[time-nuts] suitable statistics for measurements with gaps

[Michael Wouters]

The technique used for dealing with gaps is really about handling random
gaps in an otherwise uniformly sampled sequence. The idea is that you take
your sequence, pad it out with the missing data (tagging those points with
a NaN or whatever) and then when you're computing ADEV, if a data triplet
is missing a point(s), you simply drop it from the summation.
This works nicely for ADEV and TOTDEV but not so well for MDEV.

There are a couple of implementations of this around: allantools (Python)
and tftools (Matlab/Octave) are at least two on GitHub.

Cheers
MIchael

On Sunday, 24 April 2016, Attila Kinali <attila at kinali.ch> wrote:

> Hoi Jim,
>
> On Fri, 22 Apr 2016 06:39:07 -0700
> jimlux <jimlux at earthlink.net <javascript:;>> wrote:
>
> > But what about when the observations have gaps? Say you're measuring the
> > frequency of a spacecraft oscillator, and you can only see it for 8
> > hours a day?  the description of the frequency variation at a time
> > difference of 24 hours is useful, even if the integration time for each
> > measurement is, say, 1000 seconds.
>
> I recently stumbled over a paper by Sesia and Tavella[1] that might
> be of use. I didn't read it yet, so I cannot say anything about its
> content but that it covers gaps in ADEV data for space clocks.
>
>
>                         Attila Kinali
>
> [1] "Estimating the Allan variance in the presence of long periods of
> missing data and outliers", by Sesia and Tavella, 2008
> http://stacks.iop.org/Met/45/S134
>
> --
> Reading can seriously damage your ignorance.
>                 -- unknown
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com <javascript:;>
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>