STA 290 Seminar Series
Monday April 4th, 10:00am, MSB 1147 (Colloquium Room)
Speaker: Greg Rice (University of Waterloo)
Title: “Stationarity testing and break date estimation with functional time series.”
Abstract: In fields ranging from economics and finance to energy research and climatology, data on certain continuous time phenomena are collected with high frequency. Often, such data can be parsed into natural, perhaps hourly or daily, segments that may be viewed as a time series of curves. A fundamental issue that must be addressed before an attempt is made to statistically model such data is whether these curves form a stationary functional time series. I will discuss the interpretation of stationarity in the context of function space valued random variables, and introduce testing procedures to test for stationarity with a given functional time series. The tests are developed as nontrivial extensions of the broadly used tests in the KPSS family. When stationarity is rejected, it may be due to a number of factors, including a break or "change point" in the mean. I will also discuss some new methodology that does not rely on dimension reduction for estimating break dates with functional time series that contain a mean change.