R help - Jan 2011 - piecewise regression

Hello everybody!!!!

Quick question, if you'd like to throw a little tip:
 does anyone knows a function that runs piecewise regression models with
coefficients estimation and inferences ?

Thank you

	[[alternative HTML version deleted]]

Searching the archives would surely be a better way to assemble a list  
of candidate packages but the "segmented" package is surely in the
list.

On Jan 14, 2011, at 9:42 AM, Federico Bonofiglio wrote:
> Hello everybody!!!!
>
> Quick question, if you'd like to throw a little tip:
> does anyone knows a function that runs piecewise regression models  
> with
> coefficients estimation and inferences ?
>
> Thank you
>
> 	[[alternative HTML version deleted]]
David Winsemius, MD
West Hartford, CT

Hello:


       I just found 58 help pages in 32 contributed packages containing 
the terms "piecewise" and "regression", as follows:


library(sos)
pr <- ???'piecewise regression' # for "piecewise" with
"regression"
summary(pr) # 58 matches in 32 packages
pr # view the 58 links sorted by package in a web browser

p.r <- ???"{piecewise regression}" # for "piecewise"
followed immediate
by "regression"
summary(p.r) # 15 matches in 13 packages
p.4 # to view in a web browser


       Beyond this, the "structchange" package supports looking for 
changes that would be, e.g., linear in one segment then jump to a 
different linear regime.  If you want the lines to connect, then you may 
want splines.  Probably the most common spline is cubic, but you can 
also get linear splines.  One of the most general spline packages is 
"fda" (functional data analysis), with multiple companion books by
James
Ramsay and others.  However, there are other spline packages.


       If this does not answer your question, "PLEASE do read the 
posting guide http://www.R-project.org/posting-guide.html and provide 
commented, minimal, self-contained, reproducible code", as it says in 
the signature block for each email on this list.


       Hope this helps.
       Spencer Graves


On 1/14/2011 6:42 AM, Federico Bonofiglio wrote:
> Hello everybody!!!!
>
> Quick question, if you'd like to throw a little tip:
>   does anyone knows a function that runs piecewise regression models with
> coefficients estimation and inferences ?
>
> Thank you
>
> 	[[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.
>

I wish I had a better answer.

There are two things that I know of that use a direct approach:

The segmentation package seems to work well if you are doing a few fits, but
I had problems when I tried running it on loads of data.  It's a bit tricky
to parametrize.  When I tried investigating the internals I found that the
variables are are in Italian, which makes it difficult (and comical) to
follow. However, the theory is very attractive.

There are also some examples in the archives where people have used the nls
solver.  But, as I remember, it didn't do a great job at estimating
breakpoints (maybe it didn't do it at all).  If you don't know about
markmail, that has a nice interface for searching the help archives, and you
can easily find the example I'm referencing:
http://r-project.markmail.org/

A second approach (in my mind) is to use a two step approach of smoothing,
then segmenting the smooth:

The other thing that I have found very useful is using approx in conjunction
with a smoother, like a spline or GAM.  You can get a smooth estimate, and
then do something like have a piecewise line at every 10th percentile of
your smoothed line.

I've experimented with is using a classification and regression tree (rpart
package) to classify the slope of a smooth line, and then use the regions of
slopes to reconstitute a line.  One of the challenges with this approach is
deciding when to stop segmenting, and picking an appropriate CP value.

I keep thinking that there must be some sort of annealing optimization
approach that would work the best, but I have not had the time to figure out
how to specify it.

Gene

On Fri, Jan 14, 2011 at 8:42 AM, Federico Bonofiglio
<bonoricus@gmail.com>wrote:
> Hello everybody!!!!
>
> Quick question, if you'd like to throw a little tip:
>  does anyone knows a function that runs piecewise regression models with
> coefficients estimation and inferences ?
>
> Thank you
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help@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.
>
	[[alternative HTML version deleted]]

You failed to specify 2 crucial issues: Are the join points known
(linear rgression, splines) or unknown (nonlinear regression)? And as
several have already indicated, what are the smoothness constraints?

-- Bert

On Fri, Jan 14, 2011 at 6:42 AM, Federico Bonofiglio
<bonoricus at gmail.com> wrote:
> Hello everybody!!!!
>
> Quick question, if you'd like to throw a little tip:
> ?does anyone knows a function that runs piecewise regression models with
> coefficients estimation and inferences ?
>
> Thank you
>
> ? ? ? ?[[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.
>


-- 
Bert Gunter
Genentech Nonclinical Biostatistics