## Curve smoothing using Matlab

### Shalini (view profile)

on 6 Apr 2012
Latest activity Commented on by Samuel Suakye

### Samuel Suakye (view profile)

on 6 Jun 2017
I have got an experimental curve which is a bit wavy (curve has a lot of noise and hence is wavy).
I want to smoothn the curve using Matlab.
Can anyone advise the best tool in Matlab for this?
I wanted to attach the plot of curve which needs to be smoothened but couldn't find a way to attach it here.
shalini

### Tags

Answer by Jonathan Sullivan

on 6 Apr 2012

help filter
doc filter

Shalini

### Shalini (view profile)

on 6 Apr 2012
Do you mean:
yy = smooth(y)
where; yy is the smoothed vector? And, y is the dependent variable (of the curve) that needs to be given?

Answer by Richard Willey

### Richard Willey (view profile)

on 6 Apr 2012

I have a function called FitIt on the file exchange that might prove useful.
FitIt combines
Local regression (to smooth your data set) Cross validation (to estimate an optimal spanning parameter) Bootstrap (to generate confidence intervals)
FitIt has dependencies on both Curve Fitting Toolbox and Statistics Toolbox.

Show 1 older comment
Richard Willey

### Richard Willey (view profile)

on 10 Apr 2012
Hi Shalini
http://www.mathworks.com/matlabcentral/fileexchange/31562-data-driven-fitting-with-matlab
Please note: Under the hood, this function uses the "Smmoth" function that Sean is recommending. The major difference is that I added some code that automatically computes an optimal spanning parameter for the smooth function and also gives you the option of computing confidence intervals.
There is a recorded webinar available on this topic at
http://www.mathworks.com/company/events/webinars/wbnr56627.html?id=56627&p1=961661709&p2=961661727
Taymaz Tek

### Taymaz Tek (view profile)

on 9 Jun 2012
Hi everybody,
I had been trying to use your function, fitit, in my own problem but it didn't work well. In the case of your demo, it is quite ok but when I try to run it with my signal, there is an error:
??? Error using ==> horzcat
CAT arguments dimensions are not consistent.
Error in ==> fitit at 21
sse(j) = sum(crossval(f,[X,Y],'partition',cp));
can you please let me know where this can be driven?
Taymaz Tek

### Taymaz Tek (view profile)

on 9 Jun 2012

Answer by Sean de Wolski

### Sean de Wolski (view profile)

on 9 Apr 2012

How about smooth() :)
doc smooth

Shalini

### Shalini (view profile)

on 9 Apr 2012
I'm new to matlab...How to use this smooth?Why do you say doc smooth? Is there a document?
Sean de Wolski

### Sean de Wolski (view profile)

on 9 Apr 2012
If you run:
doc smooth
at the command line it will tell you all about the SMOOTH function, (well as long as you have a somewhat recent version of MATLAB with the Curve Fitting Toolbox)
Samuel Suakye

### Samuel Suakye (view profile)

on 6 Jun 2017
figure clear; clc; delta=[0.0259, 0.0518, 0.0776, 0.1035, 0.1293]*10; %delta=0.01:0.09:0.5; delta1=0.01:0.1:0.5; hbar=6.5821220*10e-16; k=8.617385*10e-5; T=300; d=[100 150 200]*10e-10; m = 9.1093897*10e-31; v = 2.8*10e8; c = 3.0*10e8; wp = 10e12; w = 0.7071; gamma = (1-((v^2)/(c^2)))^-0.5; delta2= delta1./(k*T); %% figure hold on for i = 1:length(d) wo=(wp^2*m*d(i)^2*delta1.*besseli(1,delta2))./(hbar^2*besseli(0,delta2)).^(0.5); delta3=delta./(k*T); therta = asind(1-(w.^2./wo.^2)).^0.5; wb = gamma.*wo.*cosd(therta); %wb^2=2 rho=(hbar^2.*2.*besseli(0,delta3))./(wp^2*m*d(i)^2*k*T.*delta3.*besseli(1,delta3)); plot(delta3,rho,'LineWidth',1.5) end %% xlabel('\Delta^*') ylabel('E_b/E_s') hold on grid on All comments are not working for my codes above