R help - Dec 2013 - log transforming predictor variables in a binomial GAM?

Hi all,

I am applying a Presence/absence Generalized additive model to model the
distribution of marine algae species in R. I have found that log transforming
the environmental variables improves the explained deviance of the model
considerably. While log transforming is common practice in GLM, I have been
unable to find any papers where this is performed in a GAM. Im wondering whether
this breaks any of the rules of GAMs and is statistically acceptable?

Thanks all
Peter

	[[alternative HTML version deleted]]

It doesn't break anything - you can transform the predictors pretty much
any way you like, and it is often sensible as a way of tackling very
uneven leverage. By transforming predictors, all you are changing in the
model is what "smooth" means. e.g. smooth w.r.t. log(x) is somewhat
different to sooth w.r.t. x.

best,
Simon

On 16/12/13 04:54, PETER MITCHELL wrote:
> Hi all,
>
> I am applying a Presence/absence Generalized additive model to model
> the distribution of marine algae species in R. I have found that log
> transforming the environmental variables improves the explained
> deviance of the model considerably. While log transforming is common
> practice in GLM, I have been unable to find any papers where this is
> performed in a GAM. Im wondering whether this breaks any of the rules
> of GAMs and is statistically acceptable?
>
> Thanks all Peter
>
> [[alternative HTML version deleted]]
>
> ______________________________________________ R-help at r-project.org
> 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.
>

-- 
Simon Wood, Mathematical Science, University of Bath BA2 7AY UK
+44 (0)1225 386603               http://people.bath.ac.uk/sw283