Summary
the 2014 International Symposium on Nonlinear Theory and its Applications
2014
Session Number:B1L-B
Session:
Number:B1L-B3
Discriminating dynamic regimes using measures of causality networks from multivariate time series
Christos Koutlis, Dimitris Kugiumtzis,
pp.225-228
Publication Date:2014/9/14
Online ISSN:2188-5079
Summary:
In many applications ranging from neurophysiology to finance, the dynamics of the underlying mechanism to observed multivariate time series is believed to change and this is reflected to the inter-dependence structure of the observed variables. We consider a Granger causality index for estimating the inter-dependence structure and form causality networks with nodes the observed variables and directed connections given by the selected Granger causality index. The focus of the study is on assessing the different network measures as to their ability in discriminating different dynamic regimes of the system underlying the multivariate time series. For this, we first compute the network measures on many realizations of the coupled Mackey-Glass system under different coupling structures, and then to electroencephalogram recordings containing episodes of epileptiform discharges. The ranking of the network measures on the simulated and real data revealed the same subset of measures performing best.