Functional data analysis (FDA) deals with time series consisting of functional observations. We estimate a functional moving average process of order 1 by the projection method and by the iterative method. In addition we test the dependence order of a linear strictly stationary functional time series by two hypothesis tests, based on the concepts of the Box-Pierce test and of sample-splitting. The estimation approaches are used for forecasting and the tests help us choose an appropriate time series model, provided that the dependence structure is known. Finally we apply the former and the latter to Bavarian highway traffic data.     «

Functional data analysis (FDA) deals with time series consisting of functional observations. We estimate a functional moving average process of order 1 by the projection method and by the iterative method. In addition we test the dependence order of a linear strictly stationary functional time series by two hypothesis tests, based on the concepts of the Box-Pierce test and of sample-splitting. The estimation approaches are used for forecasting and the tests help us choose an appropriate time ser...     »