I just heard a talk about a semi-parametric model. I was quite excited by the idea. This model is fitted y= xB + g(z) + e where x is a data matrix, B a column vector, z is another data matrix, and g is a smooth model fitted by a Kernel Smoothing regression (I got the idea any smoother would do as well). The speaker said that when z is considered as a "control" variable, and there is no reason to assume linearity, then one can estimate this model and the B estimates are (in some sense I cannot say exactly) better, perhaps converging more quickly to the true value as the sample size increases. I got interested in doing this and wondered if in R it is possible. In R's MASS package I find the modreg library, which has several smoothing tools, but I don't find a way to estimate B at the same time. (Incidentally, I'm rather overwhelmed by the many different flavors of smoothers!) Does an R package exist for estimating this semi-parametric model? If this is a bad idea, you can tell me, my feelings won't be hurt :) pj ps. I just found that SAS has at least one procedure for this, called tpsplines (thin-plate splines), so I know I wasn't misunderstanding this fellow's lecture. -- Paul E. Johnson email: pauljohn at ukans.edu Dept. of Political Science http://lark.cc.ukans.edu/~pauljohn University of Kansas Office: (785) 864-9086 Lawrence, Kansas 66045 FAX: (785) 864-5700 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Prof Brian D Ripley
2001-May-07 06:07 UTC
[R] semi-parametric (partial linear?) regression
On Sun, 6 May 2001 pauljohn at ukans.edu wrote:> I just heard a talk about a semi-parametric model. I was quite excited > by the idea. This model is fitted > > y= xB + g(z) + e > > where x is a data matrix, B a column vector, z is another data matrix, > and g is a smooth model fitted by a Kernel Smoothing regression (I got > the idea any smoother would do as well). > > The speaker said that when z is considered as a "control" variable, and > there is no reason to assume linearity, then one can estimate this model > and the B estimates are (in some sense I cannot say exactly) better, > perhaps converging more quickly to the true value as the sample size > increases. > > I got interested in doing this and wondered if in R it is possible. In > R's MASS package I find the modreg library, which has several smoothingYou don't! modreg is a package in R (although mainly implemented by the R of V&R's MASS package).> tools, but I don't find a way to estimate B at the same time. > (Incidentally, I'm rather overwhelmed by the many different flavors of > smoothers!) > > Does an R package exist for estimating this semi-parametric model?mgcv, gss, sm, ... -- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272860 (secr) Oxford OX1 3TG, UK Fax: +44 1865 272595 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
From: pauljohn at ukans.edu Date: Sun, 06 May 2001 21:30:48 -0500 Organization: KU X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.4.2-2 i686) X-Accept-Language: en MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-r-help at stat.math.ethz.ch Precedence: SfS-bulk I just heard a talk about a semi-parametric model. I was quite excited by the idea. This model is fitted y= xB + g(z) + e where x is a data matrix, B a column vector, z is another data matrix, and g is a smooth model fitted by a Kernel Smoothing regression (I got the idea any smoother would do as well). The speaker said that when z is considered as a "control" variable, and there is no reason to assume linearity, then one can estimate this model and the B estimates are (in some sense I cannot say exactly) better, perhaps converging more quickly to the true value as the sample size increases. I got interested in doing this and wondered if in R it is possible. In R's MASS package I find the modreg library, which has several smoothing tools, but I don't find a way to estimate B at the same time. (Incidentally, I'm rather overwhelmed by the many different flavors of smoothers!) Does an R package exist for estimating this semi-parametric model? If this is a bad idea, you can tell me, my feelings won't be hurt :) pj ps. I just found that SAS has at least one procedure for this, called tpsplines (thin-plate splines), so I know I wasn't misunderstanding this fellow's lecture. -- Paul E. Johnson email: pauljohn at ukans.edu Dept. of Political Science http://lark.cc.ukans.edu/~pauljohn University of Kansas Office: (785) 864-9086 Lawrence, Kansas 66045 FAX: (785) 864-5700 -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._ You can use ssanova in the gss package for this, where xB is the "partial" term. -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-help-request at stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._