| T O P I C R E V I E W |
| funfine |
Posted - 07/28/2007 : 11:33:53 AM
I have edited the messege to make it more clear. Hopefully, this is less misleading.
Origin Version (Select Help-->About Origin): Origin Pro 7.5 SR5
Operating System: WinXP Home Edition
Hi everyone,
I have met a problem. I need to fit the data with linear combination. The case is like this. I have two columns of data, one called ColA, one ColB. I need to fit a new dataset (a*ColA+ b*ColB) with Function1, where a and b is also variable fitting parameter. I need to do this with Function2 too. . Anybody knows how to do this? I appreciate any idea. Thanks.
Ying
Edited by - funfine on 07/28/2007 2:21:39 PM |
| 3 L A T E S T R E P L I E S (Newest First) |
| Mike Buess |
Posted - 07/29/2007 : 10:23:06 AM
So you want to change the data even as you are performing the fit? I don't see how you could do that. Best I can suggest is to create col(C)= a*col(A) + b*col(B), fit to funcA or funcB, save params and then repeat with new a & b.
Mike Buess
Origin WebRing Member |
| funfine |
Posted - 07/28/2007 : 1:45:24 PM
Hi Mike,
Thanks for the timely reply. But I am deep sorry for the misunderstanding. What I meant is, let's say, I have two columns data, one called ColA, one ColB. I need to fit a new dataset (a*ColA+ b*ColB) with FunctionA, where a and b is also variable fitting parameter. I need to do this with FunctionB too. Thanks for your great help.
Ying |
| Mike Buess |
Posted - 07/28/2007 : 1:27:39 PM
Hi Ying,
I think you mean to simultaneously fit one column to Function A and the other column to Function B. If that is correct you must create a new function which is a linear combination of both functions...
y = mu*FuncA(x) + (1-mu)*FuncB(x)
http://www.originlab.com/www/helponlin/origin/Defining_a_New_Function.htm
Then on the Select Dataset tab check the Fit Multiple Datasets option and select both columns as dependent variables.
http://www.originlab.com/www/helponlin/origin/Selecting_the_Data_Sets.htm
When you start the fitting session you'll see a duplicate set of parameters for each column, i.e., mu_1, w_1, xc_1, mu_2, w_2, xc_2. For the column that fits to FuncA set mu_# = 1 and initialize the remaining parameters for that column. For the column that fits to FuncB set mu_# = 0 and initialize the rest.
http://www.originlab.com/www/helponlin/origin/The_Fitting_Session_Concept.htm
If FuncA and FuncB happen to be Lorentz and Gauss you can use Origin's built-in PseudVoigt2 function.
Mike Buess
Origin WebRing Member |
|
|
