Memory and betweenness preference in temporal networks induced from time series

Tongfeng Weng, Jie Zhang, Michael Small, Rui Zheng, Pan Hui

Research output: Contribution to journalArticle

10 Citations (Scopus)


We construct temporal networks from time series via unfolding the temporal information into an additional topological dimension of the networks. Thus, we are able to introduce memory entropy analysis to unravel the memory effect within the considered signal. We find distinct patterns in the entropy growth rate of the aggregate network at different memory scales for time series with different dynamics ranging from white noise, 1/f noise, autoregressive process, periodic to chaotic dynamics. Interestingly, for a chaotic time series, an exponential scaling emerges in the memory entropy analysis. We demonstrate that the memory exponent can successfully characterize bifurcation phenomenon, and differentiate the human cardiac system in healthy and pathological states. Moreover, we show that the betweenness preference analysis of these temporal networks can further characterize dynamical systems and separate distinct electrocardiogram recordings. Our work explores the memory effect and betweenness preference in temporal networks constructed from time series data, providing a new perspective to understand the underlying dynamical systems.

Original languageEnglish
Article number41951
JournalScientific Reports
Publication statusPublished - 3 Feb 2017

Fingerprint Dive into the research topics of 'Memory and betweenness preference in temporal networks induced from time series'. Together they form a unique fingerprint.

Cite this