Re: [ferret_users] fourth-order polynomial least square

[Brockmann Patrick <Patrick.Brockmann@xxxxxx>]

> I would like to fit a fourth-order polynomial least-square fit (a 
> curve) to a time-series. Is there a chance to achieve this with 
> Ferret? I know regresst.jnl, but it only does a linear regression. 
> Your effort or any help would be appreciated.

Hi Peter,
Relative to this polynomial fit, you may find useful some interpolation 
technics
from http://paulbourke.net/miscellaneous/interpolation/

I compute some of them as external functions (applied on zaxis).
Studies on parameters (biais, tension) from the Hermine interpolation 
implentation gives:
http://dods.ipsl.jussieu.fr/brocksce/ferret_ef/zinterpolate/spline_hermite_bias.swf
http://dods.ipsl.jussieu.fr/brocksce/ferret_ef/zinterpolate/spline_hermite_tension.swf

External function can be found from:
http://dods.ipsl.jussieu.fr/brocksce/ferret_ef/zinterpolate

Patrick

--
LSCE/IPSL, Laboratoire CEA-CNRS-UVSQ
Data Analysis and Visualization Engineer
--