[R] models with I(1) errors

[Jason Turner]

On Wed, 2003-09-24 at 19:31, Vito Muggeo wrote:
> Dear all,
> I'm interested in fitting time-series linear models with I(1) errors. Namely
> given
> y_t=a+b*t+u_t
> the random term u_t are such that
> u_t-u_{t-1}=e_t~iid N(0,\sigma)
> 

library(nlme)
help(lme)  #note the optional correlation argument
help(corClasses)

You can specify AR(1) (among other) correlation structures in the error
term with lme().

The cannonical reference for the nlme library is

@Book{
  PinheiroBates2000,
  author =	 {Jos\'e C. Pinheiro and Douglas M. Bates},
  title =	 {Mixed-Effects Models in S and S-PLUS},
  publisher =	 {Springer-Verlag},
  year =	 {2000},
  address =	 {New York}
}

Cheers

Jason


-- 
Indigo Industrial Controls Ltd.
http://www.indigoindustrial.co.nz
+64-(0)21-343-545