need a curve fitting package
guia at CC.UManitoba.CA
Mon Feb 21 22:02:17 EST 1994
Someone mistakenly asked for a curve-fitting program which can fit his own
formula ("model") instead of some predetermined formula. There is a
specific forum for dealing with software questions (bio-soft), but here goes.
I have used a graphing program called Origin (MicroCal Software, Inc., One
Roundhouse, Northampton, MA 01060, USA +1/413/586-2013) where I defined
the formula (the X side of it anyways) and it does a best fit of the
formula to the data using the x,y values and changing the parameters of
the formula with each iteration. You can put in approximate values of the
parameters at any point and you can lock in the values at any point and it
will find the remaining ones, or re-approximate the other parameters.
When it is done you have the choice of saving the graph with the fit line,
and saving the parameter values along with each one's chi-square values to
a separate file or to plot them with the graph.
This is a Windows program seemingly written especially for scientists.
The reason that I used it in the first place was that I refused to work
with sigmaplot (didn't want to grow too old waiting for the plots to
render), and most other packages couldn't handle the number of points that
I needed to use (a few thousand per plot).
It does some minor statistics as well (mostly just whatever relates to
graphing), but it is weak in statistics. I'm still looking for a good,
easy-to-use statistical program for Windows.
More information about the Bioforum