| Author |
 Topic  |
|
|
RafaelVS
Brazil
3 Posts |
 Posted - 03/19/2015 : 3:31:53 PM
|
Hello. I am trying to plot an exponential function like this
where V=y(x) and r=r(x), and a, Z1, Z2, e are constants I need to find.
How can I define a exponential function y = (k/x)exp(-x/a) so that Origin can give me k and a?
I'm using Origin 8.1 |
Edited by - RafaelVS on 03/19/2015 7:24:03 PM |
|
|
JacquelineHe
287 Posts |
|
|
RafaelVS
Brazil
3 Posts |
Posted - 04/17/2015 : 2:05:14 PM
|
Hi, I read and did everything in those links but still the best plot I've got was this one:
Is there any way to improve it? |
Edited by - RafaelVS on 04/17/2015 2:13:18 PM |
 |
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 04/17/2015 : 3:05:20 PM
|
Hi Rafael,
When a part of the curve is very steep (near 0 in your case), the residual's contribution would become
too high and sensitive to the parameters, even though the Chi square may be numerically minimum. To consider
both X and Y directions, you can try to define the fitting function as implicit. Please look at my trial below:
(To make my life easier, I have plugged your parameter values in the definition of this function.)
--Hideo Fujii
OriginLab
P.S. Sorry for that the orthogonal regression for implicit functions has become available in the version 9.0 and later, and only in Pro versions. |
Edited by - Hideo Fujii on 04/17/2015 3:19:38 PM |
|
|
|
RafaelVS
Brazil
3 Posts |
Posted - 04/17/2015 : 3:37:49 PM
|
Hi Hideo Fujii!
That's impressive!! But could you explain it a little more? I didn't quite understand what you did! |
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 04/17/2015 : 5:05:34 PM
|
Hi,
Here is the Origin document about the Nonlinear Implicit Curve Fitting:
http://www.originlab.com/doc/Origin-Help/Fitting-Implicit
which also contains a table of comparison between the Levenberg-Marquardt and the Orthogonal Distance Regression, and the reference of later case. Sorry I haven't originally realized that your version doesn't have this feature. You can download the evaluation version of Origin 2015 to try out this method: http://www.originlab.com/demodownload.aspx
--Hideo Fujii
OriginLab
|
|
|
|
AndreDourado
Brazil
5 Posts |
Posted - 04/24/2015 : 12:41:57 PM
|
Hi,
I wrote, on Origin 9.0, the function y=A*exp(-B*x)+C*x*exp(-B*x).
It worked twice, but with other data the report error said that it could not converge and that the reason was unknown. Once it said that the parameter were not independent and that I should try to fix one of them, what is not possible to me. What can I do?
Thank you!
André |
|
|
| |
Topic |
|
User Defined Function