R help - Dec 2010 - ideas, modeling highly discrete time-series data

Hello all,

    First of all, thanks so those of you who helped me a week or so ago
managing a time series with varying gaps between the data series in 'R'.
(My final preferred solution was to use "its" function & then
forecast(Arima( ) ).  )

    My next question is a general statistical question where I'd like some
advice, for those willing / able to proffer any wisdom:

   - I need to predict using this same time series, where the *data* are
   highly discrete.  E.g., I will have values like 1e5, 2.2e5, and 3.6e5, but I
   will never have 1.3e5 or 1.8e5, etc.
      - I could simply leave these values as discrete, similar to a binomial
      distribution, but then I am not sure how to use time series tricks like
      arima above.  For time-series analyses that I know of, an
assumption of an
      approximately normal distribution is expected.  No simple normalization
      (e.g., log(values) ) works, since the non-normality arises from
the highly
      discrete distribution more than any drastic asymmetry in the population
      spread.
      - I could leave the values as they are an work with a model where the
      assumption is violated... I am not sure how sensitive a model
such as arima
      is on the population distribution
      - Or I could... (here's where I am hoping for some collective genius).

    Thanks in advance for any help!  If this isn't the best forum, since I
know this is not specifically an 'R' question, please let me know of a
better forum to post such a question.

                                                      Thanks!
                                                               Mike



"Telescopes and bathyscaphes and sonar probes of Scottish lakes,
Tacoma Narrows bridge collapse explained with abstract phase-space maps,
Some x-ray slides, a music score, Minard's Napoleanic war:
The most exciting frontier is charting what's already here."
  -- xkcd

--
Help protect Wikipedia. Donate now:
http://wikimediafoundation.org/wiki/Support_Wikipedia/en

	[[alternative HTML version deleted]]

You could try the timeseries list at
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=TIMESERIES

kjetil

On Mon, Dec 20, 2010 at 6:26 PM, Mike Williamson <this.is.mvw at
gmail.com> wrote:
> Hello all,
>
> ? ?First of all, thanks so those of you who helped me a week or so ago
> managing a time series with varying gaps between the data series in
'R'.
> (My final preferred solution was to use "its" function & then
> forecast(Arima( ) ). ?)
>
> ? ?My next question is a general statistical question where I'd like
some
> advice, for those willing / able to proffer any wisdom:
>
> ? - I need to predict using this same time series, where the *data* are
> ? highly discrete. ?E.g., I will have values like 1e5, 2.2e5, and 3.6e5,
but I
> ? will never have 1.3e5 or 1.8e5, etc.
> ? ? ?- I could simply leave these values as discrete, similar to a binomial
> ? ? ?distribution, but then I am not sure how to use time series tricks
like
> ? ? ?arima above. ?For time-series analyses that I know of, an
> assumption of an
> ? ? ?approximately normal distribution is expected. ?No simple
normalization
> ? ? ?(e.g., log(values) ) works, since the non-normality arises from
> the highly
> ? ? ?discrete distribution more than any drastic asymmetry in the
population
> ? ? ?spread.
> ? ? ?- I could leave the values as they are an work with a model where the
> ? ? ?assumption is violated... I am not sure how sensitive a model
> such as arima
> ? ? ?is on the population distribution
> ? ? ?- Or I could... (here's where I am hoping for some collective
genius).
>
> ? ?Thanks in advance for any help! ?If this isn't the best forum, since
I
> know this is not specifically an 'R' question, please let me know
of a
> better forum to post such a question.
>
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?Thanks!
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Mike
>
>
>
> "Telescopes and bathyscaphes and sonar probes of Scottish lakes,
> Tacoma Narrows bridge collapse explained with abstract phase-space maps,
> Some x-ray slides, a music score, Minard's Napoleanic war:
> The most exciting frontier is charting what's already here."
> ?-- xkcd
>
> --
> Help protect Wikipedia. Donate now:
> http://wikimediafoundation.org/wiki/Support_Wikipedia/en
>
> ? ? ? ?[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>