exponential fitting in LabVIEW 7

[rg81]

Hi,

I have a question to ask....

Is anyone know if the exponential fitting using LaVIEW 7 is weighted or unweighted? Because I'm surprise to find the Standard Deviation as input.... so, that's mean that maybe the subVI uses the SD to make a weighted fitting.... but I couldn't find in the Help any information about that. 

Thanks

[SMeserve]

laser,

I took a look at the code and the function's deatiled help.  I noticed a particular line that might help:

"Standard Deviation is the standard deviation, sigma[i] for data point (x[i],y[i]). If they are equal or you do not know, leave this array empty. Internally, LabVIEW sets all to 1.0."

So, if you leave the standard deviation input empty, the exponential fit will be equally weighted.  I hope this helps.

Thanks,

[rg81]

Sorry, maybe I didn't explain well what  my problem is. I know that to estimate the damping and the amplitude of the exponential curve the Exponential Fit VI uses a logarithmic scale and fits the line using the linear regression. Now, what I would like to know is if the linear fitting is going to minimize the chi-square as the General LS Linear Fit VI (for example) or not. If it is like that I can understand the meaning of the SD and the resulting fitting it will be weighted.

Thanks

[rg81]

Hi smeserve, thanks for your answer.

What do you mean with "So, if you leave the standard deviation input empty, the exponential fit will be equally weighted"?

it means the exponential fit is normally weighted? (because in LabView 7 the Linear Fit is unweighted...) 

and how the VI uses the weight?

Sorry but I'm a bit confused.

Many thanks

[SMeserve]

laser,

I'm not too familiar with what each function does on the lower levels, but the help document that tells us the formula that the VI uses to calculate the residue.  If you scroll down in the context help, you can see a line that reads "The VI calculates mse using the following equation.", which is followed by an equation.

This is all I can offer right now, but I am going to check around the department and see what I can find.  I appreciate your patience.

Thanks,

[SMeserve]

laser,

In the meantime, I have a suggestion.  You might want to try setting up a sample VI to test some data.  Take data that you have tested in another program to determine the fit and put it through LabVIEW's fit with various options.  You might be able to see similar data and match up the LabVIEW fit to one of your other methods.

Additionally, I noticed in LabVIEW 8.2.1 that the exponential fit VI has become considerably more robust.  It now has 6 inputs: X, Y, Weights, tolerance, method, and refine.  You can see more about these options in the exponential fit VI in 8.2.1 by navigating to this link.

Exponential Fit (Not in Base Package)
http://zone.ni.com/reference/en-XX/help/371361B-01/gmath/exponential_fit/

Thanks,

[rg81]

Thanks mseserve.

I already tried to compair the fitting using different VIs. I used the Exponential, Linear and General LS Linear Fit VI and I was surprise about the very good results obained with the Exponential Fit VI. For this, I was curious to understand what really happens in the Exponential VI. Anyway, now I discovered that one of my co-worker has LabVIEW 8.2 and infact I'm thinking to try the Exponential Fit and to test the different kind of methods available with this VI.

One thing, can I install LabVIEW 8.2 without remove the older version? I would like to keep the LabVIEW 7 for now (just in case....)

Many Thanks

[SMeserve]

laser,

Yes, installing LabVIEW 8.2 side-by-side with LabVIEW 7 is completely doable.  I have all LabVIEW versions from 7 to 8.2.1 on my machine right now!

Thanks,