A0597
Title: Joint estimation of precision matrices for high-dimensional time series with long-memory
Authors: Jongik Chung - University of Central Florida (United States) [presenting]
Qihu Zhang - Fei Tian College - Middletown (United States)
Cheolwoo Park - Korea Advanced Institute of Science and Technology (Korea, South)
Abstract: The focus is on the simultaneous estimation of multiple precision matrices for high-dimensional time series with long memory. The motivating example is a time series of resting-state functional magnetic resonance imaging (fMRI) data collected from multiple subjects. Estimating the brain network for each subject and a common structure representative of a group of subjects is of interest. A few approaches are considered for simultaneously estimating individual and group precision matrices for long-memory time series data using weighted aggregation. The convergence rates of the precision matrix estimators for various norms and their expectations under a sub-Gaussian or heavy-tailed assumption are examined. The empirical performance is demonstrated via simulated examples and resting-state fMRI data.
