[R-sig-ME] phi values from corARMA etc. [was Re: R-sig-mixed-models Digest, Vol 71, Issue 10]

[Ben Bolker]

Arnaud Mosnier <a.mosnier at ...> writes:

> Thanks Ben, as I answer to Alain Zuur, I will try providing a starting
> value for Phi (different from the default value 0).
> I just need a way to decide what kind of value would be useful, i.e. as in
> viewing a plot of correlation vs time.
> My problem is that my data are taken at an irregular time step (Argos data).
> 
> Arnaud

  If you have irregularly spaced time steps, you probably want to
start with corCAR1.  I don't think lme has any built-in capability
to do more complex continuous-time correlation structures (i.e.
something like corCARMA).  If you really need it you might have to
write your own corStruct (ugh) ... but especially because you report
that there's not much going on, you might be able to get away with
corCAR1 ... ?

[snip]

> >
> > Arnaud Mosnier <a.mosnier at ...> writes:

 [snip]

> > > I made a mixed model of the form
> > >
> > > lme(Distance ~ season, random = ~1|yearTagged/tag, correlation =
> > > corAR1(form= ~Time|yearTagged/tag), data = dat)
> > >
> > > or
> > >
> > > lme(Distance ~ season, random = ~1|yearTagged/tag, correlation =
> > > corARMA(form= ~Time|yearTagged/tag, p=1, q=1), data = dat)
> > >
> > > In the AR1 model, the Phi value obtained for the corAR1 structure is 0.
> > > In the ARMA(1,1) model, both Phi1 and Theta1 are equal to 0.
> > >
> > > I find this strange. I believed first that it was caused by the fact that
> > > yearTagegd and tag variables were not considered as factor, but it did
> > not
> > > change anything if I convert them to factor.
> > > The temporal autocorrelation is not obvious in the residuals of the model
> > > without a temporal structure included (i.e. without corAR1 or corARMA),
> > but
> > > can I conclude that there is not correlation based on Phi value ?

 [trimmed to make Gmane happy]