[R] Constrained fits: y~a+b*x-c*x^2, with a,b,c >=0

[Berwin A Turlach]

G'day Alex,

On Wed, 27 May 2009 11:51:39 +0200
Alex van der Spek <amvds at xs4all.nl> wrote:

> I wonder whether R has methods for constrained fitting of linear
> models.
> 
> I am trying fm<-lm(y~x+I(x^2), data=dat) which most of the time gives
> indeed the coefficients of an inverted parabola. I know in advance
> that it has to be an inverted parabola with the maximum constrained to
> positive (or zero) values of x.
> 
> The help pages for lm do not contain any info on constrained fitting.
> 
> Does anyone know how to?

Look at the package nnls on CRAN.

According to your subject line, you are trying to solve what is known
as a quadratic program, and there are at least two quadratic
programming solvers (ipop in kernlab and solve.qp in quadprog)
available for R.

HTH.

Cheers,

	Berwin

=========================== Full address =============================
Berwin A Turlach                            Tel.: +65 6516 4416 (secr)
Dept of Statistics and Applied Probability        +65 6516 6650 (self)
Faculty of Science                          FAX : +65 6872 3919       
National University of Singapore     
6 Science Drive 2, Blk S16, Level 7          e-mail: statba at nus.edu.sg
Singapore 117546                    http://www.stat.nus.edu.sg/~statba