For what it is worth, I would have thought that expressing the lag coefficients in a B-spline expansion would be preferable to going back to Almon approach. This would give a relatively simple lm() application.
url: www.econ.uiuc.edu/~roger/my.html Roger Koenker email [EMAIL PROTECTED] Department of Economics vox: 217-333-4558 University of Illinois fax: 217-244-6678 Champaign, IL 61820 On Tue, 21 Oct 2003, Thomas Lumley wrote: > On Tue, 21 Oct 2003, Spencer Graves wrote: > > > Have you checked "www.r-project.org" -> search -> "R site search"? I > > just got 15 hits for "polynomial lag". If you haven't already tried > > this, I'd guess that some of these hits (though certainly not all) might > > help you. > > > > Only one of these is about polynomial distributed lag models, and it's an > apparently unsuccessful request for information. > > I have code > http://faculty.washington.edu/tlumley/pdl.R > and documentation > http://faculty.washington.edu/tlumley/pdlglm.html > for a version for generalised linear models. > > Note that I have not used this for some years, so it may run into problems > with changes in R. Also note that it is just a generalised linear model > -- it doesn't do anything about residual autocorrelation (which isn't a > problem in the application I was working on). > > > -thomas > > ______________________________________________ > [EMAIL PROTECTED] mailing list > https://www.stat.math.ethz.ch/mailman/listinfo/r-help > ______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
