[R] What does LOESS stand for?

[David Winsemius]

On May 31, 2010, at 11:44 AM, Peter Neuhaus wrote:

> Thanks a lot...
>
> ... makes it a bit difficult to explain, though...

We drink no wine before its time. Somewhat like trying to explain  
splines to non-technical types:

http://www.duckworksmagazine.com/03/r/articles/splineducks/splineDucks.htm

-- 
David.

>
> Peter
>
> Quoting Joris Meys <jorismeys at gmail.com>:
>
>> This is the paper on which the loess algorithm is based in general:
>> http://www.econ.pdx.edu/faculty/KPL/readings/cleveland88.pdf
>>
>> The explanation about the origin of the term LOESS is given on page  
>> 597.
>>
>> Cheers
>> Joris
>>
>> On Mon, May 31, 2010 at 11:33 AM, Peter Neuhaus  
>> <pneuhaus at pneuhaus.de>wrote:
>>
>>> Dear R-community,
>>>
>>> maybe someone can help me with this:
>>>
>>> I've been using the loess() smoother for quite a while now, and for
>>> the matter of documentation I'd like to resolve the acronym LOESS.
>>> Unfortunately there's no explanation in the help file, and I didn't
>>> get anything convincing from google either.
>>>
>>> I know that the predecessor LOWESS stands for "Locally Weighted
>>> Scatterplot Smoothing". But what does LOESS stand for, specifically?
>>> "Locally Weighted Exponential Scatterplot Smoothing"? As far as
>>> I understand LOESS is still a local polynomial regression, so that
>>> would probably make no sense.
>>>
>>> Any help appreciated!
>>>
>>> Thanks in advance,
>>>
>>> Peter
>


David Winsemius, MD
West Hartford, CT