| Author |
 Topic  |
|
|
MrDavid
Posts |
 Posted - 11/21/2007 : 10:10:48 AM
|
Origin Version (Select Help-->About Origin):
Operating System:
Hello,
I am trying to fit a curve to my data, and I have tried using the two 'spline' options. Ideally, I would like a smooth looking line, which accurately represents the data points but looks more attractive than straight line connectors.
Spline does not give me a smooth enough line, but B-Spline gives me a curve that misses some of the data points by quite a way. Is there any way of controlling how 'tightly' the curve fits the data points?
Thank you, David |
|
|
stoelzer@p
Germany
5 Posts |
Posted - 04/02/2008 : 06:03:13 AM
|
Did you find a solution for your problem? I'm really interested in that solution.
While looking through the programming help I found this:
Function: Curve_bspline
Prototype: int Curve_bspline(curvebase* pcrvData, curvebase* pcrvResults, double dSmoothness = BSPLN_AUTO, Dataset* pdsWeights = NULL)
Changing the parameter dSmoothness and recalculating the B-spline curve is probably what we want to do. But I don't know if and how that is possible without much programming.
Is there any expert who can answer this question?
Many thanks,
Mathias
|
 |
|
|
PeterKunhardt
USA
Posts |
Posted - 04/11/2008 : 10:14:32 AM
|
The central difference between spline and b-spline is that with spline the curve is forced to fall exactly on every data point, but with b-spline the algorithm used to determine the curve allows the y-values at given x's to float. This is what allows the b-spline algorithm to appear so "smooth."
The Origin C command suggested above will draw a b-spline curve. If you were to set the parameter dSmoothness to "0" then you would get the effect you are looking for: drawing a b-spline curve that connects through every single data point. However, if you compare the result with a line and symbol plot with line type "spline," you will see the results are exactly the same.
You can also demonstrate this by running an interpolation of a symbol plot. You can choose b-spline as your method with a smoothness of 0 in the interpolation dialog, but you will see again that it produces the same result as interpolation with spline. |
|
|
| |
Topic |
|
|
|
B Spline does not fit data tightly enough