Combining statistical methods from discretely observed stochastic processes with regression, inverse problem and bootstrap techniques (with [A1], [B1]-[B3]) our aim in this project is to derive new estimating procedures for the different quantities of stochastic volatility models such as the volatility process, the integrated volatility and a possible jump component. This will lead to a better understanding of the relationship between volatility and jump components which might provide better forecasts and a more realistic risk assessment. Furthermore, we aim to build a bridge between time series models and continuous time models by considering limits of complex time series models such as ACD and long memory models (with [A3]) and analyze their features in the context of market microstructure.
DFG Programme
Research Units
