[R] Time Series ts() Objects

[Jim Lindsey]

> 
> On Mon, Feb 11, 2002 at 08:53:22AM +0100, Jim Lindsey wrote:
> > 
> > > 
> > > On Mon, Feb 11, 2002 at 03:08:55PM +1300, Ko-Kang Kevin Wang wrote:
> > > > Is it possible to create a ts() object, whose data is daily based BUT
> > > > measured only on working days?
> > > 
> > > No, you can't. R only has timeseries with constant frequency or delta-t. 
> > 
> > What you mean is "ts only has timeseries with constant frequency or delta-t."
> 
> Exactly, thank you for making that clearer. I also omitted to say that one
> can work around the problem by using, e.g., a delta of 1 day and flagging
> non-business days as NA. Adrian Trapletti has examples for that inside the
> tseries package.
> 
> > There are libraries in R for continuous time series, including my carma
> > function in my growth library.
> 
> Interesting. Can these have "gaps"? In other words, could one "discretize" a
> continuous period and fill only certain segments?  As an example take a week
> composed of 7*24*12 = 2016 five-minute intervals, and then "fill" only those
> matching Mon - Fri and 9:30 to 16:00 ?

In continuous time ARMA, observations can have any spacing. So the
answer is yes. CARMA is based on Kalman filtering so should have no
problem handling long series. Jim

> 
> Dirk
> 
> -- 
> Good judgment comes from experience; experience comes from bad judgment. 
> 							    -- Fred Brooks
> 

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._