| Author |
 Topic  |
|
|
causticreagent
USA
1 Posts |
 Posted - 07/29/2014 : 10:58:44 AM
|
Origin Ver. 9.1.0 and Service Release 2
Operating System: Windows7
I am using Origin to deconvolute experimental spectra into a series of gaussian functions. When doing this it is common to perform several models using different numbers of functions, and compare the results. But there appears to be no good way to do this!
Changing the number of replicate functions to use always reinitializes all the parameters, even if the 'Auto Parameter Initialization' box is unchecked. I also had the idea to duplicate a graph with its fitted curves and then click the little lock to change some parameters, only to find that this also reset the original fitting graph.
I've ended up sitting here with a notebook writing down all the curve parameters, and then typing them back into the fields to start each new model. Days have been wasted on a task that should take ~2 hrs. Why not make a way to reload parameters back into the fitting routine, or save them, or cut and paste into NLFit, or change the number of peaks without starting over?
The mode in which your software functions stands in sharp contradistinction to the norms of modern computer use. |
|
|
JacquelineHe
287 Posts |
Posted - 07/30/2014 : 01:36:19 AM
|
Hi,
Please try to follow the below tutorial to do the fittings for multiple dataset,
http://www.originlab.com/doc/Tutorials/Fitting-MultiDataset
Moreover, you can save a theme to remember the paramater values: After set the initial values for the parameters, uncheck the Auto Parameter Initialization , then click the arrow button at the right-top of NLFit dialog to select Save As to open the Theme Save as... dialog. In this dialog, check Auto Parameter Initialization status and Parameter Values check box. Enter Theme Name and then Click OKbutton. After that, you can load this theme with initial parameter values to do further fittings.
Originlab
Jacqueline |
Edited by - JacquelineHe on 07/30/2014 01:41:28 AM |
 |
|
| |
Topic |
|
|
|
changing number of fitting functions