Re: correlation length of a series

[Frederic Parrenin <parrenin gmail com>]

Hi Andreas,

This is how I did it but I thought maybe it is suboptimal and there is
a generic procedure to do that.

I resampled (X_i,Y_i) by linear interpolation ('interpolation'
function) to (X'_i,Y'_i) where the X'_i are regularly sampled with a
step \DeltaX.
Then I calculated the Pearson correlation coefficient ('correl'
function) of (Y'_i,Y'_{i+n}) which is the correlation for n*\DeltaX.

Frédéric




2012/2/22 Andreas J. Guelzow <aguelzow pyrshep ca>:
> On Wed, 2012-02-22 at 12:30 +0100, Frederic Parrenin wrote:
> > Dear all,
> >
> >
> > I have a serie (X_i,Y_i).
> > I would like to draw a correlation diagram as a function of \Delta X.
> > How to proceed?
> Hi Frederic,
>
> please be a bit more precise of what kind of diagram you would like to
> create. None of Walsh diagrams, Tanabe-Sugano diagrams and Orgel
> diagrams seems to make sense in this context.
>
> Andreas



-- 
http://parrenin.frederic.free.fr/