[R] testing an ARFIMA model for structural breaks with unknown breakpoint

[Achim Zeileis]

On Sun, 28 Jun 2009, mihaela wrote:

>
> Dear R users,
>
> I'm trying to use the "strucchange" package to determine structural breaks
> in an ARFIMA model.

ARFIMA models are not supported in "strucchange" at the moment. They 
typically require a different asymptotic theory due to the fractional 
integration.

> Unfortunately I'm not so familiar with this topic (and worse, I'm a beginner
> in R), so I don't know exactly how to specify my model so that the
> "Fstats","sctest" and "breakpoint" functions to recognize it and to
> calculate the potentially breakpoints.
> Could anyone give me a sugestion?
>
> I tried, however, (following the example specified in R Help) the functions:
>
> fs.x <- Fstats(x ~ 1))
> sctest(fs.x)
> plot(fs.x)
> breakpoints(fs.x)
>
> But I understand that this example only test for a sudden change in the mean
> of the series.
>
> Anyway, I have another problem related to the example above mentioned. My
> series is a "numeric" class (I only have one column with daily returns) an
> therefore I get the following error related to breakdates:

If you transform x to a "ts" object, in the simplest case
   x <- ts(x)
this error should not occur.

>> breakpoints(fs.x)
>
>         Optimal 2-segment partition:
>
> Call:
> breakpoints.Fstats(obj = fs.x)
>
> Breakpoints at observation number:
> 2441
>
> Corresponding to breakdates:
> Error in if (format.times) breakdates <- format.time(breakdates,
> obj$datatsp[3]) :
>  argument is of length zero
>
>
> So, I added to my returns column a new column containing the Date and my
> file to be imported looks like:
>
> date	                return
> 9/22/1997	-0.890957263
> 9/23/1997	-1.505530482
> 9/24/1997	-4.234587983
> 9/25/1997	0.385007594
> ............................................
>
> I tried to convert the new data in a "zoo" object or  "ts" , but after this
> the functions in the strucchange package don't work anymore..

"zoo" is not supported yet, unfortunately, but "ts" is.
Z

> Moreover, the "irts" function to create an irregular time-series object does
> not seems to work.
>
> Any help would be highly appreciated !
>
>
> Thank you in advance,
> M.
> -- 
> View this message in context: http://www.nabble.com/testing-an-ARFIMA-model-for-structural-breaks-with-unknown-breakpoint-tp24241675p24241675.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> 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.
>
>