| Author |
 Topic  |
|
|
bühler
France
Posts |
 Posted - 03/22/2005 : 10:16:41 AM
|
Origin Version : 6.0 SR1
Operating System: Windows 2000
Hello,
I would like to ask a question. I think it must be quite easy but I'm not able to find the way how to solve my problem, and I'm already searching for some days… (I have to mention that I'm a beginner in working with Origin...).
I will specify my problem: I have got a x-y-diagram with several points. My aim is to draw an interpolation curve through this points. In other words, one curve that describes the developing of this points (a best fit straight line!).
But the problem is: these points are not coming from one data sheet! I had some data sheets and I wanted to have one diagram including the several data sheets, so I created one only diagram (with the points of the several data sheets). Furthermore, my points are quite specific: I will simplify the problem:
I have three data sheets:
1)
A(X1)
1E10
1E10
1E10
B(Y1)
1.0
1.01
1.02
2)
C(X2)
1E11
1E11
1E11
D(Y2)
0.94
0.95
0.98
3)
E(X3)
1E12
1E12
1E12
F(Y3)
0.59
0.70
0.66
That means I have three times three points vertically over each other.
Now I want Originlab to create a function describing the developing of the points (from 1E10 to 1E12). As I don't know the function I can't use one of the given functions. Is it possible that Originlab create this function for me and tell me the parameters of the found function? Or do I have to specify the function that I suppose to be the function?
I hope you understand what I mean…
Greetings
|
|
|
easwar
USA
1965 Posts |
Posted - 03/22/2005 : 11:25:46 AM
|
Hi,
I think the best option is to first combine the data into one worksheet by averaging the y values at each x, so you then have the following:
X Y YErr
1E10 1.01 0.01
1E11 0.95667 0.02082
1E12 0.65 0.05568
where the last column is standard error on the Y values computed as part of the averaging process.
You can then set the third col to YErr and plot the data in a graph and then fit the data.
Now, about your question on fitting, you need to decide what is the best equation for representing your data, by either fitting with various functions available in Origin, and/or creating your own user-defined function.
Easwar
OriginLab
|
 |
|
|
Mike
USA
357 Posts |
Posted - 03/22/2005 : 11:31:52 AM
|
Hi bühler:
Origin does not suggest a model to use, so you will have to find one or more that you feel might be suitable and test them with your data. I assume that you know that you do not have to use Origin's built-in fitting functions; you can define your own function.
Once you have such a model, you can use the nonlinear fitter's Simulate page controls to test various models and parameter combinations and narrow your choices.
If I understand you correctly, it sounds like you would like to fit multiple datasets to a common fitting function and this is certainly possible -- either using separate parameter values for each dataset or using a common set of parameter values for all datasets. I would suggest that you might get started by searching through our knowledge base for information on the NLSF.
(http://www.originlab.com/www/support/querytech.aspx?keywords=nlsf&boolean=and&scope=All&category=All&type=All&version=7.5,7.0&language=English)
Mike
OriginLab |
|
|
|
bühler
France
Posts |
Posted - 03/23/2005 : 07:10:03 AM
|
Thank you for our answers! I think now I will be able to solve my problem...
|
|
|
| |
Topic |
|
|
|
fitting problem