Bug in MSE output of General LS Linear Fit.vi?

[Dragon64]

Hi: In the General LS Linear Fit.vi, output MSE. Help for the vi describes the formula as (I attached a picture, but I'm not sure if it will appear in the post), but it looks like something is missing, because if I use the formula to my input values, I get a match only if all input Standard Deviations are 1's. As Standard Deviations increase, this formula gives me smaller MSE, higher MSE if Standard Deviations decrease; but I noticed MSE of the vi doesn't change so much, and every time all input Standard Deviations have the same value (no matter how large or small they are), MSE returns to the same original value.
According to the formula, MSE is inverse proportional to the square of Standard Deviations, no way to return to original value if they're changed. Is this formula correct or shuld it be different?

[MehakD]

Hi

Could you guide me to the source of the formula you obtained for the mse. The formula mentioned in the online help can be seen in the bmp image attached to this post.

The link for this information is:

http://zone.ni.com/reference/en-XX/help/371361A-01/lvanlsconcepts/curve_fitting_in_labview/

Also, I noticed that the formula you have mentioned is the formula for Chi^2 in the context help for the General LS Linear Fit.vi:

http://zone.ni.com/reference/en-XX/help/371361A-01/gmath/general_ls_linear_fit/

Please let me know the source of the formula that you sent in your previous post so that I can correct any misrepresented information. 

I hope this helps!

[Dragon64]

Hi:

     In the second link you provided, I cannot find any formula for the MSE, in the first link it shows a formula, but it's the "normal" formula I know when Standard Deviation is an output of the function (General LS Linear Fit in this case) and therefore is not in the formula of MSE.
     However, General LS Linear Fit.vi has individual Standard Deviations (SD) as inputs (not outputs), and MSE varies as I change any SD. This indicates me that those input SD are in the formula to calculate the output MSE.

     I attached a picture of the Help Menu where I obtained the formula in my first post. This is from LV5, after help window is open I click on the Question mark, then it pops up another window with Help Topics and an Index, when I select "General LS Fit Theory" it shows me details of General LS Linear Fit including the formula for MSE. But even I include the input SDs as stated in this formula, MSE outputs are not the same outputted by the vi if I change SD values.

     Please let me know if there's a bug or some other formula is applied.

[MehakD]

Can you provide me with the dataset that you are using (text or xls format). I would like to use those to compare the values obtained in LabVIEW 5.0 to LabVIEW 8.5. The documentation on 'General LS linear fit.vi'  is different on both the versions and this VI seems to have been updated since 5.0. Comparison of the outcome of the two functions will help  determing whether this is a bug and will help in narrowing down the issue.

[Dragon64]

Of course. Here they are.
The way I noticed something is that, keeping other input values unchanged, I change Standard Deviation individual values, whenever they are different, MSE changes and show something different (as if they're part of the formula to calculate MSE), but whenever they're the same, output of MSE becomes as if all were 1.

[Karunya_R]

Hi Dragon,

Can you post your exact program to this thread? I would like to run it and see the outputs.

[Dragon64]

Hi, Karunya:

     I actually rechecked my stuff, and I found that the formula that Mehak_D showed me in

MSE formula.JPG

is correct and really works. and the one I found in one of the LV 5 help files were wrong

untitled6.jpg

I'm attaching 2 pictures showing the LV5 Help file where I found it, as well as the Index where I clicked to open it. Just let me know if the formula were corrected in the Help file for later LV versions.

     Thank you very much.

[Kameralina]

Hello,

Thank you so much for your feedback. I'll need some clarification regarding the mistake in the formula that you are speaking of.  The formula in the picture untitled6.jpg does not appear to be in the screenshot of the help file which you posted.  Can you pinpoint where you see this mse formula?

This is what the current help file looks like:
General LS Linear Fit Theory - LabVIEW 8.5 Help

Please let me know whether you still see the error and if so on which line.


Message Edited by FieldKam on 03-04-2008 01:09 AM

[Dragon64]

Hi, FieldKam:

     I checked in the webpage you show me, and effectively there's no formula for MSE (or at least not obvious). The one I got were from LV5, which I just expect is gone and updated already in LV8.5.
     But doing a search in the Internet, using the frase "General LS Linear Fit Theory" I found a PDF document for Labview named Analysis Concepts, dated as July 2000 Edition, Part Number 370192A-01, at page 4-6 (44 of 82), in the bottom, it shows the formula as I found it in the LV5 documentation. Just verify if the current equivalent document still shows that formula or the one Mehak_D gave me.
     A link to the PDF file is here: http://www.engr.uconn.edu/~ewebhk/buttons/data/labview/manuals/analysis-concepts.pdf.

     Thank you very much.

[MehakD]

Dragon64,

I researched this issue before in the current documentation and have not found the formula that you found in LabVIEW 5.0. This issue doesnt seem to exist in the documentation for the current version LabVIEW(8.5). I do appreciate your feedback on this matter and am sorry for the inconvenience the mistake might have caused you.

Thanks!