ABSTRACT

Causality analysis is one of the central topics in the field of data driven research. This talk will present our recent results on detection of causality from time series produced by dynamical systems. Not only time delays for causation transitivity are detected, but also indirect causations in multivariate time series are observed and eliminated for network construction. Several representative benchmark systems are used to validate the practical usefulness of our proposed methods.

[Light refreshments will be served at 3:00-3:30 pm]