MathGroup Archive 2004

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

Search the Archive

simultaneous fit of multiple data sets


I am a chemist using a technique called dielectric spectroscopy, the important
point is that I get 3 sets of data from a single sample by using different
frequencies.  The model to fit the data is as follows:

model = Table[{t, .000057 + Sum[(.08/t)*w*Exp[en/(k*t)]/(1 + w^2*Exp[en/(k*t)]
^2*1/Cosh[a/(k*t)]^2*Gaussian[en, ampl, x0, sigma], {en, 4, 12, .1}]}, {t, 85,
325,5}]

I would like to use Mathematica to simultaneously fit for ampl, xo, sigma, and
a for the three data sets since the parameters should remain constant over the
differ frequencies (w).  I am having trouble just getting one frequency to
fit, but can't decide even if I get that to work, that Mathematica will be
able to fit three separate curves simultaneously.  I do have the additional
Experimental Data Analysis (EDA) package.  I can (and have) done this more or
less by a visual trial and error, but it would be nice to get a chi-squared
value for my results and to have this be more automated for the future.

Thank you,

Mary Beth Mulcahy


University of Colorado


  • Prev by Date: Matrix Inversion
  • Next by Date: Re: Working with binaries
  • Previous by thread: Re: Matrix Inversion
  • Next by thread: Re: simultaneous fit of multiple data sets