[R] autoregression

Rolf Turner rolf.turner at xtra.co.nz
Sat Nov 19 01:31:24 CET 2011

On 19/11/11 05:54, Keith Thompson wrote:
> Hi,
> I am new to R and looking to do auto-regression / ARIMA type modeling.  My
> data has both date and time which I need to combine into a single date-Time
> value.  The time steps are unequal.  What package is best for doing the
> regression and plotting the predicted values against the actual data?
>   Also, what format does my data need to be in when I use the package?  For
> example, I looked at the cts package documentation and it has examples for
> data(V22174) *but I have no idea how the data in V22174 looks*.
Well, have a look at it!  If you don't know how to do *that* then you
have no idea how to drive R.  So take some driving lessons.  (I.e.
read the --- readily available --- introductory manuals.)

Looking at the *help* (type ?V22174) would, uh, help.  (The help
page is sloppily written, but more-or-less comprehensible.)
> Is it a ts
> object, a zoo object ... etc?
Actually it's a matrix.
> Lastly, if I have two different measurements
> for the time series data (e.g., heading and displacement) can I do a
> multivariate regression?

You can do anything you like.  It's a free country.  Well, it used
to be, anyway.

If you want a more useful answer than that, express your question
more clearly.


         Rolf Turner

More information about the R-help mailing list