Using Multipeak fit
ezmmrmnn
Tue, 06/19/2012 - 11:33 pm
I'm trying to write a code to fit x-y data with one peak to a combination of ExpModGauss and linear curves. My problems arise when I want to plot the fit result against the original data. The resulting fitted curve does not resemble a ExpModGauss and linear fit.
1) I would first like to verify that the four fitting parameters contained in the PeakCoefs file are the resulting fit, where PeakCoefs[0] = peak location, PeakCoefs[1] = Gaussian width, PeakCoefs[2] = amplitude , and PeakCoefs[3] = exponential decay constant.
2) After fitting the curve (i.e., calling MPF2_AutoMPFit), within my own code, I am just calling the functions fExpGauss(t,r,s) and ncdf(t), which were originally from the Peak Functions procedure, to obtain the 'y' values for the fitted curve at each 'x' values from the original data. When I use this procedure for a combination of Gauss and linear fits, the procedure I wrote works... so I think the problem must lie within my understanding of the fitting parameters.
Any help would be appreciated!
Cheers, Liz
Sounds like a sophisticated and highly customized use of Multipeak Fit 2!
The file PeakFunctions2.ipf has a function EMGFunction() that you can use to generate individual Y values from the ExpModGauss peak fit. If you look at the code for that function, you will see that MPF2 uses a somewhat different definition of the ExpModGauss function than is used by the older Multipeak Fit 1.4. I think your procedure is using the older peak shape formula.
If that doesn't answer the question, please post an example Igor experiment file that I can use to understand the problem better.
John Weeks
WaveMetrics, Inc.
support@wavemetrics.com
June 20, 2012 at 09:11 am - Permalink