Forecasting functional time series

22 February 2008
 Download pdf
  • Where: Australian Frontiers of Science

Abstract: Functional time series are curves that are observed sequentially in time. For example, the curve of death rate as a function of age is observed annually. Yield curves in finance (essentially interest rates as a function of the term of investment) are observed each week or each day. Electricity consumption as a function of temperature is observed every month. These are all high dimensional functional data, indexed by time. I will briefly discuss methods for describing, modelling and forecasting such functional time series data observed at discrete times and possibly with observational error. Challenges include developing useful graphical tools, dealing with outliers (e.g., death rates have outliers in years of wars or epidemics), cohort effects, synergy between groups, and deriving prediction intervals for forecasts. I will describe some applications including forecasting fertility rates and breast cancer incidence rates.

« Population forecasting and the importance of being uncertain | Exponential smoothing and non-negative data »