Poor Curve Fitting issues

I tried to fit a curve in LV 8.2.
I don't know why the result (attachment) was so bad and remained the same in the different weights of "general polynomial fit.vi".
 
Although I can't find the manual about mathematical functions, I expected the curve was well fit in start/end region with some weight.
In order to do, I made the weight (0.5 nearby middle and 1 start/end region) , but the gap in start/end X region of the attachment was rather big.
 
In Origin, commercial fitting program, the result was sufficient to use in other application.
 
What's the different between LV and Origin?
 
*)Annotation of red in color was incorrect in the attachment.???? 07-08-2008 09:08 AM? labmaster? ?? ?????


result.jpg:
http://forums.ni.com/attachments/ni/...4/2/result.jpg

Can you upload your data, and the snippet of code that you did. It seems to me that this Origin program isn't doing a curve fit at all, but rather a tracing, which is something completely different.

Thank you for the interesting.
 
Attachments are the demo data and my simple code in LV8.2.
(Code has a garbage for my various tryings such as derivative.)
 
The fitting was not changed according to weight and method in poly.vi.
 
The purpose of fitting is to get a base line of the graph.
The final target is to detect the position and shape (residual) of the small bumps in the graph.
(First and high order derivative was not helpful in my tryings.)???? 07-08-2008 06:33 PM? labmaster? ?? ?????


data_analysis.vi:
http://forums.ni.com/attachments/ni/...ta_analysis.vi


demo_data.txt:
http://forums.ni.com/attachments/ni/.../demo_data.txt

I want to cancel my posting.
There was a scaling in a graph.

OK. So presumably the LabVIEW curve-fitting worked for you then?