From Time Series to Complex Networks

Deniz Eroglu

Time series are very helpful to understand complex systems, e.g., for developing models, predicting the future, comparing different dynamics, etc. Besides well-known standard methods, approaches based on complex networks can provide complementary knowledge about the complexity of dynamical systems as derived from time series. There are several approaches to generate complex networks from time series. Here we will present an approach of generating complex networks from phase space reconstructions and discuss the selection of meaningful parameters for generating such complex networks. Synthetic data sets and real world applications will be used to demonstrate the potentials of this approach.