[Rd] arima (PR#7092)

stoffer at pitt.edu stoffer at pitt.edu
Thu Jul 15 02:13:05 CEST 2004

Full_Name: david stoffer
Version: 1.9.0
OS: win xp pro
Submission from: (NULL) (

arima() doesn't fit the model correctly if there is differencing and a constant
is still needed.  also, the description saying that it's not necessary to ever
fit a constant is incorrect.

if you fit a pure AR using arima(), it tells you it's giving you the intercept
term, but it's actually reporting the estimate of the mean.

More information about the R-devel mailing list