| T O P I C R E V I E W |
| viggozhang |
Posted - 03/19/2012 : 08:46:30 AM
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System:
Hehe, questioning time again:
Firstly, I'd like to ask whether it is possible to do a numerical integral in Origin.
Secondly, if not,then I do the numerical integral with some other softwares,e.g. Mathematica or Matlab.Then, I import the calculation results into Origin and obtain a column of data. The question is whether I can use this column of data as part of a self-defined fitting function.
Anyway, many thanks in advance as usual. |
| 4 L A T E S T R E P L I E S (Newest First) |
| Sam Fang |
Posted - 04/12/2012 : 04:05:36 AM
If you just wants to do some operations on Z to approximate Y, then it a Linear Fit issue indeed, Z as independent variable and Y as dependent variable.
If I misunderstood your problem, you can click Send File to Tech support button on the top right of the forum, and send your project file to us, then it may help us understand your question better.
Thanks.
Sam
OriginLab Technical Services |
| viggozhang |
Posted - 04/11/2012 : 08:53:40 AM
Hey Fujii,
thanks for your info as usual.
Unfortunately, that is not what I meant to ask... Maybe I can rephrase my question as follows,
I have a set of data, i.e., (X,Y). I'd like to fit this set of data with a model in which there is a definite integral without analytic solution. Since it has no analytic solution, I did a numerical approximation for the integral with "Mathematica", and I obtained the results of the model as (X,Z). I was expecting that Z can approximately equal to Y, while it is now not the case. Thus, I'd like to do some operations, e.g., f(Z) on Z to make f(Z)=a*Z+b ~ Y, where a & b are fitting parameters. Now, my problem is to fit the curve (X,Y) with (X,f(Z)), and the difficulty is that my "Z" is a whole column of numbers rather than an expression. So, I am wondering whether there is a way to write something like f(Z)=a*Col(Z)+b as a self-defined fitting function and let "Origin" get me the values of a & b.... |
| Hideo Fujii |
Posted - 03/19/2012 : 6:00:05 PM
Hi viggozhang,
Origin can perform the numerical integration as self-contained.
Regarding the linear transformation by using the slope and intercept from the fitting pre-calculation, sure, you can use them by performing linear fit preceded; But, I don't understand the advantage of this because once you have the values of these parameters, the integration result would be simply the result of linear transformation of your original raw result of the integration. (Maybe a different story, if you use a non-linear fitting.) You don't have to perform the integration on the transformed data by P1+col(C)+P2. When the original integration is A, then the integration of the transformed curve would be simply Xrange*P1+A*P2 .
Possibly I have misunderstood your problem.
--Hideo Fujii
OriginLab
|
| viggozhang |
Posted - 03/19/2012 : 09:24:36 AM
to specify my questions,
I have a set of experimental data [Col(A),Col(B)], and I built up a model within which there is a definite integral with no analytic solution. I did numerical integral with other software and obtained the calculation results as [Col(A),Col(C)]. However, it turns out that my model is not as sexy as I thought and thus, I'd like to induce a few fitting parameters into the model by doing some basic calculations, i.e. P1*Col(C)+P2 where P1 and P2 are fitting parameters, and Col(C) corresponds to my argument Col(A) element by element. |
|
|
