Activity Number:
|
408
- SPAAC Poster Competition
|
Type:
|
Topic Contributed
|
Date/Time:
|
Tuesday, July 31, 2018 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Scientific and Public Affairs Advisory Committee
|
Abstract #329632
|
|
Title:
|
Analysis of Non-Stationary Time Series Using Copula-Based Dependence Measures
|
Author(s):
|
Yongxin Zhu* and Charles Fontaine and Hernando Ombao
|
Companies:
|
King Abdullah University of Science and Technology and King Abdullah University of Science and Technology and King Abdullah University of Science and Technology
|
Keywords:
|
Non-stationary;
Time series;
Cross-dependence;
Parametric copulas;
Fourier transform;
Brain signals
|
Abstract:
|
Dependence between brain regions is complex and cannot be fully explained by the classical correlation and coherence-based measures. In addition, brain signals often exhibit non-stationarity as the nature and strength of dependence can change over time. We have developed a new approach to characterize the dependence between time series using copulas. Compared to classical measures, fitting parametric bivariate/multivariate copulas allows us to specify the models for marginal distributions of each time series, separately from the possible non-dependence structure among these distributions. This offers greater flexibility beyond commonly used distributions in model specification and estimation. An exploratory analysis of local field potential (LFP) from rats recorded during an induced stroke will be presented. We will show that the proposed copula-based dependence measures capture the evolution and changes in generalized dependence measures as the brain reorganizes itself as it responds to a stroke which is a serious shock to the system.
|
Authors who are presenting talks have a * after their name.