MathGroup Archive 1997

[Date Index] [Thread Index] [Author Index]

Search the Archive

non-negative least squares

  • To: mathgroup at smc.vnet.net
  • Subject: [mg6524] non-negative least squares
  • From: onm010 at uni-essen.de (Dr. Stefan P. Mueller)
  • Date: Mon, 31 Mar 1997 23:01:26 -0500 (EST)
  • Organization: Nuklearmedizin Uniklinik Essen
  • Sender: owner-wri-mathgroup at wolfram.com

For a current data analysis project with noisy data we need to fit
parameters which are physically constrained to positive values. There
seems to be an advantage to using a non-negative least squares algorithm,
e.g. Lawson & Hanson's NNLS, which is available, e.g. in Fortran or as a
MATLAB function. For various reasons we would like to use Mathematica but
we haven't been able to find an implementation of either this algorithm or
something similar.

Do you have or know the whereabouts of NNLS in Mathematica?

Thank you very much!

Stefan
____________________________________________________________________________
Dr. Stefan P. Mueller, Nuklearmedizin, Universitaetsklinikum, D-45122 Essen,
 FR Germany, e-mail: stefan.mueller at uni-essen.de, FAX: <49> (201) 723-4146

-- 
Dr. Stefan P. Mueller, Nuklearmedizin, Universitaetsklinikum, Essen,
 e-mail: onm010 at uni-essen.de, phone: <49> 201/ 723-4179, FAX: -5964


  • Prev by Date: Re: Solve and Plot
  • Next by Date: Re: NDSolve problem
  • Previous by thread: Re: Integer Solutions
  • Next by thread: Re: NDSolve problem