[R] Difference between gam() and loess().

[Rolf Turner]

On 21/03/2009, at 3:19 AM, Ravi Varadhan wrote:


	<snip>

> I also tried a number of other things including changing the  
> "family", and parameters in
> "loess.control", but to no avail.  I looked at the Fortran codes  
> from both loess and gam.
> They are daunting, to say the least. They are dense, and there are  
> absolutely no comments
> whatsoever.  But one thing is clear - they are using different  
> Fortran codes.

	Thanks for doing this digging.  It would on this basis be not  
unreasonable
	to expect there to be numerical differences in the result.
>
> So, the best bet might be to get Trevor Hastie or Bill Cleveland to  
> help you out.
>
> But, before that:  why is this an issue, Rolf?  Is it important  
> that these two results
> be identical?

	It just makes me nervous when two procedures which I believe to be  
doing the
	same thing give answers which are not identical.  Such a phenomenon  
makes me
	wonder if there is something which I am not understanding, and if  
thereby my
	lack of understanding might lead to my making serious errors.

	A lack of understanding is an all-to-frequent occurrence in my  
circumstances
	(i.e. being thick as two short planks) and I therefore need to be  
constantly
	on my guard against such a lack.

	In the current setting I think I will desist from pursuing the issue  
any further
	and just assume that different coding accounts for the different  
results.  And
	take a valium.

		cheers,

			Rolf Turner

######################################################################
Attention:\ This e-mail message is privileged and confid...{{dropped:9}}