Characterization of interactions' persistence in time-varying networks

TL;DR

This paper introduces a measure called temporality to quantify the persistence of interactions in time-varying networks, analyzing empirical data and the effects of temporal resolution on network similarity.

Contribution

It proposes a novel descriptor for interaction persistence, assesses network similarity, and examines the impact of temporal resolution on network dynamics.

Findings

Empirical networks show higher similarity than randomized sequences.

Temporality varies with interaction time resolution.

The measure helps understand network dynamics and persistence.

Abstract

Many complex networked systems exhibit volatile dynamic interactions among their vertices, whose order and persistence reverberate on the outcome of dynamical processes taking place on them. To quantify and characterize the similarity of the snapshots of a time-varying network -- a proxy for the persistence,-- we present a study on the persistence of the interactions based on a descriptor named \emph{temporality}. We use the average value of the temporality, $\overline{\mathcal{T}}$, to assess how ``\emph{special}'' is a given time-varying network within the configuration space of ordered sequences of snapshots. We analyze the temporality of several empirical networks and find that empirical sequences are much more similar than their randomized counterparts. We study also the effects on $\overline{\mathcal{T}}$ induced by the (time) resolution at which interactions take place.