When observations are curves over some natural time interval, the field of functional data analysis comes into play. In this thesis, we investigate the case where the observed curves are dependent in time. We model the temporal dependence using functional linear processes. Extending the Innovations Algorithm well-known from multivariate time series analysis, we construct consistent estimators and predictors for functional moving average models. We apply our methodology to high-dimensional highway traffic data.
«
When observations are curves over some natural time interval, the field of functional data analysis comes into play. In this thesis, we investigate the case where the observed curves are dependent in time. We model the temporal dependence using functional linear processes. Extending the Innovations Algorithm well-known from multivariate time series analysis, we construct consistent estimators and predictors for functional moving average models. We apply our methodology to high-dimensional highwa...
»