[R] maximum likelihood

[Ben Bolker]

Alexandre Bonnet <bonnet <at> gmail.com> writes:

> 
> *hi,*
> 
> *using articial data, i'm supposed to estimate model*
> 
> *y(t) = beta(1) + beta(2)*x(t) + u(t), u(t) = gamma*u(t-1) + v(t), t =
> 1,...,100*
> 
> *which is correctly specified in two ways: ML ommiting the first
> observation, and ML using all 100 observation.*
> 

 [etc.]

  Alexandre,

  I think you're not getting any answers to this question
because it's a bit too vague (we have no context as to
_why_ you want to do this -- it also sounds a bit like
a homework question ...), and because it's stated much more
as a general statistical methodology question than as
an R question.  Please read the posting guide ...

  Doing this without specifying any parametric forms
would be tricky.  You may be able to do this by
searching for autoregressive model methods in RSiteSearch ...

  Ben Bolker