[R] Constrained Log-Likelihood with SQP Solver

From: Diethelm Wuertz <wuertz_at_itp.phys.ethz.ch>
Date: Wed 14 Dec 2005 - 00:18:16 EST

Dear R-Users,

I'm searching for somebody who can support me or even likes to collaborate with
me in setting up an R-package for "constrained maximim log-likelihood" parameter
estimation.

For example fitting the parameters of a MA(1)-APARCH(1,1) model for a time series
of 17'000 points (e.g. the famous Ding-Granger-Engle mode) takes about 10 minutes
with the existing optimization algorithms available under R.

Modern state of the art algorithms, like SQP algorithms as implemented in Gauss,
Matlab, Ox, take about a few seconds. I tested this finding with a free constrained
SQP solver written in FORTRAN under R and found these results confirmed. I got the results in a few seconds instead of a few minutes!

Now I'm looking for a collegue who has the experience in implementing FORTRAN
Optimization Code in R, calling the objective function and optionally gradient and
hessian from R functions. I have already inspected a lot of Fortran, C, and R sources
from the base package, but I didn't succeed so far with a reasonable effort.

Many thanks in advance
Diethelm Wuertz



R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html Received on Wed Dec 14 01:13:29 2005

This archive was generated by hypermail 2.1.8 : Wed 14 Dec 2005 - 02:33:55 EST