[R] Holt-Winters code / time series

[Paul Gilbert]

John Aitchison wrote:
> I am also working through "Non-linear time series models in empirical
> finance" (Franses & van Dijk) and slowly implementing some bits and pieces
> .. I'd be glad to contribute these as and when.

R and S are great for cobbling together quick solutions, and for a small problems it
is often easier to just write code from scratch than it is to figure out someone
else's  code, even when that code is well documented. If you are thinking of putting
some effort into this development I would encourage you to look at my dse package.
As examples dse implements multivariate ARMA (ARMAX, ARIMA, VAR, ...) and state space
models, but more importantly there is a whole structure for dealing with time series,
time series databases (e.g. Fame), testing time series estimation techniques,
forecasting, and testing forecasting models. I have used the structures to implement
very different kinds of time series models (e.g. neural nets and  structural models
that run Troll). There is some overhead, so if you are thinking of a quick solution
to a small problem then I only recommend the library for the already implemented ARMA
and state space models, but if you have in mind something that is a substantial
programming effort then you will probably find there are many useful tools in the
package.

Paul Gilbert


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._