| Author |
 Topic  |
|
|
labjunkie
USA
4 Posts |
 Posted - 10/14/2020 : 06:16:26 AM
|
I am new to this origin software and struggling to fit a histogram using Gaussian function. I will appreciate if somebody can please explain me how to fit the curve. I have attached the data to fit and an example how I want the data to fit.
Thank you
https://my.originlab.com/ftp/forum_and_kbase/Images/origin-gaussian%20fit.txt |
Edited by - labjunkie on 10/14/2020 06:22:26 AM |
|
|
Shirley_GZ
China
Posts |
|
|
labjunkie
USA
4 Posts |
Posted - 10/14/2020 : 06:49:29 AM
|
Thank you very much for your reply.
I tried using the method you are talking about it just says
"Fit did not converge. Maximum iteration setting of 400 was reached."
Please let me know if you can suggest what I can do.
quote:
Originally posted by Shirley_GZ
Hi labjunkie,
Maybe you can refer to this tutorial,
https://www.originlab.com/doc/Tutorials/MultiFittedCurve-in-Histogram
Please let us know if any problem.
Thanks,
Shirley
Originlab Technical Service Team
as |
 |
|
|
Chris D
428 Posts |
Posted - 10/14/2020 : 09:49:14 AM
|
Hi,
First- failure to converge does not mean a bad fit. You could very well decide that the fit is good enough at any time- it can be very subjective.
That being said, you can increase the number of allowed iterations in the NLFit dialog (access it via Change Parameters from the green lock icon).
In the dialog, select the "Settings" tab and then "Advanced". On the right side of the dialog, you will see a control group entitled "Fit Control". Expand it and you will see controls that allow you to change both number of iterations and tolerance.
Perhaps increasing the number of iterations or playing with the tolerance will help your fit.
Thanks,
Chris Drozdowski
Originlab Technical Support
|
|
|
|
labjunkie
USA
4 Posts |
Posted - 10/14/2020 : 10:23:20 AM
|
Thank you very much for your reply Chris.
I tried as you suggested. It doesn't give that message any more.
However, I have two quick questions
As my x axis is based on log 10 scale, do I need to change anything about in the formula for derived parameters, like xc= 10^xc.
And second question is that how can I get a curve like gaussian fitting as shown in the image I have attached the blue and red curve fitting.
I really appreciate your help.
quote:
Originally posted by Chris D
Hi,
First- failure to converge does not mean a bad fit. You could very well decide that the fit is good enough at any time- it can be very subjective.
That being said, you can increase the number of allowed iterations in the NLFit dialog (access it via Change Parameters from the green lock icon).
In the dialog, select the "Settings" tab and then "Advanced". On the right side of the dialog, you will see a control group entitled "Fit Control". Expand it and you will see controls that allow you to change both number of iterations and tolerance.
Perhaps increasing the number of iterations or playing with the tolerance will help your fit.
Thanks,
Chris Drozdowski
Originlab Technical Support
as |
|
|
|
Chris D
428 Posts |
Posted - 10/14/2020 : 11:14:30 AM
|
Hi,
You need to create a custom fitting function to deal with the fact that your X data is logarithmic.
If you will send us the simplest project with data and a graph, we can help create the fitting function. Send it to tech@originlab.com
Thanks,
Chris Drozdowski
Originlab Technical Support
|
|
|
|
labjunkie
USA
4 Posts |
Posted - 10/14/2020 : 11:56:22 AM
|
Thank you again Chris! Just sent the data to you. Please let me know if you received it.
quote:
Originally posted by Chris D
Hi,
You need to create a custom fitting function to deal with the fact that your X data is logarithmic.
If you will send us the simplest project with data and a graph, we can help create the fitting function. Send it to tech@originlab.com
Thanks,
Chris Drozdowski
Originlab Technical Support
as |
|
|
| |
Topic |
|
|
|
Gaussian fitting