摘要：In this paper we develop a Functional Data Model for forecasting Italian Deposits Time Series. Bank deposits play an important role in ensuring the banks borrowing capacity and for this reason its correct modeling and forecasting represent an interesting task for policy makers. As it is well known,deposit series are affected by seasonality. In the Central Banks and other research institutions the standard procedure applied to this kind of monetary time series is to operate a preliminary seasonal adjustment in order to filter out typical calendar effect and within-year fluctuations. We assume a different starting point in modeling and forecasting seasonal time series,taking into account how the seasonality evolves across the years and trying to incorporate this feature in the model via functional data analysis. We utilize the Phase Plane Plot in order to show the evolution of the seasonality of the Italian Deposits from 1998 to 2008,working on a monthly time series and producing different plots for each year. We fit the data on the historical values using principal component techniques and construct forecast intervals projecting the model components with ARIMA process. The empirical results are presented using a range of graphical analysis.
关键词：deposits;forecasting;functional data analysis;seasonal time series;smoothing.
