R help - Apr 2009 - Forcing the extrapolation of loess through the origin

Hi all,

I'm fitting a line to my dataset. Later I want to predict missing values
that exceed the [min,max] interval of my empirical data, therefore I choose
surface="direct" for extrapolation.

l1<-loess(y1~x1,span=0.1,data.frame(x=x1,y=y1),control=loess.control(surface="direct"))

In my application it is highly important that the fitted line intercepts at the
point of origin. Is it possible to do this in R?

Thanks in advance.

Cheers,
Torsten
--

Below. 


Bert Gunter
Genentech Nonclinical Biostatistics
650-467-7374

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On
Behalf Of jimm-panse at gmx.de
Sent: Tuesday, April 14, 2009 10:08 AM
To: r-help at r-project.org
Subject: [R] Forcing the extrapolation of loess through the origin

Hi all,

I'm fitting a line to my dataset. Later I want to predict missing values
that exceed the [min,max] interval of my empirical data, therefore I choose
surface="direct" for extrapolation.

l1<-loess(y1~x1,span=0.1,data.frame(x=x1,y=y1),control=loess.control(surface
="direct"))

In my application it is highly important that the fitted line intercepts at
the point of origin. Is it possible to do this in R?

-- No. Indeed, this appears to me to contradict the nature of loess as a
locally adaptive fit.
Bert

Thanks in advance.

Cheers,
Torsten
--

______________________________________________
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.

On Tue, Apr 14, 2009 at 1:08 PM,  <jimm-panse at gmx.de>
wrote:
> I'm fitting a line to my dataset. Later I want to predict missing
values that exceed the [min,max] interval of my empirical data, therefore I
choose surface="direct" for extrapolation.
>
>
l1<-loess(y1~x1,span=0.1,data.frame(x=x1,y=y1),control=loess.control(surface="direct"))
>
> In my application it is highly important that the fitted line intercepts at
the point of origin. Is it possible to do this in R?
Well, you could always add lots of artificial data points x=0, y=0
..., like this:

l1<-loess(y1~x1,span=0.1,data.frame(x=c(rep(0,100),x1),y=c(rep(0,100),y1)),control=loess.control(surface="direct"))

which will eventually drive f(0) to near 0, but surely that will
create fitting artifacts.

          -s

Hi Torsten,

If you are fitting a line, why are you using "loess"?  Why not simply 
use "lm" to fit a regression line that goes through the origin? (i.e. 
with no intercept).

Julian

jimm-panse at gmx.de wrote:
> Hi all,
>
> I'm fitting a line to my dataset. Later I want to predict missing
values that exceed the [min,max] interval of my empirical data, therefore I
choose surface="direct" for extrapolation.
>
>
l1<-loess(y1~x1,span=0.1,data.frame(x=x1,y=y1),control=loess.control(surface="direct"))
>
> In my application it is highly important that the fitted line intercepts at
the point of origin. Is it possible to do this in R?
>
> Thanks in advance.
>
> Cheers,
> Torsten
> --
>
> ______________________________________________
> 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.
>

On Tue, 14 Apr 2009, jimm-panse at gmx.de wrote:
> Hi all,
>
> I'm fitting a line to my dataset. Later I want to predict missing
values that exceed the [min,max] interval of my empirical data, therefore I
choose surface="direct" for extrapolation.
>
>
l1<-loess(y1~x1,span=0.1,data.frame(x=x1,y=y1),control=loess.control(surface="direct"))
>
> In my application it is highly important that the fitted line intercepts at
the point of origin. Is it possible to do this in R?

Well, yes, but as Burt suggests it may not be sensible.

There are several ways. For one, include a reflection of the 
(x,y) data into opposite quandrants as well as the original data.

Something like

l1<-loess( y ~ x , span=0.1, data.frame(x = c(-x1,x1), y = c(-y1,y1)),
 	control=loess.control(surface="direct"))

will force the prediction through the origin.

(I corrected what seems to be as typo in your code, too.)

Of course, if any( x < 0 ) it is hard to see how the substantive results 
would make any sense.

HTH,

Chuck
>
> Thanks in advance.
>
> Cheers,
> Torsten
> --
>
> ______________________________________________
> 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.
>
Charles C. Berry                            (858) 534-2098
                                             Dept of Family/Preventive Medicine
E mailto:cberry at tajo.ucsd.edu	            UC San Diego
http://famprevmed.ucsd.edu/faculty/cberry/  La Jolla, San Diego 92093-0901