| T O P I C R E V I E W |
| tjuxjliu |
Posted - 05/09/2005 : 04:36:14 AM
Origin Version (Select Help-->About Origin): 7.5
Operating System: winXP
How to do the curve fitting with y=f(x.y)?
In the function of y=f(x.y), y is not only the dependent variable but also the independent variable. |
| 6 L A T E S T R E P L I E S (Newest First) |
| altomare |
Posted - 08/14/2005 : 3:42:58 PM
Hi Hideo,
thank you for your suggestion. I was able to reformulate the problem so to avoid this issue.
I will keep your suggestion in mind if I will need to tackle a similar problem again
Thanks a lot
Fabio |
| Hideo Fujii |
Posted - 08/09/2005 : 2:40:01 PM
Hi Xinjun and Fabio,
Have you tried the following way...?
Your function is: y=f(x,y); Define z=g(x,y)=f(x,y)-y=0;
So, say, when you have 2 columns of A(X) and B(Y), then add an extra column C(Z), and put 0's to the rows of (x,y). So, you can now proceed the 3D fitting by the way described in e.g.,
http://www.originlab.com/index.aspx?s=9&pid=434
or
http://www.originlab.com/www/support/resultstech.aspx?ID=57&language=English
I'm not sure about this methods's performance such as how easily the fitting can be trapped by local minima, etc. but anyway you may want to experiment it....
--Hideo Fujii
OriginLab
Edited by - Hideo Fujii on 08/09/2005 2:40:51 PM |
| altomare |
Posted - 08/08/2005 : 11:05:44 PM
Hi Xinjun,
were you able to solve your problem? I am in a very similar situation and I would beinterested in any help
Thanks
Fabio |
| tjuxjliu |
Posted - 05/09/2005 : 9:24:48 PM
Hi, Leo, Pete,
Thank you for your replies.
Yes, my function is complicated and impossible to present in the form of y = f(x). It concludes hyperbolic cotangent function, eg. y=cosh(x,y). This is from Brillouin function for paramagnetic thoery in Physics. In addition, Pete, woud you like to tell me the other software to resolve this fitting? I hope this problem will be considered in the new edition Origin.
Xinjun
|
| peter.cook |
Posted - 05/09/2005 : 09:10:06 AM
Hi, Xinjun,
Origin cannot fit to implicit functions ie of the type y=f(x,y).
It would be a great new feature though, as a lot of more complex pharmacological (compartmental) models cannot be resolved into the form y=f(x).
There is other software out that can do this...eg other NAG algorithms not incorporated with Origin ... but hopefully this would be in the Origin pipeline.
Cheers,
Pete
|
| Leo_Li |
Posted - 05/09/2005 : 05:31:06 AM
Hi, Xinjun,
My understanding is that a variable cannot be both dependent and independent. For expression y=f(x,y), you can usually convert it to f1(x,y)=0, then y = f2(x). Do you mean that your function is complicated and impossible to present in the form of y = f(x) ?
Leo |
|
|
