In empirical studies, financial market data show a number of so-called stylized facts, such as "volatility clustering" or "fat tails".

A variety of models (e.g. AR, ARCH, GARCH, HAR, ...) exist to model these peculiarities in the returns of financial market securities.

The latter model is the Heterogeneous Autoregressive model of realized volatility according to Corsi(2009), which tries to explain future volatility by aggregating past volatilities. The underlying idea is to aggregate volatilities over different time periods to account for different planning horizons of market participants in the model.

Within the scope of a final thesis, different models can be compared with respect to their forecasting quality. The thesis will include theoretical as well as empirical aspects.

Corsi, F. (2009) A simple long memory model of realized volatility, Journal of Financial Econometrics 7, 174–196.