On Sat, 18 Feb 2006, ohinata manabu wrote:
I want to forecast the volatility for the next day.
So the equation shoud be the following:
h_{t+1} = alpha_0 + alpha_1 * e_t^2 + beta_1 * h_t
My question is how I can get h_t, even though I don't know e_{t-1} and
h_{t-1} for the h_t equation.
You can retrieve the residuals and the conditional variance from the
GARCH model like this:
genr et = $uhat
genr ht = $h
Or if you estimate the model using the graphical interface, you can
add these variables to the dataset from the "Model data" menu in the
model window, under the item "Add to data set".
And I also want to know how to get e_t, then, and what is y_t? Is
it the actual volatility?
No, y_t is the level of the time-series in question.
Allin Cottrell