| Author |
 Topic  |
|
|
vmukund
USA
4 Posts |
 Posted - 08/12/2011 : 6:44:48 PM
|
Origin Ver.8.5 and Service Release 2(Select Help-->About Origin):
Operating System:Windows7
Hi,
I am trying to fit my data with a user-defined function y=y0+b*x+lorentzian as my data's background is changing(slope-wise).I am sharing the slope for the 2 peaks i am fitting and fix the Yo. I try running this theme on one column, it works. But when i tried batch processing for 3 columns, its taking so much time(1 hour).As i have files with 1000 columns to fit, looks like its going to take forever. Please do let me know as to how i can speed up the batch processing.
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 08/15/2011 : 1:49:17 PM
|
Hi vmukund,
If your multiple columns (three or thousand!) are in a single worksheet (with proper XY designations), then how about highlighting all columns, and run NLFit fitter in the Independent Fit setting for the Multi-Data Fit Mode. This may reduce the overhead of restaging NLFit at every iteration. Please try with three column sample (XYYY or XYXYXY), and let me know the result. (Batch processing with the analysis template is designed and capable to handle much more complex analysis routine beyond a single fitting.)
--Hideo Fujii
OriginLab |
Edited by - Hideo Fujii on 08/15/2011 1:53:36 PM |
 |
|
|
vmukund
USA
4 Posts |
Posted - 08/17/2011 : 05:49:07 AM
|
Hi Hideo,
Thanks for the reply. I have 3 more questions.
1.In between, i was trying to use the NLFit-dialog to fit my data with a built-in function lorentzian(Under peak function). After i select the 3 columns, I was not able fit the data to 2 peaks, i.e the replica option in advanced settings was grey and was not able to change from 0 to 1 to carry out the fitting. how can i change the number if replicas?
2.Is it possible to change the number of iterations in NLFit dialog?
3.When the fit is not converged, the reduced chi-squared is more than 1e-6. is it possible to change this number(say 1e-3)? |
|
|
|
easwar
USA
1964 Posts |
Posted - 08/17/2011 : 11:34:31 AM
|
Hi Vineetha,
You had previously sent your data and exchanged e-mails with tech support, so I suggest you continue with e-mail/call to tech support. It is difficult to address same issue in forum and in our support when it is not clear who is posting and if there is some history such as we have the data.
I tried the data you sent with a theme in Peak Analyzer and was able to find anchor points, fit with a straight line, then use that as baseline and find and fit the peaks. So we are not sure why PA is not working for you. From our records it appears you have OriginPro so you should have access to PA with fitting.
The user defined function you are trying to define cannot support replica fitting in NLFit.
Another approach to your data could be to separately first find and subtract baseline (and in that process compute the slope) and then fit the subtracted data with a lorentzian function with replica.
Maybe you can send more of your data to tech support, as with the data you sent, we could fit with PA, so we do not know yet why that is not working for you.
Easwar
OriginLab |
|
|
| |
Topic |
|
|
|
batch processing with user defiend function