I think you're looking for segmented regression models.
Somethings you may find useful:
(Use of Two-segmented Logistic Regression to Estimate Change-points in
Epidemiologic Studies Roberto Pastor and Eliseo GuaJlar )
http://aje.oxfordjournals.org/content/148/7/631.full.pdf?origin=publication_detail
(Using Segmented Regression Models to Fit Soil Nutrient and Soybean Grain
Yield Changes
due to Liming Xiufu SHUAI,Zhijun ZHOU, and Russell S. YOST)
http://download.springer.com/static/pdf/935/art%253A10.1198%252F1085711031580.pdf?auth66=1395993238_17e1f7f575a50838ea4a94bf59c72605&ext=.pdf
as you can probably guess, even guessing where to 'split' it is an issue.
You could try change point detection & then fit your own models. also
there's the whole idea of Pattern Recognition in Time Series which is its
own field
On Wed, Mar 26, 2014 at 3:15 AM, Dieter Jansen <[email protected]> wrote:
> I have a time series of (x,y) values.
>
> When plotted as a scattergram with a line showing the points in sequence
> it is possible to see that the data points tend to fall along a rough line
> (with possible retracements) for some time before settling into a different
> line. This happens multiple times in a data set.
>
> If I visually estimate the start and end date of each set of contiguous
> points that form a rough line I can obviously perform a linear regression
> on those points but for a large number of points I would like to
> automatically partition the sequence into multiple regression lines meeting
> some sort of maximum error criterion and perform a regression on each
> partition.
>
> Is there support for this sort of thing in scikit-learn? I've searched
> for this info but I'm not very familiar with this area yet and I'm probably
> ignorant of the most relevant search terms.
>
> Cheers,
> Dieter.
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> Learn Graph Databases - Download FREE O'Reilly Book
> "Graph Databases" is the definitive new guide to graph databases and their
> applications. Written by three acclaimed leaders in the field,
> this first edition is now available. Download your free book today!
> http://p.sf.net/sfu/13534_NeoTech
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and their
applications. Written by three acclaimed leaders in the field,
this first edition is now available. Download your free book today!
http://p.sf.net/sfu/13534_NeoTech
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
