Prev: Probability matching
Next: polyfit in 3D
From: Olivier on 4 May 2008 14:55
Hi,
I 'm looking for a way to get a smooth curves plot with
Matlab from a cadence data set. So far my curves looks like
the following:
http://www.cems.uvm.edu/~olemaire/matlabExchange/curve.jpg
and my data set is:
http://www.cems.uvm.edu/~olemaire/matlabExchange/dataset.mat
(This is a parametric analysis from cadence)
can someone tell me if there is a way to get from this set a
plot with smooth shape. I've seen this thread:
http://www.mathworks.com/matlabcentral/newsreader/view_thread/159255
but it seems to works for only one curve and and I have from
5 to 20 to plot on the same figure.

Thank you very much for your help

Olivier

From: Olivier on 5 May 2008 20:30
Hi,
Obviously my question can't get any answers, maybe I need to
precise. In fact I want something know under gnuplot as
csplines or bezier:
http://t16web.lanl.gov/Kawano/gnuplot/plot2-e.html#5.8

Thank you very much for your help.

Olivier
From: NZTideMan on 5 May 2008 22:28
On May 6, 12:30 pm, "Olivier " <m.olivier.lema...(a)gmail.com> wrote:
> Hi,
> Obviously my question can't get any answers, maybe I need to
> precise. In fact I want something know under gnuplot as
> csplines or bezier:http://t16web.lanl.gov/Kawano/gnuplot/plot2-e.html#5.8
>
> Thank you very much for your help.
>
> Olivier

I have absolutely no idea what a "cadence data set" is. Cadence is a
phrase at the end of a piece of music isn't it?

Your data would be a challenge to smooth because the rising and
falling limbs are very steep. Indeed, there appear to be only a
handful of points defining these parts of the curves. Most smoothing
algorithms will struggle with this.

Anyway, if I had to draw a smooth curve through your data, I'd use
orthogonal wavelet decomposition. I'd try out various mother wavelets
at various levels of decomposition, discarding low timescale wavelet
details until the curve was smooth enough, while still retaining its
basic rectangular shape.
This involves two commands from the wavelet toolbox:
Decomposition:
[C,L]=wavedec(y,nlev,mother);
and recomposition of the approximation at a particular level:
app=wrcoef('a',C,L,mother,nlev);
 | 
Pages: 1
Prev: Probability matching
Next: polyfit in 3D