[R] missing values for MARS (was mda package)

[Prof Brian Ripley]

On Fri, 9 Apr 2004, Wang, Zhu wrote:

> I basically wanted to use MARS to reproduce results using the dataset
> "Marketing" in the following book
> 
> http://www-stat-class.stanford.edu/~tibs/ElemStatLearn/
> 
> The authors actually provided S-Plus functions for mars, bruto ,etc. I
> used all default options of mars in R but there was an error due to NAs
> and I could not find any option to handle missing values.

Friedman originated MARS and has code for it.  The code in mda by 
Hastie/Tibshirani is different, and the code on that website is a direct 
ancestor of the mda package for R.  I see no option in the code for mars 
there to handle missing values, so you would do better to ask the authors 
how they did it (if you really believe they have such an option).

And PLEASE read the posting guide and try to learn to ask precise
questions with enough background information!

-- 
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