| Author |
 Topic  |
|
|
tantiger
Singapore
41 Posts |
 Posted - 11/06/2002 : 03:44:09 AM
|
I have a fitting like this:
y=P1/(P2*x+P3)
and I would like to set P3 as a constant.
I did according to the rules in the knowledge base:
1)Declare P3 in the Constants box and use a ";" to end that line,
2)When define my equation as above, only declare P1 and P2 but enter the whole equation in the equation field box.
3)Assign x and y values and start the fitting.
Problem comes up, poor initialization happened.
However, if I declare P3 as the 3rd variable but still put it in the constants declaration box, everything will be fine.
Then, why I cannot get the same result if I follow the procedures in knowledge base? |
|
|
easwar
USA
1964 Posts |
Posted - 11/06/2002 : 09:22:00 AM
|
Hi,
I could reproduce the problem in ver 6.1 and ver 7 when using the Equation form to perform the fit. It appears that the name P3 is still interpreted as a parameter even though one sets the number of parameters to 2.
If you change the name of your constant to something else (like PP3), then it works fine.
This problem does not exist in ver 7 when using Origin C for fitting .ie. you can use the name P3 for the constant and it works fine.
Easwar
OriginLab.
Edited by - easwar on 11/06/2002 09:23:43 AM |
 |
|
|
tantiger
Singapore
41 Posts |
Posted - 11/08/2002 : 02:15:18 AM
|
Actually, I am using version 7.
I have found my mistake. I checked "Compile before use" when define my equation and whenever I did my fitting, the script window will pop up and say "Error ..."
Thanks a lot for your reply, easwar! |
|
|
| |
Topic |
|
|
|
Using Constants in the NLSF