| T O P I C R E V I E W |
| amit_concordia |
Posted - 10/19/2005 : 12:10:05 AM
Hello,
I am new to this software and the user forum. I know that we can fit a gaussian model to a particular data or fit a user defined function to get a desired fit for any data.
Is it possible to fit an user defined equation to a dataset uising least square fitting and find one particular parameter of that user defined equation. For example if we have say the plume dispersion data (concentration on y axis vs. distance on x axis) whose distribution looks gaussian, can we find the standard deviation by fitting a gaussian plume dispersion equation using least squares.
thanks
|
| 3 L A T E S T R E P L I E S (Newest First) |
| easwar |
Posted - 11/07/2005 : 10:20:07 PM
Hi Amit,
Apologies for the delay in response. Yes, you can create your own user-defined function, and with data such as this, you can further fix the ymax and xc values and just let sigmax vary in the fitter.
See screen shots below. Please contact tech support if you need further help.
Easwar
OriginLab
NLSF screen shot showing user-defined function definition:
Screen shot showing fit result and NLSF fit page with ymax, xc fixed and only sigmax allowed to vary (just start with some initial value for sigmax such as 1. Fitter will iterate and converge to optimal value).
Edited by - easwar on 11/07/2005 10:21:02 PM |
| amit_concordia |
Posted - 10/19/2005 : 12:43:21 PM
Hello,
I am still not 100% sure but I think what you told me might work. Let me try to be more clear buy giving an example.
if we have a data set something like this
x y
-4 0
-3 1
-2 2
-1 3
0 4
1 3
2 2
3 1
4 0
the distribution looks close to a gaussian, If I use origin to fit a guassian model it uses it own equation. Now if I want to fit my equation say
y(x) = ymax. exp[-(x-xc)^2/(sigmax^2)]--------eq1
where ymax is the maximum value in y set
xc is the central value which is 0 in this case
Now can I fit my equation eq1 using least squares and find the value sigma x.
Also my email id is amit_concordia@yahoo.com.
thanks
Amit |
| easwar |
Posted - 10/19/2005 : 09:21:05 AM
quote:
Is it possible to fit an user defined equation to a dataset uising least square fitting and find one particular parameter of that user defined equation.
Hi,
By the above if you are asking whether only one parameter can be varied and all others set to be fixed to known values, yes that is possible. You can use the NLSF tool and for each parameter in the equation there is a corresponding check box that determines if that parameter should be varied or not during the fitting process.
If this is not what you mean, please explain.
Easwar
OriginLab
|
