Hi,I have a question about the fracdiff.var function (package fracdiff) which
goal is to recompute more precise confidence intervals for the parameters
estimated by fracdiff (or arfima). More precisely, it deals with the standard
error of the "d" coefficient :
Is it normal that the standard error of the "d" coefficient can be brought
customly close to zero by decreasing the "h" argument ?
The example of the help file shows this fact :
{Quote}
## Generate a fractionally-differenced ARIMA(1,d,1) model :ts.test <-
fracdiff.sim(10000, ar = .2, ma = .4, d = .3)## estimate the parameters in an
ARIMA(1,d,1) model for the simulated seriesfd.out <- fracdiff(ts.test$ser, nar=
1, nma = 1)
## Modify the covariance estimate by changing the finite-difference
interval(fd.o2 <- fracdiff.var(ts.test$series, fd.out, h = .0001))## looks
identical as print(fd.out),## however these (e.g.) differ
:vcov(fd.out)vcov(fd.o2)
{end of quote}
Looking at vcov(fd.o2) gives that sd's are indeed lower than in vcov(fd.out).
However, by testing :
vcov(fracdiff.var(ts.test$series, fd.o2, h = 0.00000000001))or
evenvcov(fracdiff.var(ts.test$series, fd.o2, h = 1e-90))
the sd of "d" doesn't stop decreasing, while the others have reached their
"computational" limit.
Is that normal ? If not, which value of sd(d) can I use ?
Hopping I have been clear enough,Many thanks !
[[alternative HTML version deleted]]
______________________________________________
R-help@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.
