I have an ARCH model like this:
Delta y = y_{t+1} - t_t =
k (theta - y_t) Delta t + v_t N(0,1) sqrt(Delta t)
where
v_t^2 = a_0 + a_1 (y_{t-1} - E[y_{t-1}])^2
Is there any R function that, given a series y_i, determines
k, theta, a_0 and a_1 by maximum likelihood?
I tried to use garch from tseries, but didn't understant how
the output from its summary relate to the above quantities.
Alberto Monteiro