Convergence Study – Smoothing

[joemParticipant]

Hi,

I’m trying to plot some points of a convergence study. When I select ‘Line type: Smooth’ the outcome doesn’t feel accurate.
I also tried to use the ‘Plot action-> Smooth -> Average’. But the values are quite distant from the real ones.

Is there a way to solve this problem?
Attached (one the left) the graph made with Datagraph (standard plot in red and with averaged values in black) and the ‘desired’ outcome made with Excel (right picture).
Thank you for your support,
Best regards
Joe

Ps. This is the table with the real values:
x, y
0.25, 1868.7
0.5, 2042.2
1, 2069.7
2, 1917.6
4, 1898.9
8, 1900.6

 

[dgteamModerator]

The smoothing method we use is a normal spline.  It is continuous in the first and second derivative.  This is a C2 approximation.

It looks like the smoothing function in excel is continuous in the first derivative only.  This is a C1 approximation. C1 approximations are not as smooth but can be less oscillatory, as seen with your data.

There are a lot of different smoothing approaches and we need to spend some time investing what other smoothing options to include.  If you find any specific information on the smoothing approach excel used that would be interesting.  We did some searching but did not find this yet on line.

 

[Author]

Posts