[R] Zero-inflated predictor

Andrew Robinson A.Robinson at ms.unimelb.edu.au
Fri Mar 9 01:58:38 CET 2007


Hi Jeff,

I'm not sure why there would be a problem with a zero-inflated count
predictor.  As far as I am aware, no regression assumptions are
predicated on the distribution of the predictor variables.

However, should you wish to, it seems to me that you should be able to
do anything you like with it.  It might be worth taking the
traditional hurdle approach, and breaking the predictor into two
predictors, one of which is zero/not zero, and the other of which is
NA at 0.  I'm not sure that you'd get a wholly satisfactory analysis
though.

I hope that this helps, 

Andrew



On Thu, Mar 08, 2007 at 07:13:37PM -0500, Jeff Miller wrote:
>  
> Hi all,
> 
> Does anyone know how to deal with a zero-inflated count PREDICTOR? I know we
> can use ZIP, Hurdle, etc for zero-inflated response variables, but what if
> the problem occurs with one of the covariates?
> 
> I have already found that the literature is correct in stating that any
> transformation will just lead to a distribution with a different inflated
> value, so the arcsin is out.
> 
> Thanks,
> Jeff
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

-- 
Andrew Robinson  
Department of Mathematics and Statistics            Tel: +61-3-8344-9763
University of Melbourne, VIC 3010 Australia         Fax: +61-3-8344-4599
http://www.ms.unimelb.edu.au/~andrewpr
http://blogs.mbs.edu/fishing-in-the-bay/



More information about the R-help mailing list