[S] BOUNCE s-news@wubios.wustl.edu: taboo header: /^received:\s

Jean Adams (Jean_Adams@usgs.gov)
Sun, 25 Oct 1998 01:53:40 -0700



Approved: 9SNews9
Received: from utstat.toronto.edu (utstat.toronto.edu [128.100.73.1])
by wubios.wustl.edu (8.9.0/8.9.0) with SMTP id XAA01548
Sat, 24 Oct 1998 23:30:33 -0500 (CDT)
From: andrey@utstat.toronto.edu
Message-Id: <199810250430.XAA01548@wubios.wustl.edu>
Date: Sun, 25 Oct 98 00:30 EDT
To: andrey, s-news@wubios.wustl.edu
Subject: [S] Time series

William Checkley asked:
Does anyone know whether SPLUS can fit multivariate time series regression
models with ARMA errors? (what commands?)

As far as I know, S-Plus function `arima.mle' will fit a univariate ARIMA
model (using gaussian MLE) and allows for regressor variables via the `xreg'
optional argument.

Mulivariate versions of this do not seem to be possible with the usual
S-Plus functions in a straightforward way. However the S+GARCH module
contains the function `mgarch' which does fit multivariate time series.
the garch feature can be disengaged by setting the `formula.var' parameter
appropriately. Doing so leaves a function which fits multivariate time
series regression models.

[Don't know the answer to the second question that had been posed..]

Andrey Feuerverger / Dept Statistics / Univ of Toronto /
andrey@utstat.toronto.edu

-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news