Paulo,
I have run into similar problems when wanting to work with multiple
seasonality. You are probably already familiar with the work by James Taylor
in this area. I contacted him and he was kind enough to share some of his
(Gauss) code, and I have tried to implement it in R, but due to my lack of
programming skill I don't really trust the results and it is very slow. I
will try to get it to the point where it could be useful for others.
best,
Spencer
On 6/14/07, Paulo Tanimoto <tanimoto@u.arizona.edu>
wrote:>
> Dear R community,
>
> I have a project with electricity load forecasting, and I got hourly
> data for system load. If you haven't worked with electricity before,
> seasonality comes in many flavors: a daily pattern, with a peak at
> around 7pm; a weekly pattern, in which we use more electricity on
> weekdays in comparison to weekends; a winter-summer pattern, with air
> conditioning and heaters playing an important role, etc.
>
> In SAS, I could specify an ARIMA with multiple seasonality terms, say
> (1,0,1) with period=24, and (1,0,1) with period=168 (I don't remember
> the orders I was using). This is how many studies I found on IEEE
> were modeling electricity load and it worked pretty well with the data
> I have.
>
> However, I can't seem to find how to do that in R. I've read the
help
> files a few times, but to no avail. Is it possible? I hope I'm
> missing something straightforward.
>
> Thank you,
>
> Paulo Tanimoto
> University of Arizona
>
> ______________________________________________
> R-help@stat.math.ethz.ch 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.
>
[[alternative HTML version deleted]]