[SciPy-User] splines again - a menagerie ?

Travis Oliphant travis@continuum...
Tue Jan 3 23:35:22 CST 2012


On Jan 3, 2012, at 9:44 PM, josef.pktd@gmail.com wrote:

> I did a bit of a literature search on splines, mainly to get some
> overview on its use in statistics.
> 
> There are quite a few different versions of splines and I don't have
> too much of an idea which is which.

> 
> Does anyone now a good reference that gives an overview of different
> splines? (So far I only know b-splines since they are in scipy.)

This book by De Boor is a standard.    http://books.google.com/books?printsec=frontcover&id=m0QDJvBI_ecC#v=onepage&q&f=false
I remember using it 5 years ago when I added some low-level spline-calculation pieces to interpolate.   

> 
> I don't really want to get into the gory details of splines, but I
> would like to have a collection of basis functions for different
> splines, similar to the polynomials vander functions that Chuck added
> or is adding in numpy.  Just something to feed to a (penalized) least
> squares estimation.
> 
> semi-aside: I just saw for the first time a reference that uses a
> polynomial up to a fixed order and then adds spline terms, which looks
> like an interesting combination of polynomial and (piecewise) spline
> fitting.
> 

That sounds interesting.     It sounds like fitting a general trend to the data and then using splines to fit the difference between the data and the "trend".   Do you have the reference handy? 

-Travis


> 
> Thanks,
> 
> Josef
> _______________________________________________
> SciPy-User mailing list
> SciPy-User@scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/scipy-user/attachments/20120103/9b8511c3/attachment.html 


More information about the SciPy-User mailing list