GARCH Estimation by Gibbs Sampling
This estimates a (univariate) GARCH model using Gibbs sampling, with a combination of independence chain Metropolis for the mean model parameters and random walk Metropolis for the GARCH variance parameters. An example with a multivariate (DCC) model is provided at http://www.estima.com/forum/viewtopic.php?f=8&t=1591. It's probably that a simpler random walk Metropolis for the full parameter set would also work with this example---it just might take a bit of tuning of the scale. A closely related example with does importance sampling rather than Gibbs sampling is provided at http://www.estima.com/forum/viewtopic.php?f=8&t=1609&p=6042