You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@commons.apache.org by lu...@univr.it on 2013/05/08 19:54:23 UTC

[math] smoothing techniques

Dear all,
I would like to know if in math commons is implemented an algorithm that can be used for cleaning a noisy time series.
I have the following problem: in my Java application I need to run some calculations on a set of time series of experimental data. Before applying my algorithms, I would like to apply a suitable smoothing technique for reducing the noise in the data.
Usually, when I work with Matlab, I solve the problem by means of the fitting toolbox and, more specifically, by substituting the noisy time series with those obtained by applying the Smoothing spline fitting algorithm (which is, more or less, the same of the smooth.spline algorithm implemented in R).
Is there something similar implemented in math commons?


Thank you very much for the support,


Luca

Re: [math] smoothing techniques

Posted by Thomas Neidhart <th...@gmail.com>.
On 05/08/2013 07:54 PM, luca.marchetti@univr.it wrote:
> 
> Dear all,
> I would like to know if in math commons is implemented an algorithm that can be used for cleaning a noisy time series.
> I have the following problem: in my Java application I need to run some calculations on a set of time series of experimental data. Before applying my algorithms, I would like to apply a suitable smoothing technique for reducing the noise in the data.
> Usually, when I work with Matlab, I solve the problem by means of the fitting toolbox and, more specifically, by substituting the noisy time series with those obtained by applying the Smoothing spline fitting algorithm (which is, more or less, the same of the smooth.spline algorithm implemented in R).
> Is there something similar implemented in math commons?

Hi Luca,

to smooth the output of a stochastic linear process, you can use the
KalmanFilter implementation (see
http://commons.apache.org/proper/commons-math/userguide/filter.html).
You may need to setup your process model and define the estimated
(gaussian) noise to get good results, just ask if you need help.

There are also several (polynomial) fitters in the fitter package, which
may do what you have in mind.

Thomas

---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
For additional commands, e-mail: user-help@commons.apache.org