| Author |
 Topic  |
|
|
giordandue
Italy
Posts |
 Posted - 02/21/2006 : 10:14:16 AM
|
Origin Version 6.0
Operating System XP professional
Dear Friends,
I get as the best-fitting of my data a Lorentz function with these parameters:
Area=A= 0.27588
Center=xc= 1.0388
Width=w= 0.12717
Offset=y0=0.0065
Heigth=(yc?)= 1.3811.
In the Origin Manual I found that the basic Lorenz function is
y=y0+2Aw/(pigreek*4*(x-xc)^2+w^2)
The problem is that using the above parameters, I am not able to reproduce the best-fit curve in Excel spreadsheet. I suppose that
is depending to the parameter "Heigth (yc?)" that do not seem included in the basic Lorentz equation.
Do you have perhaps some suggestion?
Thank you in advance
Giordano |
|
|
easwar
USA
1964 Posts |
Posted - 02/21/2006 : 2:32:26 PM
|
quote:
The problem is that using the above parameters, I am not able to reproduce the best-fit curve in Excel spreadsheet.
Hi Giordano,
How are you trying to reproduce the fit curve in Excel? Did you use column formula with the exact same formula as in Origin, or some other definition of Lorentz that perhaps Excel uses?
Note that height is not required when there is Area parameter. The Lorentz function could however be written in a different form so that A is height at x=xc and not area, similar to how there is GaussAmp function in Origin instead of Gauss. If that is what you want, you can define your own user-defined function and fit.
One other thing to note is that when parameter values are printed to the graph etc, there could be rounding off, and thus loss of significant digits. That could also affect the process of trying to reproduce the values in another program using the printed parameter values. In tbe NLSF tool, under the Options->Control page, you can set the significant digits for each parameter.
Easwar
OriginLab
|
 |
|
|
giordandue
Italy
Posts |
Posted - 02/21/2006 : 5:40:42 PM
|
Thank you Easwar for your replay: in reality it is not only a problem with Excel, but also of reproducing the same curve in Origin.
I try to be more clear: (a) I fitted some point and I got as best-fit the Lorentz function, with the parameters I showed in my previous message; (b) I found the Lorentz equation in Origin manual and I used this equation with the parameters obtained, to verify the fitting curve. I observed that the fitting curve and the added Lorentz curve are not coincident (the latter seems a little right shifted) and this is the reason of my question: entering the resulting parameters in the Lorentz equation I am not getting the fitting curve.
Sorry, I am not a specialist and probably I am not understanding completely your suggestion. Thanks for some additional help.. Ciao |
|
|
|
easwar
USA
1964 Posts |
Posted - 02/22/2006 : 09:39:44 AM
|
Hi Giordano,
Can you please send your data to tech support, or paste the data here if not too large?
Thanks,
Easwar
|
|
|
|
giordandue
Italy
Posts |
Posted - 02/22/2006 : 11:16:28 AM
|
ok: in the attached graph in the e-mail , the red line is the Lorentz fitting and the black line is curve obtained putting in the Lorentz equation the parameters of the fitting.
Than you again for your help and kind regards from Alassio (on the Ligurian sea, in north-west Italy)
Giordano |
|
|
|
giordandue
Italy
Posts |
Posted - 02/22/2006 : 11:25:18 AM
|
I am trying again to send the file as an image..
Of course is a .opj file, but it is not loaded (?)
and the same occurs with a .org file..(??)
Do you have an e-mail address?
Giordano |
|
|
|
easwar
USA
1964 Posts |
Posted - 02/22/2006 : 1:28:24 PM
|
Hi Giordano,
Please send it to tech@originlab.com and please send OPJ with the data.
Thanks,
Easwar
|
|
|
| |
Topic |
|
|
|
Lorentz function