[Rd] suggestion for ARMAacf()

Prof Brian Ripley ripley at stats.ox.ac.uk
Tue Aug 17 12:31:16 CEST 2004


Please supply a patch (diff -c or diff -u) against the current R-devel
sources, so there is no ambiguity as to what you are suggesting, together 
with examples and the intended output.

Verbal descriptions like this are not precise enough to be cross-checked.

On Tue, 17 Aug 2004, Hiroyuki Kawakatsu wrote:

> in 1.9.1, the return value from ARMAacf(pacf=TRUE) is not named by lags,
> contrary to ?ARMAacf. the simple fix is to move names(Acf) <-
> down after if(pacf), with an appropriate starting lag as pacf=TRUE appears
> to start at lag 1 (whereas pacf=FALSE starts at lag 0).
> 
> for consistency, one could argue to append 1 for lag 0 for pacf=TRUE
> (or start pacf=F at lag 1). however, given the inconsistency for the
> sample counterparts acf() and pacf(), i am not too bothered with this
> behavior as long as the lags are named for both pacf=T and pacf=F.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595



More information about the R-devel mailing list