Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: Re: st: Penalised splines


From   n j cox <n.j.cox@durham.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: Re: st: Penalised splines
Date   Tue, 22 May 2007 21:59:56 +0100

These look like what are perhaps more commonly known
as (cubic) smoothing splines. They are implemented in
R, MATLAB, etc. What's quite nice for some purposes,
but not all, is that the knots are just all the distinct
x values, so there's no arbitrariness about that.

Well-banged on but old-fashioned Fortran is at Carl
de Boor's website in Wisconsin

http://www.cs.wisc.edu/~deboor/pgs

Nick
n.j.cox@durham.ac.uk

Maarten Buis replied to Adrian Mander

> Has anyone implemented a penalised cubic spline?

I tried some of that some time ago, and got stuck on the fact that
simple spline terms tend to become highly collinear and these penalized

splines typically require a large number of knots (and thus spline
terms). I could have tried figuring out how B-splines would have fitted
in a penalized spline function, but I didn't. However, an approach that
is similar can be found in (Royston and Sauerbrei 2007)

Patrick Royston and Willi Sauerbrei (2007) "Multivariate modeling with
cubic regression splines: A principled approach" The Stata Journal,
7(1), pp. 45-70.

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index