 
Brian D. Ripley,                  ripley at 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 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595




More information about the R-help mailing list