|
Activity Number:
|
220
- High-Dimensional Analysis of Complex Dependent Data
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 11, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
|
Abstract #316737
|
|
|
Title:
|
High-Dimensional Spectral Analysis
|
|
Author(s):
|
Efstathios Paparoditis and Jonas Krampe*
|
|
Companies:
|
University of Cyprus and University of Mannheim
|
|
Keywords:
|
High Dimensional Tim Series;
Spectral Analysis;
Frequency Domain Bootstrap
|
|
Abstract:
|
We consider frequency domain inference for high dimensional time series and we focus on the matrix of partial coherencies. Partial coherencies are important quantities which measure the linear dependence between two time series after removing the linear influence of the other time series of the system. De-biased partial coherencies are introduced and their limiting distribution is investigated under certain sparsity and dependence assumptions on the underlying high dimensional process. Using an appropriate developed frequency domain bootstrap procedure for high dimensional time series, statistical inference for a variety of testing problems concerting partial coherencies is developed.
|
Authors who are presenting talks have a * after their name.
