Event Date
Speaker: Sebastian Kühnert, Visiting Assistant Professor, Statistics, UC Davis
Title: "Advances in Parameter and Covariance Estimation of Functional Time Series"
Abstract: Functional time series (FTS) with applications in economics, physics and medicine, among others, become increasingly important as modeling of randomly occurring functions becomes easier with increasing computing power.
In this talk, the basic concepts of FTS will be introduced, and motivating examples will be shown. Common FTS models will be explained and (lagged) covariances and weak dependence notions in the functional world will be discussed. We also introduce (lagged) (cross-)covariance operators and Yule-Walker estimators. We present our findings on both the explicit asymptotic rates for estimation errors of lagged (cross-)covariance operators of processes in (Cartesian products) of general, separable Hilbert spaces under mild conditions, and the asymptotic upper bounds of estimation errors when using Yule-Walker estimates for functional (G)ARCH and AR(MA) operators.
Seminar Date/Time: Thursday February 16, 2023, at 4:10pm
Refreshments: 3:30pm
Location: MSB 1147 (Colloquium Room)