Wednesday, October 13, 2010

Smoothing noisy data with GNU Octave/Matlab

Every once in a while, I find the need to summarize simulation or experimental data that are noisy using a smooth function. Smoothing splines try to pose the required regression as a least-squares problem.

A nice free program is the excellent Matlab version of Jerome Friedman's "supersmoother" by Douglas Schwarz. The program works fine without any tweaks on GNU Octave.

Take it out for a ride.

No comments: