Statalisters
Our program uses restricted (natural) cubic splines. The user can
either specify the knot values or allow the program to choose default
values. In the latter case we use the knots advocated by Frank Harrell
in his text.
These default knot locations are determined by the percentiles of the
independent variable and are derived without regard to the independent
variable. This has the advantage of allowing the same spline covariates
to be used in a number of different regressions with different dependent
variables. A disadvantage is that it does not take advantage of the
y-values to determine the optimal placement of the knots. In my
opinion, this is less important when fitting cubic splines than it is in
fitting linear splines where we have a specific interest in estimating
the location of an inflection point (and where we would like to place a
knot).
Bill Dupont
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Kit Baum
Sent: Tuesday, November 30, 2004 7:29 AM
To: [email protected]
Subject: st: Re: splines
Bill Dupont recently posted rc_spline on SSC. I haven't looked at it
all that closely, but I believe that it might have that flexibility.
Kit Baum, Boston College Economics http://ideas.repec.org/e/pba1.html
On Nov 30, 2004, at 2:33 AM, Michael wrote:
> Does anyone have code for estimating a spline regression with a given
> number of knots, but where the location of the knots is to be
> estimated along with the slopes of the separate spline segments?
> Michael Klein
*
* 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/
*
* 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/
