A0666
Title: Inference for nonstationary timeseries using optimal Gaussian approximation with explicit construction
Authors: Sayar Karmakar - University of Florida (United States) [presenting]
Abstract: Inference problems for time series, such as curve estimation for time-varying models or testing for the existence of change-point, have garnered significant attention. However, these works are restricted to the limiting assumption of independence and/or stationarity at their best. The main obstacle is that the existing optimal Gaussian approximation results for nonstationary processes only provide an existential proof, and thus they are difficult to apply. A clear path is provided to construct such a Gaussian approximation. The proposed Gaussian approximation results encapsulate a very large class of nonstationary time series, obtain the optimal rate and yet have good applicability. Building on such a Gaussian approximation, theoretical results for changepoint detection and simultaneous inference in the presence of nonstationary errors are shown. The theoretical results with extensive simulation studies and some real data analyses are substantiated.
