| T O P I C R E V I E W |
| ma0jae |
Posted - 03/27/2014 : 9:11:52 PM
Origin Ver. and Service Release (Select Help-->About Origin): OriginPro 8.0.63.988 SR6
Operating System: windows 7
Hi.
I am using the "Origin Pro 8.0" to fit the experiment data. And I am faced with some problems.
For simplicity, I give you an following formula.
y = a * x + b
x is independent variable
y is dependent variable
a,b are fitting parameters
When the fit precessure done. Two sheets appear. The name of the sheets are "FitNL1" and "FitNLCurve1".
In the first sheet "FITNL1", there are two calculated fitting parameter "a" and "b". Using these two parameters and the formula "y = a * x + b", I can calculate the y values --- ич
And In the second sheet "FitNLCurve1", there are fitted y values --- иш
I think that the two y values, ич and иш, should be same. But there is a difference between two y values. I know the difference is very small in this case. But when I use more complicated formula, the difference is very high. For example, in my real formula, the difference is about 15%.
please explain why this difference exists.. |
| 3 L A T E S T R E P L I E S (Newest First) |
| jasonzhao |
Posted - 04/10/2014 : 05:31:31 AM
The procedure of your understanding is right. If your have problems when using Fitting please contact us again.
Jason Zhao
OriginLab Tech Service |
| ma0jae |
Posted - 04/08/2014 : 9:15:29 PM
I have tried again in case of the preceding formula using the method you told me. It really works well. The two y values(one is in the sheet "FitNLCurve1" and the other one is calculated using the parameters which are in the sheet "FitNL1") are perfectly same.
I have one more question. I want to ensure the following fitting procedure. When <OriginPro8.0> fit the data. First, <OriginPro8.0> finds the parameters that minimize the chi^2. And the parameters are indicated in the sheet "FitNL1". Second, Using the parameters in the sheet "FitNL1", <OriginPro8.0> calculates the fitted dependent variable y which is indicated in the sheet "FitNLCurve1".
If the procedure is correct. Any kind of fomula, whatever the formula is very long or complicated, should satisfy that the two y are perfectly(?) same. Right? If I misunderstand about the fitting procedure, please give me advice. |
| jasonzhao |
Posted - 03/28/2014 : 01:00:51 AM
In order to fix this problem, please increase the number of significant figures of fitting parameters "a" and "b" in the first sheet "FITNL1". You can follow the steps below:
Tools in the main menu=>Options=>Numeric Format=>select a higher value for Number of Diecimal Digits=>pressing OK
Then, clicking Yes for the pop-up Window "Save as Origin startup Options"
You can use the new a, b values with more significant figures to calculate the fitting line, the results will be much closer to the fitted y values in the second sheet "FitNLCurve1".
Best regards,
Jason Zhao
OriginLab Tech Service |
|
|
