UCPH Statistics Seminar: Zhou Zhou
Title: Auto-Regressive Approximations to Non-Stationary Time Series, with Inference and Applications
Speaker: Zhou Zhou from University of Toronto
Abstract: Understanding the time-varying structure of complex temporal systems is one of the main challenges of modern time series analysis. In this talk, I will demonstrate that a wide range of short-range dependent non-stationary and nonlinear time series can be well approximated globally by a white-noise-driven auto-regressive (AR) process of slowly diverging order. Uniform statistical inference of the latter AR structure will be discussed through a class of high-dimensional L2 tests. I will further discuss applications of the AR approximation theory to globally optimal short-term forecasting, efficient estimation, and resampling inference under complex temporal dynamics.