B0933
Title: Likelihood analysis of continuous-time Gaussian moving average processes having scaling properties
Authors: Tetsuya Takabatake - Hiroshima University (Japan) [presenting]
Abstract: Recent studies in mathematical finance and econometrics suggest that properties of the roughness of the sample path and the persistency of the auto-covariance function would be important factors in constructing better forecasting models of the volatility process of the asset price. Continuous-time Gaussian moving average processes are considered, having scaling properties including the roughness and long-memory properties, as models of the log-volatility process, and then the likelihood/quasi-likelihood analysis of discretely observed continuous-time Gaussian moving average processes is discussed.