Functional Time Series Driven by Dynamic Systems

Traditionally, time series analysis deals with scalar or vector observations obtained each time period, using linear, nonlinear or nonparametric approaches. Modern data collection capability has led to broader definition of 'data' and more and more observations are in the form of functions, images, and distributions. When such observations are observed over time and when they exhibit dynamic behaviours, time series models in the functional space become a necessary and useful modeling and prediction tool. In this talk we present a new approach of modeling functional time series through a dynamic system approach. Modeling, estimation and prediction issues will be discussed, with several applications.