- Joined
- 1/16/16
- Messages
- 1
- Points
- 11
I am using the following bayesGARCH (CRAN - Package bayesGARCH) package in R. I am interested in forecasting $h_t$, the model setup is given bellow.
$r_t$ = $\varepsilon_t(\frac{v-2}{v}\omega_th_t)^{1/2}$ $\quad$ with $\quad$ $t=1,...,T$
$\varepsilon_t \overset{iid}{\sim}N(0,1)$
$\omega_t \overset{iid}{\sim}IG(\frac{v}{2},\frac{v}{2})$
$h_t = \alpha_0 + \alpha_1r^{2}_{t-1}+\beta h_{t-1}$
The package only provides simulated estimates of the parameter coefficients, namely $\alpha_0$, $\alpha_1$, $\beta$ and $v$. From my understanding the BayesGARCH does not have a function to forecast $h_t$, so I will have to forecast this manually. Any advice on forecasting this conditional volatility would be much appreciated.
$r_t$ = $\varepsilon_t(\frac{v-2}{v}\omega_th_t)^{1/2}$ $\quad$ with $\quad$ $t=1,...,T$
$\varepsilon_t \overset{iid}{\sim}N(0,1)$
$\omega_t \overset{iid}{\sim}IG(\frac{v}{2},\frac{v}{2})$
$h_t = \alpha_0 + \alpha_1r^{2}_{t-1}+\beta h_{t-1}$
The package only provides simulated estimates of the parameter coefficients, namely $\alpha_0$, $\alpha_1$, $\beta$ and $v$. From my understanding the BayesGARCH does not have a function to forecast $h_t$, so I will have to forecast this manually. Any advice on forecasting this conditional volatility would be much appreciated.